Magnetic Fabric: Methods and Applications (Geological Society Special Publication No. 238)

Magnetic Fabric: Methods and Applications

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 J. 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: MARTIN-HERNANDEZ, F., LUNEBURG, C. M., AUBOURG, C. & JACKSON, M. (eds) 2004. Magnetic Fabric: Methods and Applications. Geological Society, London, Special Publications, 238. POKORNY, J., SUZA, P. & HROUDA, F. 2004. Anisotropy of magnetic susceptibility of rocks measured in variable weak magnetic fields using the KLY-4S Kappabridge. In: MARTINHERNANDEZ, F., LUNEBURG, C. M., AUBOURG, C. & JACKSON, M. (eds) 2004. Magnetic Fabric: Methods and Applications. Geological Society, London, Special Publications, 238, 69-76.

GEOLOGICAL SOCIETY SPECIAL PUBLICATION NO. 238

Magnetic Fabric: Methods and Applications

EDITED BY F. MARTIN-HERNANDEZ Faculty of Geosciences, Utrecht University, The Netherlands

C. M. LUNEBURG Department of Geology and Geophysics, University of New Orleans, USA

c. AUBOURG Laboratoire de Tectonique, Universite de Cergy-Pontoise, France

and

M. JACKSON Institute for Rock Magnetism, University of Minnesota, USA

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)1225445046 Fax +44(0)1225442836) Online bookshop: http://www.geolsoc.org.uk/bookshop The publishers make no representation, express or implied, with regard to the accuracy of the information contained in this book and cannot accept any legal responsibility for any errors or omissions that may be made. © 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 item-fee code for this publication is 0305-8719/04/S15.00. British Library Cataloguing in Publication Data A catalogue record for this book is available from the British Library. ISBN 1 -86239- 170-X

Typeset by J W Arrowsmith Ltd, Bristol, UK Printed by Antony Rowe Ltd, Chippenham, UK Distributors USA AAPG Bookstore PO Box 979 Tulsa OK 74101-0979 USA Orders: Tel. + 1 918 584-2555 Fax +1 918560-2652 E-mail [email protected] India Affiliated East-West Press Private Ltd Marketing Division, G-l/16 Ansari Road, Daryaganj, New Delhi 110002 India Orders: Tel. +91 112327-9113, +91 112326-4180 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 MARTIN-HERNANDEZ, F., LUNEBURG, C. M., AUBOURG, C. & JACKSON, M. Magnetic fabric: methods and applications - an introduction Magnetic fabric characterization methods and mineral sources JEZEK, J. & HROUDA, F. Determination of the orientation of magnetic minerals from the anisotropy of magnetic susceptibility

1

9

POTTER, D. K. A comparison of anisotropy of magnetic remanence methods - a user's guide for application to palaeomagnetism and magnetic fabric studies

21

MUXWORTHY, A. R. & WILLIAMS, W. Distribution anisotropy: the influence of magnetic interactions on the anisotropy of magnetic remanence

37

HROUDA, F. Problems in interpreting AMS parameters in diamagnetic rocks

49

NAKAMURA, N. & BORRADAILE, G. J. Metamorphic control of magnetic susceptibility and magnetic fabrics: a 3-D projection

61

POKORNY, J., SUZA, P. & HROUDA, F. Anisotropy of magnetic susceptibility of rocks measured in variable weak magnetic fields using the KLY-4S Kappabridge

69

Sedimentary fabrics DEBACKER, T. N., ROBION, P. & SINTUBIN, M. The anisotropy of magnetic susceptibility (AMS) in low-grade, cleaved pelitic rocks: influence of cleavage/bedding angle and type and relative orientation of magnetic carriers

77

HIRT, A. M., LOWRIE, W., LUNEBURG, C., LEBIT, H. & ENGELDER, T. Magnetic and mineral fabric development in the Ordovician Martinsburg Formation in the Central Appalachian Fold and Thrust Belt, Pennsylvania

109

LARRASOANA, J. C., PUEYO, E. L. & PARES, J. M. An integrated AMS, structural, palaeo- and rock-magnetic study of the Eocene marine marls from the Jaca-Pamplona basin (Pyrenees, N Spain); new insights into the timing of magnetic fabric acquisition in weakly deformed mudrocks

127

MATASOVA, G. G. & KAZANSKY, A. Yu. Magnetic properties and magnetic fabrics of Pleistocene loess/palaeosol deposits along west-central Siberian transect and their palaeoclimatic implications

145

ANDERSON, M. W. & MORRIS, A. The puzzle of axis-normal magnetic lineations in folded lowgrade sediments (Bude Formation, SW England)

175

PARES, J. M. How deformed are weakly deformed mudrocks? Insights from magnetic anisotropy

191

Igneous fabrics CANON-TAFIA, E. Anisotropy of magnetic susceptibility of lava flows and dykes: a historical account

205

CANON-TAPIA, E. & CHAVEZ-ALVAREZ, M. J. Theoretical aspects of particle movement in flowing magma: implications for the anisotropy of magnetic susceptibility of dykes

227

PETRONIS, M. S., HACKER, D. B., HOLM, D. K., GEISSMAN, J. W. & HARLAN, S. S. Magmatic flow paths and palaeomagnetism of the Miocene Stoddard Mountain laccolith, Iron Axis region, Southwestern Utah, USA

251

ARCHANJO, C. J. & LAUNEAU, P. Magma flow inferred from preferred orientations of plagioclase of the Rio Ceara-Mirim dyke swarm (NE Brazil) and its AMS significance

285

vi

CONTENTS

Tectonic fabrics BORRADAILE, G. J. & JACKSON, M. Anisotropy of magnetic susceptibility (AMS): magnetic petrofabrics of deformed rocks

299

CHADIMA, M., HANSEN, A., HIRT, A. M., HROUDA, F. & SIEMES, H. Phyllosilicate preferred orientation as a control of magnetic fabric: evidence from neutron texture goniometry and low and high-field magnetic anisotropy (SE Rhenohercynian Zone of Bohemian Massif)

361

GIL-IMAZ, A. & BARBERO, L. Anisotropy of magnetic susceptibility in the Montes de Toledo area (Hercynian Iberian Belt, Spain) and its petrostructural significance

381

PUEYO, E. L., ROMAN-BERDIEL, M. T., BOUCHEZ, J. L., CASAS, A. M. & LARRASOANA, J. C. Statistical significance of magnetic fabric data in studies of paramagnetic granites

395

AUBOURG, C., KLOOTWIJK, C. & KORSCH, R. J. Magnetic fabric constraints on oroclinal bending of the Texas and CofTs Harbour blocks: New England Orogen, eastern Australia

421

LOPEZ DE LUCHI, M. G., RAPALINI, A. E., SIEGESMUND, S. & STEENKEN, A. Application of magnetic fabrics to the emplacement and tectonic history of Devonian granitoids in central Argentina

447

Complex fabrics: superposition and alteration KADZIALKO-HOFMOKL, M., MAZUR, S., WERNER, T. & KRUCZYK, J. Relationships between magnetic and structural fabrics revealed by Variscan basement rocks subjected to heterogeneous deformation—a case study from the Klodzko Metamorphic Complex, Central Sudetes, Poland

475

DE WALL, H. & WARR, L. N. Oblique magnetic fabric in siderite-bearing pelitic rocks of the Upper Carboniferous Culm Basin, SW England: an indicator for palaeo-fluid migration?

493

JUST, J., KONTNY, A., DE WALL, H., HIRT, A. M. & MARTIN-HERNANDEZ, F. Development of magnetic fabrics during hydrothermal alteration in the Soultz-sous-Forets granite from the EPS-1 borehole, Upper Rhine Graben

509

HAMILTON, T. D., BORRADAILE, G. J. & LAGROIX, F. Sub-fabric identification by standardization of AMS: an example of inferred neotectonic structures from Cyprus

527

Index

541

Magnetic fabric: methods and applications - an introduction F. MARTIN-HERNANDEZ1, C. M. LUNEBURG2, C. AUBOURG3 & M. JACKSON4 1

Paleomagnetic Laboratory 'Fort Hoofddijk', Faculty of Geosciences, Utrecht University, 3584 CD Utrecht, The Netherlands 2 Department of Geology and Geophysics, University of New Orleans, 2000 Lake Shore Drive, New Orleans, LA 70148, USA ^Laboratoire de Tectonique, UMR 7072, Universite de Cergy-Pontoise, 95031, Cergy Cedes, France 4 Institute for Rock Magnetism, University of Minnesota, Minneapolis MN 55455, USA

Fifty years have now passed since Graham (1954) published his seminal paper advocating the use of anisotropy of magnetic susceptibility (AMS) as a rapid and sensitive petrofabric tool. During these five decades, Graham's 'underexploited' method has become standard, and AMS and related techniques are now routinely applied to characterizing fabrics in a wide variety of geological materials (e.g. the GEOREF database lists over 500 journal publications with 'magnetic anisotropy' as keywords). Magnetic anisotropy works as a petrofabric tool because individual grains of most minerals are magnetically anisotropic, i.e. easier to magnetize in certain orientations, which are governed primarily by crystallography and/or grain shape. Magnetic anisotropy at the bulk rock scale results from the preferred crystallographic orientation (PCO) and/or preferred dimensional orientation (PDO) of anisotropic mineral grains. AMS can also result from magnetostatic interactions among closely spaced, strongly magnetic grains that are heterogeneously distributed in a matrix of more weakly magnetic minerals. In either case, magnetic anisotropy is directly related to some aspects of rock fabric, and thus it provides a quick, simple and effective characterization tool, even though the relationship between magnetic fabric and petrofabric is quite complex in detail. The present collection of papers originated, in part, at a special session on magnetic fabrics at the Joint Assembly of the EGS-AGU-EUG (April 2003) in Nice, which highlighted recent methodological advances, theoretical and experimental studies, and characterization of flow and deformation fabrics in rocks and sediments. A similar session at the AGU Fall Meeting (December 2003) in San Francisco underscored the continuing breadth of interest in magnetic

fabric research, and suggested that the time was ripe for a comprehensive critical assessment of the field on this golden anniversary of Graham's influential publication. It is our goal in this volume to present a view of the current state of the art in magnetic fabric analysis, from the physical foundations to the geological applications, and to illustrate as well some of the important limitations, unresolved problems and directions for future research. The first observations of AMS in natural samples were presented before 1954 (e.g. Ising 1943). The early works focused on establishing correlations between the AMS principal directions and the structural features in sediments (Rees 1961, 1965; Graham 1966), igneous and metamorphic rocks (Stacey 1960; Stacey et al 1960; Khan 1962; Stone 1962). Contemporary advances in palaeomagnetism led to a need for methods to assess the fidelity of the stable remanence, increasing the interest in AMS and other techniques for measuring magnetic anisotropy (Fuller 1960, 1963; Rees 1961). The technique was refined with the appearance of rigorous measurement schemes (Girdler 1961) and mathematical theories to explain the origin of AMS (Nagata 1961; Stacey 1963; Uyeda et al. 1963; Bathal 1971). The ultimate goals of magnetic fabric research in the geosciences have remained essentially unchanged since these early studies: determining flow/emplacement directions in sediments, intrusive and extrusive rocks; establishing principal orientations (as well as estimating magnitudes) of finite or incremental strains; and ascertaining the extent of possible deviations of NRM vectors from the palaeofield orientation due to anisotropic acquisition and/or subsequent deformation. With these consistent goals, the field has evolved primarily in its recognition of the

From: MARTIN-HERNANDEZ, F., LUNEBURG, C. M., AUBOURG, C. & JACKSON, M. (eds) 2004. Magnetic Fabric: Methods and Applications. Geological Society, London, Special Publications, 238, 1-7. 0305-8719/04/$15.00 © The Geological Society of London 2004.

2

F. MARTIN-HERNANDEZ ET AL.

complexity of natural magnetic fabrics and in the development of increasingly sophisticated methods for extracting meaningful information from them. This evolution may be traced through a series of excellent reviews (Bhathal 1971; Hrouda 1982; MacDonald & Ellwood 1987; Rochette et al 1992; Tarling & Hrouda 1993; Kodama 1995; Borradaile & Henry 1997). Magnetic fabric characterization methods and mineral sources The first measurements of AMS were performed using low-field torque magnetometers (Ising 1943; Granar 1958; Stone 1962; King & Rees 1966). The crystalline anisotropy can also be determined by means of high-field torque magnetometry in natural samples (Stacey 1960; Banerjee & Stacey 1967) and ferromagnetic single crystals such as pyrrhotite (Mikami et al. 1959; Bin & Pauthenet 1963; Sato et al 1964) or hematite (Townsend 1916; Townsend 1920; Lin 1959). Measurements are tedious and time-consuming and for this reason an alternative was found, with measurements based on astatic magnetometers (Johnson & Steiner 1937; Collinson 1967; Deutsch et al 1961; Roy 1971; Fujiwara & Yoshida 1981) or spinner magnetometers (Jelinek 1966; Noltimier 1971; Schmidt et al. 1988). The technique became a standard measurement with the appearance of a.c. susceptibility bridges (Fuller 1960; Girdler 1961; Graham 1964; Jelinek 1973), which are relatively fast, inexpensive, simple to use and very precise (Pokorny et al., this volume). Parallel to the development of the instrumentation, the measurement technique became more refined and additional work helped to evaluate accurately the magnetic susceptibility tensor and analytical uncertainty, for individual specimens and for groups of samples (Owens 2000a; Owens 20006; Borradaile 2003 and references therein). The statistical treatment of the AMS data requires analysis of the shape of the ellipsoid, degree of anisotropy and their correlation with bulk properties as the magnetic susceptibility (Nakamura & Borradaile, this volume). The AMS measures the anisotropy of all minerals present in the samples, weighted according to their specific susceptibilities (which vary by roughly six orders of magnitude), concentrations, grain-scale anisotropies (controlled by crystallography, grain shape and/or stress) and degrees of preferred orientation. In most rocks and sediments, AMS is mainly due to the heavily weighted trace ferromagnetic phases, and often it reflects a single dominant event or

process. In such favourable cases it is possible to quantify the orientation distribution of the magnetic minerals (Jezek & Hrouda, this volume). But in many cases the AMS has been proved to consist of multiple superposed components, carried by different mixtures of ferro-, para- and even diamagnetic minerals (Hrouda; Just et al., this volume), with different preferred orientations imposed by different mechanisms (e.g. Daly 1967). It has become necessary, in order to unravel the history of deformation and alteration of natural samples, to separate these individual magnetic subfabrics. Several techniques are now available for this purpose, based on measurements at different temperatures (Richter & van der Pluijm 1994; Hrouda et al. 1997; Liineburg et al. 1999; Pares & van der Pluijm 2002), measurements at different fields (Hrouda & Jelinek 1990; Martin-Hernandez & Hirt 2001; Kelso et al. 2002; Martin-Hernandez & Hirt 2004) or a combination (Rochette & Pillion 1988; Richter & van der Pluijm 1994). Comparison between magnetic fabrics or subfabrics and the results of other techniques such as neutron texture goniometry (Chadima et al, this volume), acoustic waves (Louis et al 2003), or X-ray texture goniometry (Debacker et al., Hirt et al., this volume) allow a better understanding of the nature, origin and significance of measured magnetic fabrics. The separation of superposed magnetic subfabrics can also be achieved through measurements of the anisotropy of magnetic remanence (AMR) alone (reviewed by Potter, this volume). The AMR is a measurement of the anisotropy of ferromagnetic minerals that can be evaluated using different types of magnetization, including anhysteretic remanence (ARM) (McCabe et al. 1985; Jackson et al. 1988; Jackson 1991), isothermal remanence (IRM) (Daly & Zinsser 1973; Stephenson et al. 1986; Borradaile & Dehls 1993; Jelinek 1993), thermoremanence (TRM) (Cogne 1987) and gyroremanence (GRM) (Stephenson 1981). In some cases the ferromagnetic mineral anisotropy can be isolated without recourse to laborious AMR measurements, by analysing the variation of AMS with mean susceptibility (e.g. Henry & Daly 1983; Hamilton et al., this volume). The results from AMS and AMR measurements can be combined in order to effectively compute the different subfabrics (Hrouda 2002). One as-yet poorly understood aspect of AMR is its sensitivity to magnetostatic interactions between neighbouring grains (Muxworthy & Williams, this volume). In samples with weak anisotropy, it has in some cases proven effective to 'enhance' the ferromagnetic fabric (i.e. increase the susceptibility

INTRODUCTION

and/or its anisotropy) by thermal treatment of the samples. In many cases the new ferromagnetic phases that appear while heating mimic the preexisting fabric of the rocks (Dunlop 1974; Henry et al. 2003; de Wall & Warr, this volume). From sedimentary fabrics to tectonic fabrics in sedimentary rocks In the five decades since the seminal paper of Graham (1954), numerous studies have documented a general consistency between magnetic fabric and petrofabric in sedimentary rocks (see Borradaile & Jackson, this volume). This consistency is essentially qualitative, i.e. a parallelism between magnetic fabric axes and sedimentary or tectonic structures or finite strain axes. Additionally, the shape of the magnetic fabric ellipsoid (oblate, triaxial or prolate) often generally corresponds in a qualitative way with the shape of the petrofabric ellipsoid. However, few successful quantitative correlations (involving direct correlation of axial ratios) have been documented. This is due in essence to the variable intrinsic properties of magnetic carriers, the variable proportions of different carriers in different samples, and the bulk (non-mineral-specific) sensitivity of magnetic fabric. In sedimentary rocks, the magnetic foliation results from a combination of depositional processes and diagenetic compaction. The magnetic lineation can result from sedimentary currents in marine conditions or wind in continental condition (see Matasova & Kazansky, this volume). When subjected to strain, the magnetic fabric of sedimentary rocks rapidly starts to record an imprint (see Pares, this volume). Studies of the palaeomagnetic remanence and AMS can help in the understanding of the magnetic fabric acquisition process at early stages of deformation (Larrasoaiia et al., this volume). The magnetic lineation carries the first imprint of shortening or extension. This imprint seems to resist subsequent deformation and can behave therefore as a passive marker. The magnetic foliation is more resistant to early strain. However, the development of magnetic foliation strongly oblique to bedding has been documented in clastic rocks and carbonates during horizontal shortening (the so called layer parallel shortening). Interestingly, this tectonic magnetic foliation is cryptic and it is not necessarily accompanied by its macroscopic equivalent plane in the field (Sun et al. 1993). In low-grade metamorphic rocks (e.g. Anderson & Morris; Aubourg et al.; Debacker et al., this volume), the consistency between magnetic

3

fabric and petrofabric is observed either at the scale of the macroscopic elements (cleavage, lineation) or thin section. Magnetic foliation, if carried by phyllosilicates, is closely parallel to macroscopic cleavage. Magnetic lineation is often the result of microfolding of phyllosilicates and therefore reflects the fold axis or intersection lineation (Hirt et al., this volume). The plungeattitude of magnetic lineation (strike-parallel to down-dip) with respect to magnetic foliation can be a useful indicator of the degree of deformation (Aubourg et al. this volume). In flyschderived metamorphic rocks, Debacker et al. (this volume) suggest that a sedimentary lineation carried by coarse ferromagnetic grains is preserved. While magnetic fabric is generally very well defined in metamorphic rocks, the frequent mixture of several ferromagnetic phases (pyrrhotite, magnetite, hematite) interplaying with paramagnetic phyllosilicates is the cause of complex magnetic fabric (Aubourg et al., this volume), rendering difficult a quantitative characterization of strain. In addition, different deformation mechanisms on the microscale are dominant with increasing strain and strongly influence magnetic - and petrofabric - ellipsoids, hampering correlations with finite strain (Hirt et al., this volume) or macroscopic structures. However, examples exist where, despite a complex tectonic history, the magnetic fabric is rather consistent with macroscopic structures (Ka_dzialko-Hofmokl et al., this volume). Plutonic & igneous fabrics Magnetic fabric analysis is a powerful approach for studying granites because it provides magmatic to strain patterns at a regional scale, in rocks where fabric is difficult to characterize (e.g. Bouchez 2002). We present in this volume a view of the current state of the art in magnetic fabric applications, together with other techniques (Gil-Imaz & Barbero-Gonzalez, Lopez de Luchi et al., Pueyo et al., this volume). Similarly to extensive studies performed in granites, there is a growing interest in magnetic fabric studies of igneous rocks, which allow definition at different scales of the direction and the sense of magma flows (trapp, dyke swarm). Khan (1962) presented the first interpretation of AMS in igneous rocks and dykes and the technique has been used since then in order to understand lava flow and emplacement mechanism of igneous rocks (Hargraves et al. 1991; Tauxe et al. 1998). Canon-Tapia (this volume) provides a historical review of AMS applications in volcanic rocks.

4

F. MARTIN-HERNANDEZ ET AL.

The interpretation of AMS analysis in igneous rocks is based on the development of the preferred orientation of magnetic particles during the flow of material (Canon-Tapia & ChavezAlvarez, this volume). Ferromagnetic minerals, in particular titanomagnetites, are the main carriers of AMS in volcanic rocks, and due to the high intrinsic susceptibility of titanomagnetite with respect to paramagnetic phases, it typically overwhelms the signal from other sources (Tarling & Hrouda 1993; Raposo 1997). Optical observations of the mineral fabrics in lava flows have revealed that the mineral fabric of paramagnetic phases such as plagioclase correlates with the principal directions given by AMS as well (Archanjo et al 2002; Ferre et al 2002; Archanjo & Launeau, this volume). The methodology to retrieve the magma flow has changed considerably during the last decade (Canon-Tapia, this volume). From the early interpretation of magnetic lineation parallel to the flow (Ellwood 1978), it appears that imbrication of magnetic lineations provides in addition the sense of flow (Knight & Walker 1988). To account for the large occurrence of inverse magnetic fabrics in volcanic rocks, Geoffrey et al. (2002) proposed to use in addition the imbrication of magnetic foliation to retrieve the sense of flow. Additionally to the emplacement mechanism, the combination of AMS studies and palaeomagnetic directions provide information on the postemplacement deformation (Petronis et al., this volume). Outlook The papers in this volume collectively portray the current state of research on magnetic fabric analysis, its physical basis and its geological applications. In practice, the terms magnetic fabric and AMS remain nearly synonymous. AMS continues to be by far the most widely applied magnetic method for characterizing fabrics because AMS measurements are fast, nondestructive and extremely precise, enabling regional mapping of sedimentary features, strain and magma flow in almost all rock types. We anticipate that this will continue to be the case in the coming decades, but that more comprehensive methods, including field- and temperature-dependent measurements and remanence anisotropy at higher applied fields, will play an ever-increasing role in magnetic fabric characterization. Together with growing integration of magnetic and nonmagnetic approaches (e.g. neutron and X-ray goniometry, Scanning Electron Microscopy with EBSD),

these more detailed approaches will enable more accurate identification and characterization of the separate components of composite fabrics, their mineral sources and geological significance. On the geophysical side, two of the important research frontiers where we anticipate major advances concern inverse fabrics and magnetostatic interactions. Inverse magnetic fabrics involve a transposition or permutation of principal axes with respect to petrofabric (e.g. magnetic lineation normal to bedding in undeformed sediments), arising from the magnetocrystalline and magnetostatic peculiarities of a relatively small number of magnetic phases (e.g. single-domain magnetite). When 'expected' orientations are unknown, inverse magnetic fabrics cannot be recognized a priori and they provide misleading indicators of petrofabric. Yet it is precisely when 'expected' orientations are unknown that magnetic fabrics have the greatest potential importance, and more reliable methods of establishing mineral PDOs and PCOs from a combination of remanence anisotropy, temperature- and field-dependent AMS measurements will be of considerable value. Similarly, interaction anisotropy (also known as 'distribution anisotropy') is generally understood theoretically and experimentally, but remains a significant complicating factor in some geological applications. An important challenge for the future will be the development of methods for partitioning the AMS of natural materials into components related to PCO, PDO and interactions, The scope of magnetic fabric applications, which already comprises all natural materials, cannot expand, but we may expect a continuing shift from qualitative to more quantitative applications. We would like to thank all the people who have contributed to this volume, authors, reviewers and colleagues. The following people were asked to review one or more of the submitted papers, they are all kindly acknowledged: I. Abad, C. Archanjo, J. Bascou, J. Becker, K. Benn, G. J. Borradaile, J. P. Callot, E. Canon-Tapia, D. Czeck, L. Geoffroy, H. de Wall, B. B. Ellwood, E. C. Ferre, M. Fuller, B. Henry, A. M. Hirt, J. Hodych, B. Housen, M. Hounslow, F. Hrouda, J. Jesek, P. Kelso, K. Kodama, F. Lagroix, P. Launeau, W. Lowrie, M. Mattei, J. M. Miranda, B. Moskowitz, M. Ort, Averbuch, J. M. Pares, B. van der Pluijm, D. Potter, I. Raposo, C. Richter, P. Robion, P. Rochette, L. Sagnotti, S. Siegesmund, M. Sintubin, P. Souquet, S. Spassov, J. Y. Talbot, D. H. Tarling, L. Tauxe, R. Trindade, K. Ullemeyer, X. Wang, T. Werner. We also want to express our sincerely thanks to Angharad Hills and Andy Morton from the Geological Society of London who have been helpful along all the editorial process.

INTRODUCTION

References ARCHANJO, C. J., ARAUJO, M. G. S. & LAUNEAU, P. 2002. Fabric of the Rio Ceara-Mirim mafic dike swarm (northeastern Brazil) determined by anisotropy of magnetic susceptibility and image analysis. Journal of Geophysical Research, 107, 10.1029/2001JB000268. BANERJEE, S. K. & STAGEY, F. D. 1967. The high-field torque-meter method of measuring magnetic anisotropy in rocks. In: COLLINSON, D. W., CREER, K. M. & RUNCORN, S. K. (eds) Methods in Palaeomagnetism. Elsevier, Amsterdam, New York, 470-476. BHATHAL, R. S. 1971. Magnetic anisotropy in rocks. Earth Science Reviews, 7, 227-253. BIN, M. & PAUTHENET, R. 1963. Magnetic anisotropy in pyrrhotite. Journal of Applied Physics, 34, 1161-1162. BORRADAILE, G. J. 2003. Statistics or Earth Science Data, Space and Orientation. Springer, 351 pp. BORRADAILE, G. J. & DEHLS, J. F. 1993. Regional kinematics inferred from magnetic subfabrics in Archean rocks of the Northen Ontario, Canada. Journal of Structural Geology, 15, 887-894. BORRADAILE, G. J. & HENRY, B. 1997. Tectonic applications of magnetic susceptibility and its anisotropy. Earth-Science Reviews, 42, 49-93. BOUCHEZ, J. L. 2002. Magnetic susceptibility anisotropy and fabrics in granites. Earth and Planetary Science Letters, 330, 1-14. COGNE, J.-P. 1987. TRM deviations in anisotropic assemblages of multidomain magnetites. Geophysical Journal of the Royal Astronomical Society, 90, 1013-1023. COLLINSON, D. W. 1967. The design and construction of astatic magnetometers. In: COLLINSON, D. W., CREER, K. M. & RUNCORN, S. K. (eds) Methods in Palaeomagnetism. Elsevier, Amsterdam, New York, 47-59. DALY, L. 1967. Possibilite d'existence dans les roches, de plusieurs anistropies magnetiques superposees et leur separation. Comptes rendus hebdomadaires des seances de P Academic des Sciences (Paris), SerieB, 264, 1377-1380. DALY, L. & ZINSSER, H. 1973. Etude comparative des anisotropies de susceptibilite et d'aimantation remanente isotherme. Consequences pour 1'analyse structurale et le paleomagnetisme. Annales de Geophysique, 29, 189-200. DEUTSCH, E. R., ROY, J. L. & MURTHY, G. S. 1967. An improved astatic magnetometer for paleomagnetism. Canadian Journal of Earth Sciences, 5, 12701273. DUNLOP, D. J. 1974. Thermal enhancement of magnetic susceptibility. Journal of Geophysics, 40, 4339-4351. ELLWOOD, B. B. 1978. Flow and emplacement direction determined for selected basaltic bodies using magnetic susceptibility anisotropy measurements. Earth and Planetary Science Letters, 41, 254— 264. FERRE, E. C., BORDARIER, C. & MARSH, J. S. 2002. Magma flow inferred from AMS fabrics in a

5

layered mafic sill, Insizwa, South Africa. Tectonophysics, 354, 1-23. FUJIWARA, Y. & YOSHIDA, M. 1981. An automatic astatic magnetometer for paleomagnetic studies. Journal of the Faculty of Science, Hokkaido University, Series IV, 19, 519-526. FULLER, M. D. 1960. Anisotropy of susceptibility and the natural remanent magnetization of some Welsh slates. Nature, 186, 791-792. FULLER, M. D. 1963. Magnetic anisotropy and paleomagnetism. Journal of Geophysical Research, 68, 293-309. GEOFFROY, L., CALLOT, J. P., AUBOURG, C. & MOREIRA, M. 2002. Magnetic and plagioclase linear fabric discrepancy in dykes: a new way to define the flow vector using magnetic foliation. Terra Nova, 14, 183-190. GIRDLER, R. W. 1961. The measurement and computation of anisotropy of magnetic susceptibility in rocks. Geophysical Journal of the Royal Astronomical Society, 5, 34-44. GRAHAM, J. W. 1954. Magnetic susceptibility anisotropy, an unexploited petrofabric element. Bulletin of the Geological Society of America, 65, 12571258. GRAHAM, J. W. 1964. Preliminary account of a refine technique for magnetic susceptiblity anisotropy measurements of rocks. In: COLLINSON, D. W., CREER, K. M. & RUNCORN, S. K. (eds) Methods in Palaeomagnetism. Elsevier, Amsterdam, New York, 409-424. GRAHAM, J. W. 1966. Significance of magnetic anisotropy in Appalachian sedimentary rocks. In: STEINHART, J. S. & SMITH, T. J. (eds) The Earth Beneath the Continents. American Geophysical Union, Geophysical Monograph Series, Washington, 627-648. GRANAR, L. 1958. Magnetic measurements on Swedish varved sediments. Arkiv for Geofysik, 3, 1-40. HARGRAVES, R. B., JOHNSON, D. & CHAN, C. Y. 1991. Distribution anisotropy: the cause of AMS in igneous rocks? Geophysics Research Letters, 18, 2193-2196. HENRY, B. & DALY, L. 1983. From qualitative to quantitative magnetic anisotropy analysis: the prospect of finite strain calibration. Tectonophysics, 98, 327-336. HENRY, B., JORDANOVA, D., JORDANOVA, N., SOUQUE, C. & ROBION, P. 2003. Anisotropy of magnetic susceptibility of heated rocks. Tectonophysics, 366, 241-258. HROUDA, F. 1982. Magnetic anisotropy of rocks and its application in geology and geophysics. Geophysical Surveys, 5, 37-82. HROUDA, F. 2002. The use of the anisotropy of magnetic remanence in the resolution of the anisotropy of magnetic susceptibility into its ferromagnetic and paramagnetic components. Tectonophysics, 347,269-281. HROUDA, F. & JELINEK, V. 1990. Resolution of ferrimagnetic and paramagnetic anisotropies in rocks, using combined low-field and high-field measurements. Geophysical Journal International, 103, 75-84.

6

F. MARTIN-HERNANDEZ ET AL.

HROUDA, F., JELINEK, V. & ZAPLETAL, K. 1997. Refined technique for susceptibility resolution into ferromagnetic and paramagnetic components based on susceptibility temperature-variation measurement. Geophysical Journal International, 129, 715-719. ISING, G. 1943. On the magnetic properties of varved clay. Arkiv for Matematik, Astronomi och Fysik Bd., 29A, 1-37. JACKSON, M. J. 1991. Anisotropy of magnetic remanence: a brief review of mineralogical sources, physical origins, and geological applications, and comparison with susceptibility anisotropy. Pure and Applied Geophysics, 136, 1-28. JACKSON, M. J., GRUBER, W., MARVIN, J. A. & BANERJEE, S. K. 1988. Partial anhysteretic remanence and its anisotropy: applications and grainsize dependence. Geophysics Research Letters, 15, 440-443. JELINEK, V. 1966. A high sensitivity spinner magnetometer. Studia Geophysica et Geodetica, 10, 58-77. JELINEK, V. 1973. Precision A.C. bridge set for measuring magnetic susceptibility of rocks and its anisotropy. Studia Geophysica et Geodetica, 17, 36-48. JELINEK, V. 1993. Theory and measurement of the anisotropy of isothermal remanent magnetization of rocks. Travaux Geophysique, 37, 124-134. JOHNSON, E. A. & STEINER, W. F. 1937. An astatic magnetometer for measuring susceptibility. Review of Scientific Instruments, 8, 236-238. KELSO, P. R., TIKOFF, B., JACKSON, M. & SUN, W. 2002. A new method for the separation of paramagnetic and ferromagentic susceptibility anisotropy using low field and high field methods. Geophysical Journal International, 151, 345-359. KHAN, M. A. 1962. Anisotropy of magnetic susceptibility of some igneous and metamorphic rocks. Journal of Geophysical Research, 67, 28732885. KING, R. F. & REES, A. I. 1966. Detrital magnetism in sediments: an examination of some theoretical models. Journal of Geophysical Research, 71, 561-571. KNIGHT, M. D. & WALKER, G. P. L. 1988. Magma flow directions in dikes of the Koolau Complex, Oahu, determined from magnetic fabric studies. Journal of Geophysical Research, 93, 4301-4319. KODAMA, K. P. 1995. Magnetic fabrics. Reviews of Geophysics, 33, supplement (IUGG Report). LIN, S. T. 1959. Magnetic properties of hematite single crystals. 1. Magnetization isotherms, antiferromagnetic susceptibility, and weak ferromagnetism of a natural crystal. Physical Review, 116, 14471452. Louis, L., DAVID, C. & ROBION, P. 2003. Comparison of the anisotropic behaviour of undeformed sandstones under dry and saturated conditions. Tectonophysics, 370, 193-212. LUNEBURG, C. M., LAMPERT, S. A., LEBIT, H. K., HIRT, A. M., CASEY, M. & LOWRIE, W. 1999. Magnetic anisotropy, rock fabrics and finite strain in deformed sediments of SW Sardinia (Italy). Tectonophysics, 307, 51-74.

MACDONALD, W. D. & ELLWOOD, B. B. 1987. Anisotropy of magnetic susceptibility: sedimentological, igneous, and structural-tectonic applications. Reviews of Geophysics, 25, 905-909. MARTIN-HERNANDEZ, F. & HIRT, A. M. 2001. Separation of ferrimagnetic and paramagnetic anisotropies using a high-field torsion magnetometer. Tectonophysics, 337, 209-221. MARTIN-HERNANDEZ, F. & HIRT, A. M. 2004. A method for the separation of paramagnetic, ferrimagnetic and hematite magnetic subfabrics using high-field torque magnetometer. Geophysical Journal International, 157, 117-127. MCCABE, C., JACKSON, M. & ELLWOOD, B. B. 1985. Magnetic anisotropy in the Trenton limestone: results of a new technique, anisotropy of anhysteretic susceptibility. Geophysics Research Letters, 12, 333-336. MIKAMI, L, HIRONE, T., WATANABE, H., MAEDA, S., ADACHI, K. & YAMADA, M. 1959. On the magnetic anisotropy of a pyrrhotite crystal. Journal of the Physical Society of Japan, 14, 1568-1572. NAGATA, T. 1961. Rock Magnetism. Maruzen, Tokyo. NOLTIMIER, H. C. 1971. Determining magnetic anisotropy of rocks with a spinner magnetometer giving in-phase and quadrature data output. Journal of Geophysical Research, 76, 4849-4854. OWENS, W. H. 20000. Error estimates in the measurement of anisotropic magnetic susceptibility. Geophysical Journal International, 142, 516526. OWENS, W. H. 20006. Statistical applications to second-rank tensors in magnetic fabric analysis. Geophysical Journal International, 142, 527-538. PARES, J. M. & VAN DER PLUIJM, B. A. 2002. Phyllosilicate fabric characterization by Low-Temperature Anisotropy of Magnetic Susceptibility (LT-AMS). Geophysics Research Letters, 29, art. no.-2215. RAPOSO, M. I. B. 1997. Magnetic fabric and its significance in the Florianopolis dyke swarm, southern Brazil. Geophysical Journal International, 131, 159-70. REES, A. I. 1961. The effect of water currents on the magnetic remanence and anisotropy of susceptibility of some sediments. Geophysical Journal of the Royal Astronomical Society, 5, 235-251. REES, A. I. 1965. The use of anisotropy of magnetic susceptibility in the estimation of sedimentary fabric. Sedimentology, 4, 257-271. RICHTER, C. & VAN DER PLUIJM, B. A. 1994. Separation of paramagnetic and ferrimagnetic susceptibilities using low temperature magnetic susceptibilities and comparison with high field methods. Physics of the Earth and Planetary Interiors,^ 111-121. ROCHETTE, P. & PILLION, G. 1988. Identification of multicomponent anisotropies in rocks using various field and temperature values in a cryogenic magnetometer. Physics of the Earth and Planetary Interiors, 51, 379-386. ROCHETTE, P., JACKSON, M. & AUBOURG, C. 1992. Rock magnetism and the interpretation of anisotropy of magnetic susceptibility. Review of Geophysics, 30, 209-226.

INTRODUCTION ROY, R. L. 1971. The use of negative feedback with astatic magnetometers for paleomagnetic studies. Canadian Journal of Earth Sciences, 8, 1595-1597. SATO, K., YAMADA, M. & HIRONE, T. 1964. Magnetocrystalline anisotropy of pyrrhotite. Journal of the Physical Society of Japan, 19, 1592-1595. SCHMIDT, V. A., ELLWOOD, B. B., NAGATA, T. & NOLTIMIER, H. C. 1988. The measurement of anisotropy of magnetic susceptibility using a cryogenic (SQUID) magnetometer and a comparison with results obtained from a torsion-fiber magnetometer. Physics of the Earth and Planetary Interiors, 51, 365-378. STAGEY, F. D. 1960. Magnetic anisotropy of igneous rock. Journal of Geophysical Research, 65, 24292442. STAGEY, F. D. 1963. The physical theory of rock magnetism. Advances in Physics, 12, 45-133. STAGEY, F. D., JOPLIN, G. & LINDSAY, J. 1960. Magnetic anisotropy and fabric of some foliated rocks from S.E. Australia. Geofisica pur a e applicata, 47, 30-40. STEPHENSON, A. 1981. Gyromagnetic remanence and anisotropy in single-domain particles, rocks, and magnetic recording tape. Philosophical Magazine B, 44, 635-664.

7

STEPHENSON, A., SADIKUN, S. & POTTER, D. K. 1986. A theoretical and experimental comparison of the anisotropies of magnetic susceptibility and remanence in rocks and minerals. Geophysical Journal of the Royal Astronomical Society, 84, 185-200. STONE, D. B. 1962. Anisotropic magnetic susceptibility measurements on a phonolite and on a folded metamorphic rock. Geophysics, 62, 375-380. SUN, W.-W., JACKSON, M. J. & CRADDOCK, J. P. 1993. Relationship between remagnetization, magnetic fabric, and deformation in the midcontinental Paleozoic carbonates. Tectonophysics, 221, 361— 366. TARLING, D. H. & HROUDA, F. 1993. The Magnetic Anisotropy of Rocks. Chapman & Hall, London. TAUXE, L., GEE, J. S. & STAUDIGEL, H. 1998. Flow directions in dikes from anisotropy of magnetic susceptibility data: the bootstrap way. Journal of Geophysical Research, 103, 17775-90. TOWNSEND, T. 1916. The magnetic properties of hematite. Review of Geophysics, 8, 721-737. TOWNSEND, T. 1920. Magnetization and hysteresis in hematite crystals. Physical Review, 15, 345-364. UYEDA, S., FULLER, M. D., BELSHE, J. C. & GIRDLER, R. W. 1963. Anisotropy of magnetic susceptibility of rocks and minerals. Journal of Geophysical Research, 68, 279-291.

This page intentionally left blank

Determination of the orientation of magnetic minerals from the anisotropy of magnetic susceptibility J. JEZEK1 & FRANTISEK HROUDA1 2 1

Faculty of Sciences, Praha, Czech Republic (e-mail: [email protected]) 2 AGICO Ltd., Brno, Czech Republic (e-mail: [email protected])

Abstract: The anisotropy of magnetic susceptibility (AMS) of rocks is controlled by preferentially oriented magnetic mineral grains that carry AMS and, therefore, it contains information about both the grain susceptibilities and the grain orientations. Under certain conditions, information about the grain orientations can be deduced from the AMS. For a multigrain system composed of identical grains that are magnetically uniaxial (for the grain principal susceptibilities it holds that K\ > K2 = K3, or K\ = K2 > £3), an exact relationship exists between the AMS and the orientation tensor. We investigate the extent to which the theoretical relationships can be used when grains are generally triaxial. The parallelism of the principal directions of the susceptibility tensor and those of the orientation tensor are well preserved in all basic grain configurations. If grain leading axes have polar or girdle distributions and the two other axes have balanced distributions (similar orientation tensors), the parameters of intensity /and shape Abased on the eigenvalues of the orientation tensor are well estimated. For unbalanced distributions, formulas are found for possible errors of 7 and T estimates.

Preferred orientation of minerals in rocks develops during various geological processes, such as water flow in sediments, lava or magma flow in volcanic and plutonic rocks, and ductile deformation in metamorphic rocks, and these processes can in turn be assessed from it. Also magnetic minerals, mostly occurring in rocks in accessory amounts, show preferred orientation and this can be advantageously investigated by means of the anisotropy of magnetic susceptibility (AMS). Modern instruments for measuring the AMS are sensitive enough to be able to measure almost all rock types with sufficient accuracy. In some geological interpretations and in mathematical modelling geological processes, precise quantitative relationship between the magnetic mineral preferred orientation and the AMS is needed. In theory, the normalized bulk magnetic susceptibility tensor is considered as a sum of oriented magnetic grains

magnetic grains are equal, having the same grain susceptibility tensor K, then equation (1) may be written as

(i)

is equal to the orientation tensor computed from the x axes, and similarly the second and third terms correspond to the orientation tensors of y and x axes, respectively. Therefore,

where O is the grain orientation matrix and K is the tensor of grain susceptibility

In this equation, the products of orientation matrices represent orientation tensors (Scheidegger 1965) of grain magnetic axes x, y, z. For example, in the first term on the right-hand side, the sum

(2)

with the principal grain susceptibilities normalized to KI -f K2 4- KI = 3 (we use bold symbols for matrices and italic for scalars). When all

where Tx, T7, Tz are the orientation tensors of grain magnetic axes x, y, z. In this way, the bulk magnetic susceptibility contains information about the orientation tensors of individual grain axes. Further simplification of the relation (2) is possible for magnetically uniaxial grains.

From: MARTIN-HERNANDEZ, F., LUNEBURG, C. M., AUBOURG, C. & JACKSON, M. (eds) 2004. Magnetic Fabric: Methods and Applications. Geological Society, London, Special Publications, 238, 9-20. 0305-8719/04/S15.00 © The Geological Society of London 2004.

10

J. JEZEK & F. HROUDA

When prolate grains are perfectly uniaxial with K2 = K3 = K and KI - K = A > 0 then

will be described by grain magnetic parameters Pg and C/g,

(3)

where I is unity matrix. It means that for this case we can exactly estimate the orientation tensor T^ from susceptibility (Jezek & Hrouda 2000). Similarly, for magnetically oblate grains that are perfectly uniaxial with K{ = K2 = K and K - K3 = A > 0, we obtain

(2) (3)

(4)

and the orientation tensor Tz can be exactly estimated from the susceptibility. Finally, in the case of'perfectly' triaxial grains, or 'ideal' triaxial grains, which we define by K! = K + A, K2 = K, K3=K-&, A > 0, it holds (5)

where T0 is the Lisle (1989) orientation tensor, T0 = Tx — Tz, which differs from the Scheidegger (1965) orientation tensor used above. For perfectly triaxial grains, this orientation tensor can be exactly estimated from the susceptibility (Jezek & Hrouda 2000). In reality, magnetic carriers in rocks are neither identical nor perfectly uniaxial or perfectly triaxial. However, the equations (3), (4) and (5) can in fact be used in cases when the assumptions are not exactly fulfilled. In the following text we show how deviations from perfect uniaxiality influence the estimates of orientation tensors from equations (3) and (4). Cases with triaxial grains will not be treated. They were treated partially by Jezek & Hrouda (2000) and we will deal with them in another publication. The results of the present study address domains where AMS is used as indicator of preferred orientation in rocks (see, e.g. Cogne 1988; Cogne & Perroud 1988; Henry 1989; Hrouda & Schulmann 1989; Borradaile & Henry 1997; Bouchez 2000; Jezek & Hrouda 2002).

(4) (5)

(6)

(7)

We prescribe a type of parameters distribution for these parameters. We choose a type and intensity of angular distribution for grain orientations. From this angular distribution, we generate a random sample of size ng (i.e. ng differently oriented grains) and compute the corresponding orientation tensor of long axis (T x ). This 'true' orientation tensor will be estimated from the bulk magnetic susceptibility. From the prescribed parameters distribution, we generate a random sample of size Np of grain magnetic parameters Pg and Ug. For each of the Np simulations, we compute bulk susceptibilities by equation (1) and estimate orientation tensors using equation (3): (a) from Pg and t/g we find the normalized grain susceptibilities K{, K2, K3, (b) we put K=(K2 + K3)/2 and A = KI -K, (c) from equation (1) we compute the bulk susceptibility k, (d) and from equation (3) the estimate Txe = (k - Kl)/A. We repeat the steps 3 to 5 N0 times, so that we have in total N0 x Np simulations of similar orientations and grain properties, for which we know the true orientation tensors and their estimates. We compare the true orientation tensors with their estimates, to show the dispersion caused by variability of angular distributions and grain magnetic properties. We do so by comparing the directions of eigenvectors and parameters of intensity 7 and symmetry T:

(6)

Insight into the problem by numerical modelling To gain an insight into this problem we first show some examples of numerical modelling. Figures la to Id are results of the same simulation procedure: (1) We consider a system composed of similar magnetic grains. The grains are magnetically prolate, not perfectly uniaxial, with x being the longest axis. Their magnetic properties

where eh i = 1, 2, 3 are the orientation tensor eigenvalues. The reason for the use of two different forms of the symmetry parameter will be explained later. The intensity parameter I is Lisle (1985) parameter used as a measure of fabric strength. It attains values from 0 (isotropic fabric) to 5 (all grains aligned parallel to one direction). The

ORIENTATION OF MAGNETIC MINERALS FROM AMS

symmetry parameters T and T* span from -1 (constriction type of fabric, e{ > e2 — £3) to 1 (flattening type of fabric, e\ = e2 > e3). For the set of simulations in Figure 1, we used Marchian distributions corresponding to constriction, axial flattening and plane strain. For the parameters Pg and Us we used Gaussian distribution (for C/g only right half of the Gaussian distribution was used). In the first row of Figures la to Id we always present the first generated angular distribution and histograms of the first generated distribution of grain magnetic properties. The shown types of distributions were used in all simulations in these figures. In the second row of each figure, we summarize simulation results in graphs of true versus estimated values of intensity / and symmetry T (SOT-X implies Scheidegger orientation tensor of x axes, TA.). The third row of Figures la to Id shows the angular errors of estimation of the first and third eigenvectors of the orientation tensors. From the simulations, we observe only small errors in estimating the parameters / and T, and especially in the orientation of the leading eigenvectors. In case of constriction-type angular distribution (polar distribution of x axes; Fig. la) we catch the first eigenvector (lineation) with an error of about 1°. A similar result is found for flattening-type angular distribution (girdle of x axes) and its third eigenvector (pole to foliation, Fig. Ib). In the case of plane strain-type of angular distribution, all eigenvectors are well estimated (Fig. Ic). Similar results were obtained by simulations using equation (4) where we estimated orientation tensors of short axes of oblate magnetic grains. To assess the variability caused by a smaller sample size, we show in Figure Id a situation comparable to Figure la except that it has ten times fewer grains (ng = 100 vs 1 000). Theoretical considerations We now try to explain the above simulation results and to find some limits for errors of orientation tensors estimates caused by deviations from perfect uniaxiality. First, we use the fact that due to the orthogonality of grain axes, it holds that (7)

Further, we restrict the considered grain axis orientations to polar distributions with circular or elliptic symmetry of the probability density function and girdles compatible with them. This is not a very restricting choice - it includes,

11

for example, all Fisherian, Binghamian and Marchian distributions. Such choice has the advantage that the statistical mean direction (expectance) of first eigenvectors of the orientation tensors TX9 T^, Tz defines a coordinate system (COx) in which all of these orientation tensors are diagonal. This property is based on the fact that the direction of minimum concentration of x axes represents the direction of maximum concentration of z axes and vice-versa. This is strictly valid asymptotically, i.e. for an infinite number of grains. When the number of grains is finite, the off-diagonal terms of orientation tensors in the coordinate system CQT are non-zero. However, they would be small and decrease as the number of grains increases, while the sum of diagonal terms remains equal to one. As the eigenvectors of the orientation tensors Tx, T^, Tz estimated from a finite number of grains are not orthogonal, the Lisle (1989) orientation tensor eigenvectors can be used to define the coordinate system COT. In Table 1 we list the median and maximum offdiagonal terms based on 10, 100 and 1 000 simulations for three angular distribution types: (a) polar distribution with circular symmetry of both x and z axes, (b) polar distribution of x axes with circular symmetry and perfect girdle of z axes (or vice-versa), and (c) intermediate between cases (a) and (b). From the point of view of our investigation, the distributions (a) and (b) are end-members of considered distributions. In estimating orientation tensors from bulk susceptibility by equations (3) and (4), there are two potential sources of errors. The first is the coincidence or superposition of effects of angular and magnetic parameters distributions. This error can be estimated for multigrain systems composed of a large number of grains. We call them large-sample errors, and we will deal with them in the following paragraphs. The second cause is finite sample size (small statistic sample variability). This error is hard to assess without knowledge or assumptions about type of distributions and related parameters, and, therefore, we shall not treat it.

Imperfect prolate uniaxial grains We now investigate prolate uniaxial grains that deviate from perfect uniaxiality. Consider an imperfect prolate uniaxial grain, KI = K + A, K2 = K + £, K3 = K — £, where e is the deviation from perfect uniaxiality. The range of possible deviations e spans over an interval 0 < s < A, where e = 0 represents perfectly prolate grain,

12

J. JEZEK & F. HROUDA

Fig. 1. Estimating of orientation tensors from magnetic susceptibility. Parameters /, T and the first and third eigenvectors estimated from a Marchian-type distribution of 1000 grains corresponding to (a) constriction, (b) axial flattening and (c) plane strain; (d) is a constriction for 100 grains. See text for details.

ORIENTATION OF MAGNETIC MINERALS FROM AMS

Fig. 1. (continued)

13

14

J. JEZEK & F. HROUDA

Table 1. An analysis of the off-diagonal terms of the orientation tensors Tx, Ty andTz in the coordinate system COT Number of grains ng

Median (and largest) values of off-diagonal terms of Tx, T,, Tz (a) polar x and y axes

(b) polar x, girdle z

(c) both polar, z less concentrated

10 100 1000

0.1131 (0.2970) 0.0429(0.1414) 0.0134(0.0422)

0.0986 (0.2607) 0.0329 (0.0928) 0.0119(0.0370)

0.0963 (0.2798) 0.0324 (0.0965) 0.0101 (0.0269)

(a) Fisherian distribution of both x and z axes (Fisher's concentration parameter of both axes qx = qz — 3), (b) Fisherian distribution of x axes and girdle of z axes (qx = 3, qz ~ 0), (c) Fisherian distribution of both axes, z axes less concentrated (qx = 3, qz = 1).

and e = A represents a perfectly oblate grain. From equation (2) it follows (8)

Balanced distributions of y and z axes Equation (8) suggests that estimation errors would be small when the orientation tensors of the y and z axes are similar. Assume that the long x axis is the most concentrated and the distributions of short y and z axes are balanced, as was the case for the Marchian distributions of prolate uniaxial grains in Figure 1. Then in the coordinate system COx, the last two terms ATX + e(ly - T z ) on the right-hand side of equation (8) have asymptotically (for an infinite number of grains) a form

(9)

where c + d + e = 1. Therefore, for an infinite number of grains, the term in brackets in equation (9) and the term in parentheses in equation (8) are zeros. For a finite number of grains, there are small non-zero off-diagonal values that decrease with increasing number of grains (see Table 1), and the right-hand term in brackets should be almost a zero-matrix, and, therefore, negligible. The error made by estimating T^ by equation (3) should be small and almost independent of the degree of deviation from perfect

uniaxiality e. In Figure 2, we illustrate this property by simulating the previous cases of constriction, plane strain and flattening type distributions. For every angular distribution, we consider identical magnetic grains, all having the same Pg and £7g parameters. We fix Pg and change Ug through all possible values, £7g = —7, -0.8,..., 1. This might correspond to a situation where the true susceptibility is produced by imperfect uniaxial grains with C7g > — 1 but we neglect this fact and estimate the orientation tensor of long axes by equation (3). Formally, we can go through all possible values of C/g. By increasing C/g we change sequentially the grain shape from perfect prolate to perfect oblate (when Ug = 1). In each figure, we show again the type of angular distribution. In graphs in the second row, points show true values and circles show estimates of the 7 and T parameters. We conclude that the simulations confirm our expectations. In the case of balanced distributions of y and z axes the errors of 7 and T estimates are small and almost independent of the degree of deviation from perfect uniaxiality. (Larger numbers of simulations were computed but we present smaller samples to avoid having too many overlapping points in the graphs.) We have not yet dealt with grain geometry (shape): the x, y and z axes represented magnetic axes. The question now becomes how to attribute a considered angular distribution of magnetic axes to a distribution of grain geometric axes? The simplest assumption is that the magnetic and geometric axes are parallel. Then our interpretation of balanced distribution of magnetic y and z axes might be that the geometrical short axes were equal or the process that reoriented grains was not sensitive to small differences in axis lengths. Polar distribution of both x and z-axes Assume now that the angular distribution of both x and z magnetic axes is polar and the

ORIENTATION OF MAGNETIC MINERALS FROM AMS

15

Fig. 2. The influence on orientation tensors estimation caused by deviations from perfect grain uniaxiality. We consider Marchian-type distributions with balanced distributions of y and z axes as in Figure 1 and identical magnetic grains.

concentration of z-axes is equal to the concentration of x-axes. When estimating possible errors caused by neglecting deviations from perfect uniaxiality, such a distribution may serve as a

boundary case because a sum of many small differences in principal susceptibilities along the y and z axes may significantly influence the estimate of the leading x axis orientation tensor

16

J. JEZEK & F. HROUDA

Fig. 2. (continued}

TV Such an angular distribution may correspond to a more complicated tectonic history and/or significant influence of real grain shapes on the reorienting process. In Figure 3 we show simulation results that can be compared with those in Figure 2. It is evident that the errors of / and T

estimates are larger due to the simultaneous effect of grain asymmetry and orientation, while the estimates of the first eigenvector remain very good. To understand this behaviour we look again at asymptotic forms. In the coordinate system COT> the last two terms

ORIENTATION OF MAGNETIC MINERALS FROM AMS

17

Fig. 3. For unbalanced distributions of y and z axes (here a polar distribution of both x and z axes) the errors of / and T estimates can be larger due to the simultaneous effect of grain asymmetry and orientation, while the estimates of the first eigenvector remain very good. Compare to Figure 2.

AT^ + £(Ty - T z ) on the right-hand side of equation (8) would have an asymptotical form

by means of parameter t/g, er = (1 4- t/ g )/ (3 - t/g). It makes no sense to consider cases where the third eigenvalue on the right-hand side of equation (10) is negative or zero, so we can focus only on cases where the third eigenvalue is positive, which is given by the condition 0 < £r < (1 - c)/(3c - 1). Another possibility (a practical choice) is to replace a non-positive eigenvalue in equation (10) by a very small positive value and to continue estimating. We adopted this approach and such cases in Figure 3 are indicated by grey circles. From equation (10), the relative error of the estimate of parameter / can be expressed by a surprisingly simple formula,

(ii) (10) where c is the first eigenvalue of the orientation tensor. It represents concentration of x axes and lies in the interval 1/3 < c < 1, with c = 1 for perfect orientation along a line and c = 1/3 for isotropic orientation (no preferred orientation). We denote £r = e/A, the relative degree of deviation from perfect uniaxiality, with values 0 < er < 1. The value er = 0 means perfectly uniaxial prolate, £r = 1/3 perfectly triaxial, er = 1 perfectly uniaxial oblate. For prolate grains, this parameter can be expressed

Therefore, for this particular distribution type, the error of the parameter 7 is independent of the concentration parameter c. This is not true for the parameter T, where the estimate of its relative error is

(12) Due to the fact that the theoretical value of T — — 1, the relative error is equal to — Te. As equation (11) is rather complicated and contains

18

J. JEZEK & F. HROUDA

Fig. 4. Simulations versus estimates of the relative errors of / and T* for a polar distribution of both x and z axes. Lower graphs: points - simulated values of the relative errors, black crosses - mean values of all simulations in column, black curves - the estimates, equations (11) and (13), respectively.

the concentration parameter c, we use the parameter T*, for which

20% lower value for the symmetry of the distribution.

(13)

Girdle of x axes and a polar distribution of z axes A distribution composed of a girdle of x axes and polar concentration of z axes is the second important boundary case for estimating possible errors caused by neglecting deviations from perfect uniaxiality. Like the previous distribution, it may correspond to a more complicated tectonic history and influence of real grain shapes on the reorienting process. In the coordinate system COT? the last two terms on the right-hand side of equation (8) have an asymptotic form

Equations (11) and (13) represent approximations of the errors for the estimates of parameters / and r* from bulk susceptibility. They were found for perfectly polar distributions of both x and y axes. Figure 4 shows results of simulations confirming our estimates of relative errors of 7 and T*. In the lower left-hand graph, the points correspond to the relative error of the / parameter estimate, black crosses show the mean value of all simulations for given C/g, and the black curve is the theoretical error corresponding to equation (11). Grey points show cases where we replaced a negative value for the third eigenvalue of (10) by a small positive number. Formula (11) reflects well the trend of increasing errors of the / estimate. In the lower right-hand graph of Figure 4 we see a similar situation for T* and error equation (13). The practical conclusion for the considered angular distribution is that the large-sample error of / estimate is negligible. For instance, if t/g = -0.8, which means er = 0.0526, Ie/I = 1.0028. This error will be overridden by finite sample variability. This is not true for T* estimate, where T*/T* = 0.8, and we estimate

(14)

where c is again the largest eigenvalue of the orientation tensor T x . As above, we can omit

ORIENTATION OF MAGNETIC MINERALS FROM AMS

19

Fig. 5. Simulations versus estimates of the relative errors of / and T for a girdle of x axes and a polar distribution of z axes. The black curve for /corresponds to equation (15), and for the shape parameter Tf = T=l.

cases where the third eigenvalue on the righthand side of (14) is negative or zero and restrict our considerations to 0 < er < (1 — c)/(3c— 1), or we can set the third eigenvalue to a small positive number. From equation (14), the relative error of the estimate of the / parameter for a girdle of x axes plus a polar distribution of z axes can be expressed by a formula (15)

For this particular distribution type, the error is independent of the concentration parameter c. For instance, if C/g = -0.8 then Ie/I = 1.2230. Asymptotically, the shape parameter Te = T = 1 and the only source of errors in estimating this parameter is small sample variability. In Figure 5 we show results of simulations confirming our estimates of the relative errors of /. In the lower left-hand graph, points correspond to the relative error of the parameter / estimate, black crosses show the mean value of all simulations for given t/g, and the black curve is the theoretical error, equation (15). Grey points show cases where we replaced negative third eigenvalues in equation (14) with a small positive number.

Imperfect oblate uniaxial grains After investigating prolate grains in detail it is straightforward, due to symmetry of our problem, to obtain results for oblate grains. We do not repeat here the mathematical treatment. The results mirror previous results for prolate grains. The only formal difference is that for oblate grains, the relative deviation from perfect uniaxiality is defined as er = (1 - C7g)/ (3 + C/g). Conclusions In this study, we consider multigrain systems composed of either identical or similar magnetic grains that deviate from perfect uniaxiality. Typical grains have one leading axis (the longest for prolate and the shortest for oblate grains) and two axes of similar but unequal length (we shall call them similar axes). We studied how a simple procedure of estimating the orientation tensor from the bulk susceptibility tensor can be used for such grains. The estimate of the orientation tensor was represented by estimates of the intensity parameter /, the shape parameter T and the estimates of eigenvectors. The main

J. JEZEK & F. HROUDA

20

results are: For any kind of grain orientation (preferred orientation), the most stable estimated characteristics are the eigenvectors of the orientation tensor. We therefore confirm the correctness of using magnetic susceptibility as a directional indicator. When the similar grain axes have a balanced orientation distribution (similar orientation tensors), the errors of estimating / and T are asymptotically zero, i.e. they would be negligible for a large number of magnetic grains in a measured rock sample. The simulated errors reflect a variance that decreases with the number of grains and are acceptable for

wg - 1 000. We found two orientation distributions which represents the worst-possible configuration for estimating of the orientation tensor. For these distributions, the average large-sample relative errors of T* and 7 can be approximated by equations (13) and (15). For other distributions the errors of estimating should be lower. Thus for a rock dominated by one magnetic mineral, the presented estimation of the orientation tensor of this mineral leading axis provides an acceptable estimate together with an estimate of large-sample errors. The research was supported by the Grant Agency of the Czech Republic, project 205/03/0336. We thank reviewers Massimo Mattei and Philippe Robion for their constructive comments.

References BORRADAILE, G. J. & HENRY, B. 1997. Tectonic applications of magnetic susceptibility and its anisotropy. Earth-Science Reviews, 42, 49-93. BOUCHEZ, J. L. 2000. Anisotropie de susceptibilite magnetique et fabrique des granites. Comptes Rendus de I'Academic de Science de Paris, Earth and Planetary Science, 330, 1-14. COGNE, J. P. 1988. Strain, magnetic fabric and paleomagnetism of deformed redbeds of the Ordovician Pont-Rean formation (Brittany, France). Journal of Geophysical Research, 93, 13673-13687. COGNE, J. P. & PERROUD, H. 1988. Anisotropy of magnetic susceptibility as a strain gauge in the Flamaville granite, NW France. Physics of the Earth and Planetary Interiors, 51, 264-270. HENRY, B. 1989. Magnetic fabric and orientation tensor of minerals in rocks. Tectonophysics, 165, 21-27. HROUDA, F. & SCHULMANN, K. 1989. Conversion of magnetic susceptibility tensor into orientation tensor in some rocks. Physics of the Earth and Planetary Interiors, 63, 71-77. LISLE, R. J. 1985. The use of the orientation tensor for the description and statistical testing of fabrics. Journal of Structural Geology, 7, 115-117. LISLE, R. J. 1989. The statistical analysis of orthogonal orientation data. Journal of Geology, 97, 360-364. JEZEK, J. & HROUDA, F. 2000. The relationship between the Lisle orientation tensor and the susceptibility tensor. Physics and Chemistry of the Earth, 25, 469-474. JEZEK, J. & HROUDA, F. 2002. Software for modeling the magnetic anisotropy of strained rocks. Computers & Geosciences, 28, 1061-1068. SCHEIDEGGER, A. E. 1965. On the statistics of the orientation of bedding planes, grain axes, and similar sedimentological data. U.S. Geological Survey, Professional Paper 525-C, 164-167.

A comparison of anisotropy of magnetic remanence methods - a user's guide for application to palaeomagnetism and magnetic fabric studies DAVID K. POTTER Centre for Geophysical and Petrophysteal Magnetism, Institute of Petroleum Engineering, Heriot-Watt University, Edinburgh, EH 14 4 AS, United Kingdom (e-mail: [email protected]) Abstract: Anisotropy of magnetic remanence (AMR) is increasingly being applied to palaeomagnetic and structural fabric studies. AMR techniques measure the anisotropy of the remanence carrying particles, and thus are directly relevant to palaeomagnetic studies concerned with computing the direction and intensity of the Earth's ancient magnetic field from the natural remanent magnetization (NRM) recorded in anisotropic rocks. This paper provides a comparison of several AMR methods, including some of the less wellknown techniques, and highlights the relative merits of each. Results from a strongly anisotropic rock and a pottery sherd are presented. The anisotropies of anhysteretic remanent magnetization (AARM) and isothermal remanent magnetization (AIRM) are currently the most commonly applied types of AMR, since they have provided reasonably good analogues of the anisotropy of thermoremanent magnetization (ATRM) acquired in the Earth's field. They have also helped to correct for inclination shallowing of detrital remanent magnetization (DRM) in sediments. IRM anisotropy is the most rapid AMR technique, and is particularly useful for very low concentrations of remanence carrying particles. The gyroremanences, gyroremanent magnetization (GRM) and rotational remanent magnetization (RRM), are preferentially acquired by stable single-domain (SD) particles, and are thus directly relevant to the particles of major interest in palaeomagnetism. GRM anisotropy is the most sensitive AMR method. It is essentially the remanence equivalent of the anisotropy of magnetic susceptibility (AMS) delineator, since a single application of an alternating field will only produce a GRM in a sample containing an anisotropic distribution of particles. Static ARM methods need to take account of components of GRM that are simultaneously acquired.

Anisotropy of magnetic susceptibility (AMS) has been used for many years because it is very rapid, sensitive and non-destructive. A full 3-dimensional AMS ellipsoid can be acquired in about a minute with some systems. Increasingly, anisotropy of magnetic remanence (AMR) is being used, particularly in palaeomagnetic studies of anisotropic rocks, to compute the ancient field vector from the deflected natural remanent magnetization (NRM) vector (Stephenson et al. 1986; Jackson 1991; Trindade et al. 2001; Gattacceca & Rochette 2003) as well as aiding accurate determinations of palaeointensity (Selkin et al. 2000). While the above studies were mainly concerned with correcting thermoremanent magnetizations (TRM) acquired by igneous and metamorphic rocks in the Earth's field, AMR methods have also found applications in sediments for correction of inclination shallowing of detrital remanent magnetization, DRM (Jackson et al. 1991; Kodama & Sun 1992; Kodama 1997; Hodych & Bijaksana 1993, 2002). AMR methods are directly relevant to all these palaeomagnetic studies, since they only measure the anisotropy of the remanence carrying particles, AMS, on the other hand, represents the sum of

the susceptibility anisotropies of all the mineral components in the rock, including the diamagnetic and paramagnetic fractions. Moreover, AMS is inappropriate for palaeomagnetic purposes, since it depends critically on the size of the remanence carrying particles. A multidomain (MD) particle of a ferrimagnetic mineral such as magnetite has a maximum susceptibility parallel to its long (easy) axis, while a stable singledomain (SD) particle has a maximum susceptibility perpendicular to its long axis. This is not a problem with AMR techniques, since the long axis is the axis of maximum remanence irrespective of particle size. These differences between AMS and AMR can sometimes give rise to so called 'inverse' fabrics in samples containing predominantly SD particles (Stephenson et al. 1986; Potter & Stephenson 1988; Rochette 1988; Borradaile & Puumala 1989; Stephenson & Potter 1989; Winkler et al. 1996), where the maximum AMS axis is perpendicular to the maximum AMR axis. Perhaps more significantly, as Potter & Stephenson (1988) first pointed out, many samples contain at least some SD particles and will have a mixture of SD and MD particles in the rock. In these cases the AMS ellipsoid

From: MARTIN-HERNANDEZ, F., LUNEBURG, C. M., AUBOURG, C. & JACKSON, M. (eds) 2004. Magnetic Fabric:

Methods and Applications. Geological Society, London, Special Publications, 238, 21-35. 0305-8719/04/S15.00 © The Geological Society of London 2004.

22

D. K. POTTER

resulting from these two groups of particles is unlikely to describe accurately the shape of an NRM anisotropy ellipsoid (such as a TRM acquired in the Earth's field). It is even possible theoretically for the AMS magnitude to be zero in samples containing appropriate mixtures of similarly aligned MD and stable SD particles (Potter & Stephenson 1988), and yet the particle alignments and AMR anisotropy could be very high. For example, figure 3.8b of Tarling & Hrouda (1993) shows results from several strongly magnetic rocks where the AMS is very low, but the AMR (in this case IRM anisotropy) can be substantially greater. Note that there may also be other contributing reasons for the latter results. For instance, Stephenson el al. (1986) demonstrated that the anisotropy of TRM (and other remanence anisotropies) should always be greater than that of AMS for multidomain particles on the basis of Stacey's (1963) theory for multidomain TRM acquisition. Other situations can theoretically occur due to the influence of mixtures of stable SD and MD grains, such as an interchange of maximum and intermediate axes between the AMS and AMR results (see Rochette et al. 1992, Fig. 8c region 2). Recently, Ferre (2002) has extended the theoretical models of the influence of mixtures of SD and MD particles on the AMS ellipsoid. The issues described above mean that the identification of anisotropic samples requiring palaeomagnetic corrections (and the process of correction itself) should not be based on AMS alone. Nevertheless, AMS is still useful for other purposes (for a review see Tarling & Hrouda 1993). Furthermore, comparisons between AMS and AMR can give additional information, such as estimates of the ferrimagnetic domain state and particle size (Stephenson et al. 1986; Stephenson & Potter 1989; Potter & Stephenson 1988). The above effects are also contributing factors explaining why AMS does not exhibit universal relationships with strain, even in the same rock type. Thus AMR may find uses in other magnetic fabric studies, such as helping to characterize specific relationships with strain due to particular grain size populations of remanence carrying particles. The purpose of the present paper is to provide a simple comparison of the variety of possible remanence anisotropy methods available, which could potentially be utilized for palaeomagnetic and other magnetic fabric analyses. Up to now AMR studies have almost exclusively applied the anisotropies of ARM or IRM. This paper will also describe how some other forms of remanence can be used for magnetic anisotropy

studies, and will discuss the relative merits of each. Examples from a comparative analysis of a strongly anisotropic rock sample and an anisotropic pottery sherd will be given. Sample description The rock sample was a metamorphic schist, with a strong visible fabric. Thermomagnetic analysis gave a Curie temperature of 565 °C, indicating a ferrimagnetic composition close to magnetite. Particle size analysis using a method based on a comparison of anhysteretic and rotational remanent magnetizations (Potter & Stephenson 1986) indicated that the rock was likely to contain a mixture of stable SD and small MD particles, with an average particle size of around 2 um. Experiments detailing the viscous decay of IRM also indicated that the rock contained a small proportion of particles with very short relaxation times, and therefore was likely to contain some superparamagnetic (SP) particles. The pottery sherd was a piece of replica Samian ware, and thermomagnetic analysis gave a Curie temperature of 550 °C, again indicating a ferrimagnetic composition close to magnetite, or a low titanium titanomagnetite. Particle size analysis (Potter & Stephenson 1986) again suggested the presence of stable SD and small MD particles with an average particle size of under 2 um. However, viscous decay of IRM experiments indicated a greater proportion of particles with short relaxation times, suggesting a higher proportion of SP particles than for the rock sample. Anisotropy of magnetic remanence (AMR) methods A summary of the necessary equipment, time taken, advantages and disadvantages of each of the described methods is given in Table 1. For each AMR method, unless otherwise stated, an anisotropy ellipsoid can be generated by applying an appropriate field successively along the x, y and z sample axes and measuring the 3 components ( x , y , z ) of remanence acquired after each field treatment. This gives 9 components of remanence: a single estimate of each diagonal tensor element, and two estimates for each pair of corresponding off-diagonal terms. Each pair of these off-diagonal terms is averaged thereby giving 3 off-diagonal coefficients, which, together with the 3 diagonal coefficients (i.e. a total of 6 independent coefficients), are then used to compute a remanence anisotropy ellipsoid (see

Table 1. A summary comparison of different magnetic anisotropy techniques Anisotropy method

Equipment

Time

Advantages

Disadvantages

Anisotropy of TRM

Furnace and Helmholtz cage controlling DF Pulse magnetizer

Several hours About 20 minutes

May generate chemical changes. Time consuming. Non-linear with low applied DF.

Helmholtz cage or shielded DF coil

Between 1 hour and several days

Direct simulation of TRM in the earth's field. Rapid. Large signal even for low concentrations. Preferentially measures anisotropy of particles close to SP size.

AF coil (as in a demagnetizer) and DF coil or Helmholtz cage Rotational magnetizer or air turbine in shielded AF coil Shielded AF coil (as in a demagnetizer) and a tumbler system Rotational magnetizer or air turbine in a shielded AF coil Shielded AF coil (as in a demagnetizer)

About 30 minutes to 1 hour

AMS delineator and bulk susceptibility bridge

About 1 minute

Anisotropy of IRM Anisotropy of VRM Anisotropy of ARM

Static ARM Rotational ARM Tumbling ARM

Anisotropy of gyroremanences

RRM GRM

AMS

About 30 minutes About 30 minutes About 30 minutes Between 1 hour and 1-2 days

Assumed linear with low DF. Partial ellipsoids possible due to different grain sizes. No static GRM produced. Theoretically the best ARM method, since no GRM or RRM should be produced. RRM is preferentially acquired by stable SD particles. Very sensitive; no GRM means no anisotropy. Gives stable SD anisotropy. Rapid, sensitive. Sum of all mineral components in the sample.

Measures limited particle size range. Can be time consuming. GRM simultaneously produced. RRM simultaneously produced. Requires tumbler large enough for sample holder generating the DF Requires specialist equipment. A full sequence of measurements can be time consuming. Palaeomagnetically less useful (MD versus SD differences; it is the sum of all minerals).

24

D. K. POTTER

Stephenson et al. 1986) comprising the magnitude and direction of the principal anisotropy axes a, b and c. The deviation between the values of each corresponding pair of off-diagonal terms (which should be equal) can be used in part to ascertain slight orientation errors between the sample axes and the applied field axes. The sample is ideally tumble AF demagnetized between each field application, and the residual remanence components should be subtracted from any subsequent remanence that is imparted. Note that there are alternative AMR measurement schemes (McCabe et al. 1985) that involve magnetizing and measuring in more than just 3 directions (at least 9 and preferably 18 directions), and the anisotropy ellipsoid is calculated using a least-squares fitting program. While this approach is more time consuming, it can allow more accurate ellipsoid determinations in weakly anisotropic samples. The GRM method described in the present paper also uses this approach.

Anisotropy of thermoremanent magnetization (ATRM) In order to simulate the acquisition of a natural thermoremanent magnetization (TRM) due to the Earth's field in an anisotropic sample, ATRM ellipsoids can be obtained by heating a sample three times to a temperature higher than the Curie point of the remanence carrying particles, and cooling in a field generally of the same order of magnitude as the Earth's field. The first heating is applied along the x axis of the sample, and, after cooling, the x, y and z components of remanence are measured. The second and third heatings and coolings are applied along the y and z axes respectively, and each time the 3 components of remanence are measured. The method merely requires a furnace and a means of controlling the applied direct field (e.g. by using a Helmholtz coil system). The effects of different cooling rates can be determined. A relatively fast heating and cooling cycle may still take around 1-2 hours, and so the method can be quite time consuming. While this method can potentially provide a direct simulation of a TRM in the Earth's field, it can introduce chemical changes, particularly oxidation, which can affect the results.

Anisotropy of isothermal remanent magnetization (AIRM) This method provides a rapid, non-destructive alternative to ATRM. Unlike TRM, and in

common with all the other laboratory remanences described below, it does not introduce chemical changes as the measurements are made at room temperature. A direct field (DF) is applied in turn along the x, y and z sample axes, and the nine components of remanence measured. This is most conveniently done using a pulse magnetizer, which normally generates a rapid (around 100ms) pulsed DF. The sample is demagnetized, preferably by tumbling in an alternating field (AF), between each DF application. Note that static AF demagnetization can introduce unwanted components of gyroremanence (GRM), which need to be accounted for by appropriate processing of the results (see Stephenson 1993). IRM gives the largest signal of any of the remanences described here, and therefore is particularly useful for samples containing very low concentrations of remanence carrying particles. Determination of the IRM ellipsoid at different field strengths can provide information about the anisotropy of different grain size populations. The equations from which the IRM ellipsoid is calculated make the assumption that remanence is proportional to field. While this assumption is generally assumed to be valid for low field TRM, it is not the case for IRM. However, despite this apparent disadvantage, experimental studies have shown a strong correspondence between the low field AIRM and ATRM ellipsoids (Stephenson et al. 1986).

Anisotropy of viscous remanent magnetization (AVRM) Essentially this can be thought of as an IRM in a low field given over a longer time period. It has the advantage of not requiring a pulse magnetizer. The direct field can be successively applied along the x, y and z sample axes by putting the sample in a Helmholtz coil system and cancelling appropriate components of the Earth's field. Alternatively, the direct field can be applied to the sample using a small coil placed within a Mumetal container to shield it from the Earth's field. The sample would again be AF demagnetized (ideally by tumbling) between each field application. The acquisition time of the VRM could be varied depending upon how viscous the sample was. This could be as little as 15 minutes for a strongly viscous rock, but one might want to leave the sample for a day or more for a less viscous sample. Therefore the method may take significantly longer than an AIRM determination, but for highly viscous

COMPARISON OF REMANENCE ANISOTROPY METHOD

samples it may be possible to obtain a VRM ellipsoid in about an hour. The resulting VRM is mainly acquired by SD particles with short relaxation times close to the superparamagnetic (SP) boundary. This is an advantage if one is interested in these particles. However, for palaeomagnetic purposes the stable SD particles and pseudo-single-domain (PSD) or small multidomain particles are of prime interest. Nevertheless, the shorter relaxation SD particles may be preferentially oriented in the same direction as the stable SD particles.

Anisotropy of anhysteretic remanent magnetization (AARM) ARM is commonly thought of as the best room temperature laboratory analogue of a TRM in the Earth's field. The sample is exposed simultaneously to an AF and a small bias DF. The AF is meant to simulate thermal agitation, and the small DF represents the Earth's field. Low field ARM is generally regarded as being linear with respect to the applied DF, although it may be non-linear depending on particle interactions (Cisowski 1981), and ARM is still non-linear as a function of applied AF. However, one major drawback of using ARM as a TRM analogue is that the ratio of ARM to TRM has been shown to vary depending on the grain size (Levi & Merrill 1976), and this ratio is further dependent on the concentration of the remanence carrying particles (Sugiura 1979). ARM anisotropy can be measured in a variety of ways. The sample can be stationary, rotating about one axis, or tumbled in the applied fields as described below: Anisotropy of static longitudinal ARM (AARMSL) In this method the sample is stationary, and the AF and DF are both applied along the same axis for each remanence acquisition step (firstly along x, then along y and z as before). This can be done by placing a small DF coil within a shielded AF demagnetizer coil. Alternatively, the AF demagnetizer coil could be placed within a Helmholtz coil system, and the DF applied by appropriate cancellation of components of the Earth's field. Both of these systems also allow partial ARM anisotropy ellipsoids to be determined, due to different grain size populations (Jackson el al. 1988; Trindade et al. 1999), by only switching on the DF over different AF windows. AF demagnetization is again performed, ideally by tumbling, between each ARM acquisition step.

25

The main disadvantage of the static AARM technique, using the above measurement scheme, is that the sample will simultaneously acquire a gyroremanent magnetization (GRM). For a longitudinal ARM the following remanence components will be produced:

The GRM is often ignored, but accurate values of static longitudinal ARM will only be obtained by determining the GRM components and subtracting them from the combined measured ARM + GRM values. One way to determine the GRM components directly is to merely apply an AF successively along the x, y and z sample axes, and measure the resulting remanence, while cancelling all direct fields. Tumble AF demagnetization would be employed between successive field treatments. Note that each pair of corresponding off-diagonal terms in the subsequently computed ARM values (A2x and A^ etc.) should ideally be equal. An alternative, but more time consuming, approach that theoretically avoids the problem of GRM produced with ARM is to use a measurement scheme involving magnetizing in more than just 3 directions (at least 9 and preferably 18 directions), and to use only the parallel component of magnetization acquired. The GRM for these components, as shown above, is theoretically zero. Note that the static ARM method does not need to have the AF and DF coincident. A static transverse ARM anisotropy ellipsoid could be determined (AARMST) by having the AF perpendicular to the DF. The DF is applied successively along the x, y and z sample axes in this case. The disadvantage of this type of ARM is that below saturation it is weaker than static longitudinal ARM (Stephenson 1983; Dunlop & Ozdemir 1997). Anisotropy of rotational ARM (AARMR) For this type of ARM the sample is rotated about a single axis. The remanence can easily be given using a rotational magnetizer (Stephenson & Molyneux 1987). The sample sits in a holder and is rotated within a split AF coil. This in turn is surrounded by a DF coil. In this rotational magnetizer the axes of the AF and DF are perpendicular to one another, with the DF applied along the sample's rotation axis (shown schematically in Fig. Ic). In the absence of a rotational magnetizer the sample could

26

D. K. POTTER

Fig. 1. Schematic diagram showing the relative orientations of the AF, DF, the sample rotation, and the rotational remanences (RRM and ARM) that are produced in a rotational magnetizer. (a) Clockwise rotation of the sample at rotation rates greater than twice the AF frequency produces an RRM that is parallel to the rotation vector co. Typically a rotation rate of 95r.p.s. is used as the RRM is large in this case for most palaeomagnetically important minerals. The DF (Bz), which is a combination of the Earth's field and the field due to the motor, is cancelled as closely as possible by a field B0 produced by a coil surrounding the AF coil. (b) Same conditions as for (a) but the sample is now rotated anticlockwise. An RRM is again produced parallel to the rotation vector. The ability to rotate the sample clockwise or anticlockwise enables an accurate average RRM value to be obtained without the need to exactly cancel the DF Bz. (c) Here the sample is rotated anticlockwise, but the DF Bz is not cancelled (B0 is switched off). This produces a rotational ARM and an RRM as shown. The rotational ARM can be determined by subtracting the remanence acquired in step (b) from that in step (c).

instead be placed in an air turbine (Stephenson 1980a) within an AF demagnetizer coil. This set-up could in turn be placed within a Helmholtz coil system, which can apply the DF by cancelling appropriate components of the Earth's magnetic field. For either the rotational magnetizer or the air turbine system the sample is first rotated about its x axis and the 3 components of remanence are measured. The process is then repeated by rotating the sample about its y and z axes. Tumble AF demagnetization is employed between each rotational remanence acquisition. The main drawback of these measurements is that a form of gyroremanence called a rotational remanent magnetization (RRM) is simultaneously produced during the ARM acquisition process (Fig. Ic). This, however, can be used to one's advantage to also determine an anisotropy of RRM ellipsoid (ARRM, see below) in addition to the ARM ellipsoid. The RRM can be determined by rotating the sample in the absence of any applied DF (Fig. Ib), and the results can be subtracted from the RRM + ARM values obtained with the DF switched on (Fig. Ic) in order to calculate the rotational ARM. The

RRM to ARM ratio can also be used as a particle size indicator (Potter & Stephenson 1986). Anisotropy of tumble ARM (AARMTU) This method relies on tumbling the sample in an AF coil. Theoretically this is the best ARM method, since an ideal tumbler should not introduce any gyroremanence, because it ideally randomizes any gyromagnetically produced moments. The sample is placed in a holder containing small permanent magnets which produce the DF (see Collinson 1983, p. 176, Fig. 6.7), and this is then tumbled in an AF demagnetizer coil. The sample is oriented initially with its x axis in the direction of the DF, and the 3 components of remanence measured. The sample is then tumble AF demagnetized, and the same procedure repeated for the y and z axes. A slight drawback to the tumble ARM method is that the AF coil and tumbler holder need to be large enough to contain the sample holder containing the permanent magnets producing the DF. Also, in the tumble ARM system described in Collinson (1983) above, it is not possible to

COMPARISON OF REMANENCE ANISOTROPY METHOD

determine partial ARM ellipsoids, since the DF is permanently on. Note that theoretically if the ARM is not saturated then for the same value of DF the magnitude of tumble ARM should be greater than that of rotational ARM, which should be greater than that of static longitudinal ARM, which in turn should be greater than that of static transverse ARM (Stephenson 1983, 1985; Stephenson & Potter 1987). At saturation the magnitude of the ARM should be independent of the ARM acquisition method.

Anisotropy of rotational remanent magnetization (ARRM) This method is not commonly used at present, but is potentially very useful, since RRM is preferentially acquired by stable SD particles (Potter & Stephenson 1986). The high temporal and thermal stability of RRM has also been documented (Mahon & Stephenson 1997). This means that ARRM potentially measures the anisotropy of palaeomagnetically important particles. However, it is still uncertain whether RRM is exclusively acquired by stable SD particles, since it is not clear if the small values of RRM observed in nominally larger MD particles are due mainly to small adhering SD particles (Potter & Stephenson 1986), or whether small MD particles themselves can also acquire RRM. It is likely that very small multidomain grains are capable of acquiring RRM (and GRM), since cubic grains that SEM analysis showed were principally in the range 0.2-0.8 um did exhibit RRM (Potter & Stephenson 1986). As mentioned above, this remanence can be easily measured using a rotational magnetizer (Stephenson & Molyneux 1987), or an air turbine (Stephenson 19800). With the rotational magnetizer the sample can be mechanically rotated clockwise (Fig. la) or anticlockwise (Fig. Ib) using the mechanical motor. By using a DF coil (providing a field B0) the combined DF (Bz) due to both the Earth's field and the field produced by the motor can be cancelled reasonably well. The fact that the sample can be rotated clockwise or anticlockwise means that an average RRM value can be obtained, by subtracting the clockwise produced RRM from the anticlockwise one and dividing by two. This allows an accurate RRM value to be obtained irrespective of whether any small ARM is produced from incomplete nulling of any residual DF present. This means that the DF does not need to be perfectly cancelled by the surrounding coil. The RRM ellipsoid is obtained (like the rotational

27

ARM ellipsoid) by rotating the sample about its x, y and z axes in turn. The sample is generally rotated at around 95r.p.s., where the RRM is large and parallel to the rotation vector (co) for most palaeomagnetically important minerals (Potter & Stephenson 1986). The orientation of the RRM and ARM at this rotation rate for clockwise and anticlockwise rotations is shown in Figures la-c. If one uses an air turbine that rotates in one direction, instead of a rotational magnetizer, then the DF would need to be perfectly cancelled to prevent any simultaneous acquisition of rotational ARM.

Anisotropy of gyroremanent magnetization (AGRM) Gyroremanent magnetization is acquired by a static sample that is merely subjected to an alternating field. For a single application of the AF a GRM is only produced if there is an anisotropic distribution of particles. GRM anisotropy is potentially very useful since the magnitude of the GRM is proportional to the anisotropy of the sample in the perpendicular plane. GRM anisotropy can essentially be thought of as the remanence equivalent of an AMS meter (which measures the susceptibility anisotropy in the sample plane perpendicular to the rotation axis; see the AMS section below). GRM anisotropy is also extremely sensitive, in that very small anisotropies can produce a GRM (Stephenson & Potter 1987; Potter & Stephenson 1988). Like RRM, GRM is also preferentially acquired by stable SD particles, and thus is a sensitive measure of the anisotropy of palaeomagnetically important particles. The GRM is produced perpendicular to both the particle easy axis and the applied AF axis. The measurements are undertaken using an AF coil, which is shielded from the Earth's field, so no ARM should be produced. Occasionally a small ARM can be introduced due to the coil along the AF axis. This can be eliminated by doing a repeat measurement in an anti-parallel direction and taking an average. This will cancel any small ARM components in the field direction. GRM anisotropy can be computed from the GRM components produced from at least six applications of the AF along two different sets of orthogonal sample axes: x, y and z and jjq, yi, and z\. Tumble AF demagnetization is preferably employed between each single-axis AF treatment, or alternatively (if one does not have a tumbler) the sample can be given a single-axis 'cyclic state' AF treatment prior to

28

D. K. POTTER

the measurements along the sample axes (Stephenson 1993). The results can be used in a least-squares fitting program to determine the orientation of the principal anisotropy axes. The method does not determine an ellipsoid, since there is no 'bulk' measurement that can be added to the anisotropy differences determined from the GRM measurements (in contrast to AMS delineator results where one adds the bulk susceptibility measurement). However, since the magnitude of the GRM relates to the magnitude of the anisotropy in the perpendicular plane of the sample, the program can calculate suitable coefficients Q, C2 and C3 (Stephenson & Potter 1987; Potter & Stephenson 1988) that are related to the magnitude of the anisotropy in the yz, zx and xy planes respectively of the sample. For accurate, sensitive measurements on weakly anisotropic distributions of SD particles, it is preferable to undertake a more complete set of GRM measurements, involving AF applications in several orientations in 3 mutually perpendicular planes, and again using a leastsquares fitting program to calculate the orientations of the principal axes. The disadvantage of this is that the measurements for one sample are more time consuming, and may take one to two days to complete. Anisotropy of magnetic susceptibility (AMS) The AMR methods were also compared with AMS measurements for the two samples studied in this paper. The AMS measurements were undertaken using a Molspin anisotropy delineator, where a sample is placed between two orthogonal, square Helmholtz coils whose axes are horizontal. One of these coils carries an alternating current of 10 kHz, producing a field of around 500 ji T at the sample. The sample is rotated about a vertical axis at 7 Hz, and if it is anisotropic a sinusoidal voltage of twice the rotational frequency is induced in the other coil. The amplitude of this voltage is proportional to the difference between the maximum and minimum susceptibilities in the sample plane perpendicular to the rotation axis. The data for each of 3 spins, in orthogonal orientations, are used to compute the orientations of the sample's principal AMS axes. The magnitudes of the principal axes are given by adding a 'bulk' susceptibility measurement (generally the z axis measurement determined using a susceptibility bridge) to the differences determined by the AMS delineator. The anisotropy measurements from the AMS delineator take around 1 minute to perform,

and the bulk susceptibility measurement (which consists of a background reading and a sample reading) takes under 5 seconds in the Molspin susceptibility bridge. Note that there are alternative measurement position schemes and methods of AMS estimation (Girdler 1961; Borradaile & Stupavsky 1995). Results Table 2 shows comparative results for several of the different AMR methods, together with AMS, for a strongly anisotropic metamorphic rock sample. All the methods gave very similar values for the orientation of the principal anisotropy axes. The AMR axes are consistent with the visual macroscopic fabric seen in the sample. The results shown in Table 2 are the average of 3 determinations. The directions of the a and b principal axes varied by no more than ±2° from the values shown. For the c principal axis the inclination did not vary by more than ±2°, but the declination was more variable since the inclination is so steep. The shape of the anisotropy ellipsoids from the AMR methods involving IRM, VRM and ARM (both the static longitudinal and the rotational ARM) measurements are all much closer to the shape of the TRM anisotropy ellipsoid acquired in the Earth's field than that of the AMS ellipsoid. In particular, the low field (5mT) IRM anisotropy and VRM anisotropy (acquired in the Earth's field) are very close to the TRM normalized values. If one considers the percentage anisotropy [(max. — min.)/total] x 100% then the TRM anisotropy was 46% compared to 51% for the low field IRM, 44% for the VRM, and 38% for the ARM (both the static and rotational ARM). The AMS ellipsoid by comparison was only 26%. The standard deviation of the normalized AMR magnitudes were no more than ±0.01 in all cases. The values of the standard deviation for the low field IRM anisotropy were very slightly larger than for the other forms of remanence, due mainly to some viscous decay of IRM during the measurement time. The low field AMR ellipsoids are all more anisotropic than the measurements made at higher fields. The higher field measurements for IRM, ARMSL, ARM R , and RRM determined in a DF or AF of 60 mT are all very similar for this sample. Note that the average dimension of the IRM ellipsoid acquired in 60 mT is substantially larger than for any of the other remanence ellipsoids. The GRM results from a comprehensive set of multidirectional AF applications (Fig. 2) also

Table 2. A comparison of anisotropy of magnetic remanence (AMR) methods, and anisotropy of magnetic susceptibility (AMS),for a strongly anisotropic rock sample Method

Directions

a

TRM IRM(5mT) IRM (60 mT) VRM ARMSL ARMR RRM GRM AMS

Average dimension of ellipsoid

Normalized magnitudes

b

c

DEC

INC

DEC

INC

DEC

INC

94 95 103 99 97 98 91 94 97

2 1 3 1 4 5 0 1 1

4 5 13 9 8 9 0 4 7

-1 -1 -4 -1 -3 _7 -6 2 _3

158 140 163 129 168 156 177 338 164

-88 -89 -85 -89 -86 -82 -84 -88 -87

a

b

c

0.56 0.58 0.48 0.55 0.50 0.52 0.47 max. 0.44

0.34 0.35 0.38 0.34 0.38 0.34 0.37 int. 0.38

0.10 0.07 0.14 0.11 0.12 0.14 0.16 mm. 0.18

324.8 x 380.9 x 12.1 x 60.0 x 176.1 x 151.7 x 122.7x

10~ 6 Am 2 kg -i 10~ 6 Am 2 kg -i 10"3Am2kg -i 10~ 6 Am 2 kg-i 10~ 6 Am 2 kg -i 10~ 6 Am 2 kg -i 10- 6 Am 2 kg -i

193.7 x 10" 8 m 3 kg -1

The directions are in degrees declination (DEC) and inclination (INC). The direct field used to produce the TRM, VRM (acquired in one day), ARMSL and ARMR was 44 jiT (close to the magnitude of the vertical component of the Earth's field at the latitude of the laboratory). The TRM was acquired by heating to 700 °C and cooling in the direct field. An AF of peak value 60mT was used for all the ARM, RRM and GRM measurements.

30

D. K. POTTER

Fig. 2. The GRM components (circles) relating to the results for the rock specimen in Table 2. A peak AF of 60 mT was applied in the xy, yz and zx planes. The components of GRM are M^ (Oxy\ Mgy (Oxy), and Mgz (6xy) for the AF applied in the xy plane (see Figure 3 for a schematic example of the orientation of the AF axis with respect to the sample axes), and likewise the components as indicated for the yz and zx planes. The curves are best fit theoretical ones from a least-squares fitting program.

identified similar orientations of the principal anisotropy axes as the other AMR methods (Table 2). While it is not possible to obtain the shape of the anisotropy ellipsoid from the GRM results, a good sensitive indication of the magnitude of the anisotropy in different sample planes can be obtained by looking at the raw

GRM results as a function of the orientation of the applied AF. Figure 2 shows the GRM components for a full sequence of measurements in the xy, yz and zx planes for the rock sample of Table 2. A schematic is given in Figure 3 to illustrate the orientation of the AF axis with respect to the sample x, y and z axes for an AF applied

COMPARISON OF REMANENCE ANISOTROPY METHOD

Fig. 3. Diagram illustrating a typical single-axis AF application in the generation of GRM components, the results of which are shown for the rock sample in Figure 2. The diagram shows the orientation of the AF axis with respect to the sample x, y and z axes for an AF applied in the xy plane. This represents a positive Oxy AF field application in Figure 2. See text for further details. This process is repeated for several different orientations.

in the xy plane. The convention is that positive values of the angle 9xy lie between x and y, and negative values of the angle 9xy lie between x and —y. Therefore, for the xy plane 9xy = 0° is the x direction, 9xy — 90° is the y direction, and 9xy = -90° is the —y direction. The GRM components in Figure 2 pick out the anisotropy in different planes very clearly. From GRM theory (Stephenson 1980b) the component Mgz (9xy) in the bottom left portion of Figure 2 shows that the stable SD particles are primarily aligned to a slightly greater extent along the y sample axis rather than the x axis, since Mgx (Oxy) is negative for positive values of 9yz and positive for negative values of 9yz. Similarly, the results for component M^x (9yz) in the middle upper portion of Figure 2 show that the stable SD particles are significantly more aligned along the y axis than along the z axis. Likewise the values for component Mg>, (9ZX) show that the stable SD particles are significantly more aligned along the x axis than along the z axis. The amplitude of the GRM signal for both the Mgx (9yz) and Mgy (9ZX) components is very similar, meaning that the magnitude of the SD anisotropy in the yz plane is very similar to that in the zx plane. Putting all this information together enables one to deduce that there is approximately a plane of maximum SD anisotropy very close to the xy plane, and within that plane the SD particles are preferentially aligned slightly more towards the y axis than the x axis. Quantitatively,

31

the least-squares fitting program calculated the coefficients Q, C2 and C3 (Stephenson & Potter 1987, as detailed earlier) to be 124, -128 and -21 x 10~6 Am2 kg*1, which are related to the anisotropy magnitude in the yz, zx and xy planes respectively. Table 3 gives a comparison of the different methods for the pottery sherd. All the AMR methods, together with AMS, gave similar values for the orientation of the principal anisotropy axes. The minimum principal AMR and AMS axis coincides exactly with the axis that is visually perpendicular to the plane of the pottery sherd. The rotational remanences RRM and ARM R , where the sample is usually spun at around 95r.p.s., were not undertaken in this case due to the fragile nature of the pottery sample. The low field IRM anisotropy and VRM anisotropy ellipsoids are again of almost identical shape to the TRM ellipsoid acquired in the Earth's field. In this case the percentage anisotropies are 16% for TRM, 16% for low field IRM, 14% for higher field (60 mT) IRM, 18% for VRM, and 16% for static longitudinal ARM. All the room temperature AMR results are significantly closer to the TRM anisotropy than the AMS value of 8%. Note that for this sample the higher field IRM anisotropy magnitude is only slightly lower than the low field magnitude. For the GRM results the coefficients Q, C2 and C3 were 41.4, -70.8 and 35.4 x lO^An^kg- 1 . Discussion The results for the examples shown in Tables 2 and 3 show that the low field AMR anisotropy methods are good analogues of the TRM anisotropy ellipsoid acquired in the Earth's field, even though (as in the case of IRM) the low field remanence can be non-linearly acquired with respect to the applied DF. The equivalence of low field IRM and TRM ellipsoids was also observed in a previous study (Stephenson et al. 1986). The fact that IRM gives a large signal, for a given applied field, means that the IRM method would be particularly useful for samples with low concentrations of remanence carrying particles, where the low field ARM methods may give too low a signal for accurate anisotropy measurements. For both the rock sample and pottery sherd the magnitude of AMR was greater than that of AMS, as theoretically expected on the basis of MD TRM theory and also if each sample contained a mixture of stable SD and MD particles with the same alignment direction (Stephenson

Table 3. A comparison of anisotropy of magnetic remanence (AMR) methods, and anisotropy of magnetic susceptibility (AMS).for an anisotropic pottery sample (same conditions as for Table 2) Method

Directions

Normalized magnitudes

a

TRM IRM(5mT) IRM(60mT) VRM ARMSL GRM AMS

c

b

DEC

INC

DEC

INC

DEC

INC

43 42 45 47 44 42 44

-4 -10 __3 -11 _4 —4 -2

120 104 122 115 122 119 129

72 70 75 63 71 73 66

316 315 315 323 315 313 315

17 18 15 25 18 16 24

Average dimension of ellipsoid

a

b

c

0.41 0.41 0.40 0.41 0.42 max. 0.37

0.34 0.34 0.34 0.36 0.32 int. 0.34

0.25 0.25 0.26 0.23 0.26 min. 0.29

418.2 x 386.7 x 12.3 x 85.0 x 217.4x

10~ 6 Am 2 kg~ i 10"6Am2kg- i 10~ 3 Am 2 kg- i 10~ 6 Am 2 kg~ i 10~ 6 Am 2 kg~ i

220 x KT^kg- 1

COMPARISON OF REMANENCE ANISOTROPY METHOD

el al. 1986; Potter & Stephenson 1988; Rochette et al. 1992; Ferre 2002). In a previous study (Potter & Stephenson 1988) the shapes of the AMR and AMS ellipsoids for an igneous rock sample that contained predominantly stable SD particles were very different, with AMR varying from 6-7% (in this case IRM, ARMSL, ARM R , RRM, and GRM) while the AMS was only 0.8%. Moreover, the maximum AMR axis corresponded to the minimum AMS axis, as expected from stable SD theory. For the rock sample in the present study the magnitude of AMR decreased slightly for higher applied fields, as observed previously for artificially prepared anisotropic samples containing small MD magnetite particles (Stephenson et al. 1986). The pottery sherd, however, exhibited less of a difference between low and high field measurements. This would result if this sample has a narrower particle size range (comprising SP, stable SD and small MD magnetite particles) than the rock sample. The results from Tables 2 and 3 show that for these samples different particle size fractions from the various AMR methods exhibit similar orientations. The particles close to SP size (from the VRM measurements), the stable SD particles (from the RRM and GRM measurements), and the stable SD through to MD particles from the other forms of remanence (IRM measurements at different fields, TRM and ARM) all appear to be oriented similarly. This may not necessarily be the case in other samples, and the different AMR techniques can be used to distinguish the anisotropies of the different particle size fractions accordingly. Furthermore, partial anisotropies using an individual technique such as ARM (Jackson et al. 1988) can also be employed for this purpose. Interestingly, the results from the gyroremanences for the rock sample seem to suggest a slightly lower magnitude for the anisotropy of the stable SD particles. This is evident from the RRM results of Table 2. Also, the small values of the Mgz (0xy) components of the GRM in Figure 2 suggest that the difference between the a and b principal axes (which are very close to the y and x sample axes respectively in this case) is not as great as that suggested from the other forms of remanence. This difference in the magnitude of the anisotropy for different particle size fractions may be related to different particle shapes of the various size fractions as well as their degree of alignment. The fact that gyroremanences are preferentially acquired by stable SD particles may be exploited for further palaeomagnetic applications such as in DRM studies, including those involving inclination shallowing, where

33

the stable SD and small MD particles exhibit the highest DRM signal. It is clear from the results of Figure 2 and Table 2 that the GRM in that rock sample can be up to 30% of the maximum static longitudinal ARM. Therefore, depending upon the orientation of the applied AF with respect to the principal anisotropy axes, the influence of the GRM on the static ARM ellipsoid could be very significant unless the GRM is subtracted, or one utilizes a methodology which only uses the parallel component of magnetization acquired, which should not contain any GRM. For the present rock sample, where the GRM components have been subtracted from the static longitudinal ARM results, the uncorrected measurements would not have been greatly different. This is merely because the principal anisotropy axes a, b and c in this case are close to the sample axes y, x and z respectively. In this situation the GRM components along the sample axes should be close to zero. Figure 2 confirms that these GRM components are low along the x, y, and z axes, which are the orientations in which the AF producing the static longitudinal ARM ellipsoid was applied. If the principal anisotropy axes were in orientations very different from the sample axes (as is likely to occur in many other samples) then the influence of the GRM on the uncorrected ARM ellipsoid would have been significant. Conclusions The results of Tables 2 and 3 indicate that, for the samples analysed, the room temperature AMR ellipsoids provided good analogues for the TRM anisotropy ellipsoid acquired in the Earth's field. This is particularly true for the low field AMR ellipsoids (such as IRM), since their shape and orientation was almost identical to that of the TRM ellipsoid. This demonstrates the potential usefulness of these AMR methods in obtaining the corrected ancient direction of the Earth's field from the NRM recorded in anisotropic samples (rocks, pottery etc). It also adds support to previous work (Stephenson et al. 1986), which suggested that IRM anisotropy could potentially be an analogue of TRM anisotropy. The shape of the AMR ellipsoids were dependent to varying degrees on the applied field strength. Even at higher applied fields (60 mT) the AMR results gave magnitudes of the normalized principal anisotropy axes that were significantly closer to the TRM anisotropy than that of the AMS ellipsoid, which exhibited a significantly lower anisotropy magnitude.

34

D. K. POTTER

This would be expected theoretically both from the domain state dependence of AMS (a mixture of uniaxial SD and MD particles with the same alignment axis can have a high AMR magnitude, but a low AMS magnitude), and also from MD TRM theory. The results presented here therefore support previous suggestions (Stephenson et al. 1986) that AMR is more appropriate than AMS for the above palaeomagnetic purposes. AMR methods should also be more appropriate than AMS for (a) palaeointensity determinations in anisotropic samples, (b) DRM studies and inclination shallowing, and (c) correlations between magnetic anisotropy and strain in samples containing remanence carrying particles. IRM is the quickest, and arguably easiest, of the AMR techniques. The results reported here, along with other previous studies, show that the average dimension of the IRM ellipsoid (for a given applied field) is generally considerably higher than for other types of remanence. Therefore IRM anisotropy will be useful for determining anisotropic distributions in low concentrations of remanence carrying grains. The different AMR methods allow the anisotropy of different particle size or domain state fractions to be determined. VRM anisotropy preferentially measures the anisotropy of the particles close to SP size. The gyroremanences, RRM and GRM, on the other hand, preferentially measure the anisotropy of the stable SD particles, and are thus directly relevant to palaeomagnetically important particles. GRM has been shown here to be a sensitive indicator of anisotropy, and is the remanence equivalent of the AMS delineator. An isotropic distribution of SD particles will not produce a GRM when subjected to a single application of an AF. The results shown here indicate that GRM, which is simultaneously produced during the ARM acquisition process, can have a magnitude that is significant compared to the ARM. Therefore care must be taken if ones uses a rapid static ARM method (longitudinal or transverse), where the DF is applied in turn along the x, y and z sample axes, to ensure that the off-diagonal components of GRM are subtracted from the results. Alternatively, one can utilize a longer measurement scheme involving a multidirectional least-squares method using only the parallel components of magnetization acquired. Likewise, if one uses rotational ARM, the components of RRM need to be subtracted. Theoretically the best ARM method is by tumbling the sample in the AF. This ideally randomizes any gyromagnetically produced moments. The wide variety of AMR methods currently available, as summarized in Table 1, should

allow the most appropriate technique to be chosen for any particular sample. This may depend on several factors, including the intrinsic anisotropy of the sample, the concentration of remanence carrying particles, and the domain state or particle size range involved. This study has helped to demonstrate that there are several possible alternatives for providing good analogues of NRM anisotropy for application to palaeomagnetic and other studies, even when one may be constrained by available equipment and time. The thorough and constructive reviews of two referees, Joseph Hodych and Ricardo Trindade, together with suggestions by co-editor Mike Jackson, improved the manuscript and are gratefully acknowledged.

References BORRADAILE, G. J. & PUUMALA, M. A. 1989. Synthetic magnetic fabrics in a plasticene medium. Tectonophysics, 164, 73-78. BORRADAILE, G. J. & STUPAVSKY, M. 1995. Anisotropy of magnetic susceptibility - measurement schemes. Geophysical Research Letters, 22, 1957-1960. CISOWSKI, S. 1981. Interacting versus non-interacting single domain behavior in natural and synthetic samples. Physics of the Earth and Planetary Interiors, 26, 56-62. COLLINSON, D. W. 1983. Methods in Rock Magnetism and Palaeomagnetism. Chapman and Hall, London, pp. 503. DUNLOP, D. J. & OZDEMIR, O. 1997. Rock magnetism: fundamentals and frontiers. Cambridge University Press, Cambridge, pp. 573. FERRE, E. C. 2002. Theoretical models of intermediate and inverse AMS fabrics. Geophysical Research Letters, 29, art. no. 1127. GATTACCECA, J. & ROCHETTE, P. 2003. Anisotropies of remanence and magnetic susceptibility: application to the correction of palaeomagnetic directions in igneous rocks. Geophysical Research Abstracts, 5, 04027. GIRDLER, R. W. 1961. The measurement and computation of anisotropy of magnetic susceptibility in rocks. Geophysical Journal of the Royal Astronomical Society, 5, 34-44. HODYCH, J. P. & BIJAKSANA, S. 1993. Can remanence anisotropy detect palaeomagnetic inclination shallowing due to compaction - a case study using Cretaceous deep sea limestones. Journal of Geophysical Research - Solid Earth, 98, 22429-22441. HODYCH, J. P. & BIJAKSANA, S. 2002. Plastically deforming clay rich sediment to help measure the average remanence anisotropy of its individual magnetic particles, and correct for palaeomagnetic inclination shallowing. Physics and Chemistry of the Earth, 21, 1273-1279. JACKSON, M. J. 1991. Anisotropy of magnetic remanence - a brief review of mineralogical sources,

COMPARISON OF REMANENCE ANISOTROPY METHOD physical origins, and geological applications, and comparison with susceptibility anisotropy. Pure and Applied Geophysics, 136, 1-28. JACKSON, M. J., GRUBER, W., MARVIN, J. & BANERJEE, S. K. 1988. Partial anhysteretic remanence and its anisotropy: applications and grain size dependence. Geophysical Research Letters, 15, 440-443. JACKSON, M. J., BANERJEE S. K., MARVIN, J. A. & GRUBER, W. 1991. Detrital remanence, inclination errors, and anhysteretic remanence anisotropy quantitative model and experimental results. Geophysical Journal International, 104, 95-103. KODAMA K. P. 1997. A successful rock magnetic technique for correcting paleomagnetic inclination shallowing: Case study of the Nacimiento Formation, New Mexico. Journal of Geophysical Research - Solid Earth, 102, 5193-5205. KODAMA K. P. & SUN, W. W. 1992. Magnetic anisotropy as a correction for compaction caused paleomagnetic inclination shallowing. Geophysical Journal International, 111, 465-469. LEVI, S. & MERRILL R. T. 1976. A comparison of ARM and TRM in magnetite. Earth and Planetary Science Letters, 32, 171-184. MAHON, S. W. & STEPHENSON, A. 1997. Rotational remanent magnetization (RRM) and its high temporal and thermal stability. Geophysical Journal International, 130, 383-389. MCCABE, C., JACKSON, M. & ELLWOOD, B. B. 1985. Magnetic anisotropy in the Trenton Limestone: results of a new technique, anisotropy of anhysteretic susceptibility. Geophysical Research Letters, 12, 333-336. POTTER, D. K. & STEPHENSON, A. 1986. The detection of fine particles of magnetite using anhysteretic and rotational remanent magnetizations. Geophysical Journal of the Royal Astronomical Society, 87, 569-582. POTTER, D. K. & STEPHENSON, A. 1988. Single-domain particles in rocks and magnetic fabric analysis. Geophysical Research Letters, 15, 1097-1100. ROCHETTE, P. 1988. Inverse magnetic fabric in carbonate-bearing rocks. Earth and Planetary Science Letters, 90, 229-237. ROCHETTE, P., JACKSON, M. J. & AUBOURG, C. 1992. Rock magnetism and the interpretation of anisotropy of magnetic susceptibility. Reviews of Geophysics, 30, 209-226. SELKIN, P. A., GEE, J. S., TAUXE, L., MEURER, W. P. & NEWELL, A. J. 2000. The effect of remanence anisotropy on palaeointensity estimates: a case study from the Archaen Stillwater Complex. Earth and Planetary Science Letters, 183, 403-416. STACEY, F. D. 1963. The physical theory of rock magnetism. Advances in Physics, 12, 45-133. STEPHENSON, A. 19800. The measurement of the magnetic torque acting on a rotating sample using an

35

air turbine. Journal of Physics E: Scientific Instruments, 13, 311-314. STEPHENSON, A. 1980/7. Gyromagnetism and the remanence acquired by a rotating rock in an alternating field. Nature, 284, 48-49. STEPHENSON, A. 1983. Changes in direction of the remanence of rocks produced by alternating field demagnetization. Geophysical Journal of the Royal Astronomical Society, 73, 213—239. STEPHENSON, A. 1985. The angular dependence of rotational and anhysteretic remanent magnetizations in rotating rock samples, Geophysical Journal of the Royal Astronomical Society, 83, 787-796. STEPHENSON, A. 1993. 3-axis static alternating field demagnetization of rocks and the identification of natural remanent magnetization, gyroremanent magnetization, and anisotropy. Journal of Geophysical Research - Solid Earth, 98, 373-381. STEPHENSON, A. & MOLYNEUX, L. 1987. The rapid determination of rotational remanent magnetization and the effective field which produces it. Geophysical Journal International, 90, 467-471. STEPHENSON, A. & POTTER, D. K. 1987. Gyroremanent magnetizations in dilute anisotropic dispersions of gamma ferric oxide particles from magnetic recording tape. IEEE Transactions on Magnetics, MAG-23, 3820-3830. STEPHENSON, A. & POTTER, D. K. 1989. Some aspects of the measurement of magnetic anisotropy. In: F. J. LOWES ETAL. (eds) Geomagnetism andPalaeomagnetism NATO ASI Series C261, pp. 271-278. Kluwer Academic Publishers. STEPHENSON, A., SADIKUN, S. & POTTER, D. K. 1986. A theoretical and experimental comparison of the anisotropies of magnetic susceptibility and remanence in rocks and minerals. Geophysical Journal of the Royal Astronomical Society, 84, 185-200. SUGIURA, N. 1979. ARM, TRM and magnetic interactions - concentration dependence. Earth and Planetary Science Letters, 42, 451^455. TARLING, D. H. & HROUDA, F. 1993. The Magnetic Anisotropy of Rocks. Chapman and Hall, pp.217. TRINDADE, R. I. F., RAPOSO M. I. B., ERNESTO, M. & SIQUEIRA, R. 1999. Magnetic susceptibility and partial remanence anisotropies in the magnetite bearing granite pluton of Tourao, N. E. Brazil. Tectonophysics, 314, 443-468. TRINDADE, R. I. F., BOUCHEZ, J. L., BOLLE, O, NEDELEC, A, PESCHLER, A. & POITRASSON, F. 2001. Secondary fabrics revealed by remanence anisotropy: methodological study and examples from plutonic rocks. Geophysical Journal International, 147, 310-318. WINKLER, A., FLORINDO, F., SAGNOTTI, L. & SARTI, G. 1996. Inverse to normal magnetic fabric transition in an upper Miocene marly sequence from Tuscany, Italy. Geophysical Research Letters, 23, 909-912.

This page intentionally left blank

Distribution anisotropy: the influence of magnetic interactions on the anisotropy of magnetic remanence ADRIAN R. MUXWORTHY & WYN WILLIAMS Grant Institute of Earth Science, University of Edinburgh, Kings Buildings, West Mains Road, Edinburgh, EH9 3JW, UK (e-mail: [email protected]) Abstract: The anisotropy of magnetic remanence (AMR) is often used as a tool for examining magnetic anisotropy of rocks. However, the influence of magnetostatic interactions on AMR has not been previously rigorously addressed either theoretically or experimentally, though it is widely thought to be highly significant. Using a three-dimensional micromagnetic algorithm, we have conducted a systematic numerical study of the role of magnetostatic interactions on AMR. We have considered both lineation and foliation, by modelling assemblages of ideal single domain grains and magnetically non-uniform magnetite-like cubic grains. We show that magnetostatic interactions strongly affect the measured AMR signal. It is found that depending on the orientation of the single-grain anisotropy and grain spacing it is possible for the AMR signal from a chain or grid of grains to be either oblate or prolate. For non-uniform grains, the degree of anisotropy generally increases with increasing interactions. In the modelling of AMR anisotropy, saturation isothermal remanence was chosen for numerical tractability. The influence of interactions on other types of more commonly measured AMR, are considered in light of the results in this paper.

Anisotropy of magnetic remanence (AMR) is commonly used as an alternative to anisotropy of low-field magnetic susceptibility (AMS) as a method of determining the magnetic fabric of a rock sample. Measuring AMR has some distinct advantages over AMS, for example, if the anisotropy of partial anhysteretic remanence (AARM) is measured, then the anisotropy of different fractions of the coercivity spectrum can be assessed (Jackson et al 1988; Aubourg & Robion 2002). However, because AMR is still very much the junior partner of AMS, some theoretical aspects of AMR have not been rigorously tackled. In particular the role of distribution (or textural) anisotropy. Distribution anisotropy occurs when magnetic grains are both unevenly distributed and close enough to interact magnetostatically, producing an asymmetric magnetic interaction field which contributes to the bulk magnetic anisotropy (Hargraves et al. 1991). Several theoretical models have examined distribution anisotropy's contribution to AMS (Stephenson 1994; Canon-Tapia 1996). These models have shown that when grains become closer and interact magnetostatically, the distributions of grains rather than their individual orientations dominate AMS. Differences between these model predictions and experimental results (Gregoire et al. 1995; Gregoire et al. 1998) have recently been attributed by Canon-Tapia (2001) to oversimplifications in the previous models. The influence of interactions on AMR cannot be directly inferred from the AMS behaviour and

models, due to numerous theoretical reasons (Jackson 1991). In this paper we investigate the influence of magnetostatic interactions on AMR in assemblages of ideal single domain (SD) and pseudo-SD (PSD) magnetite-like grains, i.e. grains which display non-uniform internal magnetic structures, using for the first time a rigorous numerical model. We consider both lineation and foliation for assemblages of grains with a range of both interaction spacing and single-grain anisotropy. A numerical model for distribution anisotropy One of the main difficulties in modelling interacting assemblages is that it is a highly non-linear problem, unlike non-interacting uniform SD grains, which can be very well explained by analytical theories (e.g. Stoner & Wohlfarth 1948). With the rapid advancement in computing power, it has become possible to model this non-linear behaviour directly by implementing Brown's (1963) micromagnetic formalism to study non-uniform magnetic phenomena (e.g. Williams & Dunlop 1989; Muxworthy et al. 2003). The micromagnetic algorithm In this study we have implemented a new micromagnetic algorithm. This new algorithm differs

From: MARTIN-HERNANDEZ, F., LUNEBURG, C. M., AUBOURG, C. & JACKSON, M. (eds) 2004. Magnetic Fabric: Methods and Applications. Geological Society, London, Special Publications, 238, 37-47. 0305-8719/04/S15.00 © The Geological Society of London 2004.

38

A. R. MUXWORTHY & W. WILLIAMS

from the algorithm used in previous micromagnetic studies conducted by the Edinburgh group (e.g. Wright et al. 1997; Muxworthy et al. 2003). The algorithm is a combination of both a minimum energy conjugate-gradient (CG) algorithm (as used in previous studies) and a dynamic algorithm that follows the torque of a magnetic moment according to the Landau-LifshitzGilbert (LLG) equation (Suess et al 2002). The reasoning behind this approach is that the dynamic algorithm gives the more rigorous solution since the magnetization between stable states must follow a physically reasonable path dictated by the LLG equation of motion; however, it is relatively slow compared to the CG method. In this combination algorithm, we use the CG algorithm to generate rapidly an initial guess for the magnetic structure, which is then put into the dynamic solver. This increases the efficiency of the algorithm by roughly an order of magnitude compared to the dynamic solver alone. For the case of ideal SD assemblages each grain is represented by a simple cube, that is, each cube represents the averaged magnetization direction of many hundreds of atomic magnetic dipole moments, or simply each cube is an ideal SD grain. The orientation of each magnetic grain can vary in direction. The grain assemblage structure is initially calculated with the CG algorithm by minimizing the total magnetic energy Etot, which is the sum of magnetostatic energy EA and the anisotropy £anis (Brown 1963). Etoi is calculated using fast-Fourier transforms (FFT), to give a local energy minimum (LEM) for the assemblage. The calculation of the energy terms and the implementation of the FFT are exactly the same as in the work of Wright et al. (1997). After the LEM state has been estimated, the structure is optimized using the LLG algorithm. In effect, instead of minimizing the total energy, the LLG solver minimizes the torque on each magnetic moment by calculating the total effective field. Since the LLG solver ensures that the magnetization path between stable states is physically realistic the method is less susceptible to becoming trapped in a 'false' LEM state. Additionally the LLG convergence criterion is more rigorous since it examines the torque on each magnetization vector rather than simply the gradient of the total energy. For the PSD models, to model the nonuniform internal structures each grain is represented by 7 x 7 x 7 cubes or cells compared to 1 in the ideal SD case, and the exchange interaction within each grain accommodated, i.e. the

exchange interaction between the 7 x 7 x 7 cells representing each grain.

Modelling distribution anisotropy There are several different types of magnetic remanence and consequently several different types of AMR. The most commonly examined is AARM, as it is rapidly, isothermally induced and it is possible to access distinct fractions of the coercivity distribution. However, AARM is practically impossible to model using a micromagnetic approach due to the difficulty in simulating the alternating field. Instead we have considered the anisotropy of isothermal remanence (AIRM), which is utilized when the samples are too weak to consider using AARM (Stephenson et al 1986; Potter & Stephenson 1988). As a first approach to resolving the initial starting-state problem that exists for nonsaturating IRM models, we have considered anisotropy of saturation IRM (ASIRM). For linear anisotropic magnetizations like AARM or thermoremanent AMR the directional variability is mathematically described by a second-rank symmetric tensor (Tarling & Hrouda 1993). This tensor can be represented geometrically as a triaxial ellipsoid, with the principal axes parallel to the eigenvectors of the second-rank tensor and with the principal axis lengths equal to the corresponding eigenvalues of the tensor. However, for strong field remanences like SIRM, the relationship between the magnetization and field is non-linear, meaning that the magnetization and the field strength are not related by a second-rank tensor. It might be argued then, that modelling ASIRM is not the best choice of remanences to consider; however, from a computational point of view, the other types are impractical. The distribution anisotropy is treated in a similar manner to that reported in similar AMS studies (e.g. Stephenson 1994; Canon-Tapia 1996; Gregoire et al. 1998). We consider both lineation (chains) and foliation (planes/grids) in assemblages of ideal SD and PSD grains. For the SD models we considered three singlegrain anisotropy (SGA) distributions: 'aligned' (AR) and 'side-by-side' (SBSR) regimes (Fig. 1), as well as randomly distributed SGA regimes (RR). The AR and SBSR configurations are the extreme cases and were considered in previous AMS models (Stephenson 1994; Canon-Tapia 1996, 2001). For the PSD models only AR and SBSR were considered. To conform with previous anisotropy studies we consider uniaxial anisotropy. Most previous

AMR AND DISTRIBUTION ANISOTROPY

Fig. 1. Schematic showing the lineation (chain) arrangement in the model. Two single-grain anisotropy distributions are shown: AR and SBSR. Each grain has a uniaxial anisotropy. For AR the anisotropy is in the same direction as the chain, for SBSR the anisotropy direction is perpendicular to the chain direction. Random anisotropy distributions were also modelled, though a schematic of this is not depicted. Only part of the chain is shown.

studies have used elongation ratios to describe the uniaxial anisotropy in their models (e.g. Stephenson 1994; Canon-Tapia 1996, 2001). In order to maintain the computational efficiencies of the FFT used in the CG algorithm, rather than varying the elongation ratio we added an additional energy term of the form /sanis = K\j sin2 0, where 0 is the angle between the elongation axis and the magnetization and K\j is a parameter related to the elongation ratio q of the grain. The value of K\j is determined by using the standard formula Kv = ^0M|7V(^)/2, where ^0 is the permeability of free space, Ms the spontaneous magnetization and N(q) is the demagnetizing factor which is simply a function of
Lineation for uniform SD grain assemblages Lineation due to distribution anisotropy is where the interaction field is primarily in one direction

39

only (Tarling & Hrouda 1993). Such interaction fields are often modelled as simple chain of grains (Stephenson 1994; Canon-Tapia 1996). We have modelled ASIRM for chains of 25 identical grains with identical interaction spacing. It has been shown experimentally that chain lengths of ~10 particles are as a first approximation sufficient to represent infinite chains (Fuller 1961). For AR, the magnetostatic interaction field (H^ and the SGA field (//SGA) are parallel, whereas for SBSR //t and //SGA are perpendicular (Fig. 1). HI is in the direction of the chain. Due to symmetry, when calculating ASIRM for AR only two field directions, i.e. Fx and Fy, were determined as ASIRM in the y and z directions, i.e. ASIRMARy and ASIRMARz, are identical (Fig. 1). The remanence tensor is calculated by simulating a SIRM in three perpendicular directions and measuring the intensity of the component of remanence that is acquired parallel to the applied field. For ASIRMARx, ASIRMARy and ASIRMARz the magnetic moments are aligned in the direction parallel to x9 regardless of interaction spacing d/r or q (Fig. 2a). Both HI and //SGA are parallel and have the same effect, i.e. rotating the magnetic moments into the x direction. For fields applied in the y and z directions, i.e. Fy and Fz, the particles rotate into the x direction as the field is decreased from saturation. The shape parameter T is equal to -1 for this highly anisotropic configuration. T is given by (Jelinek 1981), (1)

where K^ K2 and K^ are the maximum, intermediate and minimum values of SIRM for the three calculated directions. If -1 < T < 0, then the system shows lineation (prolate), whereas, if 0 < T < 1, then the sample displays foliation (oblate) (Jelinek 1981; Hrouda 1982). In contrast to the AR case, the SBSR is strongly influenced by interactions (Fig. 2b); //r and //SGA are perpendicular. At large interaction spacing (approaching the true non-interacting state), i.e. d/r —»0, SGA dominates, and ASIRM is close to 1 for Fz (ASIRMSBSRz), and 0 for Fy and Fx. However, as the interactions start to dominate ASIRMSBSRz —»0 and ASIRMSBSRx —> 1. ASIRMSBSRy remains close to 0 and is invariant to d/r. The interaction spacing at which HI starts to dominate depends on the size of KU9 which is proportional to the size of the anisotropy effective field. The larger q, the less the magnetic structure is affected by the interaction field. We have not considered the extreme situation where Ku = 0; however,

40

A. R. MUXWORTHY & W. WILLIAMS

Fig. 2. SIRM versus d/r for chains of ideal SD grains with three different single-grain anisotropy distributions (a) AR, (b) SBSR and (c) RR. Three single-grain anisotropies are shown on each diagram; these are related to the elongation ratio q. q — 1.1 has the smallest anisotropy, and q — 1.5 largest. For each single-grain anisotropy distribution, SIRM was calculated for a field simulated in the x, y and z directions, i.e. Fxt Fy and Fz. Due to symmetry considerations, some field directions were identical. For simplicity and clarity in (b) the Fy data is not shown; it equalled ~0 for all d/r. When d/r — 0 the grains are non-interacting, and when d/r = 1 the particles are physically touching.

AMR AND DISTRIBUTION ANISOTROPY

Fig. 3. Shape parameter T versus d/r for the SD SBSR lineation (chain) model. Three single-grain anisotropies are shown on each diagram, i.e. q = 1.1, 1.3 and 1.5. When T < 0, the assemblage has a prolate signature, and when T > 0 an oblate one. The aligned and random models both gave a shape parameter of T = — 1 for all d/r.

in this case for ideal SD grains the interaction field would control the behaviour even at d/r = 0. For non-interacting particles, i.e. d/r — 0, the linear //SGA dominates the magnetic signal and T = — 1 (Fig. 3). For strongly interacting systems, i.e. d/r > 0.8, the linear H{ dominates and T is also equal to -1. However, for intermediate states, HI and //SGA are m competition with each other, and depending on the balance of the energies, it is possible for T > 0, i.e. an oblate signal. This is seen for q = 1.1 (Fig. 3). The behaviour of the randomly distributed anisotropy regime (RR) is markedly different, primarily because //SGA nas no preferred direction. For large interaction spacings, i.e. d/r —> 0, ASIRMRR approaches ~0.5, i.e. the expected MRS/MS value for randomly distributed non-interacting SD grains (Dunlop & Ozdemir 1997). As the assemblage is only for 25 grains the small differences between 0.5 and the calculated values can be attributed to the low number of grains not giving a true random distribution. As d/r —> 1, the moments align in the direction of H\, i.e. ASIRM RRx —>• 1 and ASIRMRRy and ASIRMRRz -»0. Note ASIRMRRy and ASIRMRRz are symmetric as in the AR case. Again there is a dependency on q. The smaller q, the more the particles are influenced by HI. The shape parameter T is -1 for all values of d/r > 0. A random assemblage at d/r = 0 is isotropic, for which T is not defined. For the random regime the Jelinek's (1981) corrected anisotropy degree Pj (sphere to ellipsoid; 1 —» oo) is plotted as a function of d/r in Figure 4. Clearly the degree of anisotropy

41

Fig. 4. Degree of anisotropy Pj for a random assemblage of uniform SD grains in a chain (lineation). Three single-grain anisotropies are shown on each diagram; these are related to the elongation ratio q. q— 1.1 has the smallest anisotropy, and q = 1.5 largest.

increases as the d/r increases. This effect is most pronounced for q — 1.1.

Foliation for uniform SD grain assemblages In a similar set of simulations, foliation was examined by simulating ASIRM for grids of 10 x 10 arrays of SD particles. In the foliation model the direction of the HI is more localized and variable than in the lineation system; it lies within the x-y plane rather than along an axis. AR is now for anisotropy aligned in either the x or y directions, and SGSR is for SGA out of the x-y plane, i.e. in the z direction (Fig. 5). Generally the effect of interactions is to decrease ASIRM for anisotropies parallel to the field, and to cause small increases in ASIRM for anisotropies perpendicular to the fields (Figs 6a and 6b), for example ASIRMSBSRz decreases from 1 to 0 with increasing interactions, while ASIRMSBSRx and ASIRMSBSRy increase slightly from 0. In the AR model the H{ and //SGA are within the same plane. The effect of interactions is more pronounced in SBSR model, where HI rotates the moments into the x-y plane, i.e. the hard uniaxial SGA plane (Fig. 6b). Similarly to the lineation models, the effect of interactions is highly dependent on q. For q = 1.1, ASIRMARy increases slightly as d/r increases from 0.8 to 1. This increase is repeatable, and is due to the structure being unable to break the symmetry for large H\. The behaviour of the random assemblage is markedly different (Fig. 6c). For small d/r, ASIRM approach ~0.5 for Fx, Fy and Fz as

42

A. R. MUXWORTHY & W. WILLIAMS

Fig. 5. Schematic showing the foliation (grid) arrangement in the model. Two single-grain anisotropy distributions are shown: AR and SBSR. For AR the anisotropy is in the same plane as the grain distribution, for SBSR the anisotropy direction is perpendicular to this plane. Random anisotropy distributions were also modelled, though a schematic of this is not depicted. Only part of the grid is shown.

expected for randomly distributed non-interacting uniaxial SD particles. For the out-of-plane field FZ9 the effect of interactions is to decrease ASIRMRRz. ASIRMRRz -> 0 as d/r -> 1. Interactions initially increase both ASIRMRRx and ASIRMRRy as d/r increases, due to the rotation of the moments into the x-y plane. As 1 ~ 0, both ASIRMRRx and ASIRMRRy then decrease to zero as d/r —> 1 ~ 0, giving rise to a peak in ASIRMRRx and ASIRM RRy at intermediate values of d/r. These peaks are clearly dependent on q. This decrease in ASIRMRRx and ASIRMRRy as d/r-*l, is due to the rearrangement of the moments in the x-y plane as they respond to the magnetostatic interaction field. As d/r —> 1 some moments will lie parallel and others anti-parallel to the applied field direction in the x-y plane, giving a net magnetization of ~0. This contrasts sharply with models of susceptibility where in-plane and mean susceptibilities continue to increase (Canon-Tapia 1996). For the susceptibility

model of Canon-Tapia (1996), the effect of interactions was to decrease the micro-coercivity of each grain, which in turn led to an increase in the susceptibility. This difference between remanence and susceptibility is due to their intrinsically different nature; one is a zero-field state and the other an in-field state. As d/r increases, the more planar H\ starts to dominate over the more linear //SGA- Consequently for AR and SBSR T switches from negative to positive, i.e. from a prolate to an oblate signal (Figs 6d and 6e). For RR, T = 1 for all d/r > 0 due to //SGA being isotropic and HI having high-planar symmetry (Fig. 6f). A random assemblage at d/r = 0 is isotropic, for which T (equation (1)) is not defined, though it could be argued that it should take the value zero. For RR with q= 1.1, Pj exhibits a peak at d/r ~ 0.6 (Fig. 7). This peak is caused by the weak SGA field being unable to break the symmetry of the magnetic structure as HEXT decreases, resulting in large values of ASIRMRRx and ASIRM RRy and a peak in Pj. RR with q = 1.3 and 1.5 displays the same gradual increase with increasing d/r as in the lineation model (Fig. 4). The magnitude of Pj is less. Distribution anisotropy and non-uniform magnetic structures In the previous section it was assumed that each particle has a uniform internal magnetic structure. This assumption is valid for very small grains. However, larger grains of soft magnetic minerals like magnetite have complicated nonuniform magnetic structures. As the SD to multidomain (MD) or PSD grain size threshold d0 is approached, SD grains display non-uniform flower structures (Fig. 8a), while in magnetite the smallest MD grains display vortex structures (Fig. 8b). The magnetic behaviour of these vortex structures is significantly different to SD grains (Williams & Dunlop 1989). We investigate their AMR here. The size of a non-uniform grain is critical to its magnetic behaviour. We model only one size of grain: cubic grains with a length of 90 nm. This size was carefully selected, because it is known that the magnetic structure of grains just above 4) are very strongly influenced by magnetostatic interactions (Muxworthy et al. 2003). For 90 nm grains the vortex structure is the lowest energy state in a zero-field environment. However, in the presence of small external fields (either HEXT or H1), the SD or flower state becomes

AMR AND DISTRIBUTION ANISOTROPY

43

Fig. 6. SIRM versus d/r for grids (foliation) of ideal SD grains with three different single-grain anisotropy distributions (a) AR (SGA in x direction), (b) SBSR and (c) RR, and the corresponding shape-parameter T versus d/r for (d) AR, (e) SBSR and (f) RR. Three single-grain anisotropies are shown on each diagram, i.e. q = 1.1, 1.3 and 1.5. For each single-grain anisotropy distribution, SIRM was calculated for a field simulated in the x, y and z directions, i.e. Fx, Fy and Fz. Due to symmetry considerations, some field directions were identical.

energetically more favourable. Muxworthy et al. (2003) showed that magnetostatic interactions effectively increase d0. Although we are modelling only the smallest MD grains, by considering the role of interactions in assemblages of such grains just above 4), we are arguably examining MD grains, which are most strongly influenced by magnetostatic interactions. Modelling both the internal structure as well as the interaction fields makes the problem

computationally more intensive than for uniform SD models. This significantly limits the number of grains and variations in d/r that can be modelled. This type of computational limit is not uncommon. In previous numerical models, which assessed the role of interactions on AMS, it was commonly assumed that calculating only the magnetic signal of a grain in the middle of a short chain was sufficient to predict the behaviour of an infinite chain. This approximation is not applicable to models for ASIRM. For

44

A. R. MUXWORTHY & W. WILLIAMS

Each grain is modelled with a resolution of 7 x 7 x 7 cells. This is a little lower than that predicted to be the minimum to model magnetite-like cubes 90 nm in size, i.e. 13 x 13 x 13 (Rave et al. 1998). By implementing a slightly lower resolution, we potentially underestimate the exchange energy. However, as we are primarily interested in changes with interaction spacing, our model will still show the same trends.

Lineation of non-uniform magnetic structures Fig. 7. Degree of anisotropy Pj for a random assemblage of uniform SD grains in a grid pattern (foliation). Three single-grain anisotropies are shown on each diagram; these are related to the elongation ratio q. q = 1.1 has the smallest anisotropy, and q = 1.5 largest.

example, if one considers a SIRM for a 3 x 3 grid of grains, then for certain interaction spacings and field orientations, the magnetic moment of the middle grain is reversed in direction with respect to the applied field, i.e. a negative SIRM, yet the entire assemblage has a positive SIRM. Clearly, using the behaviour of the middle grain to predict the behaviour of a larger assemblage is not valid for SIRM models. For our lineation model, chains of 15 grains were modelled, while for foliation model grids of 5 x 5 grains were considered. Due to the small number of grains within the models only AR and SBSR are considered.

The internal magnetic structures and consequent SIRM values of the grains within the chains are strongly influenced by interactions. For example, consider the normalized ASIRMAR for q = 1.1: ASIRMARx = 0.17 for d/r = 0.5, but for d/r = 0.875, ASIRMARx = 0.95. This is because for d/r = 0.5 all the particles are in a vortex state, which has a lower remanence, but as the interactions increase the particles reside in a flower structure (Fig. 9a); HI increases d0 to >90nm, and ASIRMARx similarly increases. Similar effects are seen in the domain structures of all the models. For non-interacting particles, i.e. d/r = 0, the linear //SGA dominates the magnetic signal and T = — 1 (Fig. lOa). For strongly interacting systems, i.e. d/r > 0.875, the linear H{ dominates and T is also equal to -1. However, for intermediate, states, for the SBSR HI and //SGA are in competition with each other, and it is possible for T > 0, in a similar manner as seen for the SBSR SD particles (Fig. 3). This is seen for both
Fig. 8. Domain states occurring in cubic grains of magnetite-like minerals at room temperature for a grain with edge length of 90 nm (a) single domain (flower state) and (b) single vortex state. In this paper the term 'SD state' refers not just to homogeneous magnetization structures as in Neel theory, but also to non-uniform domain structures as shown in (a), which are basically SD-like with a degree of flowering towards the edges of the grain.

AMR AND DISTRIBUTION ANISOTROPY

45

Fig. 9. Two middle-segments from the lineation model, showing the contribution of magnetostatic interactions to remanence domain structures of an aligned regime. The upper chain has a closer interaction spacing (d/r = 0.875), the lower has a larger spacing and d/r — 0.5. For d/r = 0.875 the grains are clearly in the flower state (Fig. 7a) and for d/r = 0.5 the grains are all in a vortex state (Fig. 7b). The effect of interactions is to increase dQ. The direction of the applied field (Fx) and the single-grain anisotropy field (AR) are also highlighted. The grains depicted are grains 7, 8 and 9 from a chain of 15.

Ideally, more points are needed in Figure lOa. However, this problem is computationally intensive. Nevertheless, the trends can still be seen.

Foliation of non-uniform magnetic structures As in the uniform SD models, HI lies within the x-y plane for the foliation scheme (Fig. 4). Similar to the lineation model, the interaction field effectively increases dQ making the flower state more favourable than the vortex state as d/r —>• 1. However, this does not affect the magnitude of the SIRM as it does for the lineation model since the domain state and domain orientation can change. For example, consider the zero-field state for the aligned assemblage with q= 1.3, with the field applied in the ^-direction, i.e. ASIRMARx. For d/r — 0.5 all 25 grains are in a vortex state, whereas for d/r = 0.875 each grain is in a flower state, though the reduced remanences are quite similar: ASIRMARx = 0.13 for d/r = 0.5 and 0.19 for d/r = 0.875. Even though the SIRM magnitude does not change significantly, the symmetry of the system is strongly affected. As d/r increases, the planar HI dominates over the linear #SGA> and the shape parameter switches from —1 to >0 (Fig. lOb). However, d/r = 0.875 for the SBSR model, T < 0. It is believed that this is an artefact of the small assemblage size, i.e.

only 25 grains, and that for a larger assemblage size or a less symmetric distribution that T > 0. Discussion and conclusions The situation for the foliation regimes is easily summarized. For systems with a linear anisotropic //SGAJ regardless of the exact anisotropy distribution as d/r increases T will increase from negative to positive values as the planar HI starts to dominate. The interaction spacing at which T becomes positive depends on the size and orientation of T/SGA and HI - These two fields are in direct competition with each other. If //SGA *s is°tropic, i.e. a random distribution, then it is shown that for the SD assemblage for d/r > 0, T > 0. How would a random distribution of non-uniform grains (or PSD) grains behave? Unfortunately it was not possible to model large distributions to obtain genuine random distributions, however, it is possible to predict their behaviour qualitatively. For foliation, the behaviour is quite simple, for d/r > 0, then T would also be >0, due to the planar HI. For the lineation models, the behaviour is a little more complicated. Assuming some linear anisotropic //SGA» then f°r small d/r, T < 0, similarly for large d/r, due to the linearity of HI, T < 0. However, at intermediate states depending on d/r, anisotropy, relative directions

46

A. R. MUXWORTHY & W. WILLIAMS

Fig. 10. Shape parameter T versus d/r for the PSD models: (a) lineation (chain) and (b) foliation model. Two single grain anisotropies are shown (q = 1.1 and 1.3). For the SBSR model for d/r = 0.875, T < 0. It is believed that this is an artefact of the small assemblage size, and for larger assemblage models T would be larger > 0.

of //SGA and HI , it is possible to obtain values for T > 0. That is, what would appear to be a highly linear system, can yield a foliation signal. It can be hypothesized for a random assemblage of PSD grains that T would similarly always be <0. For both the lineation and foliation models the size of Pj roughly increases with increasing d/r. There are no direct experimental results to which to compare the model results. However, Gregoire et al. (1995) measured the shape parameter for AMS for two large MD magnetite crystals. They found that for an SBS chain configuration T could switch from T < 0, to T > 0 and then back to T < 0 as d/r increased towards 1, in a similar manner to that seen in Figures 3 and 9a for AMR. It is difficult to draw too strong comparisons due to the differences in grain size and between AMS and AMR. ASIRM is not commonly measured due to the desire to measure a linear tensorial regime. A question that might be raised after reading this paper is: what can be speculated from these results about the role of interactions on the

more commonly measured AARM? As the alternating field decreases, there will be some critical point dependent on d/r, where the interaction field dominates over the alternating field, and the behaviour of the assemblage controlled by this interaction field. For highly interacting regimes, i.e. where d / r — > l , the interactions will dominate the system at large alternating fields and the results will be very similar to those for ASIRM. For weak and moderate interacting regimes, the role of interactions is a little harder to predict, due to the non-linearity of the problem. It is likely that the features found for the ASIRM are less pronounced for AARM, as the symmetry seen in the highly uniform ASIRM model structures will be broken by the alternating field resulting in the 'transition' from non-interacting to interacting behaviour occurring at higher values of d/r. The natural extension of this study would be to model the role of interactions on the anisotropy of non-saturating IRM, and to assess the role of the initial starting state of the assemblage on the behaviour.

AMR AND DISTRIBUTION ANISOTROPY We thank both M. Jackson and H. de Wall for independently suggesting the concept of this paper, and M. Fuller, B. Moskowitz and M. Jackson for constructive reviews. This work was funded through NERC grant NER/A/S/2001/00539 to W.W.

References AUBOURG, C. & ROBION, P. 2002. Composite ferromagnetic fabrics (magnetite, greigite) measured by AMS and partial A ARM in weakly strained sandstones from western Makran, Iran, Geophysical Journal International, 151, 729-737. BROWN, W. F., JR. 1963. Micromagnetics, John Wiley, New York. CANON-TAPIA, E. 1996. Single-grain versus distribution anisotropy: a simple three-dimensional model, Physics of the Earth and Planetary Interiors, 94, 149-158. CANON-TAPIA, E. 2001. Factors affecting the relative importance of shape and distribution anisotropy in rocks: theory and experiments, Tectonophysics, 340,117-131. DUNLOP, D. J. & OZDEMIR, 6. 1997. Rock Magnetism: Fundamentals and Frontiers, Cambridge University Press, Cambridge. FULLER, M. D. 1961. Magnetic Anisotropy of Rocks, Ph.D. Thesis, University of Cambridge. GREGOIRE, V., BLANQUAT, M. D., NEDELEC, A. & BOUCHEZ, J. L. 1995. Shape anisotropy versus magnetic-interactions of magnetite grains - experiments and application to AMS ingranitic-rocks, Geophysical Research Letters, 22, 2765-2768. GREGOIRE, V., DARROZES, J., GAILLOT, P., NEDELEC, A. & LAUNEAU, P. 1998. Magnetite grain shape fabric and distribution anisotropy vs rock magnetic fabric: a three-dimensional case study, Journal of Structural Geology, 20, 937-944. MARGRAVES, R. B., JOHNSON, D. & CHAN, C. Y. 1991. Distribution anisotropy: the cause of AMS in igneous rocks?, Geophysical Research Letters, 18, 2193-2196. HROUDA, F. 1982. Magnetic anisotropy of rocks and its application in geology and geophysics, Geophysical Surveys, 5, 37-82. JACKSON, M. 1991. Anisotropy of magnetic remanence - a brief review of mineralogical sources, physical origins, and geological applications, and comparison with susceptibility anisotropy, Pure Applied Geophysics, 136, 1-28.

47

JACKSON, M. J., GRUBER, W., MARVIN, J. A. & BANERJEE, S.K. 1988. Partial anhysteretic remanence and its anisotropy: applications and grainsize dependence, Geophysical Research Letters, 15, 440-443. JELINEK, V. 1981. Characterization of the magnetic fabrics of rocks, Tectonophysics, 79, T63-T67. MUXWORTHY, A. R., WILLIAMS, W. & VIRDEE, D. 2003. The effect of magnetostatic interactions on the hysteresis parameters of single-domain and pseudo-single domain grains, Journal of Geophysical Research, 108, B11, 2517. NAGATA, T. 1961. Rock Magnetism., revised edition, Mazuren Company Ltd, Tokyo. POTTER, D. K. & STEPHENSON, A. 1988. Single-domain particles in rocks and magnetic fabric analysis, Geophysical Research Letters, 15, 1097-1100. RAVE, W., FABIAN, K. & HUBERT, A. 1998. The magnetic states of small cubic magnetic particles with uniaxial anisotropy, Journal of Magnetism and Magnetic Materials, 190, 332-348. STEPHENSON, A. 1994. Distribution anisotropy: two simple models for magnetic lineation and foliation, Physics of the Earth and Planetary Interiors, 82, 49-53. STEPHENSON, A., SADIKUN, S. & POTTER, D. K. 1986. A theoretical and experimental comparison of the anisotropies of magnetic susceptibility and remanence in rocks and minerals, Geophysical Journal of the Royal Astronomical Society, 84, 185-200. STONER, E. C. & WOHLFARTH, E. P. 1948. A mechanism of magnetic hysteresis in heterogeneous alloys, Philosophical Transactions of the Royal Society of London, Series A, 240, 599-602. SUESS, D., TSIANTOS, V., SCHREFL, T., FlDLER, J., SCHOLZ, W., FORSTER, H., DlTTRICH, R. &

MILES, J. 2002. Time resolved micromagnetics using a preconditioned time integration method, Journal of Magnetism and Magnetic Materials, 248,298-311. TARLING, D. H. & HROUDA, F. 1993. The Magnetic Anisotropy of Rocks, Chapman & Hall, London. WILLIAMS, W. & DUNLOP, D. J. 1989. Three-dimensional micromagnetic modelling of ferromagnetic domain structure, Nature, 337, 634-637. WRIGHT, T. M., WILLIAMS, W. & DUNLOP, D. J. 1997. An improved algorithm for micromagnetics, Journal of Geophysical Research, 102, 1208512094.

This page intentionally left blank

Problems in interpreting AMS parameters in diamagnetic rocks FRANTISEK HROUDA AGICO Inc., Jecnd 29a, Box 90, CZ-621 00 Brno, Czech Republic (e-mail: [email protected]) and Institute of Petrology and Structural Geology, Charles University, Albertov 6, CZ-128 43 Praha, Czech Republic Abstract: The AMS parameters widely used to characterize the magnetic fabric of rocks (L, F, P, T) are not directly applicable to diamagnetic rocks. Namely, there are two principal methods of their definition: (1) from the signed principal susceptibilities or (2) from the absolute values of the principal susceptibilities. The first method results in L, F and P values less than one in contrast to materials whose susceptibility arises from paramagnetism and/or ferromagnetism sensu lato, where they are greater than one by definition. The ratio parameters decrease with increasing intensity of alignment of the magnetic minerals; the calculated shape parameters are, however, correct. The second method gives L, F and P values greater than one; they increase with increasing preferred orientation, however the ellipsoid shape parameters are inverse. The development of AMS parameters with increasing preferred orientation is modelled mathematically for quartzite/evaporite (para- and/or ferromagnetic AMS superimposed on the isotropic diamagnetic matrix) and for marble/limestone (preferred orientation of calcite in a rock). A rational set of AMS parameters is then recommended.

Modern instruments for measuring the anisotropy of magnetic susceptibility (AMS) of rocks (e.g. the KLY-4S Kappabridge) are highly sensitive and are precise enough to be able to measure with sufficient accuracy even diamagnetic rocks such as quartzites, evaporites, marbles and limestones (e.g. Borradaile et al. 1999; de Wall et al. 2000; Hrouda et al. 2000). Examples are shown in Figures 1 and 2 in which the principal directions of individual specimens either create cluster or girdle patterns that are related to the mesoscopic fabric elements. In quartzites and evaporites, the whole-rock AMS is controlled by the fabric of paramagnetic or ferromagnetic admixtures, because quartz and salt are virtually isotropic magnetically (Nye 1957; Owens & Bamford 1976; Bleil & Petersen 1982; Hrouda 1986; Borradaile et al. 1999; Hrouda et al. 2000). In marbles and limestones, the rock AMS can be substantially affected by the preferred crystallographic orientation of calcite (Krishnan et al. 1933; Owens & Rutter 1978; de Wall et al. 2000). Whereas the interpretation of the directions of the principal susceptibilities is more or less straightforward, problems arise in the interpretation of magnetic parameters characterizing the eccentricity and shape of the susceptibility ellipsoid. The purpose of the present paper is to show some of these problems and to suggest ways of solving them.

Approaches to describe the AMS of diamagnetic rocks There are two possible approaches to describing the AMS of diamagnetic rocks. The first one orders the signed magnitudes of the principal susceptibilities (Fig. 3). With this approach, the largest absolute magnitude is the minimum susceptibility, while the smallest absolute magnitude is the maximum susceptibility. The parameters used most frequently in the AMS studies are defined as follows (e.g. Nagata 1961; Hrouda 1982; Jelinek 1981; Tarling & Hrouda 1993):

(1) (1) shape parameter where k\ > k2> k3 are the principal susceptibilities. As the principal susceptibilities of diamagnetic rocks are negative, the T parameter cannot be calculated using the above formula. This can be rearranged into the form enabling work with negative principal susceptibilities (2)

From: MARTIN-HERNANDEZ, F., LUNEBURG, C. M., AUBOURG, C. & JACKSON, M. (eds) 2004. Magnetic Fabric: Methods and Applications. Geological Society, London, Special Publications, 238, 49-59. 0305-8719/04/S 15.00 © The Geological Society of London 2004.

50

F. HROUDA

Fig. 1. Anisotropy of magnetic susceptibility in diamagnetic evaporite in a locality of the Hormoz Island (Persian Gulf), (a) Plot of the degree of AMS, Pa, vs. mean susceptibility, Km, (b) Orientation of magnetic lineations (k\, closed square), magnetic foliation poles (£3, closed circle), and mesoscopic foliation poles (open circle). Equal-area projection on lower hemisphere.

In diamagnetic rocks, the P, L and F values are always less than 1 in contrast to the paramagnetic and/or ferromagnetic sensu lato rocks, and they decrease with increasing eccentricity of the susceptibility ellipsoid. The reformulated shape parameter (7") correctly describes the shape of the susceptibility ellipsoid. A second approach considers the absolute (nonsigned) values of the principal susceptibilities (Fig. 3). The strongest diamagnetic susceptibility, with the largest absolute magnitude, corresponds to the maximum susceptibility, while the minimum susceptibility corresponds to the smallest absolute magnitude susceptibility (Kv = \k3\ >K2 = \k2\ >K3 = |fc,|). The AMS parameters are then

(3)

The Pa9 La and Fa values are always higher than

AMS PARAMETERS IN DIAMAGNETIC ROCKS

51

Fig. 2. Anisotropy of magnetic susceptibility in diamagnetic limestone (locality of Veselice in the Moravian Karst, Bohemian Massif), (a) Plot of the degree of AMS, Pa, vs. mean susceptibility, Km, (b) Orientation of magnetic lineations (fc1? closed square), magnetic foliation poles (k3, closed circle), and mesoscopic foliation poles (open circle). Equal-area projection on lower hemisphere. 1, and they increase with increasing eccentricity of the susceptibility ellipsoid. The values of the shape parameter unfortunately indicate the inverse ellipsoid susceptibility shape with respect to the real one.

The quartzite/evaporite model The AMS of quartz is very weak. Nye (1957) presents values of —15.1 x 10~6 [SI] along the c axis and —15.2 x 10~6 along the ab plane. Hrouda

Fig. 3. Possible definitions of the AMS of diamagnetic rocks.

52

F. HROUDA

(1986), who investigated 19 single crystals of four quartz varieties, found the mean susceptibility to be —13.44 x 10~6 with standard deviation being 0.80 x 10~6. The degree of AMS of these crystals was very low, with the mean value 1.014 ± 0.012. However, in some cases the absolute value of the susceptibility along the c (trigonal) axis was slightly higher than that in the ab plane and, in other cases it was lower. The mean value of the ratio of the susceptibility in the ab plane to that along the c axis was 1.000. The AMS of quartz single crystals is considered to be near isotropic, which is in agreement with Nye (1957) and Owens & Bamford (1976). Even though the AMS of salt has not been investigated experimentally, it can be assessed theoretically through the application of Neumann's principle, which requires that the physical property includes the symmetry elements of the point-group of the crystal (e.g. Nye 1957). In another words, the symmetry of the physical property must be, in terms of point-group, the same as or higher than the symmetry of the crystal. The symmetry of salt is cubic. Among symmetries exhibited by symmetric second-rank tensors only one obeys the Neumann principle and this is the spherical symmetry (e.g. Shubnikov 1949; Nye 1957; Hrouda 1973) indicating isotropy. One can therefore deduce that salt is isotropic magnetically. Bleil and Petersen (1982) present the empirical mass susceptibility value -0.376 x 10~6 cm3 g"1, which corresponds to the bulk susceptibility -10.3 x 10"6 [SI], using a salt density of 2 170kgm~ 3 . The rock susceptibility can be described with sufficient accuracy by the following model (Henry 1983; Henry & Daly 1983): (4)

where k is the rock susceptibility tensor, K f , K p , Kd are tensors of ferromagnetic (sensu lato), paramagnetic and diamagnetic susceptibilities, respectively; Cf, cp, Q are the respective percentages; and k f , kp, kd are called the respective susceptibility contribution tensors. If we consider ferro-/paramagnetic admixtures in an isotropic quartzite/evaporite diamagnetic matrix, the susceptibility of the quartzite/evaporite model can be written as follows: (5)

where k f / p is the susceptibility contribution tensor of the ferro-/paramagnetic admixture and k& is the isotropic diamagnetic susceptibility of the matrix. It is obvious from equation (5), especially if considered in the coordinate system of principal directions, that the signed principal axes of

quartzite/evaporite correspond to those of the ferro-/paramagnetic admixture. For example, if the minimum susceptibility of the ferro-/paramagnetic admixture is perpendicular to the mesoscopic foliation of quartzite/evaporite and the maximum susceptibility is parallel to the mesoscopic lineation, the signed minimum susceptibility of quartzite/evaporite is also perpendicular to the mesoscopic foliation and the signed maximum susceptibility of quartzite/evaporite is also parallel to the mesoscopic lineation. Hence, it is advantageous to work with signed principal susceptibilities in plotting principal directions, because they correspond to those of the anisotropic component of quartzite/evaporite, namely the ferro-/paramagnetic admixture. The variation in the degree of AMS for quartzite/evaporite with the degree of AMS for the ferro-/paramagnetic admixture was modelled using approach 1 (Fig. 4a) and approach 2 (Fig. 4b). The mean susceptibility of the matrix was kd = —15 x 10~6 and the mean susceptibility of the ferro-/paramagnetic admixture was allowed to vary from +8 x 10"6 to +2 x 10~6. In approach 1, the degree of AMS of quartzite/ evaporite is less than 1 and it decreases with increasing degree of AMS of ferro-/paramagnetic admixture. In approach 2, the degree of AMS of quartzite/evaporite is higher than 1 and increases with increasing degree of AMS of ferro-/paramagnetic admixture. In addition, a variation of the shape parameter of quartzite/evaporite was modelled with the shape parameter of the ferro-/paramagnetic admixture for approach 1 (Fig. 5a) and approach 2 (Fig. 5b). In approach 1, the shape parameter of quartzite/evaporite is very near the shape parameter of the ferro-/paramagnetic admixture, even though both parameters are not precisely equal (this follows from eqn (5)). In approach 2, the shape of the susceptibility ellipsoid of quartzite/evaporite is indicated to be roughly inverse to that of the ferro-/paramagnetic admixture. The marble/limestone model Calcite is clearly anisotropic magnetically. Nye (1957) presents susceptibility values of -13.8 x 10~6 [SI] along the c axis and — 12.4 x 10~6 along the ab plane, giving an AMS degree Pfl = l.ll (Daly & d'Albissin 1968). Owens & Rutter (1978) recalculated the data by Krishnan et al (1933) and Hellwege & Hellwege (1967) considering calcite density of 2 700 kg nT3, with the degree of AMS thus being Pa = 1.12.

AMS PARAMETERS IN DIAMAGNETIC ROCKS

53

Fig. 4. Model relationship between the whole-rock degree of AMS (P, Pa) and the degree of AMS of para-/ ferromagnetic fraction (Pf) for the quartzite/evaporite model. Numbers of the individual curves denote the mean susceptibility of the para-/ferromagnetic fraction on the order of 1CT6 [SI], (a) whole-rock degree of AMS, P, (b) whole-rock degree of AMS, Pa.

The AMS of the diamagnetic fraction (as indeed of any and all fractions) should be lower than the intrinsic anisotropy of the single crystal under consideration, because perfect alignments are rare (Borradaile & Henry 1997). Normally, the crystallographic axes of individual mineral grains have a considerable scatter in a rock (e.g. Turner & Weiss 1963). The models assume simple distributions of calcite c axes, having either a cluster distribution, a girdle distribution or a cluster/girdle combination. The orientation tensor concept is employed for the models. The orientation tensor is defined as a 3 x 3 matrix of the sums of cross products of the direction cosines of the linear fabric elements

evaluated (see, e.g. Scheidegger 1965). (6)

where //, mh nt are the direction cosines of the f-th linear element represented by unit vector and n is the number of the linear elements considered. The principal values of this tensor (E\ > £2 > ET,) have the following property: (7)

Consequently, E\ > E2 — £3 represents a cluster type of distribution, whereas E\ = E2 > E3 corresponds to a girdle type pattern. In a perfect

54

F. HROUDA

Fig. 5. Model relationship between the whole-rock shape parameter and the shape parameter of para-/ferromagnetic fraction 7} for the quartzite/evaporite model. Individual curves were calculated for the same mean susceptibilities as in Fig. 4. (a) whole-rock shape parameter, T, (b) whole-rock shape parameter, Ta.

cluster, in which all linear elements are parallel, Ei = l, E2 = E3 = 0. In perfect girdle, E{ = E2 = 0.5, E3 = 0. There is a straightforward relationship between the orientation tensor and the susceptibility tensor (Henry 1989; Hrouda & Schulmann 1990; Jezek & Hrouda 2000). For a rock whose AMS is carried by one magnetic mineral with grains of equal magnetic properties and grain AMS being represented by an oblate susceptibility ellipsoid the relationship between the L, Fand P parameters and the principal values of the orientation tensor are as follows (8)

where A = 3(1 - PC)/(2PC + 1) and B = 3PC/(2PC +1), and Pc is the grain degree of AMS calculated from signed values of the principal susceptibilities. It is obvious from equations (8) that in approach 1 the minimum susceptibility is parallel to the maximum value of the orientation tensor and the maximum susceptibility is parallel to the minimum value of the orientation tensor. In approach 2 the maximum and minimum susceptibilities are parallel to the maximum and minimum values of the orientation tensor, respectively. In calcitic tectonites (mostly marbles), the calcite e-twin lamellae are usually parallel to the tectonic foliation resulting in small circle preferred orientation of calcite c axes (e.g. Hejtman 1962). In this case, the magnetic foliation is parallel to the tectonic foliation

AMS PARAMETERS IN DIAMAGNETIC ROCKS

55

Fig. 6. Model relationship between the whole-rock degree of AMS and the intensity of the preferred orientation of the c axes of calcite / for marble/limestone model, (a) whole-rock degree of AMS, P, (b) whole-rock degree of AMS, Pa.

in approach 1 and perpendicular to it in approach 2. Consequently, it is advocated to work with the signed values (approach 1) in plotting principal axis orientations. Variation in the degree of AMS for the marble/limestone was modelled according to the intensity of orientation of calcite c axes for approach 1 (Fig. 6a) and approach 2 (Fig. 6b). The intensity of the orientation of lines can be defined by means of the orientation tensor (Lisle 1989) as follows (9)

It can vary from 0 for isotropic fabrics to 5 for fabrics in which all elements are perfectly parallel

to one another. In approach 1, the degree of AMS of marble/limestone is less than 1 and decreases with increasing intensity of orientation of calcite c axes. In approach 2, the degree of AMS is greater than 1 and increases with increasing intensity of orientation of calcite c axes. In addition, the shape parameter of marble/ limestone was modelled to vary according to the shape factor calculated from the orientation tensor of the c axes for approach 1 (Fig. 7a) and approach 2 (Fig. 7b). In approach 1, the shape parameter of marble/limestone is very near the shape parameter calculated from the orientation tensor of the c axes, even though both parameters are not precisely equal. In approach 2, the relationship between both shape parameters is inverted.

56

F. HROUDA

Fig. 7. Model relationship between the whole-rock shape parameter and the shape parameter calculated Te from the preferred orientation of the c axes of calcite for the marble/limestone model, (a) whole-rock shape parameter, T, (b) whole-rock shape parameter, Ta.

Discussion The preferred orientation of minerals in rocks is traditionally represented by means of density diagrams presented on equal-area projections. In addition, it can be represented purely mathematically, by the orientation tensor introduced by Scheidegger (1965) and then widely used in structural geology (e.g. Woodcock, 1977). There is a straightforward relationship between the orientation tensor and the susceptibility tensor (Henry 1989; Hrouda & Schulmann 1990; Jezek & Hrouda 2000). For determining the orientation tensor of calcite c axes from the AMS data of diamagnetic marble/limestone in order to compare magnetic and non-magnetic data on c axis distribution as made for instance by de Wall et al (2000), the Hrouda &

Schulmann (1990) formulae for rotational oblate grain susceptibility ellipsoids and the grain degree of AMS should be calculated from signed values of the principal susceptibilities. Some specimens of both quartzite/evaporite and marble/limestone may show mixed values of principal susceptibilities, i.e. one negative and two positive or two positive and one negative. In this case, neither approach 1 nor approach 2 is applicable, because the former results in negative values of some AMS parameters and the latter gives rise to incorrect values of the shape parameter. It is therefore recommended to avoid quantitative interpretation of AMS of specimens with mixed principal susceptibilities. Realizing that the mean susceptibility of such specimens is very near zero and accuracy of measurement of specimens with almost zero

AMS PARAMETERS IN DIAMAGNETIC ROCKS

57

Fig. 8. Model relationship between the whole-rock degree of AMS (P) and mean bulk susceptibility (Km) at variable degrees of AMS of the para-/ferromagnetic fraction (Pf) for the quartzite model. For the P parameter, the approach 1 values are plotted if Km > 0, while the approach 2 values are plotted if Km < 0. Note unrealistically high degree of AMS in the vicinity of zero susceptibility. After Hrouda (1986).

susceptibility is in general low, avoiding these specimens is not a real loss. Hrouda (1986), who investigated the quartzite model, showed that the degree of AMS of quartzite is higher than the degree of AMS of the para-/ferromagnetic fraction, if the quartzite susceptibility is positive. If the quartzite susceptibility is negative, the degree of AMS of quartzite is lower than that of the para-/ferromagnetic fraction. If the susceptibility of quartzite is in the vicinity of zero (e.g. in the range between -5 x 1(T6 and 5 x 1(T6 [SI]), the degree of AMS of quartzite can be much higher than that of the para-/ferromagnetic fraction (Fig. 8), and the AMS parameters then have no fabric meaning. Recent investigations have shown that similar effects may exist also in the case of limestone (Fig. 9) in which there is a contribution of calcite with a para-/ferromagnetic fraction. It is therefore recommended to avoid quantitative interpretation of AMS of rocks whose bulk susceptibility is near zero (conveniently in the susceptibility interval between —5 x 10~ and 5 x 1(T6 [SI]). The AGICO programs ANI20BAS, SUSAR, SUSAM, SUFAR, SUFAM and ANISOFT use approach 1 (signed negative values) for plotting principal directions and approach 2 (absolute values) in calculating all AMS parameters. This is advantageous in the case of the L,

F, P parameters, but problematic in the case of Tand [/parameters. Conclusions The mathematical modelling investigation of the AMS parameters of diamagnetic rocks (quartzite/evaporite and limestone/marble models) has drawn the following conclusions: For plotting principal directions, it is useful to work with the signed values (approach 1). In quartzite and evaporite, they correspond to those of the ferro-/paramagnetic admixture fraction. In calcitic tectonites (mostly marbles), the calcite e-twin lamellae are usually parallel to the schistosity resulting in small circle preferred orientation of calcite c axes; the magnetic foliation is then parallel to the schistosity. For characterizing the degree of AMS, magnetic lineation and magnetic foliation, it is advantageous to use the parameters based on the absolute values of the principal susceptibilities (indexed a, approach 2). These are higher than 1 and increase with increasing anisotropy. For characterizing the shape of the susceptibility ellipsoid in quartzite or evaporite, the T parameter based on signed (negative) principal susceptibilities should be used, because it

F. HROUDA

58

Fig. 9. Empiric relationship between the whole-rock degree of AMS (P) and mean bulk susceptibility (Km) for limestone Hranice Devonian area (East Bohemian Massif). For the P parameter, the approach 1 values are plotted if Km > 0, while the approach 2 values are plotted if Km < 0. Note high degree of AMS in the vicinity of zero susceptibility.

describes the shape of the susceptibility ellipsoid of the ferro-/paramagnetic admixture fraction. In case of marble or limestone, it is better to use the Ta parameter (based on absolute values), because it corresponds to the T parameter calculated from c axis distribution pattern. Mgr. V. Dvorak and Mgr. J. Smid are thanked for permission to publish their AMS data on limestone and evaporite. The constructive criticism by the reviewers, Dr. A. Hirt and Prof. G. Borradaile, and the editorial help of Dr. A. Hirt and Dr. M. Jackson are highly appreciated. The research was partly supported financially by the Ministry of Education of the Czech Republic (grant #2431 3005).

References BLEIL, U. & PETERSEN, N. 1982. Magnetic properties of natural minerals. In: ANGENHEISTER, G. (ed.) Landolt-Bornstein Numerical Data and Functional Relationships in Science and Technology, Volume 1, Physical Properties of Rocks, subvolume b. Springer-Verlag, Berlin, Heidelberg, New York, 308-432. BORRADAILE, G. J. & HENRY, B. 1997. Tectonic applications of magnetic susceptibility and its anisotropy. Earth-Science Reviews, 42, 49-93. BORRADAILE, G. J., FRALICK, P. W. & LAGROIX, F. 1999. Acquisition of anhysteretic remanence and tensor subtraction from AMS isolates true palaeocurrent grain alignments. In: TARLING, D. H. & TURNER, P. (eds) Palaeomagnetism and Diagenesis in Sediments, Geological Society, London, Special Publications, 151, 139-145.

DALY, L. & D'ALBISSIN, M. 1968. Correlation entre les anisotropies de susceptibilite magnetique et de dilatation thermique des roches; application en structurologie. Comptes Rendus de rAcademic de Science de Paris, 267, 473-476. DE WALL, H., BESTMANN, M. & ULLEMEYER, K. 2000. Anisotropy of diamagnetic susceptibility in Thassos marble: A comparison between measured and modeled data. Journal of Structural Geology, 22, 1761-1771. HEJTMAN, B. 1962. The petrography of metamorphic rocks (in Czech). N ESAV, Praha. HELLWEGE, K. H. & HELLWEGE, M. 1967. Magnetic properties. In: VON BORCHERS, H., HAUSEN, H., HELLWEGE, K. H., SCHAEFER, K. & SCHMIDT, E. (eds), Landolt-Boernstein: Zahlenwerte and Funktionen aus Physik. Chemie, Astronomic, Geophysik und Technik, 11/10. Springer, Berlin. HENRY, B. 1983. Interpretation quantitative de 1'anisotropie de susceptibilite magnetique. Tectonophysics, 91, 165-177. HENRY, B. 1989. Magnetic fabric and orientation tensor of minerals in rocks. Tectonophysics, 165, 21-27. HENRY, B. & DALY, L. 1983. From qualitative to quantitative magnetic anisotropy analysis: the prospect of finite strain calibration. Tectonophysics, 98, 327-336. HROUDA, F. 1973. A determination of the symmetry of the ferromagnetic mineral fabric in rocks on the basis of the magnetic susceptibility anisotropy measurements. Ger lands Beit rage fur Geophysik, 82, 390-396. HROUDA, F. 1982. Magnetic anisotropy of rocks and its application in geology and geophysics. Geophysical Surveys, 5, 37-82. HROUDA, F. 1986. The effect of quartz on the magnetic anisotropy of quartzite. Studia Geophysica et Geodaetica, 30, 39-45.

AMS PARAMETERS IN DIAMAGNETIC ROCKS HROUDA, F. & SCHULMANN, K. 1990. Conversion of magnetic susceptibility tensor into orientation tensor in some rocks. Physics of the Earth and Planetary Interiors, 63, 71-77. HROUDA, F., HENRY, B. & BORRADAILE, G. 2000. Limitations of tensor subtraction in isolating diamagnetic fabrics by magnetic anisotropy. Tectonophysics, 322, 303-310. KRISHNAN, K. S., GUHA, B. C. & BANERJEE, S. K. 1933. Investigation on magneto-crystallic action, Part I. Diamagnetics. Philosophical Transactions of the Royal Society of London, Series A, 231, 235-262. JELINEK, V. 1981. Characterization of the magnetic fabric of rocks. Tectonophysics, 79, T63-T67. JEZEK, J. & HROUDA, F. 2000. The relationship between the Lisle orientation tensor and the susceptibility tensor. Physics and Chemistry of the Earth, Part A, 25, 469-474. LISLE, R. J. 1989. The statistical analysis of orthogonal orientation data. Journal of Geology, 97, 160-364. NAGATA, T. 1961. Rock magnetism. Maruzen Tokyo. NYE, J. F. 1957. Physical properties of crystals. Clarendon Press, Oxford.

59

OWENS, W. H. & BAMFORD, D. 1976. Magnetic, seismic, and other anisotropic properties of rock fabric. Philosophical Transactions of the Royal Society of London, Series A, 283, 55-68. OWENS, W. H. & RUTTER, E. H. 1978. The development of magnetic susceptibility anisotropy through crystallographic preferred orientation in a calcite rock. Physics of the Earth and Planetary Interiors, 16,215-222. SCHEIDEGGER, A. E. 1965. On the statistics of the orientation of bedding planes, grain axes, and similar sedimentological data. US Geological Survey Professional Paper, 525-C, 164-167. SHUBNIKOV, A. V. 1949. On the symmetry of vectors and tensors (in Russian). Izvestia AN SSSR, Seria Fizicheskaja, 13, 347-375. TARLING, D. H. & HROUDA, F. 1993. The magnetic anisotropy of rocks. Chapman & Hall, London. TURNER, F. J. & WEISS, L. E. 1963. Structural analysis of metamorphic tectonites. McGraw Hill, New York. WOODCOCK, N. H. 1977. Specification of fabric shapes using an eigenvalue method. Geological Society of America Bulletin, 88, 1231-1236.

This page intentionally left blank

Metamorphic control of magnetic susceptibility and magnetic fabrics: a 3-D projection NORIHIRO NAKAMURA1 & GRAHAM J. BORRADAILE2 1

Department of Geo-Environmental Science, Tohoku University, Sendai 980-8578, Japan (e-mail: [email protected]) 2 Geology & Physics Dept., Lakehead University, Thunder Bay, P7B 5E1, Canada (e-mail: [email protected]) Abstract: Magnetic fabric changes due to progressive metamorphism are poorly understood. Bulk magnetic susceptibility (K) is known to increase with metamorphic grade but anisotropy changes have been neglected. To combine information on anisotropy with bulk susceptibility, we introduce a projection with three axes: K, ellipsoidal eccentricity (Pj, the so-called 'anisotropy degree', despite the fact that this is quantified)d) and ellipsoid symmetry (7^,) as independent variables. The projection reveals that metamorphic facies can be discriminated successfully in the 3-D projection, with distinct, significant regression surfaces for crustal metamorphic rocks metamorphosed successively in greenschist, amphibolite, and granulite facies. This emphasizes that bulk magnetic susceptibility (K) and its anisotropies ('magnetic fabric') evolve in response to metamorphic process, not just strain. Moreover, post-tectonic granitic plutons, upper mantle harzburgites and serpentinized mantle rocks also have characteristic regression surfaces relating «, Pj and 7} in the new projection.

Metamorphic processes are solid-state transformations of minerals and textures of a pre-existing rock due to changes in temperature, pressure, fluid-rock interaction, strain and recrystallization (e.g. Miyashiro 1994). Generally, magnetic susceptibility (K) may change with metamorphic grade from greenschist to granulite facies with depth (Hrouda 1982; Shive et al. 1988, Jackson &Tauxe 1991). Under some circumstances, 'magnetic isograds' may be identified, for example, in a zeolite to amphibolite facies progression recorded in lithologically monotonous black shale (Rochette 1987). The effect of metamorphic grade on magnetic susceptibility may reflect, in the broadest sense, a generalized metamorphic cross-section of the continental crust, as well as lithologically homogeneous deep crustal and upper mantle rocks. The bulk susceptibility («), anisotropy of low field susceptibility (AMS) and other forms of a magnetic anisotropy are commonly used as tools to investigate the petrofabric, mainly of technically deformed rocks (Hrouda 1982; King et al. 1982; Henry 1983; Jackson & Tauxe 1991; Rochette et al. 1992; Borradaile & Henry 1997). Whereas most AMS studies concern technically deformed rocks, the relative roles of strain and metamorphic recrystallization are not readily distinguishable. To discern multivariate dependent relationships among magnetic anisotropy and bulk susceptibility due to different metamorphic facies, we examine four distinct protoliths in the common facies series from greenschist through amphibolite to granulite

facies. The protoliths considered include: (1) Archaean greywacke, deposited in an accretionary prism (greenschist through granulite facies) (2) Oceanic mantle harzburgites (Troodos ophiolite) (3) Serpentinized oceanic mantle (Troodos ophiolite) (4) Post-tectonic I-type granite AMS may be represented by a magnitudeellipsoid whose anisotropy parameters are most effectively described by the eccentricity Pj and its shape Tj (Jelinek 1981). The PJ9 7} plot represents AMS more symmetrically than the Flinn-plot used to discriminate strain ellipsoids and fabric types in structural geology (Borradaile 2003a). Earlier studies tended to attribute the preferred orientations detected by AMS to finite strain. Principally, the shape parameter (Tj) and eccentricity (Pj) of the magnitude ellipsoid (Jelinek 1981), and the mean or bulk susceptibility (K) can be used to infer the magnetic mineralogy, for example, the role of paramagnetic minerals versus remanence-bearing minerals, grain-size variations, and their relative contributions of their orientation distributions to the AMS. However, in most technically deformed rocks and many plutonic, nominally igneous rocks, both the paramagnetic silicates and the iron oxides have recrystallized (Housen et al. 1993). This may only affect the orientation-distribution but it may equally produce a

From: MARTIN-HERNANDEZ, F., LUNEBURG, C. M., AUBOURG, C. & JACKSON, M. (eds) 2004. Magnetic Fabric: Methods and Applications. Geological Society, London, Special Publications, 238, 61-68. 0305-8719/04/$15.00 © The Geological Society of London 2004.

62

N. NAKAMURA & G. J. BORRADAILE

new mineral assemblage with a different partitioning of K among the magnetically anisotropic minerals. The prominent control on K is the concentration of accessory magnetite due to its enormous susceptibility, (e.g., common multidomain (MD) magnetite has «~3.1SI; Heider el al 1996) whereas the most susceptible rock-forming pure silicates have values <0.002 SI. Although magnetite's grain-shape-controlled anisotropy degree (Pj) may be low, its contribution to the rock-fabric may be more significant than the much more anisotropic, but less

susceptible, rock-forming silicates (Borradaile 1987). Although the paramagnetic matrix minerals of interest (chlorites, serpentine, micas, biotite, amphibole and pyroxene) show relatively low intrinsic bulk susceptibility, they may be contaminated by single-domain (SD) or pseudosingle-domain (PSD) magnetite as inclusions or exsolution grains. For pure paramagnetic silicates, (K) ranges from 50uSI to a theoretical maximum of 2000|iSI (Syono 1960; Droop 1987) but their anisotropy is much higher than magnetite, and controlled by crystal symmetry,

Fig. 1. Traditional 2-D Jelinek plot and PJ-K plot of metamorphic rocks from greenschist-facies through amphibolite-facies to granulite-facies may not always show effective visual discriminations due to complete overlap of the data.

METAMORPHIC CONTROL OF AMS: 3D PROJECTION

not by grain-shape. Moreover, magnetite inclusions in paramagnetic minerals may increase bulk susceptibility to lOOOOuSI (Borradaile 1994; Borradaile & Werner 1994; Lagroix & Borradaile 2000). However, in those cases, the orientation distribution of the inclusions is usually controlled by the host silicate lattice, thus reinforcing the matrix-fabric. This gives some insight, into the complications that metamorphism may cause, shifting the control on anisotropy between iron-oxide accessories and matrix silicates, according to mineral reactions during progressive metamorphism. On the traditional 2-D Jelinek plot, the initial example of metamorphic rocks shows almost complete overlap of greenschist-facies, amphibolite-facies and granulite-facies data (Fig. la). This traditional Pj—Tj plot also lacks information about bulk susceptibility that can discriminate the fabric-shape patterns with respect to K. Although the influences of metamorphic facies on K may be evident (Fig. Ib), AMS is also influenced by facies. Metamorphic control of magnetic fabrics is presented on a 3-D projection for the common progressive metamorphic facies series (greenschist-amphibolite-granulite), in terms of K, Pj and 7}. This is really a simplified approach to linear discriminant analysis; we choose a suitable viewing axis in coordinatespace such that clusters of data are distinguishable on the basis of some attribute, here metamorphic facies. Furthermore, most of our data groups (~ metamorphic facies) may be characterized by multiple regression surfaces in the same program used for three-dimensional viewing of data (in our case SigmaPlotS.O © software). Magnetic petrology Our AMS data is from published sources for Archaean metasedimentary rocks from subgarnet greenschist-facies slates (n = 224) through amphibolite-facies schists (n = 193) to granulite-facies lower crustal gneisses (n = 258) and late Archaean granitoids (n = 129). All of these rocks consist of a homogeneously metamorphosed similar protolith, representative of mean-continental crustal chemistry: greywacke. Also mantle harzburgite and serpentinite samples (n = 457) from the Troodos ophiolite, Cyprus are used to represent metamorphosed oceanic upper mantle. These examples outline facies control on K and AMS for typical continental crust and oceanic upper mantle. The greenschist-facies slates consist of Archaean metagreywacke subjected to one

63

penetrative deformation accompanied by a lowgrade metamorphism (Borradaile & Sarvas 1990). Magnetic susceptibility is dominated by ferromagnetic pyrrhotite and paramagnetic chlorite (thuringite), and biotite. A steep metamorphic gradient gives a smooth transition through to amphibolite facies rocks with an extensive growth of pyrrhotite (Werner & Borradaile 1996). The granulite-facies gneiss is an Archaean lower crustal granoblastic rock in the Kapuskasing Structural Zone formed at pressures between 0.7 and 1.0 GPa in the temperature range 650750 °C (Percival 1983; Bursnall et al 1994). The gneiss possesses poorly defined 'visible' petrofabric but well-defined magnetic fabrics (Borradaile et al. 1999). Major sources of susceptibility are free-MD magnetite, as accessory grains, and PSD magnetite inclusions in pyroxene host grains. The magnetite inclusions in mafic silicates occur as polycrystalline submillimetric veinlets. Paramagnetic mafic silicates, pyroxene, biotite and hornblende provide a similar combination to bulk susceptibility but show much higher anisotropy (Pj) than magnetite. Plutonic I-type granites are represented by a sample suite of late Archaean post-tectonic granitic plutons (Trout Lake and Barnum Lake plutons) in north-western Ontario (Borradaile & Kehlenbeck 1996). Susceptibility («) is controlled dominantly by MD accessory magnetite. These plutons have a bimodal frequency distribution of bulk susceptibility. The less common lower susceptibilities might be due to more extensive oxidation of magnetite or a reduced magnetite content so we have removed them from consideration. In these plutonic granitoids, AMS is dominated by the stress-induced preferred arrangement of significant magnetic domains within equidimensional multidomain magnetite, since there is no visible evidence of their preferred grain-shape alignment. The data sets of mantle rocks are serpentinized harzburgites from the Troodos ophiolite, Cyprus (Borradaile & Lagroix 2001). Although serpentinite is one of the most susceptible rocks, this is largely due to the role of abundant ('accessory') magnetite, a by-product of serpentinization (Rochette 1994; Dunlop & Ozdemir 1997; Kido et al. 2001). Their magnetite inclusions within mafic silicates and the free-MD accessory magnetite produced by serpentinization dominate K. More than 50% of the bulk susceptibility is attributed to PSD magnetite inclusions in mafic silicates (Borradaile & Lagroix 2001). This highly susceptible PSD magnetite fabric camouflages the more anisotropic, but less susceptible mafic silicate fabric.

64

N. NAKAMURA & G. J. BORRADAILE

Fig. 2. Regression surfaces for magnetic properties of selected important protoliths in 3-D space, viewed along a suitable axis. Bulk susceptibility («); eccentricity 1 < Pj < oc (sphere to ellipse) and symmetry, — 1 < 7} < +1 (oblate to prolate). Note logarithmic scale for K. Multiple regression magnetic fabric parameters A 3-D projection represents the metamorphic control of magnetic fabrics under progressive metamorphism with multiple linear regression surfaces: K = aPj + bTj + c (Fig. 2). The logarithmic axis facilitates the plotting of K due to its large range. We choose whatever orientation of axes of 3-D plot best reveals the relationship between K, P7, Tj and metamorphic grade.

Regression permits statistical comparison of fabric-shapes by using Jelinek's parameters of a symmetrical shape description: Tj (oblate to prolate; +1 to —1) and an ellipsoidal eccentricity: Pj (sphere to ellipsoid; 1 to oo) which includes reference to the intermediate value. The multiple regression surfaces are distinct and statistically significant at the 95% confidence level (Table 1). The test-statistics were used to determine whether the correlation coefficients (R) were significantly non-zero at the 95% level (see

Table 1. Multiple linear regression data and their statistics Rock type

n

R

Test-statistic (R^n)

a

b

c

>?0.05/2

Greenschist-facies slate1 Amphibolite-facies schist2 Granulite-facies gneiss I- type granite Harzburgite Serpentined Harzburgite6

224 193 258 94 436 150

0.66 0.39 0.57 0.24 0.34 0.28

9.88 5.42 9.16 2.33 7.10 3.43

589.0 954.5 31396.5 30217.0 19469.0 -2496.3

-54.5 -394.2 -2442.3 -1321.2 -1103.8 -8434.8

-347.0 -644.5 -32898.2 -7123.7 -17585.3 24608.1

OK OK OK OK OK OK

The multiple regression is defined by a linear equation: k = aPj + bTj• + c(Pj > 1; -1 < 7} < 1). The significance of the correlation coefficient is determined by comparing the test-statistic, (R^/n), with ^0.05/2 according to conventional distribution-theory statistics (Borradaile 2003a, b). OK= Significant correlation surfaces at the 95% confidence level are indicated by 'OK. References: 1. Archaean Quetico belt, Canada (Borradaile & Servas 1991). 2. Archaean Quetico belt, Canada (Werner & Borradaile 1996). 3. Kapuskasing Structural Zone (Borradaile et al. 1999). 4. Rainy and Barnum Lake, Canada (Borradaile & Kehlenbeck 1996). 5. Troodos ophiolite, Cyprus (Borradaile & Lagroix 2001). 6. Troodos ophiolite, Cyprus (Borradaile & Lagroix 2001).

METAMORPHIC CONTROL OF AMS: 3D PROJECTION

Table 1, Figs 3 & 4). Although the correlation coefficients may be relatively low (Table 1), the regression surfaces are significant at the 95% confidence level, because the test-statistics |/£-\/H| exceeds 1.96 and n is a large sample-size (Borradaile 20030, b). Formally, we should not reject the hypothesis that the regression surfaces provide a faithful generalization of the magnetic parameters with metamorphic grade. Although some data sets show considerable dispersion, particularly with regard to 7} (ellipsoid shape), these are notable differences of the multiple regression surfaces with respect to the PJ — log K axes (see Figs 3 & 4). The graphs show regression lines and 95% confidence envelopes following classical, distribution-

65

theory statistics. This emphasizes a significant discrimination from the 2-D Jelinek plot by employing bulk susceptibility, avoiding an overlap of Pj-Tj data. Figure 3a shows that the AMS data (n = 224) in greenschist facies slates lie predominantly in the oblate field (Tj > O^and the regression surface indicates the PJ — k correlation (Borradaile & Sarvas 1990; Rochette et al 1992; Borradaile & Henry 1997). In amphibolite facies schists, PJ value is less anisotropic and AMS symmetry shows more neutral ellipsoid shapes (Tj —> 0), which may be due to the growth of stubby biotite at the expense of well-aligned chlorite in greenschist facies grade (Fig. 3b). Moreover, bulk susceptibility, K, in amphibolite facies is

Fig. 3. Traditional PJ-K plot with 95% confidence envelopes (dotted lines) for the regression line (solid lines), new 3-D projection and Jelinek plot: (a) greenschist-facies slate, (b) amphibolite-facies schist, (c) granulite-facies gneiss. Note that sample size is («) and regression coefficient is (R). All 3-D regression surfaces imply correlations significantly different from zero at the 95% level.

66

N. NAKAMURA & G. J. BORRADAILE

Fig. 4. Traditional PJ-K plot, with 95% confidence envelopes (dotted lines) for the regression line, new 3-D projection and Jelinek plot: (a) granite and oxidized granite, (b) harzburgite and serpentinite. Note that sample size is (n) and regression coefficient is (R). All 3-D regression surfaces imply correlations significantly different from zero at the 95% level.

METAMORPHIC CONTROL OF AMS: 3D PROJECTION

relatively greater than in the greenschist facies. Each multiple regression surface discriminates the differences in bulk susceptibility and in AMS (see Fig. 2, Figs 3a & 3b). The lower continental crustal section exposed at Kapuskasing, Ontario (Percival & West 1994) shows granulite-facies metamorphism. At this high metamorphic grade, bulk susceptibility is higher but also Pj drastically increases more rapidly with increasing K (Fig. 3c). The fabric symmetry 7} evolves from oblate for low K specimens to neutral or prolate at higher susceptibilities. This may be due to a nucleation of needle-like veinlet magnetite inclusions in pyroxene host minerals during retrograde metamorphism. In northern Ontario, certain late Archaean granitic plutons intruded forcefully and posttectonically, and their feldspar megacryst fabric is clearly magmatic or, at least related to latemagmatic inflation. The granites' bulk susceptibilities are high (Fig. 2, Fig.4a), classifying them as magnetite-series granitoid (Ishihara 1979). The distribution of these granites in P7TJ-K space is relatively insensitive to anisotropy, Pj or TJ, although data still show distinct anisotropic fabrics. These anisotropies may be due to a stress-controlled alignment of intragranular domain walls in multidomain magnetite since there is no obvious shape alignment of grains (Borradaile & Kehlenbeck 1996), despite relatively constant susceptibility (see also Fig. 4a). These high susceptible post-tectonic granitic plutons are responsible for high induced magnetization in Earth's magnetic field, which is reflected in the aeromagnetic anomaly. The lower susceptible oxidized granite is less anisotropic but bulk susceptibility is sensitive to Pj. The Troodos harzburgite exposes oceanic upper mantle rocks, and shows high-temperature, solid-state flow textures of the serpentinized and retrogressed mafic silicates. The regression surface for harzburgite lies close to the TJ-K plane because anisotropy degree (P y ) increases slowly with K (Fig. 4b). The shape (7}) is insensitive to bulk susceptibility. It varies from oblate due to the S-fabric and intrinsic crystallographic-symmetry of the mafic minerals, to prolate in specimens in which magnetite dominates (highest K). The eccentricity Pj is lower, perhaps due to aligned PSD magnetite inclusions within host-silicates, which may be controlled by their host-crystal (Borradaile & Lagroix 2001). Serpentinized harzburgites are less anisotropic but of significantly higher susceptibility (Fig. 4b), due to MD magnetite as a by-product of serpentinization, although some lower susceptible

67

serpentinized harzburgites data are plotted in similar region as harzburgite. Data indicates that Tj disperses broadly from +1 (oblate) to -1 (prolate) and is independent of K. Summary Some earlier studies imply that the magnetic susceptibility and magnetic anisotropy of rocks did not depend on metamorphism (Williams et al. 1985; Shive & Fountain 1988; Werner & Borradaile 1996). However, their sampling may not have involved uniform protoliths, nor a wide enough range of metamorphic conditions. We sampled lithologically distinct protoliths metamorphosed under very different metamorphic facies, from greenschist to granulite for continental crust and certain granites, and for oceanic upper mantle. Three low-field magnetic parameters, («, Pj, 7}), provide a successful visual discrimination of greywackes metamorphosed in greenschist/amphibolite/granulite facies. They also characterize upper mantle rocks, their serpentinized equivalent and I-type granite. The success of discrimination in this new three-dimensional projection testifies to the fact that metamorphic facies is at least as important as strain in controlling the magnetic fabric. This is corroborated and quantified by multiple regression statistics. We thank M. Jackson and B. Housen for their helpful review. We are grateful to T. Werner for valuable review comments and for providing AMS data from an Archaean greenstone belt. The National Sciences and Engineering Research Council of Canada funded GB. N. N. is grateful to the JSPS postdoctoral fellowships for research abroad. This work was partly supported by the 21st century Center-Of-Excellence (COE) program (Earth Sciences) of Tohoku University.

References BORRADAILE, G. J. 1987. Anisotropy of magnetic susceptibility: rock composition versus strain, Tectonophysics, 138, 327-329 BORRADAILE, G. J. 1994. Paleomagnetism carried by crystal inclusions: the effect of preferred crystallographic orientation, Earth and Planetary Science Letters, 126, 171-182. BORRADAILE, G. J. 20030. Statistics of Earth Science Data -Their Distribution in Time, Space and Orientation. Springer-Verlag, New York. BORRADAILE, G. J. 20036. Viscous magnetization, archaeology and Bayesian statistics of small samples from Israel and England, Geophysical Research Letters, 30, No. 10, 1528, doi:10.1029/ 2003GL016977, (35-l)-(35-4).

68

N. NAKAMURA & G. J. BORRADAILE

BORRADAILE, G. J. & HENRY, B. 1997. Tectonic application of magnetic susceptibility and its anisotropy, Earth Science Reviews, 42, 49-93. BORRADAILE, G. J. & KEHLENBECK, M. M. 1996. Possible cryptic tectono-magnetic fabrics in 'post-tectonic' granitoid plutons of the Canadian Shield, Earth and Planetary Science Letters, 137, 119-127. BORRADAILE, G. J. & LAGROIX, F. 2001. Magnetic fabrics reveal Upper Mantle Flow fabrics in the Troodos Ophiolite Complex, Cyprus. Journal of Structural Geology, 23, 1299-1327. BORRADAILE, G. J. & SARVAS, P. 1990. Magnetic susceptibility fabrics in slates: structural, mineralogical and lithological influences, Tectonophysics, 172,215-222. BORRADAILE, G. J. & WERNER, T. 1994. Magnetic anisotropy of some phyllosilicates, Tectonophysics, 235, 233-248. BORRADAILE, G. J., WERNER, T. & LAGROIX, F. 1999. Magnetic fabrics and anisotropy-controlled thrusting in the Kapuskasing Structural Zone. Canada, Tectonophysics., 302, 241-256. BURSNALL, J. T., LECLAIR, A. D., MOSER, D. E. & PERCIVAL, J. A. 1994. Structural correlation within the Kapuskasing uplift. Canadian Journal of Earth Science, 31, 1081-1095. DROOP, G. T. R. 1987. A general equation for estimating Fe + concentrations in ferromagnesian silicates and oxides from microprobe analyses, using stoichiometric criteria. Mineralogical Magazine, 51, 431^35. DUNLOP, D. P. & OZDEMIR, 6. 1997. Rock Magnetism: Fundamentals and frontiers. Cambridge University Press, Cambridge. HEIDER, F., ZITZELSBERGER, A. & FABIAN, K. 1996. Magnetic susceptibility and remanence coercive force in grown magnetite crystals from 0.1 (im to 6 mm. Physics of the Earth and Planetary Interiors, 93, 239-256. HENRY, B. 1983. Interpretation quantitative de 1'anisotropie de susceptibilite magnetique. Tectonophysics, 91, 165-177. HOUSEN, B. A., VAN DER PLUIJM, B. A. & VAN DER VOO,

R. 1993. Magnetite dissolution and neocrystallization during cleavage formation: Paleomagnetic study of the Martinsburg formation, Lehigh Gap, Pennsylvania. Journal of Geophysical Research, 98, 13799-13813. HROUDA, F. 1982. Magnetic anisotropy of rocks and its application in geology and geophysics. Survey of Geophysics, 5, 37-82. ISHIHARA, S. 1979. Lateral variation of magnetic susceptibility of the Japanese granitoids. Journal of the Geological Society of Japan, 85, 509-523. JACKSON, M. J. & TAUXE, L. 1991. Anisotropy of magnetic susceptibility and remanence: Developments in the characterization of tectonic, sedimentary, and igneous fabric. Reviews of Geophysics, 29, supplement (IUGG Report Contributions in Geomagnetism and Paleomagnetism), 371-376.

JELINEK, V. 1981. Characterization of the magnetic fabrics of rocks. Tectonophysics, 79, T63-T67. KIDO, Y., MACHIDA, S., SATO, H. & FUJIOKA, K. 2001. Origin of magnetic dipole anomaly of Shikoku An example of utilization of Jamstec Frontier Database system. JAMSTEC Journal of Deep Sea Research, 18, \l\-m. KING, J., BANERJEE, S. K., MARVIN, J. & OZDEMIR, 6. 1982. A comparison of different magnetic methods for determining the relative grain size of magnetite in natural materials: some results from lake sediments, Earth and Planetary Science Letters, 59, 404-419. LAGROIX, F. & BORRADAILE, G. J. 2000. Magnetic fabric interpretation complicated by inclusions in mafic silicates. Tectonophysics, 325, 207-225. MIYASHIRO, A. 1994. Metamorphic Petrology. Oxford University Press, London. PERCIVAL, J. A. 1983. High-grade metamorphism in the Chapleau-Folyet area, Ontario. American Mineralogist, 68, 667-686. PERCIVAL, J. A. & WEST, G. F. 1994. The Kapuskasing uplift: a geological and geophysical synthesis. Canadian Journal of Earth Sciences, 31, 1256-1286. ROCHETTE, P. 1987. Metamorphic control of the magnetic mineralogy of black shales in the Swiss Alps: toward the use of 'magnetic isograds'. Earth and Planetary Science Letters, 84, 446-456. ROCHETTE, P. 1994. Comments on 'Anisotropic magnetic susceptibility in the continental lower crust and its implication for the shape of magnetic anomalies' by G. Florio et al. Geophysical Research Letters, 21, 2773-2774. ROCHETTE, P., JACKSON, M. J. & AUBOURG, C. 1992. Rock magnetism and the interpretation of anisotropy of magnetic susceptibility. Reviews of Geophysics, 30, 209-226. SHIVE, P. N. & FOUNTAIN, D. M. 1988. Magnetic mineralogy in an Archean crustal cross section: implications for crustal magnetization. Journal of Geophysical Research, 93, 12177-12186. SHIVE, P. N., FROST, B. R. & PERETTI, A. 1988. The magnetic properties of metaperidotitic rocks as a function of metamorphic grade: implications for crustal magnetic anomalies. Journal of Geophysical Research, 93, 12187-12195. SYONO, Y. 1960. Magnetic susceptibility of some rock forming silicate minerals such as amphiboles, biotites, cordierites garnets. Journal of Geomagnetism and Geoelectricity, 11, 85-93. WERNER, T. & BORRADAILE, G. J. 1996. Paleoremanence dispersal across a transpressed Archean terrain: Deflection by anisotropy or by late compression? Journal of Geophysical Research, 101, 5531-5545. WILLIAMS, M. C., SHIVE, P. N., FOUNTAIN, D. M. & FROST, B. R. 1985. Magnetic properties of exposed deep crustal rocks from the Superior Province of Manitoba. Earth and Planetary Science Letters, 76, 176-184.

Anisotropy of magnetic susceptibility of rocks measured in variable weak magnetic fields using the KLY-4S Kappabridge JIRI POKORNY1, PETR SUZA1 & FRANTISEK HROUDA 12 1

AGICO Inc., Jecnd 29a, Box 90, CZ-621 00 Brno, Czech Republic (e-mail: [email protected]) 2 Institute of Petrology and Structural Geology, Charles University, Albertov 6, CZ-128 43 Praha, Czech Republic Abstract: The theory of the anisotropy of magnetic susceptibility assumes a linear relationship between magnetization and magnetizing field, resulting in field-independent susceptibility. This relationship is valid in diamagnetic and paramagnetic minerals by definition, and also in magnetite in the fields used in common AMS meters. Pyrrhotite, hematite and titanomagnetite may show field variation of susceptibility in the same fields and therefore in principle the linear theory is incorrect for calculating the AMS. Fortunately, the linear theory nevertheless provides accurate determination of the orientations of the principal susceptibilities and of the AMS ellipsoid shape. It gives rise to inaccurate determination of the degree of AMS. Simple techniques are suggested to overcome this problem.

The theory of the low-field anisotropy of magnetic susceptibility (AMS) of rocks is based on the assumption of a linear relationship between magnetization and magnetizing field, resulting in field-independent susceptibility. A linear relationship is by definition valid in diamagnetic and paramagnetic minerals, while in ferrimagnetic and antiferromagnetic minerals, the relationship is in general non-linear (represented by a minor hysteresis loop or Rayleigh loop), being linear only in very weak fields in which the initial susceptibility is measured. The instruments for measuring the AMS of rocks use relatively weak fields (see Table 1) that are implicitly considered to be weak enough for the field-independent initial susceptibility to be measured in all rock types. However, it has been shown recently that this assumption is not valid in all cases. Whereas magnetite and magnetite-bearing rocks show field-independent susceptibility in the low fields used in common AMS meters (Worm et al. 1993; Jackson et al. 1998; Hrouda 2002); in some cases, pyrrhotite, hematite and titanomagnetite may show clear field variation of susceptibility in the same fields (Chlupacova 1984; Worm 1991; Zapletal 1992; Worm et al. 1993; Markert & Lehmann 1996; Jackson et al. 1998; de Wall 2000; Hrouda 2002). Consequently, strictly speaking, the use of linear theory in calculating the AMS is in principle incorrect in the last cases. On the other hand, the linear theory of the AMS is so simple, elegant and beautiful that the AMS researchers would advocate using it even though it is not fully correct, provided that the errors introduced in this way

are not too large. Therefore, it is desirable to find out how much the susceptibility field variation results in the field variation of the AMS. Several studies have investigated the field dependence of the AMS of titanomagnetitebearing rocks (de Wall 2000) and hematite single crystals (Hrouda & Quade 1997; Hrouda et al. 1998; Hrouda 2002). In general they found virtually no field variation in the orientation of the principal susceptibilities and very weak variation in the symmetry of the susceptibility ellipsoid, while in the degree of AMS a clear variation was observed. These investigations used the KLF-3 miniKappa instrument which measures the susceptibility in two fields, 300 A/m and 30A/m, and is not designed for measuring the AMS, having acceptable accuracy only for relatively strongly magnetic and anisotropic specimens. The purpose of the present paper is to investigate the effect of the variable measuring fields on the AMS using a more precise instrument, namely the KLY-4S Kappabridge upgraded from the KLY-3S Kappabridge (Jelinek & Pokorny 1997). In our investigations, the AMS characteristics obtained through the linear theory, such as orientation of principal susceptibilities, the shape of the susceptibility ellipsoid and the degree of AMS are evaluated. Instrumentation The KLY-4S Kappabridge originated through upgrading the KLY-3S Kappabridge (Jelinek & Pokorny 1997) in such a way that both the bulk susceptibility and the AMS can be measured in

From: MARTIN-HERNANDEZ, F., LUNEBURG, C. M., AUBOURG, C. & JACKSON, M. (eds) 2004. Magnetic Fabric: Methods and Applications. Geological Society, London, Special Publications, 238, 69-76. 0305-8719/04/S 15.00 © The Geological Society of London 2004.

J. POKORNY ET AL.

70

Table 1. Field intensity in various AMS and susceptibility meters Instrument

Field Intensity (A/m)

Reference

Low Field Torque Meter Molspin Anisotropy Delineator SI Sapphire Instrument KLY-1 Kappabridge KLY-2 Kappabridge KLY-3 Kappabridge KLY-3S Kappabridge KLF-3 IRM 'Roly-poly' Lake Shore MicroMag VSM

>8 400 80 150 300 300 300 30 and 300 10-1000 0.1-2000 80-2000

Tarling&Hrouda(1993) Collinson (1993) Borradaile et al. (1999) Jelinek (1973) Jelinek (1980) Jelinek & Pokorny ( 1997) Jelinek & Pokorny (1997) Sapik (1988) IRM leaflet IRM leaflet Leaflet

various magnetic fields ranging from approximately 2 A/m to 450 A/m. The KLY-4S Kappabridge is a classical a.c. bridge of transformer type. The main features of the KLY-4S Kappabridge are: Separation of the in-phase (real) and out-ofphase (imaginary) components. Automatic compensation of the real and imaginary components. Autoranging and autozeroing over the entire measurement range up to 0.2 [SI], which increases the precision of AMS measurement. Automated measurement of field variation of bulk susceptibility. High sensitivity of AMS measurement (2 x 1(T8 [SI] at the field intensity of 300 A/m). AMS measurement using a spinning method (in which the specimen slowly rotates at 0.5r.p.s., sequentially about three perpendicular axes), accelerating the measurement time and simplifying the positioning operations. The measuring field can be selected, in varying increments, between 2 A/m and 450 A/m. Precise calibration of each field is performed for both bulk susceptibility and AMS measurements. Specimens investigated, techniques used A.C. susceptibility anisotropy measurements were made on three sets of specimens: Devonian pyrrhotite-bearing schists from the Silesian Zone of the Bohemian Massif (provided by Dr M. Chlupacova); titanomagnetite-bearing volcanic rocks from Hawaii (provided by Dipl. Geol. C. Vahle) and single crystals of hematite from Sao Juliao quarry, Minas, Gerais, Brazil (provided by Prof. Dr H. Quade). The in-phase bulk susceptibility and the AMS were measured using the KLY-4S Kappabridge

at a frequency of 875 Hz in fields ranging from 2 A/m to 450 A/m. The AMS was calculated using the linear theory (Jelinek 1996). The mean bulk susceptibility, eccentricity and shape of the AMS ellipsoid can be characterized by the following parameters (Nagata 1961; Jelinek 1981):

where k\ > k2 > k3 are the principal susceptibilities, r]i = In/c 1; 772 = ln/C2, ??3 = In ^3. The parameter P, called the degree of anisotropy, indicates the intensity of the preferred orientation of magnetic minerals in a rock. The parameter T characterizes the symmetry of the AMS ellipsoid. If 0 < T < +1 the AMS ellipsoid is oblate (the magnetic fabric is planar); T = -hi means that the AMS ellipsoid is rotationally symmetric (uniaxial oblate). If — 1 < T < 0 the AMS ellipsoid is prolate (the magnetic fabric is linear); T = — 1 means that the AMS ellipsoid is uniaxial prolate. Bulk susceptibility field variation In the pyrrhotite-bearing schists, all three principal susceptibilities monotonically increase with field. The susceptibility measured in a field of 450 A/m may be up to 200% higher than that measured in 2 A/m (Fig. la). In the fields less than 80 A/m, the susceptibility increases with field in an approximately linear manner (Fig. Ib). In the titanomagnetite-bearing volcanics, the field variation of principal susceptibilities also exists, but is not as conspicuous as that in the pyrrhotite-bearing schists and the curve representing the susceptibility vs. field relationship is approximately linear (Fig. 2).

AMS IN VARIABLE FIELDS

71

Fig. 1. Field variation of the principal susceptibilities in a specimen of gneiss with pyrrhotite, (a) full range of the fields used, (b) fields between 2A/m and 80 A/m. Note approximately linear relationship between susceptibility and field intensity in this field range.

In the single crystals of hematite, the susceptibility is two to three orders of magnitude stronger along the basal plane than along the c axis. The susceptibility along the basal plane shows very strong field variation, increasing several times between the fields of 2 A/m and 450 A/m, while the susceptibility along the c axis displays much weaker, if any, field variation (Fig. 3). AMS field variation The orientations of the principal susceptibilities are virtually field-independent in all the specimens investigated. The differences in orientations of the principal susceptibilities measured in various fields are so small that they cannot be visualized using a standard equal-area projection. Instead, the following approach was

used. First, the mean susceptibility tensor was calculated for each specimen from the AMS measurements made in all fields; the technique by Jelinek (1978) was used, consisting of averaging individual components of the susceptibility tensors normalized by the mean susceptibility and measured in individual fields. Second, the differences in the principal directions (in terms of declination and inclination) between the mean tensor and the tensors measured in individual fields were calculated and these differences are presented in Figure 4. It is obvious from Figure 4 that the differences are very small, less than 2°. In addition, the differences show no systematic trends; they appear to be random. Figure 5 shows the standard deviations of the differences in declination and inclination for five different pyrrhotite-bearing specimens. Again, the deviations are very small, less than 2°.

72

J. POKORNY ET AL.

Fig. 2. Field variation of the principal susceptibilities in a specimen of basalt with titanomagnetite.

Figures 6 and 7 show the field variation of the degree of AMS and shape parameter in a pyrrhotite-bearing schist specimen and in a titanomagnetite-bearing basalt specimen, respectively. In strongly anisotropic gneiss with pyrrhotite, the degree of AMS increases with field considerably, while the shape parameter varies only very mildly. In much less anisotropic basalt with titanomagnetite, the degree of AMS also tends to increase with field, but this increase is much more modest. The shape parameter shows very little, if any, field variation. In hematite single crystals that are extremely strongly anisotropic (see Fig. 3), the minimum susceptibility calculated using the linear theory is negative. However, as shown by direct measurement, this negative value has nothing to do with diamagnetism, evidently resulting

from linear fit to non-linear data. Consequently, it is inappropriate to calculate the P and T parameters in this case. Discussion The fields used in the common AMS meters (Table 1) enable accurate measurement of the field-independent susceptibility in diamagnetic and paramagnetic minerals as well as in magnetite. The linear theory of the AMS is then fully legitimate. For pyrrhotite, hematite and titanomagnetite, these applied fields are an order of magnitude stronger than the nonlinearity threshold. Consequently, the calculation of AMS using linear theory is in principle incorrect in this case.

Fig. 3. Field variation of the susceptibility along the basal plane and along the c axis in a hematite single crystal.

AMS IN VARIABLE FIELDS

73

Fig. 4. Field variation of the orientations of principal susceptibilities in a specimen of gneiss with pyrrhotite.

Fig. 5. Standard errors in the determination of the maximum susceptibility direction for 5 specimens of gneiss with pyrrhotite.

Our investigations made using a precise AMS meter, the KLY-4S Kappabridge, fully corroborate the results by de Wall (2000) and Hrouda (2002), obtained using a less precise KLF-3 miniKappa instrument (not originally designed to measure the AMS), that using linear theory in measuring pyrrhotite-, hematite- and titanomagnetite-bearing rocks does not necessarily give rise to inaccurate results in all AMS aspects. Namely, the principal directions are virtually field-independent. The variation of the shape parameter with field is also relatively weak. Consequently, if one is primarily interested in the orientations of magnetic lineation and

foliation and in symmetry of the AMS ellipsoid (and this is the case in most geological applications), one can use the simple and illustrative linear theory without danger of the loss of accuracy. On the other hand, the degree of AMS in pyrrhotite-, hematite- and titanomagnetite-bearing rocks may show conspicuous variation with field. If one is interested only in relative changes in fabric intensity, one can use the linear theory as well, provided that all the specimens are measured in the same field, and that all have similar P(Hac) dependences. However, if one wants to make precise quantitative fabric

74

J. POKORNY ET AL.

Fig. 6. Field variation of the degree of AMS and the shape parameter in a specimen of gneiss with pyrrhotite.

Fig. 7. Field variation of the degree of AMS and the shape parameter in a specimen of basalt with titanomagnetite.

interpretation, it is desirable to work with the initial susceptibility, which is defined rigorously. The KLY-4S Kappabridge enables the AMS to be measured in very weak fields, which would theoretically solve the problem in the best and purest way. The only problem can be reduced sensitivity and measuring precision when using very weak fields. It is obvious that this solution is hardly applicable to weakly magnetic and weakly anisotropic rocks. The susceptibility of multidomain materials measured in the low fields in which the empirical Rayleigh law is valid can be described as (e.g. Hrouda 2002)

where K is the measured susceptibility, k is the initial susceptibility, a is the Rayleigh coefficient and H is the intensity of the magnetizing field. By measuring directional susceptibilities of a specimen in several fields within the Rayleigh law range, one would be able to determine the initial directional susceptibilities using the above equation. Then, one could calculate the tensor of the initial susceptibility, for instance using the Jelinek (1977) theory and program. The problem with this approach, in addition to extra measurements required, is the relatively narrow and variable Rayleigh law range for different minerals. Fortunately, the KLY-4S Kappabridge has the option of automatically measuring field variation of the directional susceptibility (in 21 fields

AMS IN VARIABLE FIELDS

ranging from 2 A/m to 450 A/m). One can easily find the extent of the Rayleigh law range and calculate the initial directional susceptibility. In the present study, the grain-size dependence of the field variation of AMS has not been investigated, because the field variation of susceptibility is inherently a multidomain phenomenon in the field range in question. This was nicely shown by Worm et al. (1993), who investigated this phenomenon in pyrrhotite. They found that while the field variation of susceptibility was very conspicuous in large grains (typically hundreds of micrometers in size), it was hardly observable in small grains (less than 30 um). For titanomagnetites the field-dependence is in general less substantial than that of pyrrhotite, but it appears to extend to lower grain sizes of a few micrometres (Jackson et al. 1998). Similar grain-size control on field-dependent AMS may reasonably be expected, but should be confirmed by further experiments.

75

fields (less than 10 A/m). This, however, could result in significantly lowering the sensitivity and precision of the AMS measurement. The other way of solving this problem would be measuring the directional susceptibilities in several fields outside the initial susceptibility range, but still within the Rayleigh law range, calculating the initial directional susceptibilities, and subsequently computing the tensor of initial susceptibility. M. Chlupacova, C. Vahle and H. Quade are thanked for providing us with the specimens of gneiss with pyrrhotite, basalt with titanomagnetite and single crystals of hematite, respectively. The help of B. B. Ellwood, D. K. Potter and M. Jackson in reviewing and editing process is highly appreciated. The research was partly supported financially by the Ministry of Education of the Czech Republic (grant #2431 3005).

References BORRADAILE, G. J., FRALICK, P. W. & LAGROIX, F.

Conclusions The investigation of the field variation of the AMS in specimens containing various different magnetic minerals has led to the following conclusions: The theory of the low-field AMS of rocks is based on the assumption of a linear relationship between magnetization and magnetizing field, resulting in field-independent susceptibility. This relationship is valid in diamagnetic and paramagnetic minerals by definition and also in magnetite in the fields used in common AMS meters. Using linear theory in calculating the AMS is fully legitimate in all these cases. In pyrrhotite-, hematite- and titanomagnetitebearing rocks, a clear field dependence of susceptibility may exist even in the low fields used in common AMS meters. Strictly speaking, the use of linear theory in calculating the AMS is in general incorrect in this case. Fortunately, the linear theory nevertheless yields accurate estimates of the orientations of the principal susceptibilities and of the AMS ellipsoid shape parameter T; it gives rise to inaccurate determination of the degree of AMS. The problem of inaccurate determination of the degree of AMS plays no role if one is primarily interested in the orientations of magnetic lineation and foliation and in the symmetry of the AMS ellipsoid, and this is the case in most geological applications. It can be solved using very weak measuring

1999. Acquisition of anhysteretic remanence and tensor subtraction from AMS isolates true palaeocurrent grain alignments. In: TARLING, D. H. & TURNER, P. (eds.) Palaeomagnetism and Diagenesis in Sediments, Geological Society, London, Special Publications, 151, 139-145. CHLUPACOVA, M. 1984. Anisotropy of magnetic susceptibility (in Czech). In: CHLUPACOVA, M., PRUNER, P., KRSOVA, M., HROUDA, F. & JELINEK, V. Magnetic properties of rocks and ores with pyrrhotite (in Czech). Unpublished report of Geofyzika, n.p. Brno, 76-106. COLLINSON, D. W. 1993. Measurement of the anisotropy of low- and high-field susceptibility. In: TARLING, D. H. & HROUDA, F. The Magnetic Anisotropy of Rocks. Chapman and Hall, London, 72-80. DE WALL, H. 2000. The field dependence of AC susceptibility in titanomagnetites: implications for the anisotropy of magnetic susceptibility. Geophysical Research Letters, 27, 2409-2411. HROUDA, F. 2002. Low-field variation of magnetic susceptibility and its effect on the anisotroy of magnetic susceptibility of rocks. GeophysicalJournal International, 150, 715-723. HROUDA, F. & QUADE, H. 1997. Non-linear magnetization in hematite basal plane and its implication for AMS determination. Annal. Geophys., 15, supplement, Cl 13. HROUDA, F., POKORNY, J. & QUADE, H. 1998. Magnetic anisotropy and low temperature susceptibility of hematite single crystals from Minas Gerais, Brazil. EOS Transactions, AGU, 79, supplement, F236. JACKSON, M., MOSKOWITZ, B., ROSENBAUM, J., KISSEL, C. 1998. Field-dependence of AC susceptibility in titanomagnetites. Earth and Planetary Science Letters, 157, 129-139.

76

J. POKORNY ET AL.

JELINEK, V. 1973. Precision A.C. bridge set for measuring magnetic susceptibility of rocks and its anisotropy. Studio, Geophysica et Geodaetica, 17, 36-48. JELINEK, V. 1980. Kappabridge KLY-2. A precission laboratory bridge for measuring magnetic susceptibility of rocks (including anisotropy). Leaflet, Geofyzika Brno. JELINEK, V. 1977. The statistical theory of measuring anistropy of magnetic susceptibility of rocks and its application. Geofyzika, n.p. Brno. JELINEK, V. 1981. Characterization of magnetic fabric of rocks. Tectonophysics, 79, T63-T67. JELINEK, V. 1996. Measuring anisotropy of magnetic susceptibility on a slowly spinning specimen - basic theory. AGICO Print No. 10 (Leaflet). JELINEK, V. & POKORNY, J. 1997. Some new concepts in technology of transformer bridges for measuring

susceptibility anisotropy of rocks. Physics and Chemistry of the Earth, 22, 179-181. MARKERT, H. & LEHMANN, A. 1996. Three-dimensional Rayleigh hysteresis of oriented core samples from the German Continetal Deep Drilling Program: susceptibility tensor, Rayleigh tensor, threedimensional Rayleigh law. Geophysical Journal International, 127, 201-214. NAGATA, T. 1961. Rock Magnetism. Maruzen, Tokyo. TARLING, D. & F. HROUDA 1993. The Magnetic Anisotropy of Rocks. Chapman and Hall, London. WORM, H.-U. 1991. Multidomain susceptibility and anomalously strong low field dependence of induced magnetization in pyrrhotite. Physics of the Earth and Planetary Interiors, 69, 112-118. WORM, H.-U., CLARK, D. & DEKKERS, M. J. 1993. Magnetic susceptibility of pyrrhotite: grain size, field and frequency dependence. Geophysical Journal International, 114, 127-137.

The anisotropy of magnetic susceptibility (AMS) in low-grade, cleaved pelitic rocks: influence of cleavage/bedding angle and type and relative orientation of magnetic carriers TIMOTHY N. DEBACKER1, PHILIPPE ROBION2 & MANUEL SINTUBIN3 1

Structural Geology & Tectonics Group, Katholieke Universiteit Leuven, Redingenstraat 16, B-30QO Leuven, Belgium (e-mail: [email protected]) 2 CNRS-UMR 7072 'Laboratoire de Tectonique', Departement des Sciences de la Terre et de VEnvironnement, Universite de Cergy-Pontoise, Avenue du Parc-le Campus- Bat I, 95031 Cergy Pontoise, France (e-mail: [email protected]) 3Structural Geology & Tectonics Group, Katholieke Universiteit Leuven, Redingenstraat 16, B-3000 Leuven, Belgium (e-mail: [email protected]) Abstract: Cambrian and Silurian, low-grade, pelitic rocks of the single-phase deformed Brabant Massif consistently have a maximum magnetic susceptibility axis (Kl) parallel to the cleavage/bedding intersection. In contrast, the minimum susceptibility axis (K3) either coincides with the bedding pole, with the cleavage pole or occupies an intermediate position. Anisotropy of anhysteretic remanence (AARM) and X-ray pole figure goniometry allow the distinguishing of the orientation distributions of the ferromagnetic and paramagnetic (white mica and chlorite) carriers, respectively. Mismatches between K3 and the poles to the macroscopic fabric elements (i.e. bedding and cleavage) are attributed to different orientations of the different magnetic (s.l.) carriers. A strong relationship exists between the cleavage/ bedding angle and the shape parameter: low, respectively high angles leading to oblate, respectively prolate susceptibility ellipsoids. However, differences are observed between the Cambrian and Silurian samples in terms of the shape parameter and the behaviour of the degree of anisotropy with changing cleavage/bedding angle. This is tentatively attributed to differences in relative orientation and mineralogy of the magnetic (s.l.) carriers. These results demonstrate the influence of the relative orientation of the different carriers on AMS and suggest that, although being a petrofabric tool, AMS cannot be used as a strain gauge in the case of composite magnetic fabrics.

Because of the fine grain-size and the common scarcity of classical strain markers (deformed pebbles, macro-fossils, reduction spots, etc.), performing quantitative strain analyses in slate belts may be difficult. As an alternative, one may apply more analytical methods such as phyllosilicate X-ray pole figure goniometry (e.g. Oertel 1983; Sintubin \994a, b; van der Pluijm et al. 1994) and the analysis of the anisotropy of magnetic susceptibility (e.g. Graham 1954; Fuller 1964; Rathore 1979; Hrouda 1982; Borradaile & Henry 1997 and references therein). However, although these petrofabric methods have been applied in structural geology for more than 20 years, the relationship with strain is still debated. The anisotropy of magnetic susceptibility (AMS) in particular not only depends on the degree of deformation but is to a large extent controlled by the lithology (rock type, type of

magnetic carriers, and orientation and concentration of different carriers) (e.g. Borradaile & Tarling 1981; Borradaile 1987, 1988). According to several studies, in foliated rocks, such as slates, the principal magnetic susceptibility axes reflect the tectonic foliation, with the minimum susceptibility axis (K3) perpendicular to foliation and the maximum susceptibility axis (Kl) parallel to the tectonic extension direction or to the cleavage/bedding intersection (e.g. Rathore 1979; Hrouda 1982; Aubourg et al 1991; Robion et al. 1995; Borradaile & Henry 1997; Hirt et al. 2000; Nakamura & Borradaile 2001; Pares & van der Pluijm 2002). Similarly, in shales, the principal magnetic susceptibility axes may reflect the bedding-parallel compaction fabric, with the minimum susceptibility axis (K3) perpendicular to bedding (e.g. Li & Powell 1993; Hirt et al. 1995). Such a coincidence between the pole to the macroscopic fabric

From: MARTIN-HERNANDEZ, F., LUNEBURG, C. M., AUBOURG, C. & JACKSON, M. (eds) 2004. Magnetic Fabric: Methods and Applications. Geological Society, London, Special Publications, 238, 77-107. 0305-8719/04/S15.00 © The Geological Society of London 2004.

78

T. N. DEBACKER ET AL.

elements (cleavage in slate or bedding in shale) and the minimum susceptibility axis (K3) suggests that the (minimum axis of) AMS is controlled by one or more types of magnetic carriers all parallel to the macroscopic fabric elements. In such cases, a qualitative and possibly even a quantitative relationship may exist between AMS and strain (compaction strain in the case of shale, tectonic shortening strain in the case of slate). However, a different situation occurs in rocks with two competing macroscopic fabric elements. Common examples of these are poorly deformed sedimentary rocks, characterized by both a bedding fabric and a cleavage fabric. In such rocks the maximum susceptibility axis (Kl) is commonly parallel to the cleavage/ bedding intersection, whereas the minimum susceptibility axis (K3) may remain perpendicular to bedding (Saint Bezar et al 2002; Pares & van der Pluijm 2002), or in other cases may be perpendicular to cleavage or occupy an intermediate position in between the pole to bedding and the pole to cleavage (e.g. Robion et al. 1995; Liineburg et al. 1999; Frizon de Lamotte et al. 2002). The work of Liineburg et al. (1999), for instance, in which the minimum (K3) and intermediate (K2) susceptibility axes often show a girdle distribution (e.g. Housen et al. 1993; Pares & van der Pluijm 2003), demonstrates the competing influence of both beddingparallel (related to sedimentation, compaction and pre-kinematic metamorphism) and cleavage-parallel (related to tectonic shortening and syn- to post-kinematic metamorphism) magnetic (s.l.) carriers. In such cases, in which the minimum susceptibility axis is not perpendicular to the cleavage, it is unlikely that a simple relationship will exist between finite strain and AMS (e.g. Housen et al. 1993; Pares & van der Pluijm 2002). Indeed, Housen et al. (1993), who studied the effect of composite magnetic fabrics on AMS by means of experiments and numerical models, conclude that in the presence of composite magnetic fabrics, quantitative measures of finite strain are limited by the ability to determine accurately the degree of anisotropy and relative susceptibility of each component fabric element. Housen et al. (1993) further conclude that in the case of two competing fabrics (1) the maximum susceptibility axis (Kl) parallels the intersection axis of the two fabrics, and (2) the degree of anisotropy and shape of the susceptibility ellipsoid changes in function of the angle between the two fabrics. Although the parallelism of the maximum susceptibility axis and the cleavage/ bedding intersection is well documented, we are not aware of studies demonstrating the

effect of variations in angle between cleavage and bedding on AMS in natural rocks. This effect in natural rocks is documented in the present study. Moreover, an attempt is made to resolve the AMS-ellipsoid orientation qualitatively in terms of the preferred orientation of paramagnetic and ferromagnetic carriers. As will be demonstrated, even with the help of additional techniques, the AMS of rocks with composite magnetic fabrics may be difficult to interpret and therefore cannot be used as a measure of strain. Geological setting and sampling The largely concealed Lower Palaeozoic Brabant Massif (Fig. 1) is a typical example of a slate belt, forming the south-eastern part of the AngloBrabant deformation belt, one of the deformation belts of eastern Avalonia (Van Grootel et al. 1997; Verniers et al. 2002). The massif consists of low-grade, mainly fine-grained, siliciclastic deposits, ranging from the lowermost Cambrian in the core of the massif to upper Silurian along the rims. An angular unconformity separates these deformed Lower Palaeozoic deposits from overlying, diagenetic, undeformed Givetian deposits (Legrand 1967; De Vos et al. 1993; Van Grootel et al. 1997; Debacker et al. 1999). At present, there is only evidence for a single progressive deformation, currently considered to have taken place between the Llandovery and the end of the Early Devonian, possibly continuing into the Eifellian (Debacker 2001; Debacker et al. 2002). The main features associated with this deformation are folds with a well-developed, cogenetic cleavage (Sintubin 1997, 1999; Debacker 2001; Debacker et al. 2002; Verniers et al. 2002). The stratigraphy and structural architecture of the Brabant Massif are well known, making it an ideal study area for the geological application of AMS in low-grade slate belts. Criteria used to select sampling localities are the fine grain size (only fine-grained siltstone and claystone), the homogeneity of the deposits (in order to avoid grain-size-dependent variations such as cleavage refraction), the presence of tectonic folds with a moderately to welldeveloped cogenetic cleavage (in order to compare magnetic fabrics with bedding and cleavage orientations around folds), a well-known structural architecture and a known degree of metamorphism. Four suitable lithostratigraphic units from three large outcrops were sampled. These are the Ripain Member and the Asquempont Member of the Lower to lower Middle

COMPLEXITY OF AMS IN CLEAVED PELITES

79

Fig. 1. Geological subcrop map of the Brabant Massif (after De Vos et al. 1993 and Van Grootel et aL 1997), showing the position of the sampled outcrops. The upper right inset shows the position of the Brabant Massif, forming the southeastern part of the Anglo-Brabant Deformation Belt (ABDB), situated within the microcontinent Avalonia along the NE-side of the Midlands Microcraton (MM).

Cambrian Oisquercq Formation in outcrop Northern Asquempont section (Sennette valley), the Wenlock Vichenet Formation in the Vichenet section (Orneau valley) and the lower Ludlow Ronquieres Formation in the Ronquieres section (Inclined Shiplift of Ronquieres, Sennette valley) (Figs 1 & 2). All the sampled lithologies are homogeneous mudstone, mainly composed of white mica, chlorite and quartz, with a minor amount of dispersed opaque material (cf. Geerkens & Laduron 1996). All the sampled lithologies underwent an anchizonal degree of metamorphism, as suggested by illite crystallinity studies (Geerkens & Laduron 1996; Van Grootel et al. 1997) and the cleavage is moderately developed, corresponding to the embryonic cleavage stage to cleavage stage of Ramsay & Huber (1983). The Ripain Member, representing the lower member of the Lower to lower Middle Cambrian Oisquercq Formation, consists of blue-grey to purplish grey, extremely homogeneous finegrained mudstone (Verniers et al, 2001). The samples used in this study were taken from outcrop Northern Asquempont section (Figs 1 & 2). In this outcrop steeply plunging tectonic folds occur, with a Z-shaped geometry and a predominance of sub-vertical to steeply ENEdipping, WSW-ward younging limbs (Debacker 2001; Debacker et aL 2004). The rocks are affected by a moderately to well-developed cleavage, cogenetic with the folds. Microscopic observations show that white mica, oriented parallel to cleavage, is present throughout the rock mass. Occasionally, vague cleavage-parallel

alignments of opaque material, resembling cleavage domains, reflect a spaced cleavage. A spaced cleavage becomes apparent locally around chlorite/mica stacks and opaque objects. Chlorite/mica stacks occur sub-parallel to bedding, as well as sub-parallel to cleavage (Debacker 2001). The Asquempont Member represents the upper member of the Lower to lower Middle Cambrian Oisquercq Formation. It consists of rather porous, greenish grey to green, very homogeneous mudstone, occasionally with laminated siltstone. In outcrop Northern Asquempont section this member contains folds with variable plunges, ranging from sub-horizontal to steeply plunging, which are cogenetic with the cleavage (Fig. 2; Debacker 2001; Debacker et al, 2004; cf. Sintubin et al. 1998). The cleavage is only moderately developed, often having an irregular, anastomosing nature. Microscopically, cleavage is mainly reflected by white mica, distributed throughout the rock mass and oriented parallel to cleavage. A spaced cleavage, reflected by vague bands of opaque material, is only apparent around chlorite/mica stacks. Chlorite/mica stacks occur sub-parallel to both bedding and cleavage (Debacker 2001). The Vichenet Formation (Wenlock, Silurian) consists of grey mudstone, siltstone and fine sandstone, often with calcareous pelite, forming an alternation of distal thick-bedded turbidites, characteristically Tde-sequences (for turbidite terminology see Bouma 1962), and thin-bedded laminated hemipelagites (De Schepper 2000;

Fig. 2. Sample positions within, and structural architecture of, the three sampled outcrops. Structural data from the Northern Asquempont section are taken from Debacker (2001) and Debacker et al. (2004), those from the Vichenet section are taken from Belmans (2000; see also Debacker 2001 and Herbosch et al. 2002), and those from the central antiform within the Inclined Shiplift of Ronquieres are taken from Debacker et al. (1999). Lower-hemisphere equal-area stereographic projections are added showing bedding, cleavage and cleavage/bedding intersections. Samples in bold were used for AMS analysis, whereas those in italic were used for phyllosilicate X-ray pole figure goniometry. Samples in bold and italic were used for both. Underlined (bold) samples were used for (p)AARM analyses, and (bold) samples marked by * were heated in the oven in order to determine the ferromagnetic mineralogy.

COMPLEXITY OF AMS IN CLEAVED FELIXES

Verniers et al. 2001). Samples were taken from the type-section of this formation, the Vichenet section in the Orneau valley (Figs 1 & 2). Only the most fine-grained, homogeneous interval was sampled, being the homogeneous, pelitic e-interval (cf. Bouma 1962). The type-section contains sub-horizontal to gently plunging open folds, affected by a well-developed cogenetic cleavage, showing a pronounced convergent cleavage fanning with a symmetrical disposition with respect to the fold hinges (Fig. 2; Belmans 2000; Herbosch et al 2002). Microscopically, the cleavage has an anastomosing disjunctive nature in the sampled intervals. The width of the microlithons ranges from 10 to 30 urn, whereas the cleavage domains, with a high concentration of opaque material, are generally 5 to lOum wide. Chlorite, as a part of chloritemica stacks, generally occurs in the microlithons, whereas white mica, oriented parallel to cleavage, is abundant in the cleavage domains (Belmans 2000; Debacker 2001). The Ronquieres Formation (lower Ludlow, Silurian) consists of grey mudstone, siltstone and fine sandstone, forming an alternation of distal turbidites, predominantly Tcde-intervals (for turbidite terminology see Bouma 1962), and laminated hemipelagites (Louwye et al. 1992; Verniers et al 1992, 2001). Samples were taken from the type-section of this formation, the Ronquieres section in the Inclined Shiplift of Ronquieres, Sennette valley (Figs 1 & 2). Also here, only the most fine-grained, homogeneous, interval was sampled, being the pelitic e-interval. Because of the availability of turbidite logs and at least two marker horizons (Verniers et al 1992), it was possible to sample a single bed across a large antiform occupying a central position within the section (Fig. 2). The sampled bed is the e-interval of turbidite sequence 121 of Verniers et al (1992), situated between 30 and 65 cm below the lower marker horizon depicted in Figure 2. The sampled antiform has a gentle to open interlimb angle, a sub-horizontal to gentle plunge and a well-developed cogenetic cleavage, showing a convergent cleavage fanning with a symmetrical disposition with respect to the fold hinge (Legrand 1967; Debacker et al 1999). Microscopically, the cleavage has a spaced nature, with microlithons in the order of 10 to 30 um wide. White mica is mainly concentrated in the cleavage domains, aligned subparallel to cleavage, whereas chlorite, as a part of chlorite/mica stacks, is mainly concentrated in the microlithons, usually statistically aligned sub-parallel to bedding. Opaque material is usually concentrated in the cleavage domains. It is not clear whether cleavage is a disjunctive

81

or a crenulation type cleavage (Debacker et al 1999; Debacker 2001).

Methodology Magnetic anisotropy studies The vast majority of the investigated samples consists of oriented hand specimens (e.g. TD001) that were cut into cubes, on average 7 per sample (cf. Table 1), with size 2 x 2 x 2cm (e.g. TDOOla, TDOOlb, etc.). Only in four cases (TD1001, TD1002, TD1003, TD1004) was sampling performed by means of a hand drill, giving cylinders with a diameter of 2.4cm and 2.1cm high. The reason for the small number of drill core samples with respect to hand specimen cubes is the presence of the cleavage, along which the rocks tend to break during drilling. Broken samples, both cubes and cylinders, were glued together using a non-magnetic glue. There are no significant differences in results between the cylinders and the cubes, neither between the intact samples and the samples that were glued together again. The anisotropy of the low field susceptibility (AMS) and of the remanent magnetization (AARM) were investigated, the former at Katholieke Universiteit Leuven (Belgium), the latter at Cergy-Pontoise University (France). Whereas the anisotropy of magnetic susceptibility (AMS) represents the contribution of all magnetic constituents of a rock, anisotropy of anhysteretic remanent magnetization (AARM) only reflects the ferromagnetic fraction without paramagnetic and diamagnetic contributions. Hence, AMS and AARM have different sources, and combining these two methods may give complementary information in cases where sources of magnetization are complex (McCabe et al 1985). AMS was measured with a KLY3S Kappabridge (Jelinek & Pokorny 1997) at Katholieke Universiteit Leuven. The susceptibility tensor is computed by using the device software. The eigenvectors of this tensor, Kl, K2 and K3, corresponding to the maximum, intermediate and minimum susceptibility, respectively, reflect the orientation and shape of the magnetic ellipsoid. Three different arrangements of the eigenvectors Kl, K2 and K3 are used: the corrected degree of anisotropy Pj (Jelinek 1981), the shape parameter T (Jelinek 1981) and the mean susceptibility Km (cf. Borradaile 1988; Tarling & Hrouda 1993). To investigate the effects of mineralogy on the susceptibility anisotropy, Pj and Km are compared. The effect of the ferromagnetic fraction on Pj is generally accompanied

82

T. N. DEBACKER ET AL.

by an increase of Km, and on a Pj/Km plot the paramagnetic contribution has an upper limit around Pj-1.2-1.3 and Km ~ 3 00500 x 10~6SI (Rochette 1987a; Rochette et al. 1992; Martin-Hernandez & Hirt 2003). Also the shape parameter T and the degree of anisotropy Pj are compared. Whereas Pj reflects the degree of preferred orientation of magnetic minerals (jJ.), T is a measurement of the shape of the ellipsoid. If -1 < T < 0, the susceptibility ellipsoid is prolate, whereas if 0 < T < 1, the susceptibility ellipsoid is oblate (Jelinek 1981). The AMS data, averaged per sample, are summarized in Table 1. AARM is based on the ability of samples to acquire a remanent magnetization when an alternating field is applied in the presence of a small direct field (McCabe et al. 1985). Magnetization is imparted along a chosen direction of the sample in an alternating field peak of lOOmT with a coaxial small direct field of lOOuT. Before processing magnetization, the sample is demagnetized with an alternating field of lOOmT. Both the magnetization and demagnetization procedures are performed by the LAD-3 AF device manufactured by AGICO. After each magnetization step the sample is measured with a JR5 spinner magnetometer. The AARM tensor is determined by using the Jelinek procedure (Jelinek 1993) that had been developed for characterizing the anisotropy of isothermal remanent magnetization. In this procedure, 12 senses (6 directions) of magnetization are used, with the first and second sense opposed (same direction), the third and fourth opposed and so forth. This measuring scheme is useful when a hard coercivity component of magnetization cannot be demagnetized with a lOOmT alternating field. The remanibility tensor R is constructed in the same way as the susceptibility tensor K, by means of the least square inversion method. By combining principal remanibility values Rl, R2, R3 (with Rl > R2 > R3) we obtain the anisotropy degree PjR and the shape parameter T R . Because of the maximum value of the alternating field applied, the main minerals controlling AARM are magnetite, with wide range of coercivities, and low-coercivity pyrrhotite. In addition, on a few selected samples we measured partial AARM (pAARM). This method consists of imparting an anhysteretic magnetization in a selected window between two specified values of alternating field, in the presence of a lOOuT direct field. On the basis of coercivity spectra of the samples, we chose two windows, the first between 0 and 50mT and the second between 50 and lOOmT (cf. Jackson et al. 1988).

Magnetic mineralogy Ferromagnetic mineralogy was investigated by applying a stepwise demagnetization of a 'three axis' isothermal remanent magnetization following the procedure described by Lowrie (1990). This coercivity/blocking temperature spectrum analysis separates ferromagnetic minerals with different magnetic properties. We applied three successive saturation fields (1.4T, 0.6T and 0.12T) along three perpendicular directions on the samples. These are then demagnetized thermally in steps of 50 °C, with finer steps around 325 °C and 580 °C, the Curie temperatures of pyrrhotite and magnetite respectively. Samples were heated up to 700 °C. During this stepwise demagnetization, bulk magnetic susceptibility was monitored with a KLY3S at room temperature in order to detect artificial changes in magnetic properties due to heating

X-ray pole figure goniometry Mica (001) (
Table 1. Main parameters of the investigated samples, averaged over the measured specimens (cubes, cylinders) Sample

n

Ripain Member, TD1001 7 TD1002 9 TD001 6 TD002 5 TD256 7

Kl

K2

K3

95% confidence E12/E23/E13

Foliation F

Angle SO-S1

Degree of anisotropy Pj

Mean suscept.

1.431 5 ±0.0265 1.401 6 ±0.0228 1.4369 ±0.0308 1.3232±0.0173 1.4631 ±0.0378

407.1 ±20.5 415.9±33.7 401.5±7.6 386.8 ±10.5 375.9 ±6.3

Sennette valley 0.4109 ±0.0356 0.5349 ±0.0856 0.7350 ±0.0691 0.5056 ±0.0348 0.6331 ±0.0599 -0.2965 ±0.0620 -0.0298 ±0.1 108 0.3594 ±0.1 323 0.6533 ±0.0467 0.7514±0.0172 0.2179±0.1356

1.1604±0.0161 1.1 626 ±0.0097 1.1932±0.0133 1.1934 ±0.0080 1.1670±0.0218 1.0755 ±0.0046 1.1036±0.0119 1.1 732 ±0.0093 1.1913±0.0124 1.2304 ±0.0067 1.1367 ±0.0235

339.7 ±14.3 43 1.9 ±23.4 327.4 ±8.2 358.2±19.1 381.1±37.5 437.6 ±22.8 335.2±9.9 318.4±18.3 333.0 ±6.7 323.2 ±10.0 316.6±10.5

Shape parameter T

Oisquercq Formation (Lower to lower Middle Cambrian), Asquempont section, Sennette valley 1.0924 1.0629 0.8448 8.14/0.97/0.87 1.0279 1.2583 0.7867 ±0.1 123 16 1.0379 1.0932 1.0534 0.8534 4.81/0.91/0.72 1.2344 0.7011 ±0.1110 25 1.2574 1.0972 1.0598 0.8430 23.33/4.07/3.48 1.0353 0.7354 ±0.0611 16 1.0671 1.0539 0.8790 16.62/1.30/1.20 1.0126 1.1990 0.8714±0.0173 25 1.1167 1.0497 0.8336 12.60/4.83/3.43 1.0640 1.2594 0.5755 ±0.1242 26

Asquempont Member, Oisquercq Formation TD1003 13 1.0506 1.0150 0.9344 TD1004 6 1.0466 1.0193 0.9342 TD003 9 1.0469 1.0297 0.9235 TD260 7 1.0560 1.0214 0.9226 TD262 8 1.0441 1.0230 0.9329 TD274 6 1.0322 0.9939 0.9739 TD266 5 1.0453 0.9985 0.9562 TD199 9 1.0576 1.0139 0.9285 TD271 6 1.0494 1.0266 0.9240 TD197 6 1.0545 1.0354 0.9102 TD248 9 1.0508 1.0074 0.9418

(Lower to lower Midc le Cambrian), Asquempont 1.0351 0.95/0.47/0.33 1.0863 1.0269 5.08/1.50/1.07 1.0911 1.0167 15.56/2.48/2.11 1.1150 1.0339 7.53/2.60/1.96 1.1071 1.0207 7.96/1.98/1.59 1.0966 1.0384 3.70/6.98/2.43 1.0206 1.0468 13.04/14.80/7.02 1.0443 1.0431 12.12/7.19/4.47 1.0920 1.0222 9.67/2.18/1.78 1.1111 1.0184 12.63/1.95/1.72 1.1376 4.27/2.73/1.62 1.0432 1.0697

Vichenet Formation (Wenlock), Vichenet section, Orneau valley TD1020 6 1.0607 0.9796 0.9597 3.95/16.07/3.18 TD1023 6 1.0581 0.9908 0.9511 4.85/9.70/3.25 TD1026 6 1.0547 1.0064 0.9389 9.58/7.05/4.08 TD1025 5 1.0555 1.0012 0.9432 4.48/4.30/2.20 TD1021 4 11.0589 0.9738 0.9673 1.40/19.05/1.33 TD1021L 3 1.0398 0.9958 0.9645 5.03/7.03/2.97 Ronquieres TD1 TD2 TD3 TD4 TD5 TD6

Lineation L

Form ation (Ludlow), Inclined Shiplift of Ronquieres, 8 1.0370 0.9848 0.9782 4.05/32.51/3.56 7 1.0380 0.9834 0.9786 2.70/30.39/2.44 6 1.0325 0.9879 0.9797 14.63/54.72/12.10 8 1.0346 0.9862 0.9791 12.89/40.76/11.50 8 1.0361 0.9879 0.9760 9.38/31.20/7.20 10 1.0363 0.9892 0.9745 4.26/13.76/3.26

section, 55 25 25 39 18 71 86 50 13 40 61

1.0207 1.0418 1.0719 1.0615 1.0068 1.0324

74 60 46 41 87 87

-0.5909 ±0.0694 -0.2336±0.1343 0.1935 ±0.0769 0.0614 ±0.0501 -0.8515 ±0.0356 -0.1 509 ±0.0477

1.1508±0.0135 1.1 397 ±0.0086 1.1430±0.0118 1.1276±0.0167 1.1467 ±0.0011 1.0921 ±0.0036

311.5±7.9 316.7±6.6 240.8 ±7.9 274.6 ±16.2 250.8 ±3.9 295.3 ±5.4

SSennette valley 1.0531 1.0067 1.0556 1.0048 1.0452 1.0083 1.0491 1.0073 1.0488 1.0121 1.0475 1.0150

76 87 88 86 86 90

-0.7730 ±0.0957 -0.8379 ±0.0758 -0.7013 ±0.1992 -0.7264 ±0.1 160 -0.5949 ±0.1 803 -0.5152 ±0.0663

1.0897 ±0.0090 1.0922 ±0.0017 1.0781 ±0.01 18 1.0837±0.0177 1.0868 ±0.0072 1.0870 ±0.0057

315.4±8.4 295.6 ±7.2 311.8±9.9 315.7±7.6 307.7 ±7.4 298.5 ±6.4

1.0828 1.0680 1.0479 1.0543 1.0873 1.0442

Table 1. (cont.) Sample

n

Kl

K2

K3

95% confidence E12/E23/E13

Lineation L

Foliation F

Angle SO-S1

Shape parameter T

Additional samples Asquempont Member, Oisquercq Formation (Lower to lower Middle Cambrian) Virginal Railway section, Sennette valley 1.0492 1.0628 50 0.1186±0.1319 TD1031 6 1.0526 1.0033 0.9441 13.08/13.27/6.93 1.0933 30 0.5231 ±0.0265 1.0283 TD1032 9 1.0482 1.0194 0.9324 7.30/2.42/1.81 Neoproterozoic turbidites, TDD1 10 1.0338 TDD3 8 1.0253 TDD4 14 1.0298 TDD5 6 1.0141 TDD6 9 1.0298

Central 1.0003 1.0134 1.0153 1.0002 1.0172

Dobrogea , Moesian Platform, 0.9659 4.17/4.43/2.11 0.9613 9.15/2.14/1.73 0.9549 8.71/2.08/1.69 0.9857 8.55/8.48/4.28 0.9530 14.84/2.94/2.44

Romania 1.0335 1.0118 1.0143 1.0139 1.0124

1.0357 1.0542 1.0633 1.0146 1.0673

88 55 22 73 27

0.0119±0.1488 0.6367 ±0.0169 0.6251 ±0.0320 0.0208 ±0.1062 0.6828 ±0.0393

Degree of anisotropy Pj

Mean suscept.

1.1319±0.0300 1.1 672 ±0.0067

254.5 ±5.6 311.8±8.8

1.0773 ±0.0167 1.0932 ±0.0039 1.1 094 ±0.0043 1.0316±0.0031 1.1 144 ±0.0057

361.8±27.3 280.2 ±4.5 367.0 ±25.8 290.3 ±10.8 333.5±13.0

N: number of specimens; Kl, K2, K3: mean maximum, intermediate and minimum principal susceptibility axis; E12, E23, E13: mean 95% confidence angles of Kl, K2 and K3; lineation L —K1/K2; foliation F = K2/K3; angle SO-S1: angle between cleavage and bedding; shape parameter T = (2% — rji —rfe)/(rji -%), with 77! = lnKl, 7/2 = lnK2, and 773 = lnK3 (Jelinek 1981); corrected degree of anisotropy Pj = exp2[(r7! - 7?m)2 ± (rj2 ~~ rjm)2 ± (773 - T^)2]1/2, with rjm = (T?^7/3)1/3 (Jelinek 1981); mean susceptibility — (Kl + K2 ± K3)/3, expressed in 10~6 SI. In the case of the shape parameter, the corrected degree of anisotropy and the mean susceptibility, standard deviations are added in order to give an idea about possible variations between different specimens of the same sample.

COMPLEXITY OF AMS IN CLEAVED PELITES

85

Fig. 3. Demagnetization curves of one sample of the Ripain Member (TD001), two samples of the Asquempont Member (TD260 and TD266), two samples of the Vichenet Formation (TD1020 and TD1026) and two samples of the Ronquieres Formation (TD1 and TD4). The dark grey band shows the demagnetization temperature interval of the dominant ferromagnetic carrier (hematite in TD001, magnetite in TD260, TD266, TD1 and TD4, and pyrrhotite in TD1020 and TD1026), the pale grey band that of the additional ferromagnetic component (magnetite in TD001 and TD1026, pyrrhotite in TD260, TD266, TD1 and TD4). Note, that in sample TD001 (Ripain Member) the presence of hematite is reflected by the soft and medium component curves, and not, as one would expect, by the hard component curve. Probably this reflects low-coereivity, multidomain hematite (cf. Robion et al. 1997).

of 670 °C), with, however, a medium to low coercivity (>0.6T). Such low-coercivity hematite was already observed in upper Lochkovian slates on the Rocroi massif in the Ardennes (Robion et al. 1997) and was attributed to the presence of coarse-grained hematite. Magnetite and possibly some goethite are also observed in these samples but only in small amounts. The

Asquempont Member of the Oisquercq Formation, in contrast, has a ferromagnetic mineralogy dominated by magnetite (blocking temperature around 580 °C and low coercivity), with a small amount of pyrrhotite (blocking temperature between 325 and 350 °C and medium coercivity). The ferromagnetic mineralogy of the Vichenet Formation is dominated by pyrrhotite. This

86

T. N. DEBACKER ET AL.

pyrrhotite shows a wide range of coercivities, but mainly controls the low coercivity component. One of the two analysed samples of this lithology shows a small amount of additional magnetite (TD1026). The sampled bed within the Ronquieres Formation has a ferromagnetic mineralogy entirely consisting of magnetite, with locally a minor amount of pyrrhotite (Fig. 3). Anisotropy of magnetic susceptibility

Orientation analysis In all analysed samples the maximum susceptibility axis (Kl) coincides with the cleavage/bedding intersection (Fig. 4). However, in the case of the minimum susceptibility axis (K3), strong variations may occur, seemingly unrelated to lithology or structural position. In samples of the Ripain Member, all characterized by a small angle between cleavage and bedding (maximum 26°), K3 generally forms a cluster in between the cleavage and bedding poles (Fig. 4). In some samples, K3 tends to approximate the bedding pole (e.g. TD001, TD1002), whereas in other samples it tends to approximate the cleavage pole (e.g. TD002, TD256, TD1001). Although in all samples a rather small angle occurs between cleavage and bedding, we do not think that this deviation from the cleavage or bedding pole is due to errors induced by orientation and cutting irregularities. In samples of the Asquempont Member, K3 generally clusters around the bedding pole (Fig. 4; samples TD274, TD266, TD248, TD003, TD271, TD260, TD197, TD1003). Seemingly, two exceptions occur. In sample TD262, the position of K3 in between the bedding and cleavage poles may be an artefact due to the small angle between cleavage and bedding, in combination with orientation or cutting irregularities. For TD1004, the position of K3, closer to the cleavage pole than to the bedding pole, may also be an artefact. Unlike in the case of cubes, cut from oriented hand specimens on which bedding and cleavage orientation were determined, in the case of cylinders (TD1004, TD1003)

cleavage and bedding orientation were not taken from the cylinder itself. Hence, a small mismatch may be expected, which, in the case of a small angle between cleavage and bedding, will not allow determining whether K3 approximates the pole to cleavage or the pole to bedding or takes up a truly intermediate position. The samples of the Vichenet Formation (TD1021, TD1020, TD1025, TD1026) appear to be characterized by a sub-horizontal K3, occupying an intermediate position between the pole to cleavage and the pole to bedding (Fig. 4). Although in sample TD1021, K2 and K3 show a marked girdle distribution, reflecting its strong prolateness (cf. Table 1), in the other samples (TD1020, TD1025, TD1026) K2 and K3 are well defined, reflecting truly triaxial ellipsoids with an orientation in between that of cleavage and bedding. Only in the case of sample TD1023 does K3 coincide with the pole to cleavage. A laminated hemipelagite, present in sample TD1021, also shows a K3 parallel to the cleavage pole. The fact that in the same sample the e-interval, with the same cleavage/ bedding angle, shows a K3 in between the pole to cleavage and the pole to bedding, points to an influence of lithology on the orientation of the minimum axis of the susceptibility ellipsoid. The samples of the Ronquieres Formation, all taken from a single bed across the central antiform (e-interval of turbidite sequence 121 of Verniers et al. 1992), are characterized by a strongly variable K3 orientation (Fig. 4). In some samples K3 clusters around the pole to cleavage (TD2, TD4, TD6), whereas in other samples K3 takes up an intermediate position (TD3, TD5, TD1), in some cases closer to the bedding pole (TD1), in other cases closer to the cleavage pole (TD5). These variable relative orientations do not appear to show any relationship with the position within the fold (cf. Fig. 2). Many samples have K2-K3-girdles, reflecting a strong prolateness (e.g. TD1, TD3, TD4; cf. table 1), and hence a rather poorly defined nature of K3. However, the confidence ellipses suggest that the mismatches between K3 and the poles to the macroscopic fabric elements are significant.

Fig. 4. Lower-hemisphere equal-area stereographic projections showing the principal magnetic susceptibility axes (Kl, K2, K3; Kl > K2 > K3), the 95% confidence ellipses for the principal magnetic susceptibility axes, bedding (pole and plane) and cleavage (pole and plane) of samples of the Ripain Member (outcrop Northern Asquempont section), the Asquempont Member (outcrop Northern Asquempont section), the Vichenet Formation (outcrop Vichenet section) and the Ronquieres Formation (central antiform in outcrop Ronquieres section; Inclined shiplift of Ronquieres). Samples are shown going from N to S along each outcrop. See Fig. 2 for sample location. LHP: laminated hemipelagite (sample TD1021, Vichenet Formation).

COMPLEXITY OF AMS IN CLEAVED PELITES

87

88

T. N. DEBACKER ET AL.

Bulk susceptibility, degree ofanisotropy and shape parameter (T)

(Pj)

Samples of the Ripain Member all have a relatively high bulk susceptibility between 350 and 460 x 10~6 SI (Fig. 5, Table 1). Samples of the Asquempont Member show a large spread in bulk susceptibility, ranging between 280 and 470 x 10~6SI, with a maximum between 300 and 400 x 10~6 SI. This large spread is reflected both by different samples as by different cubes of the same samples, indicating the possibility of a strong variation in bulk susceptibility in apparently homogeneous deposits, even over short distances. In the Vichenet Formation the bulk susceptibility ranges from 230 to 330 x 10~6 SI and in the Ronquieres Formation from 280 to 330xlO~ 6 SI. The consistently higher values of the Ripain Member may reflect a relatively important influence of ferromagnetic carriers to the magnetic susceptibility (cf. Rochette et al 1992; Hrouda 2002). The degree of anisotropy (Pj) shows a clear difference between samples of the Ripain Member on the one hand, and samples of the Asquempont Member, the Vichenet Formation and the Ronquieres Formation on the other hand (Fig. 5, Table 1). Whereas the latter have a degree of anisotropy between 1.02 and 1.21, the former all have a degree of anisotropy higher than 1.29, ranging up to 1.51. Considering that the degree of anisotropy of white mica and chlorite, the main paramagnetic carriers in the investigated samples, is 1.15 (Martin-Hernandez & Hirt 2003; cf. Rochette et al 1992), a relatively important contribution of ferromagnetic carriers

to the AMS is expected in the case of the samples of the Ripain Member. The shape parameter (T) shows a strong spread for each of the four sampled lithostratigraphic units. However, a graph of the shape parameter versus the angle between cleavage and bedding (Fig. 6a; cf. Table 1) points to a major influence of the cleavage/bedding angle on the shape parameter. Oblate susceptibility ellipsoids are obtained in the case of small angles between cleavage and bedding, and prolate ellipsoids are obtained in the case of large angles between cleavage and bedding. Although showing the same relationship, the samples of the Ripain Member and the Asquempont Member (Cambrian samples) depart slightly towards the oblate field with respect to the samples of the Vichenet and Ronquieres formations (Silurian samples; Fig. 6a). It is noteworthy that the analysed hemipelagic parts of the Vichenet Formation (sample TD1021), although showing the same cleavage/bedding angle, show a departure towards the oblate field with respect to the cubes of the turbidite e-interval of the same sample (Table 1), suggesting also an influence of lithology on the shape parameter. In order to check the relationship between shape parameter and cleavage/bedding angle, two samples of the Asquempont Member from another outcrop (Virginal Railway section, 600m to the W of outcrop Northern Asquempont section) with different cleavage/bedding angles were also analysed (Table 1). Both samples plot in the elongated cluster given by the Asquempont and Ripain members. We also

Fig. 5. Graph of degree of anisotropy (Pj) versus bulk magnetic susceptibility. The grey bands correspond to the mean of the approximate upper limits of the paramagnetic contribution given in Rochette et al. (1992). Note the high Pj-values of the samples of the Ripain Member.

Fig. 6. (a) Graph of cleavage/bedding angle versus shape parameter (T). A marked relationship becomes apparent: low angles give rise to oblate susceptibility ellipsoids whereas high angles give rise to prolate ellipsoids. Note that the samples of the Vichenet and Ronquieres formations (Silurian samples) show a slight shift towards the prolate field with respect to the samples of the Ripain and Asquempont members (Cambrian samples). In addition, note the marked difference in shape parameter between the e-interval and a laminated hemipelagite of the same sample (TD1021) of the Vichenet Formation.

Fig. 6. (b) Graph of the cleavage/bedding angle versus the degree of anisotropy (Pj). For the samples of the Asquempont Member and those of the Neoproterozoic turbidites of the Moesian Platform, the degree of anisotropy slightly increases with decreasing cleavage/bedding angle, compatible with the results of Housen et al. (1993). In contrast, in the case of the Vichenet Formation, the degree of anisotropy slightly decreases with decreasing cleavage/bedding angle.

COMPLEXITY OF AMS IN CLEAVED PELITES

included five samples with different cleavage/ bedding angles from fine-grained e-intervals of Neoproterozoic turbidites from five different outcrops of the Moesian Platform, Central Dobrogea (Romania; Table 1). Again, a similar relationship becomes apparent between the shape parameter and the cleavage/bedding angle (Fig. 6a). The one sample showing a slight departure towards the oblate field, macroscopically has a cleavage that is better developed than in the other samples. The fact that this relationship between T and the cleavage/bedding angle is also reflected by the samples of the Moesian platform indicates that it is not a regional phenomenon. A graph of the degree of anisotropy Pj versus the cleavage/bedding angle shows that this angle also has an influence on the degree of anisotropy (Fig. 6b). However, this influence is different for the samples of the Vichenet Formation on the one hand and the samples of the Asquempont and Ripain members and those of the Moesian platform on the other hand (Fig. 6b; cf. Fig. 7). In the former case, the degree of anisotropy decreases slightly with decreasing cleavage/bedding angle, whereas in the latter case, the degree of anisotropy increases with decreasing cleavage/bedding angle.

Anisotropy of anhysteretic remanent magnetization (AARM)

Orientation of principal axes with respect to macroscopic fabric elements and AMS axes Seemingly, the axes of AARM are more difficult to link to the macroscopic fabric elements (bedding, cleavage) than the axes of the magnetic susceptibility ellipsoids. Whereas in the latter case, Kl always coincides with the cleavage/ bedding intersection and only K3 shows a variation with respect to the macroscopic fabric elements, in the case of the anhysteretic remanence, both the maximum axis (Rl) and the minimum axis (R3) have variable orientations with respect to the macroscopic fabric elements (Fig. 8, Fig. 10). In all analysed samples of the Oisquercq Formation (Ripain Member and Asquempont Member), R3 consistently clusters around the bedding pole, showing a large spread in the case of the Asquempont Member, whereas Rl takes up a position within to possibly slightly oblique to the bedding plane, at high, but variable angles to the cleavage/bedding intersection (Fig. 8). The high-angle obliquity between Rl

91

and the cleavage/bedding intersection, occurring in three samples, all having completely different cleavage/bedding relationships, indicates that Rl is likely to mark a pre-deformational feature, which is compatible with R3 approximately coinciding with the bedding pole. For this reason, in each of the investigated samples bedding was unfolded to horizontal, using the local cleavage/bedding intersection as an approximation of the local fold axis. After unfolding, each of the samples, both from the Ripain Member and the Asquempont Member, shows a subhorizontal to gently N- to NNE-plunging Rl (Fig. 8). This ferromagnetic lineation, probably caused by coarse-grained hematite in the case of the Ripain Member and by magnetite in the case of the Asquempont Member, is oriented subperpendicular to sedimentary ripples recorded in deposits of the Asquempont Member. Because of the presence of two populations of ferromagnetic carriers in the samples of the Ripain Member, as indicated by the coercivity spectra (Fig. 9), pAARM was performed for the field between 0 and 50mT (low coercivity window) and for the field between 50 and lOOmT (high coercivity window) (Fig. 8). The low-coercivity R3, thought to be caused by the preferred orientation of larger ferromagnetic grains (cf. Jackson et al 1988), clusters around the bedding pole, whereas those of the high coercivity window, probably caused by smaller ferromagnetic grains (cf. Jackson et al. 1988), take up a position in between the bedding pole and the cleavage pole (Fig. 8). Although this is a very small angular difference, of which the significance might be questioned considering the accuracy of the JR5 spinner magnetometer, this difference is consistent for all analysed samples. Possibly, this difference reflects a partial rotation of the smaller ferromagnetic grains away from the bedding towards the cleavage, and hence a tectonic influence on the highcoercivity fraction, which seemingly is not reflected by the ferromagnetic fraction with a larger grain-size. This is also reflected by the Rl of one of the samples (Fig. 8): whereas Rl is oriented somewhere in the bedding plane, at high angles to the cleavage/bedding intersection in the case of AARM and low-coercivity pAARM, Rl of the high-coercivity pAARM has an orientation sub-parallel to the cleavage/ bedding intersection, thus reflecting some tectonic influence on the smaller ferromagnetic grains in one of the samples. In the Vichenet Formation, the results differ from sample to sample (Fig. 10). In sample TD1020, R3 of AARM, low-coercivity pAARM and high-coercivity pAARM all cluster

Fig. 7. Graph of the shape parameter (T) versus the degree of anisotropy (Pj). The isolated position of the Ripain Member is due to the small angle between cleavage and bedding (resulting in high T; cf. Fig. 6a) and to a relatively important contribution of ferromagnetic carriers (resulting in a relatively high Pj). Two trends can be observed in the other samples, both coinciding with a decrease in angle between cleavage and bedding. In samples of the Asquempont Member and of the Neoproterozoic turbidites, a decrease in cleavage/bedding angle results in an increase in T (cf. Fig. 6a) and a slight increase in Pj (arrow A; cf. Fig. 6b). In contrast, in samples of the Vichenet Formation, a decrease in cleavage/bedding angle also results in an increase in T (cf. Fig. 6a) but is accompanied by a slight decrease in Pj (arrow B; cf. Fig. 6b). Note that, although the cleavage/bedding angle remains the same, as well as the mean grain size, a small change in lithology/sedimentology (turbidite e-interval on the one hand versus laminated hemipelagite on the other hand) can cause large differences in Pj and T, as demonstrated by sample TD1021 of the Vichenet Formation (arrow C).

COMPLEXITY OF AMS IN CLEAVED PELITES

93

Fig. 8. Lower-hemisphere equal-area stereographic projections showing the principal axes of anisotropy of anhysteretic remanent magnetism (AARM), partial AARM (pAARM), AMS and macroscopic fabric element data of selected samples of the Oisquercq Formation. Unfolding the data around the local cleavage/bedding intersection (as an approximation of the local fold axis) gives similar orientations of the long axes of remanence for all samples, sub-perpendicular to sedimentary ribbles. See text for discussion.

around the cleavage pole, reflecting a tectonic control (cleavage-parallel) on the ferromagnetic carriers (pyrrhotite; cf. Fig. 3). The maximum axis of AARM and high-coercivity pAARM (small grains; cf. Jackson et al. 1988) coincides with the cleavage/bedding intersection, but Rl of low-coercivity pAARM, also situated within the cleavage plane, is markedly oblique to the cleavage/bedding intersection. Seemingly, this suggests that, although the majority of the ferromagnetic carriers, together with the small carriers, are tectonically controlled, but still influenced by the bedding fabric, the larger ferromagnetic carriers are not influenced by the original bedding fabric and hence fully controlled by the tectonic fabric. In sample TD1026, Rl of AARM and low-coercivity pAARM is steeply plunging within, or slightly oblique to the bedding plane, at high angles to the cleavage/bedding intersection, except for cube TD1026c, in which Rl of AARM plunges steeply within the cleavage plane. The minimum axis of AARM and low-coercivity pAARM is slightly oblique to the bedding pole, except for cube TD1026c that has an R3 of AARM that approaches the cleavage pole. In sample TD1025, the R3 of low-coercivity and highcoercivity pAARM coincides with the cleavage

pole, whereas Rl of low-coercivity and highcoercivity pAARM is steeply plunging within the cleavage plane. This suggests a tectonic control on both the small and large ferromagnetic carriers, without there being an influence of the original bedding fabric. Although taken from a single bed, the samples of the Ronquieres Formation also give results that differ from sample to sample (Fig. 10). As in the case of AMS, the orientation of the minimum axis of AARM varies from sample to sample and does not always coincide with the pole to the macroscopic fabric elements. Because the coercivity spectra are all characterized by a single pronounced peak in the lower part of the spectrum, reflecting the presence of only one population (relatively large grains; Fig. 9; cf. Jackson et al 1988), only the low-coercivity pAARM was determined. In sample TD1, Rl is situated within the bedding plane, oblique to the cleavage/bedding intersection, whereas R3 occupies a position in between the bedding pole and the cleavage pole, slightly closer to the latter. Possibly, this reflects a faint tectonic influence on the low-coercivity pAARM. In sample TD4, Rl coincides with the cleavage/bedding intersection, and R3 approximates the cleavage pole (and K3 of AMS), thus suggesting a tectonic

94

T. N. DEBACKER ET AL.

Fig. 9. Coercivity spectra of the Oisquercq Formation (above: Ripain and Asquempont members), the Ronquieres Formation and the Vichenet Formation (below). In the latter case, also a laminated hemipelagite (LHP, cube TD1021c) has been analysed, adjacent to a turbidite e-interval (cube TD1021e) of the same sample (TD1021). Note the marked difference in remanent magnetization between the Ripain Member on the one hand and the Asquempont Member, the Vichenet Formation and the Ronquieres Formation on the other hand. In addition, note the variation in remanent magnetization between different samples of the same lithology and even between different specimens of the same samples.

COMPLEXITY OF AMS IN CLEAVED PELITES

95

Fig. 10. Lower-hemisphere equal-area stereographic projections showing the principal axes of anisotropy of anhysteretic remanent magnetism (AARM), partial AARM (pAARM), AMS (cf. Fig. 4) and macroscopic fabric element data of selected samples of the Vichenet Formation and the Ronquieres Formation. See text for discussion.

control on the orientation of the ferromagnetic carriers. In sample TD3, Rl coincides with the cleavage/bedding intersection, whereas R3 takes up an intermediate position in between the bedding pole and the cleavage pole, situated within the girdle-shaped cluster of K3 axes. The above results show that often the axes of remanebility do not coincide with the axes of susceptibility. Together with the relatively low degrees of anisotropy and bulk susceptibility within the Ronquieres Formation, the Vichenet Formation and the Asquempont Member, this suggests only a minor contribution of the ferromagnetic carriers to the AMS in these lithologies (cf. Hrouda & Jelinek 1990; Rochette et al. 1992; Hrouda 2002; Martin-Hernandez & Hirt 2003). Only in the Oisquercq Formation does a consistent relationship appear to exist between the remanebility ellipsoid on the one hand and the susceptibility ellipsoid and the macroscopic fabrics on the other hand. In the case of the Ripain Member this is compatible with the relatively high degree of anisotropy and mean susceptibility, suggesting a relatively important contribution of the ferromagnetic fraction (cf. Hrouda & Jelinek 1990; Rochette et al. 1992; Hrouda 2002; Martin-Hernandez & Hirt 2003).

Degree of anisotropy and shape parameter The degree of anisotropy of AARM, PjR, shows a very large spread (Fig. 11). For nearly all samples, PjR is significantly higher than Pj (AMS), which, judging from the literature, is generally the case (Stephenson et al 1986; Jackson 1991; Hrouda 2002). Only samples TD4g (Ronquieres), TD260 and two cubes (c and d) of sample TD266 have nearly the same degree of anisotropy for AMS and AARM. The shape parameter of AARM, TR, shows a very strong variation, pointing to strongly prolate to strongly oblate AARM ellipsoids for the samples of the Ripain and Asquempont members, oblate to strongly oblate AARM ellipsoids for the Vichenet Formation, and 'plane strain' to strongly oblate AARM ellipsoids for the Ronquieres Formation (Fig. 12). A comparison between TR and T shows that there is no relationship between them. This suggests that the ferromagnetic mineralogy exerts only a minor influence on the shape of the AMS-ellipsoid. However, from this graph, a division becomes apparent between the Cambrian samples (Ripain and Asquempont members) and the Silurian samples (Vichenet and Ronquieres formations). Whereas the former range from the

96

T. N. DEBACKER ET AL.

Fig. 11. Graphs of the degree of anisotropy of AMS (Pj) versus the degree of anisotropy of (p)AARM (PjR). The two graphs show the same data set, but with a different horizontal scale (PjR-scale). The stippled line in the lower graph represents the line on which Pj equals Pj R . Note that the majority of the samples has a PjR that is much higher than Pj.

diagonal line towards the prolate AARM oblate AMS field (above the diagonal line), the latter range from the diagonal line towards the oblate AARM - prolate AMS field (below the diagonal line). This division results partly from the difference in AMS ellipsoids, being more prolate in the Silurian for similar cleavage/bedding angles (cf. Fig. 6a), but also partly from the difference in AARM ellipsoids, apparently being more commonly slightly more prolate in the Cambrian samples. X-ray pole figure goniometry Judging from the overall mineralogy of the sampled lithologies (see above), the main

paramagnetic carriers present are white mica and chlorite (cf. Geerkens & Laduron 1996; Debacker et al 1999; Debacker 2001). Hence, phyllosilicate X-ray pole figure goniometry can be used to characterize the preferred orientation of these two paramagnetic carriers. In the Oisquercq Formation, two main types of pole figure patterns can be distinguished (Fig. 13). The first type, from samples of the Asquempont Member and the Ripain Member, is characterized by a moderate (samples TD279, TD256, TD197), occasionally weak (sample TD186) preferred orientation (Fig. 13), with the maxima of both white mica and chlorite coinciding with the cleavage pole. The pole figure patterns have an axially symmetrical (e.g. sample TD279) to slightly orthorhombic shape

COMPLEXITY OF AMS IN CLEAVED PELITES

97

Shape parameter TR (AARM) Fig. 12. Graph of the shape parameter of AMS (T) versus the shape parameter of (p)AARM (TR). Although, from this graph, there appears to be no relationship between T and T R , the Cambrian samples and Silurian samples occupy different fields. The stippled diagonal line represents the line of equal values.

(e.g. sample TD256). In the case of samples TD279 and TD179, the pole figure pattern reflects a flattening fabric. In contrast, the steeply plunging short axes of the slightly orthorhombic pole figure patterns of samples TD256 and TD186 may correspond to an intersection lineation between cleavage and bedding, which complies with the macroscopically observed steeply plunging cleavage/bedding intersection. The second and dominant type of pole figure, obtained from the Asquempont Member, both from zones of sub-horizontal and steeply plunging folds, has a clear girdle pole figure pattern, a relatively weak degree of preferred orientation, which is higher for chlorite than for white mica, and different pole figure maxima for chlorite and for mica (samples TD185, TD272, TD192, TD276, TD195, TD196 and TD248). The girdle pattern and the different pole figure maxima of chlorite and mica point to an intersection fabric. The mica pole figure maxima coincide with the cleavage pole, whereas the chlorite pole figure maxima approximate the bedding pole. This difference between mica and chlorite pole figure maxima, which, thus far, seems quite unique in the Brabant Massif (see Sintubin et al 1998; Debacker et al 1999; Belmans 2000; Piessens et al. 2000; Debacker 2001; Sintubin, unpublished data), is compatible with the macroscopically determined angle between bedding and cleavage (Fig. 13). Large differences between

chlorite and mica pole figure maxima are obtained from samples with a large angle between bedding and cleavage (e.g. samples TD192, TD195, TD196), whereas small angles between chlorite and mica pole figure maxima are obtained from samples with a small angle between cleavage and bedding (e.g. sample TD272). In all samples, the short axis of the girdle corresponds to the cleavage/bedding intersection lineation. The shape and the higher amount of preferred orientation of the first type of pole figure patterns as compared to the second type of pole figure patterns might be related to the rather small angle between bedding and cleavage. Indeed, there appears to be a relationship between the cleavage/bedding angle and the degree of phyllosilicate preferred orientation (Fig. 14). As for the magnetic analyses, the samples for X-ray pole figure goniometry in the Ronquieres section were taken from the e-interval of turbidite sequence 121 of Verniers et al. (1992; i.e. 30 to 65cm below the lower, quartzitic, marker horizon depicted in Fig. 2). The results of this analysis have already been discussed in Debacker et al. (1999). There is no significant variation in degree of preferred orientation nor in pole figure pattern across the antiform. Both for mica and for chlorite an intersection pole figure pattern is apparent, reflecting the superposition of a cleavage fabric on a pre-existing,

98

T. N. DEBACKER ET AL.

Fig. 13. Phyllosilicate X-ray pole figures of mica (dOOl; left) and chlorite (d002; right) from the Oisquercq Formation in the Northern Asquempont section (Cambrian) and the Ronquieres Formation in the Ronquieres section (Silurian; see Fig. 2 for sample location), together with the projected positions of the bedding and cleavage poles. Added are the orientation of the cleavage/bedding intersection, the angle between cleavage and bedding, and the maximum degree of preferred orientation, expressed in multiples of a random distribution (mrd). Note the occasional presence of axially symmetrical, flattening fabrics in the samples of the Oisquercq Formation and the overall predominance of orthorhombic to girdle patterns, reflecting intersection fabrics both in the Silurian and Cambrian samples. This intersection fabric is interpreted as resulting from a cleavage fabric (reflected by white mica) affecting a bedding-parallel compaction fabric (reflected by chlorite) (cf. Sintubin I994a; Debacker et al. 1999). An important difference between the Cambrian samples and Silurian samples is that, whereas in the Silurian samples both the mica and chlorite maxima approximately coincide with the cleavage pole, in the Cambrian samples the intersection pole figure maximum approximately coincides with the bedding pole in the case of chlorite and with the cleavage pole in the case of mica.

bedding-parallel, compaction fabric (Fig. 13). The degree of preferred orientation of chlorite is weak. The chlorite pole figure pattern shows a clear girdle, with the cleavage/bedding intersection as symmetry axis. In contrast, mica shows a stronger degree of preferred orientation,

with an orthorhombic pole figure pattern, centred around the cleavage pole. Also for mica, the short axis of the orientation distribution coincides with the cleavage/bedding intersection, implying that there are still remnants of a bedding-parallel compaction fabric, even

COMPLEXITY OF AMS IN CLEAVED PELITES

99

Fig. 14. Graph of the angle between cleavage and bedding versus degree of phyllosilicate preferred orientation, expressed in multiples of a random distribution (mrd) from samples of the Oisquercq Formation (outcrop Northern Asquempont section). Apparently, the degree of preferred orientation increases with decreasing cleavage/bedding angle. This happens both for mica and for chlorite.

though mica is preferentially oriented parallel to the cleavage.

Discussion Comparison of AMS, A ARM and X-ray pole figures: relative orientation and contribution of paramagnetic and ferromagnetic carriers to AMS The coincidence of Kl with the cleavage/bedding intersection suggests an intersection fabric (Housen et aL 1993). Dealing with two fabric elements, a bedding fabric and a cleavage fabric, the mismatches between K3 (AMS) and the pole to one of these two fabric elements can be attributed to the presence of two or more different orientation populations of magnetic (s.L) carriers (e.g. ferromagnetic carriers on the one hand and paramagnetic carriers on the other hand), of which some may be statistically oriented along the bedding, some along the cleavage, and possibly some oblique to both cleavage and bedding (e.g. due to an incomplete rotation away from the bedding plane towards the cleavage plane, or micro-kinking or bending of phyllosilicates). In the case of the Ripain Member, K3 is sometimes perpendicular to bedding, sometimes perpendicular to cleavage, and often takes up an intermediate position in between the bedding pole and the cleavage pole. The paramagnetic carriers, mica and chlorite, exhibiting a flattening fabric, appear to be situated statistically within the cleavage plane.

In contrast, as indicated by (p)AARM, the ferromagnetic carriers (probably coarse-grained hematite) are situated within the bedding plane. Hence, the samples of the Ripain Member suggest a competition between cleavage-parallel paramagnetic carriers on the one hand and bedding-parallel ferromagnetic carriers on the other hand in controlling the orientation of the minimum AMS axis (Fig. 15). Indeed, the high degree of anisotropy (Pj) and the relatively high bulk susceptibility of the samples of the Ripain Member support a relatively strong influence of the ferromagnetic carriers on the AMS signal (cf. Rochette 1987a; Hrouda & Jelinek 1990; Hrouda 2002; Martin-Hernandez & Hirt 2003). Possibly, the difference in relative position of K3 with respect to cleavage and bedding between the different samples results from small changes in relative concentration of ferromagnetic carriers (cf, Borradaile 1987). This idea is supported by the coercivity spectra, showing changes in coercivity of up to two orders of magnitude, both between different samples and between different cubes of the same samples (Fig. 9). However, in contrast to K3, Kl always coincides with the cleavage/bedding intersection. This probably results from the cleavage- and bedding-parallel nature of the paramagnetic, respectively ferromagnetic carriers. Because of the fabric-parallel alignment of both types of carriers, and the angle between them, they will both contribute to the cleavage/bedding intersection, without the relative concentration of ferromagnetic and paramagnetic carriers having a significant influence on the orientation of Kl (cf. Housen et aL 1993). However, although

100

T. N. DEBACKER ET AL.

Fig. 15. Schematic representation of the probable magnetic (s.1.) fabric orientation with respect to bedding and cleavage in the Ripain Member of the Oisquercq Formation, the Asquempont Member of the Oisquercq Formation, the Vichenet Formation and the Ronquieres Formation. The degree of alignment shown is based on X-ray pole figure goniometry data in the case of mica and chlorite and on the variation in orientation of R3 ((p)AARM) in the case of the ferromagnetic carriers. An important difference between the Cambrian samples (Ripain and Asquempont members of the Oisquercq Formation) and the Silurian samples (Vichenet and Ronquieres formations) is the predominance of ferromagnetic carriers along the bedding in the former samples and along the cleavage in the latter samples. In order to explain the common mismatch between K3 and the pole to cleavage in the Silurian samples, we invoke the presence of paramagnetic carriers along the bedding plane. Likely candidates are large Fe-rich chlorites. Note that these bedding-parallel carriers are likely to be present also in the Cambrian samples. However, there, because of the relative orientations of the other carriers, their presence does not appear to affect the orientation of K3, and hence they are not depicted.

situated in the bedding plane, locally a slight tectonic influence on the ferromagnetic carriers is apparent, as suggested by Rl of high-coercivity pAARM of one of the cubes (TDOOle). In this cube, the coincidence of Rl with the cleavage/ bedding intersection, in combination with the position of R3 in between the cleavage pole and the bedding pole, indicates that the ferromagnetic carriers, probably with a relatively small grain-size (cf. Jackson et al. 1988), reflect an intersection fabric. This cube possibly also contains some additional ferromagnetic carriers that grew along the lattice of phyllosilicates (cleavage-parallel), thus reflecting an intersection fabric. In the case of the Asquempont Member, K3 usually coincides with the bedding pole. Mica is

statistically oriented along the cleavage plane, whereas chlorite is statistically oriented along the bedding plane. This implies that, even without the presence of ferromagnetic carriers, AMS would likely reflect an intersection fabric, with Kl parallel to the cleavage/bedding intersection and K3 either coinciding with the bedding pole, the cleavage pole, or taking up an intermediate position, all depending on the relative concentration and degree of preferred orientation of chlorite and mica. The ferromagnetic carriers are statistically situated within the bedding plane, thus accentuating the beddingparallel AMS ellipsoid of chlorite. Hence, the position of K3 around the bedding pole may be attributed to the presence of both chlorite and ferromagnetic carriers along the bedding plane

COMPLEXITY OF AMS IN CLEAVED PELITES

on the one hand and the presence of mica along the cleavage plane on the other hand (Fig. 15). Small shifts in K3 may be attributed to small changes in concentration and degree of preferred orientation of ferromagnetic carriers and chlorite on the one hand and mica on the other hand. Because of the angle between cleavage (mica) and bedding (chlorite and magnetite), all magnetic minerals (s.L) contribute to the cleavage/ bedding intersection, without the relative concentration of the carriers having a significant influence on the orientation of Kl. The orientation of Rl (AARM) within the Oisquercq Formation, having a high, but variable, angle with respect to the cleavage/bedding intersection suggests a pre-cleavage origin of the orientation reflected by the ferromagnetic carriers. Because after unfolding Rl has the same orientation in all samples, even in the presence of variable cleavage/bedding relationships (steeply plunging versus gently plunging folds), and always results in an orientation subperpendicular to sedimentary ripples, with a slight northward tilt within the bedding plane, we tentatively attribute Rl of the Oisquercq Formation to a preferred alignment due to palaeocurrents. Judging from the slight northward dip of the ellipsoids of remanence (plunge of Rl) within the bedding plane, a southward palaeocurrent is inferred in the case of an imbrication of the ferromagnetic carriers. However, instead of showing an imbrication, the ferromagnetic minerals might also be oriented along low-angle foresets, in which case the inferred palaeocurrent would be towards the north. In the Vichenet Formation, the majority of the samples has a sub-horizontal K3, situated in between the cleavage pole and the bedding pole. Unfortunately, we have no phyllosilicate X-ray pole figures of the Vichenet section. However, data from turbidite e-intervals 700 m to the north of the Vichenet section (the turbiditic, Wenlock Vissoul Formation of the Chenemont section; Belmans 2000; Sintubin, unpub. data) show the same pattern as the Ronquieres Formation in the Ronquieres section: mica and chlorite being statistically aligned along the cleavage plane, both exhibiting orthorhombic to girdle symmetries, with the cleavage/bedding intersection as symmetry axis and with a stronger degree of preferred orientation for mica. Hence, in order to explain the mismatch between K3 and the cleavage pole, one has to invoke the presence of a magnetic (s.L) carrier sub-parallel to the bedding plane. However, in the three analysed samples, R3 usually coincides with the cleavage pole. This implies the presence of an

101

unknown magnetic (s.L) carrier parallel to bedding or with an orientation close to parallelism with bedding (Fig. 15), in order to explain the orientation of K3 in between the bedding pole and the cleavage pole. Judging from the relatively low bulk susceptibilities (< 350 x 10~6 Si) and the low degree of anisotropy, this is probably a paramagnetic carrier (Rochette 1987'a; Hrouda & Jelinek 1990; Hrouda 2002). However, X-ray pole figure data suggest that the two main paramagnetic carriers, mica and chlorite, are likely to be oriented statistically parallel to cleavage. Still, optical microscopic observations show the presence of a significant number of chlorite-mica stacks parallel to bedding. Therefore, we suggest that, although chlorite is oriented statistically parallel to cleavage as suggested by X-ray pole figure goniometry, two different populations of chlorite may be present: small matrix chlorites, of a Mg-/Al-rich type, oriented parallel to cleavage, and larger, Fe-rich, chlorites, oriented parallel to bedding (e.g. chlorite/mica stacks). Whereas X-ray pole figure goniometry will preferentially show the preferred orientation of small, cleavage-parallel, matrix chlorites, AMS will be influenced mainly by the orientation of the relatively few Fe-rich chlorites, oriented parallel to bedding. As pointed out by several authors (e.g. Rochette et al, 1992; Borradaile & Werner 1994), the susceptibility of phyllosilicates is strongly controlled by their Fe-content. Hence, in combination with cleavage-parallel mica and ferromagnetic carriers, the presence of Fe-rich chlorites along the bedding plane should lead to a K3 situated in between the cleavage pole and the bedding pole. The coincidence of Kl with the cleavage/bedding intersection can be interpreted as the combined effect of the cleavage-parallel orientation of mica and ferromagnetic carriers on the one hand and the bedding-parallel orientation of Fe-rich chlorite on the other hand. In most samples, the maximum axis (Rl) of (p)AARM, moderately to steeply plunging in the cleavage plane, at moderate to high angles to the cleavage/bedding intersection, points to a tectonic control on the ferromagnetic carriers. In sample TD1026, Rl reflects a predominantly bedding-parallel ferromagnetic carrier. This possibly represents a carrier that grew along the lattice of beddingparallel chlorites. In the Ronquieres Formation, although taken from a single bed, the samples exhibit a strong variation in K3 orientation with respect to cleavage and bedding. In some cases, K3 coincides with the cleavage pole, whereas in other cases it takes up a position in between the cleavage pole and the bedding pole. X-ray pole

102

T. N. DEBACKER ET AL.

figure goniometry indicates a preferred orientation of both mica and chlorite parallel to cleavage, with a higher degree of preferred orientation for mica, without showing a significant variation across the sampled antiform. Hence, considering the cleavage-parallel orientation of the phyllosilicates, the variation in K3 may be expected to be influenced by the variation in the orientation of the ellipsoid of remanence. Indeed, in some samples R3 is parallel to the pole to cleavage (e.g. TD4), whereas in other samples it occupies a position in between the cleavage pole and the bedding pole (e.g. TD3). In the cases where K3 parallels the pole to cleavage and R3 approximates the pole to cleavage, one may suggest a sub-parallelism of both the ferromagnetic and paramagnetic carriers within the cleavage plane. In contrast, however, considering that both mica and chlorite appear to be oriented along the cleavage plane, and that the ferromagnetic carriers do not appear to be situated within the bedding plane, it is difficult to explain the samples in which K3 takes up a position in between the pole to bedding and the pole to cleavage (e.g. TD1, TD3). In such cases, by analogy with the Vichenet Formation, we invoke the presence of a magnetic carrier along the bedding plane, of which the relative concentration may strongly affect the orientation of K3 (Fig. 15). We tentatively suggest Fe-rich bedding-parallel chlorite as a possible candidate. This Fe-rich chlorite may only have a minor influence on the X-ray pole figures (which predominantly reflect the finer-grained, omni-present matrix chlorites), but may dominate the AMS signal (cf. Rochette et al 1992; Borradaile & Werner 1994). In all samples, Rl approaches the cleavage/bedding intersection. Together with the common proximity between the cleavage pole and R3, this suggests a slight tectonic control on the ferromagnetic carriers. From the data and the above discussion, the complexity of the AMS fabric in the analysed samples becomes apparent, being influenced by both ferromagnetic and paramagnetic carriers. The isolation of the ferromagnetic signal does not always allow an explanation of all observations. In the case of the Silurian samples, the presence of a bedding-parallel, probably paramagnetic, carrier is proposed, in order to explain the observed AMS fabric. As a possible candidate we tentatively suggest large Fe-rich chlorites, which, because of their small numbers, are not detected by X-ray pole figure goniometry. Such a difference in preferred orientation of Ferich phyllosilicates and Fe-poor phyllosilicates was observed by Ho et al. (1995). However, in their example, the phyllosilicates oriented

parallel to cleavage have a high iron content and those oriented parallel to bedding have a low iron content, whereas we suggest the reverse. Influence of the cleavagefbedding angle By means of experiments and numerical models, Housen et al. (1993) pointed out the influence of both the angle between two magnetic fabrics and the relative concentration of magnetic minerals along these fabrics on the shape parameter and the degree of anisotropy. Obviously, if magnetic carriers occur along both bedding and cleavage, as in the present study, these can be considered as natural examples of two magnetic fabrics (cf. Housen et al. 1993). Our data confirm the results of Housen et al (1993) in the way that the shape parameter changes with the angle between cleavage and bedding, resulting in prolate ellipsoids in the case of high cleavage/bedding angles and oblate ellipsoids in the case of small cleavage/bedding angles. Seemingly in contrast with this, Pares & van der Pluijm (2003) deduce a direct relationship between the magnetic susceptibility shape parameter and tectonic shortening in rocks having a very weakly developed cleavage (pencil structure). However, by definition, this link seems questionable, considering that the 'tectonic shortening', based on the length-to-width ratios of pencil structures, is related to the incipient cleavage (at high angles to bedding), whereas all minimum susceptibility axes are perpendicular to bedding, thus reflecting a compaction strain rather than a tectonic shortening strain. The demonstrated influence of the cleavage/ bedding angle on the shape of the susceptibility ellipsoid seemingly resembles the influence of pre-deformation compaction strain (beddingparallel fabric) and position within a fold (large cleavage/bedding angle in hinge, small angles in limbs) on the shape and the orientation of the finite strain ellipsoid in structural geology. Finite strains are not only controlled by the incremental strain during deformation (related to cleavage development), but are also controlled by the pre-deformation compaction strains (Sanderson 1976; Maltman 1981; Ramsay & Huber 1983; Paterson et al. 1995). Depending on the folding mechanism and relative degree of pre-deformation compaction, this can lead to prolate finite strains in the fold hinge zones (large cleavage/bedding angles) and oblate finite strains in the fold limbs (small cleavage/bedding angles; cf. change in strain ellipsoid across a flexural fold). As we have observed in the Ronquieres section (Debacker 1996; Debacker

COMPLEXITY OF AMS IN CLEAVED PELITES

et al. 1999), the shape of calcitic nodules changes around the folds with changing cleavage/bedding angles, resulting in prolate nodules in the fold hinges (large cleavage/bedding angle) and 'plane strain' to oblate nodules in the fold limbs (small cleavage/bedding angles). However, a similar relationship between magnetic susceptibility shape parameter and cleavage/bedding angle is hardly ever documented in natural rocks (cf. Gil-Imaz et al. 2000). Liineburg et al. (1999) noticed a change in finite strain from oblate to plane strain, going from a fold limb to a fold hinge, but did not document a similar change in magnetic susceptibility shape parameter, nor in phyllosilicate anisotropy. This implies that there is more to controlling the shape of the magnetic susceptibility ellipsoid than the simple superposition of a tectonic strain on a pre-deformation compaction strain (cf. Housen et al. 1993; Gil-Imaz et al. 2000). Indeed, apart from the degree of preferred orientation of the magnetic carriers, which can be related to strain, the type and the relative concentration of the different magnetic carriers along the fabric elements also control the shape and orientation of the susceptibility ellipsoid (e.g. Borradaile 1987, 1988; Housen et al. 1993; cf. Pares & van der Pluijm 2003). Apart from an increase in shape parameter, Housen et al. (1993) also suggest an increase in degree of anisotropy with decreasing cleavage/ bedding angle. Although we do observe an increase in degree of anisotropy for the Cambrian samples (Asquempont Member; not enough variation in angle for Ripain Member) and the five samples of the Neoproterozoic turbidites of the Moesian platform, a decrease in degree of anisotropy with decreasing cleavage/ bedding angle becomes apparent in the Silurian samples (Vichenet Formation, not enough variation in angle for the Ronquieres Formation) (Fig. 6b). This difference in behaviour of the degree of anisotropy with changing cleavage/ bedding angle between the Silurian and the Cambrian may possibly be related to the slight difference in overall shape parameter (Fig. 6a). As we have shown, for similar cleavage/bedding angles, the Silurian samples are shifted towards the prolate field with respect to the Cambrian samples (and those of the Moesian platform). This may possibly also be related to the different fields occupied by the Cambrian samples and the Silurian samples on a T-TR graph (Fig. 12). Housen et al. (1993) modelled the effect of the variation in cleavage/bedding angle by means of two differently oriented magnetic fabrics of identical composition (magnetite, with variable concentrations). In real rocks, however, the two

103

differently oriented fabrics are not likely to have the same mineralogical composition. In our examples, the shape and degree of anisotropy of the AMS-ellipsoid of the Asquempont Member, which complies with the trend modelled by Housen et al. (1993), are controlled by mica oriented along the cleavage plane on the one hand, and chlorite and ferromagnetic carriers (magnetite) oriented along the bedding plane on the other hand (Fig. 15). In contrast, in the case of the Vichenet Formation, the shape and the degree of anisotropy of the AMS ellipsoid are controlled by mica, matrix-chlorite and ferromagnetic carriers (pyrrhotite) oriented along the cleavage plane on the one hand, and another carrier, probably paramagnetic and presumably large, Fe-rich chlorite, oriented along the bedding plane on the other hand (Fig. 15). This difference in relative orientation of magnetic (s.l.) mineralogy may well be the cause of the different behaviour of degree of anisotropy with respect to changes in cleavage/bedding angle (Fig. 6b), the tendency for the Silurian samples to show a slight shift towards the prolate field with respect to the Cambrian samples for similar cleavage/bedding angles (Fig. 6a) and the fact that the Cambrian samples and the Silurian samples occupy different fields on a T-TR graph (Fig. 12). If such is the case, then obviously AMS cannot be used as a strain gauge in deformed sedimentary rocks, characterized by two magnetic fabrics, the first one being a bedding-parallel compaction fabric and the second one being a cleavage-parallel tectonic fabric. Conclusion On the basis of experiments and numerical models, Housen et al. (1993) gave an outline of the characteristics of composite magnetic anisotropy fabrics, caused by the presence of two orientation populations of magnetic (s.l.) carriers. The present study, performed on lowgrade, weakly to moderately deformed, finegrained sedimentary rocks from the Brabant Massif, characterized by a bedding fabric and a moderately developed cleavage fabric (embryonic cleavage stage to cleavage stage of Ramsay & Huber 1983) complements the results of Housen et al. (1993). A strong relationship is observed between the angle between cleavage and bedding and the magnetic susceptibility shape parameter (T). High angles between cleavage and bedding (e.g. fold hinges) give rise to prolate susceptibility ellipsoids and small angles between cleavage and bedding (e.g. fold limbs) give rise to oblate

104

T. N. DEBACKER ET AL.

susceptibility ellipsoids. Importantly, a difference is observed between Cambrian samples and Silurian samples: for similar cleavage/bedding angles the former show a shift towards the oblate field with respect to the latter. Although Housen et al. (1993) conclude an increase in degree of anisotropy with decreasing cleavage/ bedding angle, we observe an increase with decreasing cleavage/bedding angle for the Cambrian samples, but a decrease with decreasing cleavage/bedding angle for the Silurian samples. Furthermore, a difference between the Cambrian and Silurian samples is also observed on a T-TR graph. The differences in shape parameter for similar cleavage/bedding angles, the different behaviour of the degree of anisotropy with decreasing cleavage/bedding angle and the different position on a T~-TR graph of the Cambrian samples with respect to the Silurian samples are all tentatively attributed to differences in magnetic mineralogy and relative orientation and concentration of the different magnetic (s.l.) carriers. In the light of the present observations, we suggest, like Housen et al. (1993), that AMS, although being a measure of petrofabric anisotropy, cannot be used as a strain gauge in rocks having composite magnetic anisotropy fabrics (cf. Borradaile 1988, 1991). Considering the mineralogical composition of low-grade, pelitic rocks and taking into account the influence of a large number of factors on the presence, concentration and relative orientation of paramagnetic (phyllosilicates) and ferromagnetic minerals (relative degree of tectonic shortening and compaction, degree of metamorphism, relative timing between metamorphism and deformation, sediment source/composition, fluid composition during and after diagenesis, metamorphism and deformation; cf. Rochette 1987&; Borradaile 1987; Robion et al. 1999), it becomes clear that in weakly to moderately deformed, cleaved, low-grade pelitic rocks, composite magnetic anisotropy fabrics are likely to occur. In such cases, AMS will generally not reflect finite strain, either qualitatively or quantitatively. Only in very specific cases, in which AMS is effectively controlled by one orientation population of magnetic (s.l.) carriers parallel to one of the fabric elements (cleavage or bedding), can one consider the possibility to link AMS to finite strain. Examples of specific cases are the two end members of deformed pelitic rocks, being undeformed rocks which only underwent compaction (shales), on the one hand, and intensely deformed rocks in which the initial bedding fabric is completely destroyed (slates, schists), on the other hand. It is suggested that when

applying AMS to deformed pelitic rocks, special attention should be paid to the relationship between the maximum (Kl) and minimum (K3) susceptibility axis on the one hand and the macroscopic fabric elements on the other hand (cf. Rochette et al. 1992). If Kl coincides with the cleavage/bedding intersection and a mismatch occurs between K3 and the pole to bedding and cleavage, then AMS is unlikely to record finite strain (cf. Housen et al. 1993). We are grateful to A. Seghedi for guiding us to outcrops of the Neoproterozoic turbidites of the Moesian platform (Dobrogea, Romania). We kindly acknowledge K. Ullemeyer, J. M. Pares and F. Martin-Hernandez for the very constructive remarks. T. Debacker is a Postdoctoral Fellow of the F.W.O.Vlaanderen. M. Sintubin is a Research Associate of the Onderzoeksfonds K.U.Leuven. This work forms part of research project G.0094.01 of the F.W.O.Vlaanderen. The research in Romania benefits from the International scientific and technological cooperation program from the Science, Innovation and Media Department of the Ministry of the Flemish Community (BIL01/34).

References AUBOURG, C, ROCHETTE, P. & VIALON, P. 1991. Subtle stretching lineation revealed by magnetic fabric of Callovian-Oxfordian black shales (French Alps). Tectonophysics, 185, 211-223. BELMANS, M. 2000. Structured geologic van het Siluur uit de Omeauvallei, Massief van Brabant. M.Sc. thesis, Laboratorium voor Paleontologie, Universiteit Gent. BORRADAILE, G. J. 1987. Anisotropy of magnetic susceptibility: rock composition versus strain. Tectonophysics, 138, 327-329. BORRADAILE, G. J. 1988. Magnetic susceptibility, petrofabrics and strain - a review. Tectonophysics, 156, 1-20. BORRADAILE, G. J. 1991. Correlation of strain with anisotropy of magnetic susceptibility (AMS). PAGED-PH, 135, 15-29. BORRADAILE, G. J. & HENRY, B. 1997. Tectonic applications of magnetic susceptibility and its anisotropy. Earth-Science Reviews 42, 49-93. BORRADAILE, G. J. & TARLING, D. H. 1981. The influence of deformation mechanisms on magnetic fabrics in weakly deformed rocks. Tectonophysics, 77, 151-168. BORRADAILE, G. J. & WERNER, T. 1994. Magnetic anisotropy of some phyllosilicates. Tectonophysics, 235, 223-248. BOUMA, A. H. 1962. Sedimentology of Some Flysch Deposits. Amsterdam: Elsevier. DEBACKER, T. N. 1996. Structureel-geologische studie van Siluur-afzettingen te Ronquieres, Massief van Brabant. M.Sc. thesis, Laboratorium voor paleontologie, Universiteit Gent.

COMPLEXITY OF AMS IN CLEAVED PELITES DEBACKER, T. N. 2001. Palaeozoic deformation of the Brabant Massif within eastern Avalonia: how, when and why? Ph.D. thesis, Laboratorium voor paleontologie, Universiteit Gent. DEBACKER, T. N., SINTUBIN, M. & VERNIERS, J. 1999. Cleavage/fold relationships in the Silurian metapelites, southeastern Anglo-Brabant fold belt (Ronquieres, Belgium). Geologic & Mijnbouw, 78, 47-56. DEBACKER, T. N., SINTUBIN, M. & VERNIERS, J. 2002. Timing and duration of the progressive deformation of the Brabant Massif, Belgium. Aardkundige Mededelingen, 12, 73-76. DEBACKER, T. N., SINTUBIN, M. & VERNIERS, J. 2004. Transitional geometries between gently plunging and steeply plunging folds - an example from the Lower Palaeozoic Brabant Massif, Anglo-Brabant fold belt, Belgium. Journal of the Geological Society, London, 161, 641-652. DE SCHEPPER, S. 2000. Kartering, lithostratigrafie en biostratigrafie met Chitinozoa van de Siluurtypesecties van de Orneauvallei (Massief van Brabant). M.Sc. thesis, Laboratorium voor Paleontologie, Universiteit Gent. DE Vos, W., VERNIERS, J., HERBOSCH, A. & VANGUESTAINE, M. 1993. A new geological map of the Brabant Massif, Belgium. Geological Magazine, 130,605-611. FRIZON DE LAMOTTE, D., SOUQUE, C, GRELAUD, S. & ROBION, P. 2002. Early record of tectonic magnetic fabric during inversion of a sedimentary basin: examples from the Corbieres transfer zone. Bulletin de la Societe Geologique de France, 173, 461-469. FULLER, M. D. 1964. On the magnetic fabric of certain rocks. Journal of Geology, 12, 368-376. GIL-IMAZ, A., Pocovi, A., LAGO, M. & PARES, J. M. 2000. Effect of lithostatic pressure and tectonic deformation on the magnetic fabric (anisotropy of magnetic susceptibility) in low-grade metamorphic rocks. Journal of Geophysical Research, 105,21305-21317. GEERKENS, B. & LADURON, D. 1996. Etude du metamorphisme du Massif du Brabant. Unpublished BNRE-report. GRAHAM, J. W. 1954. Magnetic susceptibility anisotropy, an unexploited petrofabric element. Bulletin of the Geological Society of America, 65, 12571258. HERBOSCH, A., VERNIERS, J., DEBACKER, T. N., BILLIAERT, B., DE SCHEPPER, S. & BELMANS, M. 2002. The Lower Palaeozoic stratigraphy and sedimentology of the Brabant Massif in the Dyle and Orneau valleys and of the Condroz Inlier at Fosses: an excursion guidebook. Geologica Belgica, 5, 71-142. HIRT, A. M., EVANS, K. F. & ENGELDER, T. 1995. Correlation between magnetic anisotropy and fabric for the Devonian shales on the Appalachian Plateau. Tectonophysics, 247, 121-132. HIRT, A. M., JULIVERT, M. & SOLDEVILA, J. 2000. Magnetic fabric and deformation in the Navia-Alto Sil slate belt, northwestern Spain. Tectonophysics, 320, 1-16.

105

Ho, N.-C, PEACOR, R. & VAN DER PLUIJM, B. A. 1995. Reorientation mechanisms of phyllosilicates in the mudstone-to-slate transition at Lehigh Gap, Pennsylvania. Journal of Structural Geology, 17, 345-356. HOUSEN, B. A., RlCHTER, C. & VAN DER PLUIJM, B. A.

1993. Composite magnetic anisotropy fabrics: experiments, numerical models, and implications for the quantification of rock fabrics. Tectonophysics, 220, 1-12. HROUDA, F. 1982. Magnetic anisotropy of rocks and its application in geology and geophysics. Geophysical Surveys, 5, 37-82. HROUDA, F. 2002. The use of anisotropy of magnetic remanence in the resolution of the anisotropy of magnetic susceptibility into its ferromagnetic and paramagnetic components. Tectonophysics, 347, 269-281. HROUDA, F. & JELINEK, V. 1990. Resolution of ferromagnetic and paramagnetic anisotropies, using combined low-field and high-field measurements. Geophysical Journal, 103, 75-84. JACKSON, M. 1991. Anisotropy of magnetic remanence: a brief review of mineralogical sources, physical origins, and geological applications, and comparison with susceptibility anisotropy. PAGEOPH, 136, 1-28. JACKSON, M., GRUBER, W., MARVIN, J. & BANERJEE, S. K. 1988. Partial anhysteretic remanence and its anisotropy: applications and grainsizedependence. Geophysical Research Letters, 15, 440^443. JELINEK, V. 1981. Characterisation of the magnetic fabric of rocks. Tectonophysics, 79, 63-67. JELINEK, V. 1993. Theory and measurement of the anisotropy of isothermal remanent magnetisation of rocks. Travaux Geophysiques, 37, 124—134. JELINEK, V. & POKORNY, J. 1997. Some new concepts in the technology of transformer bridges for measuring susceptibility anisotropy of rocks. Physics and Chemistry of the Earth, 22, 179-181. LEGRAND, R. 1967. Ronquieres, documents geologiques. Memoire pour servir a I'Explication des Cartes geologiques et minieres de la Belgique, 6, 1-60. Li, Z. X. & POWELL, C. McA. 1993. Magnetic fabric in mid-Cambrian rocks of the Central Flinders Zone and implications for the regional tectonic history. Tectonophysics, 223, 165-176. LOUWYE, S., VAN GROOTEL, G. & VERNIERS, J. 1992. The stratigraphy of the type locality of the ?late Wenlock/early Ludlow Mont Godart Formation and the early Ludlow Ronquieres Formation, Brabant Massif, Belgium. Annales de la Societe Geologique de Belgique, 115, 307-331. LOWRIE, W. 1990. Identification of ferromagnetic minerals in a rock by coercivity and unblocking temperature properties. Geophysical Research Letters 17, 159-162. LUNEBURG, C. M., LAMPERT, S. A., LEBIT, H. D., HIRT, A. H., CASEY, M. & LOWRIE, W. 1999. Magnetic anisotropy, rock fabrics and finite strain in deformed sediments of SW Sardinia (Italy). Tectonophysics 307, 51-74.

106

T. N. DEBACKER ET AL.

MALTMAN, A. J. 1981. Primary bedding-parallel fabrics in structural geology. Journal of the Geological Society, London, 138, 475-483. MARTIN-HERNANDEZ, F. & HIRT, A. M. 2003. The anisotropy of magnetic susceptibility in biotite, muscovite and chlorite single-crystals. Tectonophysics, 361, 13-28. MCCABE, C, JACKSON, M. & ELLWOOD, B. B. 1985. Magnetic anisotropy in the Trenton Limestone: results of a new technique, anisotropy of anhysteretic susceptibility. Geophysical Research Letters, 12, 333-336. NAKAMURA, N. & BORRADAILE, G. 2001. Do reduction spots predate finite strain? A magnetic diagnosis of Cambrian slates in North Wales. Tectonophysics, 340, 133-139. OERTEL, G. 1983. The relationship of strain and preferred orientation of phyllosilicate grains in rocks-a review. Tectonophysics, 100, 413-447. PARES, J. M. & VAN DER PLUIJM, B. A. 2002. Evaluating magnetic lineations (AMS) in deformed rocks. Tectonophysics, 350, 283-298. PARES, J. M. & VAN DER PLUIJM, B. A. 2003. Magnetic fabrics and strain in pencil structures of the Knobs Formation, Valley and Ridge Province, US Appalachians. Journal of Structural Geology, 25, 1349-1358. PATERSON, S. R., Yu, H. & OERTEL, G. 1995. Primary and tectonic fabric intensities in mudrocks. Tectonophysics, 247, 105-119. PIESSENS, K., VIAENE, W. & MucHEZ, PH. 2000. Laboratoriumstudie van Boorkernen in het Massief van Brabant. Unpublished ANRE-report (report VLA98-3-6), Afdeling Fysico-Chemische Geologic, Katholieke Universiteit Leuven. RAMSAY, J. G. & HUBER, M. I. 1983. The techniques of modern structural geology: Volume 1: Strain analysis. Academic Press, London. RATHORE, J. S. 1979. Magnetic susceptibility anisotropy in the Cambrian slate belt of North Wales and correlation with strain. Tectonophysics, 53, 83-97. ROBION, P., FRIZON DE LAMOTTE, D., KISSEL, C. & AUBOURG, C. 1995. Tectonic versus mineralogical contribution to the magnetic fabrics of epimetamorphic slaty rocks: an example from the Ardennes massif (France-Belgium). Journal of Structural Geology, 17, 1111-1124. ROBION, P., KISSEL, C., FRIZON DE LAMOTTE, D., LORAND, J.-P. & GUEZOU, J.-C. 1997. Magnetic mineralogy and metamorphic zonation in the Ardennes Massif (France-Belgium). Tectonophysics, III,231-248. ROBION, P., AVERBUCH, O. & SINTUBIN, M. 1999. Fabric development and metamorphic evolution of Lower Palaeozoic slaty rocks from the Rocroi massif (French-Belgian Ardennes): new constraints from magnetic fabrics, phyllosilicate preferred orientation and illite crystallinity data. Tectonophysics, 309, 257-273. ROCHETTE, P. 1987#. Magnetic susceptibility of the rock matrix related to magnetic fabric studies. Journal of Structural Geology, 9, 1015-1020. ROCHETTE, P. 19876. Metamorphic control of the mineralogy of black shales in the Swiss Alps:

towards the use of 'magnetic isograds'. Earth and Planetary Science Letters, 84, 446-456. ROCHETTE, P., JACKSON, M. & AUBOURG, C. 1992. Rock magnetism and the interpretation of anisotropy of magnetic susceptibility. Reviews of Geophysics, 30, 209-226. SAINT BEZAR, B., HEBERT, R., AUBOURG, C., ROBION, P., SWENNEN, R. & FRIZON DE LAMOTTE, D. 2002. Magnetic fabric and petrographic investigation of hematite-bearing sandstones through ramp-related folds : examples from the South High Atlas front (Morocco). Journal of Structural Geology, 24, 1507-1520. SANDERSON, D. J. 1976. The superposition of compaction and plane strain. Tectonophysics, 30, 35-54. SINTUBIN, M. 19940. Phyllosilicate preferred orientation in relation to strain path determination in the Lower Palaeozoic Stavelot-Venn Massif (Ardennes, Belgium). Tectonophysics, 237, 215231. SINTUBIN, M. 19946. Textures in shales and slates. In: BUNGE, H. J., SlEGENUND, S., SKROTSKI, W. &

WEBER, K. (eds) Textures of Geological Materials. DGM Informationsgesellschaft Verlag, 221-229. SINTUBIN, M. 1997. Cleavage-fold relationships in the Lower Paleozoic Brabant Massif (Belgium). Aardkundige Mededelingen, 8, 161-164. SINTUBIN, M. 1998. Mica (110) pole figures in support of Marchian behaviour of phyllosilicates during the development of phyllosilicate preferred orientation. Materials Science Forum, 273-275, 601606. SINTUBIN, M. 1999. Arcuate fold and cleavage patterns in the southeastern part of the Anglo-Brabant Fold Belt (Belgium): tectonic implications. Tectonophysics, 309, 81-97. SINTUBIN, M., WENK, H.-R. & PHILLIPS, D. S. 1995. Texture development in materials with a platy grain morphology. A comparison between cold pressed Bi2223-powders and experimental and natural phyllosilicate fabrics. Materials Science and Engineering, A202, 157-171. SINTUBIN, M., BRODKOM, F. & LADURON, D. 1998. Cleavage-fold relationships in the Lower Cambrian Tubize Group, southeast Anglo-Brabant Fold Belt (Lembeek, Belgium). Geological Magazine, 135,217-226. STEPHENSON, A., SADIKUN, S. & POTTER, D. K. 1986. A theoretical and experimental comparison of the anisotropies of magnetic susceptibility and remanence in rocks and minerals. Geophysical Journal of the Royal Astronomical Society, 84, 185-200. TARLING, D. H. & HROUDA, F. 1993. The Magnetic Anisotropy of Rocks. Chapman & Hall, London. VAN DER PLUIJM, B. A., Ho, N.-C. & PEACOR, D. R. 1994. High-resolution X-ray texture goniometry. Journal of Structural Geology, 16, 1029-1032. VAN GROOTEL, G., VERNIERS, J., GEERKENS, B., LADURON, D., VERHAEREN, M., HERTOGEN, J. & DE Vos, W. 1997. Timing of subsidence-related magmatism, foreland basin development, metamorphism and inversion in the Anglo-Brabant fold belt. Geological Magazine, 134, 607-616.

COMPLEXITY OF AMS IN CLEAVED PELITES VERNIERS, J., LOUWYE, S. & VAN GROOTEL G. 1992. Lithostratigraphical descriptions of the Mont Godart and the Ronquieres Formations in their type localities and evaluation of the previous descriptions. Professional Paper of the Geological Survey of Belgium, 252, 1-67. VERNIERS, J., HERBOSCH, A. ET AL., 2001 CambrianOrdovician-Silurian lithostratigraphical units

107

(Belgium). Geologica Belgica, 4, 5-38. VERNIERS, J., PHARAOH, T. ET AL., 2002. The Cambrian to mid Devonian basin development and deformation history of Eastern Avalonia, east of the Midlands Microcraton: new data and a review. In: Palaeozoic Amalgamation of Central Europe. Geological Society, London, Special Publications, 201, 47-93.

This page intentionally left blank

Magnetic and mineral fabric development in the Ordovician Martinsburg Formation in the Central Appalachian Fold and Thrust Belt, Pennsylvania A. M. HIRT1, W. LOWRIE2, C. LUNEBURG3, H. LEBIT3 & T. ENGELDER4 1

Institute of Geophysics, ETH-Honggerberg, CH-8093 Zurich, Switzerland (e-mail: [email protected]) 2 Institute of Geophysics, ETH-Honggerberg, CH-8093 Zurich, Switzerland 3 Department, of Geology and Geophysics, University of New Orleans, New Orleans, LA 70148, USA ^Department ofGeosciences, Pennsylvania State University, State College, PA 16802, USA Abstract: The Martinsburg Formation at Lehigh Gap, Pennsylvania, undergoes a transition from shales to slates, reflecting local progressive deformation on an outcrop scale. The anisotropy of magnetic susceptibility (AMS) was measured in low and high fields. The high-field measurements show that the magnetic susceptibility is controlled by the paramagnetic minerals. X-ray goniometry was used to define the mineral fabrics of chlorite and mica. The phyllosilicates are initially oriented preferentially in the bedding plane and are gradually reoriented into the cleavage plane through rotation, microfolding and recrystallization. The AMS fabric mirrors this change in mineral fabric. The magnetic fabric is originally oblate in the least deformed site, with the plane of flattening parallel to bedding, and becomes prolate with increasing deformation, reflecting the development of pencil structure in the shales. In the most deformed site, shortening results in a tectonic cleavage fabric, which controls the magnetic fabric. A similar pattern of fabric development can be observed on a regional scale at other sites across the central Appalachian fold and thrust belt. The AMS and mineral fabric from the Martinsburg Formation has undergone bedding compaction in the foreland near the Allegheny Front. The AMS and textural analysis both show that, as the deformation increases towards the hinterland, prolate fabrics develop and in the most deformed sites slaty cleavage controls both the mineral and magnetic fabrics.

The anisotropy of magnetic susceptibility (AMS) has been established as a qualitative proxy of petrofabrics since the mid-1950s. Graham (1954) demonstrated that the magnetic fabric reflects petrofabric, and Balsley & Buddington (1960) showed that magnetic methods were more sensitive in detecting fabrics than traditional microscopic methods. Several studies in the early 1960s further used magnetic anisotropy to determine petrofabrics and postulated on the origin of the magnetic fabric (e.g. Stacey et al. 1960; Girdler 1961; Fuller 1963). The authors were primarily interested in investigating whether a petrofabric could be responsible for a deviation of a rock's remanent magnetization, Graham (1966) pointed out that the magnetic susceptibility of a rock is composed of contributions from all constituent minerals in the rock, and that the magnetic anisotropy will depend on the orientations of crystal lattices or magnetite grain shapes. He examined the anisotropy of magnetic susceptibility (AMS) in flat-lying and folded Palaeozoic sediments from the Valley and Ridge Province of the Appalachian

fold and thrust belt. The results from these studies led Graham (1966) to speculate on the development of a magnetic fabric during folding, where he considered the AMS to reflect the textural reorientation and grain rotation that occurs in the rock matrix during progressive deformation. The evolution of a magnetic fabric that undergoes horizontal compaction is continuous: it starts with an oblate ellipsoid that expresses flattening by the bedding compaction, and proceeds to a prolate ellipsoid with long axis parallel to the trend of the fold axis. Continued shortening leads to an oblate ellipsoid that is flattened in the cleavage plane and which can show extension down-dip in the cleavage plane at the highest stage of deformation, X-ray texture goniometry provides information about the orientation of individual mineral phases in a rock. Ihmle et al. (1989) were able to show in deformed limestones from the Morcles nappe in Switzerland that the principal axes of the AMS ellipsoid are related to the orientation of the crystallographic axes of calcite crystals. Hirt et al. (1995) demonstrated that a magnetic

From: MARTIN-HERNANDEZ, F., LUNEBURG, C. M., AUBOURG, C. & JACKSON, M. (eds) 2004. Magnetic Fabric: Methods and Applications. Geological Society, London, Special Publications, 238, 109-126. 0305-8719/04/ $15.00 © The Geological Society of London 2004.

110

A. M. HIRT ET AL.

lineation in Devonian shales from the Allegheny Plateau was sub-parallel to the orientation of the long axes of chlorite crystals. Siegesmund et al. (1995) modelled the AMS of gneissic rocks from the mica texture. They found a good agreement in the orientation of the modelled and measured magnetic fabric, but a poor agreement of shape and degree of anisotropy. Since these first studies, there has been increasing interest in how magnetic fabrics reflect mineral fabrics (e.g. Limeburg et al. 1999; Becker et al 2000; de Wall et al. 2000). This study returns to the area covered in Graham's (1966) seminal study in the Valley and Ridge Province of the Central Appalachian fold-thrust belt, to examine the development of mineral and magnetic fabrics in shales on two different spatial scales. Special attention was given to determining the minerals responsible for the AMS so that it could be compared with representative mineral fabrics. A detailed investigation of fabric development was made at Lehigh Gap on the hinterland edge of the Valley and Ridge in eastern Pennsylvania where, over the space of an outcrop, the Martinsburg Formation (locally called the Reedsville Formation) is progressively deformed from shale to slate. To establish whether the results from Lehigh Gap can be extended on a larger scale, the development of mineral and magnetic fabrics in the Martinsburg Formation was studied at four locations distributed across the Valley and Ridge of Pennsylvania. Magnetic anisotropy results were obtained from 33 additional sites in the Valley and Ridge. Because of the time-intensive nature of acquiring texture goniometry data, we selected four locations exhibiting different degrees of progressive deformation for detailed mineral fabric analysis. Geological setting The Central Appalachian Valley and Ridge Province has been extensively studied since H. D. & W. B. Rogers made the first structural geology maps in 1838-1846. It was an important area in developing the idea of thin-skinned tectonics (Rodgers 1963; Gwinn 1964). As recognized by Graham (1966), deformation in the Valley and Ridge follows two fundamental steps. The first step is a layer-parallel shortening (LPS), which is accommodated by pressure solution and other low-temperature creep mechanisms (e.g. Nickelsen 1966, 1979; Engelder & Geiser 1979; Geiser 1989; Geiser & Engelder 1983). On the Appalachian Plateau where LPS precedes major fold growth, stretch (53) by LPS is about 0.85

(Engelder & Engelder 1977). The second step is growth of first-order folds that 'behaved as a generally coherent and passive plate' (Graham 1949). The outer folds of the Valley and Ridge exhibit an LPS-related S3 of about 0.85 (Faill 1977). Only in the anthracite district of the eastern Valley and Ridge of Pennsylvania is the LPS more intense and even there it predates the growth of major anticlines (Nickelsen 1979). The Valley and Ridge behaves as two distinct mechanical units (Hatcher et al. 1989). A Cambrian-Ordovician carbonate sheet stacks as a duplex under a roof thrust in the Ordovician Martinsburg Formation (Geiser 1988). Interestingly, the carbonate thrust sheets show very little evidence of LPS where the roof exhibits well developed tectonic stretch, S3 (e.g., Ferrill & Dunne 1989). The Ordovician Martinsburg Formation was deposited in a deep foreland basin during the Taconic orogeny (Faill 1997). It consists of fine-grained dark grey shales with interlayered siltstones and sandstones. Depending on where the Martinsburg sat relative to the roof thrust, it may exhibit a well-developed LPS fabric. If the Martinsburg sat below the roof-thrust detachment it may have ridden passively on carbonate horses below and thus exhibit virtually no LPS fabric. Conodant colour alteration indicates that most Ordovician rocks in Pennsylvania were never heated above 250 °C, except in the hinterland where slaty cleavage is developed (Epstein et al. 1977). Here temperatures may have exceeded 300 °C. The geological setting of Lehigh Gap has been discussed extensively in other studies (e.g. Epstein & Epstein 1969; Holey well & Tullis 1975; Wright et al. 1979; Wright & Platt 1982; Wintsch et al. 1991; Ho et al. 1995). The outcrop consists of the upper member of the Martinsburg Formation, which is conformably overlain by the Silurian Shawangunk Formation, a massive sandstone unit (Fig. 1). The Martinsburg Formation has a penetrative cleavage in this area of the Appalachian Fold-Thrust Belt. However, at the Lehigh Gap outcrop, the slaty cleavage in the Martinsburg Formation dies out within 3060 m of the contact with the Shawangunk sandstone. Epstein & Epstein (1969) suggested that the quartzitic sandstone acts as a pressure shadow against cleavage formation in the shales. The bedding orientation remains relatively constant with dip direction of 327-336° and dip of 41-43° along the 120m of outcrop that was sampled (Fig. 1). The Martinsburg shales show only bedding structure within the first metre from the contact with the Shawangunk Formation. The first pencil structures are identifiable about 2m from the contact and are

MAGNETIC AND MINERAL FABRIC DEVELOPMENT

111

Fig. 1. (a) Map showing the sampling localities in the Appalachian fold and thrust belt and (b) schematic diagram of the Lehigh Gap outcrop in eastern Pennsylvania showing the stages of cleavage development and locations of the sampling sites relative to the contact with the Shawangunk sandstone (modified from Housen et al. 1995).

the dominant structure until the gradual appearance of cleavage between 60 and 80m. The cleavage becomes stronger further along the outcrop where a penetrative slaty cleavage dominates. The cleavage plane has a dip direction of 171-186° and a dip of 56-59°. We sampled four sites along the 120 m outcrop at different distances from the contact: PALI within the first 3m, PAL2 between 35m and 40m, PAL3 around approximately 90m, and PAL 4 at around 120m. Four additional localities were selected across the Central Appalachian fold and thrust belt that represent progressive deformation from the foreland to the hinterland (Fig. 1). The sampling locations are named after the corresponding quadrangle of the geological map of Pennsylvania and are shown with related geological information in Table 1. MLM represents a region near the Allegheny Front in the foreland where only bedding compaction is observed. Samples were drilled in fine-grained siltstones from the SSW limb of the Nittany anticlinorium. Although pencil structures are found in fine-

grained lithologies east of the Allegheny front, pervasive pencil structure occurs only in the area of FAN. Most samples at this locality came from siltier layers, since shaley units were too fractured to sample. The PL A outcrop covers a small asymmetric fold in which an incipient cleavage is visible in the more shaley units, and most samples were taken in shale and siltstone beds. A well-defined slaty cleavage is found in the hinterland at POR. The black slates show a strong lineation sub-parallel to the fold axis in the area. Sampling and analytical techniques

Optical and electron microscopy methods Microstructures of the shales and slates were studied by optical and electron microscopy in order to evaluate deformation mechanisms on the microscopic scale. Most studies were performed on a JEOL JSM 840 scanning electron microscope (SEM), which allows investigation of

A. M. HIRT ET AL.

112

Table 1. Site locations and geological information Site

Location Quadrangle Lat/Long

LehighGap

Palmerton

PALI PAL2 PALS PAL4 MLM

FAN PLA POR

Bedding Orientation Dip Direction/Dip

Cleavage Orientation Dip Direction/Dip

Lineation Dip/Dip Direction

328-335/42-43 336-338/41-45 325-333/36-42 327/43 167/42

* * 170-173/48-54 174-182/50-56 *

07/252 06/255 11/256 * *

149/33

*

02/58

150/70

152/70

*

345/80

160/54

05/075

40° 41' N/77° 36' W

Millheim 40° 49' N/77° 25.5' W Fannetsburg 40° 04.5' N/77° 48.5' W Plainfield 40° 13'N/77° 18; W Portland 40° 54' N/75° 04' W

* Indicates none present

microstructures of very small grains, such as phyllosilicates, as well as their chemical compositions. Two types of images were created: backscatter and secondary electron images. Backscatter images are performed on highly polished surfaces and reflect compositional differences since they are sensitive to the atomic number contrast. Secondary electron images are made from rough surfaces, treated with 48% hydrofluoric acid, and reflect the topographic relief of the grains. In addition Energy Dispersive X-ray Spectrometry (EDS) analysis allowed us to determine element spectra and quantitative analysis of sample areas or individual grains.

X-ray texture goniometry methods For the X-ray texture goniometry measurements oriented block samples of 12 x 10 x 24mm size were cut from adjacent parts of remaining cores from the magnetic fabric study. The measured surface was chosen as close as possible to the cleavage plane in order to obtain centred pole figures. The samples were analyzed with a Scintag-USA/DMS 2000 with a Cu X-ray source and a germanium solid-state detector, which measures the intensity of the Ka radiation directly. The mineral fabrics of mica and chlorite were determined by measuring the diffraction intensity of the dominant basal planes, (001) and (002) respectively, which determine the lattice as well as grain-shape preferred orientation (Oertel 1983; Siddans 1976). The data are illustrated in contoured pole figures in upper hemisphere projections, which represent scans

of diffraction intensity in intervals of 5° (—1387 positions) (Casey 1981). In order to cover the full range of spherical orientations the samples are measured in reflection mode (as a block) and in transmission mode (as a thin section). The eigenvalues and eigenvectors of the intensity distributions determine shapes and principal axial orientations of fabric ellipsoids (e.g. Cheeney 1983; Davis 1986). The principal axes of the mineral fabric ellipsoids are described as t\ > h > *3- The t\ axis represents the largest eigenvector, which reflects the lowest intensity defined by the smallest number of poles to basal planes (c axes). The t$ axis represents the smallest eigenvector, which reflects the highest intensity defined by the largest number of poles to basal planes (x axes). In the remainder of this text, t\ is referred to as the maximum axis of the mineral fabric ellipsoid, t2 as the intermediate axis and t3 as the minimum axis. This convention is the same as used for describing the magnetic fabric axes. For comparison with magnetic directions the principal axes of the mineral fabrics were inverted and plotted on the lower hemisphere.

Magnetic methods Eight cores, yielding nine to thirteen samples, were drilled with a portable gasoline drill at each locality at Lehigh Gap for the study of the magnetic fabric. Specimens had 2.54 cm diameter and 2.30cm length. From 9 to 31 cores, yielding 18 to 54 specimens, were taken at each of the localities across the fold belt and prepared as above. The low-field anisotropy of magnetic

MAGNETIC AND MINERAL FABRIC DEVELOPMENT

susceptibility (AMS) of all samples was measured with an AGICO KLY-2 susceptibility bridge (applied field: 0.4 mT, frequency: 920 Hz) at room temperature. Each sample was measured in 15 positions to obtain the susceptibility magnitude ellipsoid whose principal axes are k\ > k2 > £3. Cooling the samples enhances the paramagnetic component of the susceptibility. Therefore, selected samples were measured at liquid nitrogen temperature (77 K), using the method outlined in Liineburg et al. (1999). In order to separate the paramagnetic or antiferrimagnetic component of the susceptibility anisotropy from the ferromagnetic component, the high-field anisotropy (HFA) of selected samples was measured on a torsion magnetometer, using the procedure described by MartinHernandez & Hirt (2001). Samples were measured in four fields: 1200, 1400, 1600 and ISOOmT, all of which are above the saturation magnetization of the ferrimagnetic phases in the shales. The anisotropy of a partial anhysteretic remanent magnetization (ARM) was measured in two coercivity windows: 0-30 mT and 60-100 mT, in each case with a 0.1 mT d.c. bias field. We chose the low coercivity window so that we could compare our results with an earlier study of magnetic fabric at Lehigh Gap by Housen & van der Pluijm (1991). The ARM was imparted in 9 directions to over-define the anisotropy tensor and obtain an error estimate on the fit (McCabe et al. 1985), which was in the range 1-5%. Magnetic mineralogy The ferrimagnetic mineralogy was defined from (a) acquisition of anhysteretic remanent magnetization (ARM), (b) the acquisition of isothermal remanent magnetization (IRM) and (c) subsequent thermal demagnetization of a crosscomponent IRM. The ARM was acquired in a large coil in peak alternating fields up to 200 mT (Fig. 2a). The IRM was imparted with an electromagnet with a maximum field of LOT (Fig. 2b). After magnetization in the LOT field along the sample Z axis, a 0.5 T field was applied along the sample Y axis and a 0.1 T field along the sample X axis (Lowrie 1990). Samples were thermally demagnetized in a Schonstedt TSD-1 oven (Fig. 2c). Remanence was measured with a 2G 3-axis cryogenic magnetometer with RF SQUIDs. The ARM acquisition curves are still rising steeply at 30mT, but have reached most of their maximum intensity by lOOmT (Fig. 2a). The IRM acquisition curves of the shales and

113

slates show a rapid increase in magnetization below 0.15T, where they are close to saturation; the magnetization is saturated by 0.5T (Fig. 2b). Thermal demagnetization of the crosscomponent IRM is similar in most samples (Fig. 2c). The magnetization is dominated by the low coercivity component, which is fully demagnetized by 600 °C. A more rapid loss in magnetization occurs between 500 and 600 °C and a less dramatic loss between 300 and 350 °C. The medium and high coercivity components undergo little unblocking until 300 °C where the intensity drops more rapidly. The remaining magnetization is generally unblocked by 600 °C. These results suggest that magnetite is the dominant ferrimagnetic phase in the Martinsburg Formation with a smaller contribution from pyrrhotite. Lehigh Gap, Pennsylvania

Electron microscopy and texture goniometry Electron microscopy at PALI, the least deformed site, shows that the phyllosilicates, particularly the mica, are mainly oriented in the bedding plane, although some kinking can be observed (Fig. 3). Chlorite grains have been identified lying at an angle to the main bedding orientation, as also observed by Holeywell & Tullis (1975) and Ho et al (1995). The mica and chlorite mineral fabrics show a distribution of the c axes in a weak girdle structure, whose pole ^i is sub-parallel to the pencil structure lineation (Fig. 4). The average of the chlorite minimum axes t^ is at an angle of about 30° to the bedding plane, whereas the mica average is close to the bedding pole. Both mineral fabrics reflect the microscopic observations. Pencil structures are well developed at PAL2. The phyllosilicate minerals tend to lie flattened in the bedding or kinked at an angle to the bedding plane (Fig. 3). The mineral fabric is similar to PALI; the main difference is that the average of the chlorite minimum axes now lies closer to the bedding plane pole (Fig. 4). The minimum axes of the mica and chlorite are well grouped around the pole to bedding and the maximum axes are still grouped about the visible intersection lineation. At 90m from the contact at PAL3 a slaty cleavage has developed and the intersection lineation is evident from the pencil structures. SEM studies show a stronger preferred orientation of the phyllosilicate grains and increased kinking (Fig. 3). The kinking is responsible for the more pronounced girdle structure of the

114

A. M. HIRT ET AL.

Fig. 2. (a) ARM acquisition and (b) IRM acquisition in representative samples from several Appalachian sites, (c) Thermal demagnetization of a cross-component IRM for two representative samples.

MAGNETIC AND MINERAL FABRIC DEVELOPMENT

115

Fig. 3. SEM secondary electron images of the Lehigh Gap samples showing phyllosilicates parallel to bedding at PALI and increasing grain rotation, kinking and micro folding from PAL2 to PAL3 and PAL 4. Cleavage is defined by phyllosilicates parallel to the cleavage plane and the kink axial planes. Scale bar is lOjim in each image. S0 and Si are the traces of the bedding and cleavage planes, respectively.

mica fabric (Fig. 4). The chlorite fabric is very weak, and may be controlled by a few large crystals in the section. However, the minimum axes are close to the bedding pole. In contrast, the mica minimum axis lies between the bedding and the cleavage poles. The intersection lineation still controls the mineral fabrics. A pervasive slaty cleavage is developed at PAL4. Grains still show kinking, but new growth of grains and grain rotation within the cleavage plane are also visible (Fig. 3), leading to a stronger alignment of the phyllosilicates with the macroscopic cleavage plane. The mineral fabric is still dominated by the intersection lineation. As at site PALS the mica maximum axes form a girdle structure about the intermediate axis whereas the chlorite maxima form a weak cluster. Anisotropy of magnetic susceptibility in low and high fields The magnetic fabric at room temperature shows that kv maximum axes are well-grouped about the lineation, similar to the phyllosilicate maximum axes (/]), and &3 minimum axes are offset from the pole to bedding, lying between the mica and chlorite minimum axes (f 3 ) (Figs 5 & 6). PAL2 shows a similar AMS fabric. The magnetic fabrics at PALI and PAL2 are similar to the magnetic fabric reported by Housen &

van der Pluijm (1991). At site PAL3 the k{ axes are tightly grouped about the intersection lineation and the k2 and k^ axes are distributed in a plane normal to the lineation. The &3 axes from specimens of coarser-grained siltstone are oriented closer to the bedding pole, whereas specimens taken from finer-grained siltstones lie closer to the cleavage pole (Fig. 5). The magnetic fabric at PAL4 has k3 axes normal to cleavage and k\ axes grouped around a stretching lineation within the cleavage. The AMS measured at liquid nitrogen temperature shows an increase over the room temperature values that average around 3.5 for all samples from Lehigh Gap. This is close to the ratio of 3.8, which would be expected if the AMS is controlled solely by paramagnetic minerals. The directions of the principal axes of the magnetic fabric at 77 K agree within a few degrees with those measured at room temperature and shown in Fig. 5. High-field torque is linearly proportional to the squared field in all samples from Lehigh Gap, which indicates that only paramagnetic phases contribute to the susceptibility anisotropy. The principal axes of the paramagnetic component agree well with the low-field AMS measured at both room temperature and liquid nitrogen temperature, which further supports the assumption that only paramagnetic phases are responsible for the observed anisotropy (Fig. 6).

Fig. 4. X-ray texture goniometry pole figures of mica and chlorite for the Lehigh Gap samples. Contour interval is 0.5 m.r.d. (multiples of random distribution). With increasing deformation the mica fabric develops from a cluster to a girdle distribution. The chlorite fabric shows an angle to bedding and less pronounced girdle structures. The mean directions of t\ axes are shown by squares, t2 axes by triangles and t3 axes by circles; S0 and S\ are the bedding and cleavage planes, respectively.

Fig. 5. Orientations of the principal axes of the low-field AMS ellipsoid in Lehigh Gap sites, measured at room temperature. The k\ axes are shown by squares, k2 axes by triangles and k3 axes by circles. The bedding plane is shown with a solid line, cleavage plane with a dotted line, and pencil lineation with a white diamond in this and subsequent figures. Plots are lower-hemisphere, equal area projections.

Fig. 6. Orientation of the principal axes of the chlorite (black symbols) and mica (grey symbols) mineral fabrics at Lehigh Gap. Open symbols indicate the site mean of low-field AMS and open symbols with dots show the paramagnetic component of individual samples from HFA. Maximum axes of the magnetic ellipsoids and minimum axes of the mineral fabric ellipsoids are shown by squares, intermediate axes by triangles, and minimum axes of the magnetic ellipsoids and maximum axes of the mineral fabric ellipsoids are shown by circles in this and subsequent figures. All axes are plotted on lower hemisphere, equal area stereograms. SQ and Si are the bedding and cleavage planes, respectively.

Fig. 7. Orientations of the principal axes of the AARM ellipsoids for (a) the low coercivity component (0.035mT d.c. bias field, 0-30 mT a.c. field) and (b) the high coercivity component (0.5mT d.c. bias field, 60-100 mT a.c. field). The bedding plane is shown with a solid line, cleavage plane with a dotted line.

120

A. M. HIRT ET AL.

Anisotropy of anhysteretic remanence The two coercivity windows in which the ARM was acquired activate different grain-size populations of the magnetite, and possibly of pyrrhotite. The lower coercivity window will largely affect coarse detrital grains of magnetite. The higher coercivity window will activate finer grained magnetite and pyrrhotite. The anisotropy of ARM (AARM) shows similar orientations in both coercivity windows for PALI and PAL2 (Fig. 7). The minimum axes are relatively well grouped and are closer to the bedding pole than the AMS fabric. This is again in agreement with the study of Housen & van der Pluijm (1991). At sites with a higher degree of deformation the fabrics in the two coercivity windows are no longer similar. The lower coercivity component is more poorly grouped, although there is still a loose grouping of ki axes close to the intersection lineation at PAL3 and an even looser grouping of £3 axes close to the pole to cleavage at PAL4. The magnetic fabrics of the higher coercivity fractions at PAL3 and PAL4 exhibit the effects of cleavage flattening and intersection lineation. Central Appalachian Fold and Thrust Belt, Pennsylvania Electron microscopy and texture goniometry The phyllosilicates at site MLM show alignment sub-parallel to the bedding plane and some minor and irregular kinking (Fig. 8a). The rocks do not contain much chlorite and the

fabrics are very weak (Fig. 9). The minimum axes of chlorite and mica display weak clustering with some girdling of the mica axes. At site FAN, pencil structures are well developed. Both the chlorite and mica fabrics show weak girdling of the minimum axes caused by increased kinking (Fig. 9), which is also observed in the scanning electron microscope (SEM) data. The maximum axes for mica and chlorite group around the intersection lineation (Fig. 10); the minimum axes are only slightly offset from the bedding pole. A similar fabric can be observed at PLA. The SEM data at site POR reveal a strong preferred orientation of the phyllosilicates parallel to the cleavage plane (Fig. 8b). Kinking is minor and new growth of grains occurs parallel to the cleavage plane. The minimum axes of chlorite and mica show clustered distributions with a strong point-maximum for mica and a weaker asymmetric maximum for chlorite, both normal to the cleavage plane (Fig. 9). Anisotropy of magnetic susceptibility: low and high fields Both low-field AMS, measured at 77 K, and HFA indicate that the susceptibility anisotropy is solely controlled by paramagnetic phases at all the sites. Bedding clearly controls the magnetic fabric at site MLM. There is a weak grouping of the k\ and k2 axes within the bedding plane (Fig. lOa). The AMS at site FAN is a classic 'pencil' fabric. The ki axes are well grouped along the intersection lineation and the k2 and A:3 axes are spread out in a girdle. PLA also shows a pencil fabric, but there is a preferential clustering of /c3

Fig. 8. SEM secondary electron images of the Appalachian fold and thrust belt showing kinking and microfolding at MLM and increased recrystallization parallel to the cleavage plane at POR. Scale bar represents 10 jim in each image (white bar in (a), black bar in (b)). SQ and Si are the traces of the bedding and cleavage planes, respectively.

Fig. 9. X-ray texture goniometry pole figures of the Appalachian fold and thrust belt showing the transition of weak girdles to cluster distributions of the mica with increasing deformation. Chlorite pole figures are less well denned due to low chlorite content in these rocks. Contour interval is 0.5 m.r.d. at sites MLM, FAN and PLA, and 1.0 at FOR. *S0 and Si are the bedding and cleavage planes, respectively.

Fig. 10. (a) Orientations of the principal axes of the AMS ellipsoids of specimens from the fold belt, (b) Plots showing the chlorite (black symbols) and mica (grey symbols) mineral fabrics, the AMS from the sample used for the mineral fabric measurement (open symbols with dot), and the low-field site average (open symbols) for the sites across the fold belt. The bedding plane is shown with a solid line, cleavage plane with a dotted line and pencil lineation with a grey triangle. All axes are plotted on lower hemisphere, equal area stereograms. SQ and S\ are the bedding and cleavage planes, respectively.

MAGNETIC AND MINERAL FABRIC DEVELOPMENT

axes in the plane normal to the intersection lineation close to the pole to cleavage. A triaxial ellipsoid, in which all three axes are well grouped, characterizes the AMS at site FOR, where A:3 is sub-parallel to the pole to cleavage and k\ is sub-parallel to the observed lineation. The orientations of the principal axes of the low-field AMS measured at 77 K and those of the HFA are very similar to the low-field AMS at room temperature. There is an excellent agreement of the principal axes of the AMS ellipsoid of the individual sample used for the texture goniometry measurement and the axes of the chlorite and mica fabric (Fig. lOb). This agreement supports the low-temperature and highfield results that attribute the anisotropy to paramagnetic phases. Since the chlorite and mica fabrics are similar, it is not possible to distinguish which is more important in defining the AMS. Discussion Graham (1966) described the evolution of the AMS ellipsoid shape that could be expected in sedimentary rocks that undergo a horizontal tectonic compaction. This pattern of fabric development was also reported by Kligfield et al. (1981) in Permian shales from the Alpes Maritimes, where they showed that the AMS was reflecting the strain that the rocks had undergone. Although both studies commented that the AMS development must be related to the orientation and anisotropy of the individual minerals, neither study provided a detailed analysis of the minerals responsible for the AMS. A combination of magnetic anisotropy, texture goniometry and SEM allows us to evaluate the development of magnetic and mineral fabric in more detail than these earlier works. The rocks at Lehigh Gap show the development of mineral and magnetic fabric over an outcrop scale. The shales show an original bedding compaction that is gradually overprinted by a horizontal tectonic shortening to form a slate. The tectonic shortening is first accommodated primarily by rotation and kinking of platy grains, as seen from the chlorite and mica fabrics. With increasing deformation neocrystallization of these minerals in the cleavage plane becomes a more dominant mechanism of grain reorientation. In SEM it can be observed that new grains are commonly smaller and aligned in cleavage lamellae, whereas detrital grains are found between the cleavage lamellae in the microlithons and are commonly larger and reoriented by mechanical rotation and kinking.

123

The low temperature and high field measurements have shown that the AMS is controlled by the paramagnetic minerals, and this accounts for the good agreement observed between the magnetic and mineral fabrics. At PALI and PAL2 the k3 axes lie on the great circle between the /3 axes of chlorite and mica, which suggests that both minerals are contributing to the AMS. At PAL4 the magnetic fabric is controlled by the cleavage, whereas the chlorite and mica still display girdled structures. This difference may be due to the fact that the texture goniometry was made on samples that had higher sand content than the samples used for the magnetic measurements, which indicates that the finer grain sizes take up the deformation more efficiently. The AARM results indicate that the orientation of the ferromagnetic minerals is controlled by bedding compaction at PALI and PAL2, and by cleavage at PAL3 and PAL4. The remanent magnetic fabrics are better defined in the higher coercivity window than in the low coercivity window. This suggests that the finer grained magnetite and pyrrhotite particles are better aligned than the coarser, low coercivity grains. Most sedimentary rocks from the central Appalachian orogenic belt have undergone a pervasive remagnetization in the Late Permian. McCabe & Elmore (1989) and McCabe et al, (1989) found new growth of magnetite during deformation throughout the Appalachians, and attributed this to fluid flow during the orogeny. Stamatakos et al. (1996) interpreted the Late Permian remagnetization as taking place in a relatively short period of time more rapidly than the folding and thrusting propagated toward the foreland. Although we have no independent evidence for new growth of ferromagnetic phases, it is possible that the finer grains formed in the stress field during deformation, whereas the coarser grains underwent mechanical rotation together with the phyllosilicates. This leads to the development of a paramagnetic-related AMS that rotated passively during fold growth as predicted by Graham (1966). An earlier study of the anisotropy of magnetic susceptibility and anhysteretic remanence at Lehigh Gap was made by Housen & van der Pluijm (1990; 1991). They found that the AMS (measured with a Sapphire Instrument (SI-2) susceptibility bridge in an applied field of 39A/m (0.049 mT, at a frequency of 750800 Hz) showed a sudden change from beddingcontrolled fabric to cleavage-controlled fabrics at a distance from the contact with the Shawangunk sandstone corresponding to our site PAL3.

124

A. M. HIRT ET AL.

This is in clear disagreement with our AMS results, which do not show an abrupt change but mirror the progressive deformation across the outcrop. The prolate AMS fabric, which we observe at PALS, clearly reflects the observed mineral fabric. One difference between the two studies is that Housen & van der Pluijm (1990) took many small hand samples continuously along the outcrop. The disadvantage of this method is that only 2-4 specimens represent each site and the results may only reflect a very local fabric, rather than a more representative averaged fabric (cf. Luneburg et al. 1999). In order to obtain statistically meaningful results we took a large number of samples at each site and located the sites across the outcrop so as to reveal the progressive variation. Our AARM results are similar to those obtained by Housen & van der Pluijm (1991, their Fig. 4). They found that bedding clearly controls the AARM fabric in the first 55m along the outcrop from the contact, which agrees with our results at PALI and PAL2. They reported some 'pencil fabrics' between 75m and approximately 90m and cleavagecontrolled fabrics further along the outcrop, which is in agreement with our results at PALS and PAL4. A similar pattern of progressive deformation is observed in the magnetic and mineral fabrics from the foreland of the fold-thrust belt to the hinterland. Sites close to the Appalachian Front have oblate magnetic fabrics. Some kinking of the phyllosilicate grains eventually leads to a grouping of the k\ axes sub-parallel to the fold axes and in the direction of pencil structures. Pencil structure becomes better developed towards the east, and at approximately 70km normal to the trend of the Allegheny Front the fabrics are prolate. Within 80km of the Front an incipient cleavage is found in the finer grained lithologies. The Martinsburg Formation does not crop out again until further to the east in the hinterland, where a welldeveloped slaty cleavage is found. The POR locality shows evidence of significant recrystallization; pole figures exhibit high-density distributions of the phyllosilicate c axes that define triaxial AMS ellipsoids. These results imply a gradual progression of deformation across the Central Appalachian fold-thrust belt that is related to the gradual rotation and recrystallization of grains into the plane of the tectonic shortening. Both transects, Lehigh Gap on the outcrop scale, and the Central Appalachian fold and thrust belt on the regional scale, show a gradual transition from shales to slates with progressive

deformation. This transition is characterized by dominance of different deformation mechanisms, starting with mechanical reorientation of detrital grains by rotation and kinking and leading into new growth of grains parallel to the cleavage plane. The resulting microstructural patterns are reflected in the development of the mineral and magnetic fabric pole figure patterns, which range from cluster distributions, representing bedding (oblate fabric ellipsoids), to girdle distributions (prolate fabric ellipsoids) and again to cluster distributions representing cleavage. Both transects, although on different scales, exhibit the same characteristic patterns of progressive deformation and fabric development.

Conclusions The development of magnetic fabric is shown to be closely related to the orientation of phyllosilicate grains on both an outcrop and regional scale in the Ordovician Martinsburg Formation in the Central Appalachian fold-thrust belt. The anisotropy of magnetic susceptibility in the rocks is carried exclusively by paramagnetic phases, and a good agreement is found between the mineral fabric of chlorite and mica and the AMS. At both Lehigh Gap and across Pennsylvania, the least deformed sites show bedding compaction with an intersection lineation subparallel to the trend of the major folds axes. This pattern develops into a true 'pencil' fabric, with a well-grouped intersection lineation, and later to a tectonic shortening. The development is gradual and is reflected in the orientation of the mineral grains and the magnetic anisotropy. Financial support was provided by the Swiss National Science Foundation (NF Project: 20-28884.90) We thank G. Oertel, J. Stamatakos, K. Kodama and R. Nickelson for their helpful discussions. The constructive reviews of I. Abad, M. Sintubin and D. Tarling are gratefully acknowledged. We especially thank M. Sintubin for pointing out an error in an earlier figure. This is publication number 1320 of the Institute of Geophysics, ETH Zurich.

References BALSLEY, J. R. & BUDDINGTON, A. F. 1960. Magnetic susceptibility and fabric of some Adirondack granites and orthogneisses. American Journal of Science, 258-A, 6-20. BECKER, J. K., SIEGESMUND, S. & JELSMA, H. A. 2000. The Chinamora batholith, Zimbabwe: structure and emplacement-related magnetic rock fabric. Journal of Structural Geology, 22, 1837-1853.

MAGNETIC AND MINERAL FABRIC DEVELOPMENT CASEY, M. 1981. Numerical-analysis of x-ray texture data - an implementation in Fortran allowing triclinic or axial specimen symmetry and most crystal symmetries. Tectonophysics, 78, 51-64. CHEENEY, R. F. 1983. Statistical Methods in Geology. George Allen & Unwin, London, 169. DAVIS, J. C. 1986. Statistics and Data Analysis in Geology, John Wiley & Sons, New York. DE WALL, H., BESTMANN, M. & ULLEMEYER, K. 2000. Anisotropy of diamagnetic susceptibility in Thassos marble: A comparison between measured and modeled data. Journal of Structural Geology, 22, 1761-1771. ENGELDER, T. & ENGELDER, R. 1977. Fossil distortion and decollement tectonics of the Appalachian plateau. Geology, 5, 457-460. ENGELDER, T. & GEISER, P. 1979. The relationship between pencil cleavage and lateral shortening within the Devonian section of the Appalachian Plateau, New York. Geology, 7, 460-464. EPSTEIN, J. B. & EPSTEIN, A. G. 1969. Geology of the Valley and Ridge province between Delaware Water Gap and Lehigh Gap, Pennsylvania. In: SUBITZKY, S. (ed.) Geology of Selected Areas in New Jersey and Eastern Pennsylvania, and Guidebook of Excursions. Rutgers University Press, New Brunswick, 132-205. EPSTEIN, A. G., EPSTEIN, J. B. & HARRIS, L. D. 1977. Conodont color alteration - an index to organic metamorphism. U.S. Geological Survey Professional Papers, 995, 1-27. FAILL, R. T. 1977. Fossil distortion, Valley and Ridge Province, Pennsylvania. Geological Society of America Abstracts with Programs, 9, 262. FAILL, R. T. 1997. A geologic history of the northcentral Apppalachians, Part 1, Orogenesis from the Mesoproterozoic through the Taconic orogeny. American Journal of Science, 297, 551-619. FERRILL, D. A. & DUNNE, W. M. 1989. Cover deformation above a blind duplex: an example from West Virginia, U.S.A. Journal of Structural Geology, 11,421-431. FULLER, M. D. 1963. Magnetic anisotropy and palaeomagnetism. Journal of Geophysical Research, 68, 293-309. GEISER, P. A. 1988. The role of kinematics in the construction and analysis of geological cross sections in deformed terranes. In: MITRA, G. & WOJTAL, S. (eds) Geometries and Mechanisms of Thrusting, Geological Society of America Special Paper 222, 47-76. GEISER, P. 1989. Day 5: Deformation fabrics and mechanisms of an 'autochthonous roof duplex: The Valley and Ridge province of Pennsylvania and Maryland. In: Engelder, T. (ed.) Structures of the Appalachian Foreland Fold-Thrust Belt, Field Trip Guidebook T166. American Geophysical Union, Washington, D.C., 44-52. GEISER, P. & ENGELDER, T. 1983. The distribution of layer parallel shortening fabrics in the Appalachian foreland of New York and Pennsylvania: Evidence for two non-coaxial phases of the Alleghanian orogeny. Geological Society of America Memoir, 158, 161-175.

125

GIRDLER, R. W. 1961. Some preliminary measurements of anisotropy of magnetic susceptibility. Geophysical Journal of the Royal Astronomical Society, 5, 197-206. GRAHAM, J. W. 1954. Magnetic susceptibility anisotropy, an unexploited petrofabric element. Geological Society of America Bulletin, 65, 1257-1258. GRAHAM, J. W. 1966. Significance of magnetic anisotropy in Appalachian sedimentary rocks. In: Steinhart, J. S. & Smith, T. J. (eds) The Earth Beneath the Continents, Geophysical Monograph 10. American Geophysical Union, Washington, 627-648. GWINN, V. E. 1964. Thin skinned tectonics in the plateau and northwestern Valley and Ridge Provinces of the central Appalachians. Geological Society of America Bulletin, 75, 863-900. HATCHER, R. D., THOMAS, W. A., GEISER, P. A., SNOKE, A. W., MOSHER, S. & WILTSCHKO, D. V. 1989. Alleghanian orogen. In: HATCHER, R. D., THOMAS, W. A. & VIELE, G. W. (eds) The Appalachian-Ouachita Orogen in the United States. The Geological Society of America, Boulder, Colorado, The Geology of North America, F-2, 233318. HIRT, A. M., EVANS, K. F. & ENGELDER, T. 1995. Correlation between magnetic anisotropy and fabric for Devonian shales on the Appalachian Plateau. Tectonophysics, 247, 121-132. Ho, N.-C., PEACOR, D. & VAN DER PLUIJM, B. 1995. Reorientation mechanisms of phyllosilicates in the mudstone-to-slate transition at Lehigh Gap, Pennsylvania. Journal of Structural Geology, 17, 345-356. HOLEYWELL, R. C. & TULLIS, T. E. 1975. Mineral reorientation and slaty cleavage in the Martinsburg Formation, Lehigh Gap, Pennsylvania. Geological Society of America Bulletin, 86, 1296-1304. HOUSEN, B. A. & VAN DER PLUIJM, B. A. 1990. Chlorite control of correlations between strain and anisotropy of magnetic susceptibility. Physics of the Earth and Planetary Interiors, 61, 315-323. HOUSEN, B. A. & VAN DER PLUIJM, B. A. 1991. Slaty cleavage development and magnetic anisotropy fabrics. Journal of Geophysical Research, 96, 9937-9946. IHMLE, P. F., HIRT, A. M., LOWRIE, W. & DIETRICH, D. 1989. Inverse magnetic fabric in deformed limestones of the Morcles nappe, Switzerland. Geophysical Research Letters, 16, 1383-1386. KLIGFIELD, R., OWENS, W. H. & LOWRIE, W. 1981. Magnetic susceptibility anisotropy, strain and progressive deformation in Permian sediments from the Maritime Alps (France). Earth and Planetary Science Letters, 55, 181-189. LOWRIE, W. 1990. Identification of ferromagnetic minerals in a rock by coercivity and unblocking temperature properties. Geophysical Research Letters, 17, 159-162. LUNEBURG, C. M., LAMPERT, S. A., LEBIT, H. D., HIRT, A. M., CASEY, M. & LOWRIE, W. 1999. Magnetic anisotropy, rock fabrics and finite strain in deformed sediments of SW Sardinia (Italy). Tectonophysics, 307, 51-74.

126

A. M. HIRT ET AL.

MARTIN-HERNANDEZ, F. & HIRT, A. M. 2001. Separation of ferrimagnetic and paramagnetic anisotropies using a high-field torsion magnetometer. Tectonophysics, 337, 209-222. McCABE, C. & ELMORE, D. 1989. The occurrence and origin of Late Paleozoic remagnetization in the sedimentary rocks of North America. Reviews in Geophysics, 27, 411-494. McCABE, C., JACKSON, M. & ELLWOOD, B. B. 1985. Magnetic anisotropy in the Trenton limestone: results of a new technique, anisotropy of anhysteretic susceptibility. Geophysical Research Letters, 12, 333-336. McCABE, C., JACKSON, M. & SAFFER, B. 1989. Regional patterns of magnetic authigenesis in the Appalachian basin: Implications for the mechanism of Late Paleozoic remagnetization. Journal of Geophysical Research, 94, 1042910443. NICKELSEN, R. P. 1966. Fossil distribution and penetrative rock deformation in the Appalachian plateau. Journal of Geology, 74, 924-931. NICKELSEN, R. P. 1979. Sequence of structural stages of the Allegheny orogeny at the Bear Valley Strip Mine, Shamokin, Pennsylvania. American Journal of Science, 279, 225-271. OERTEL, G. 1983. The relationship of strain and preferred orientation of phyllosilicate grains in rocks: a review. Tectonophysics, 100, 413-447. RODGERS, J. 1963. Mechanics of Appalachian foreland folding in Pennsylvania and West Virginia.

American Association of Petroleum Geologists Bulletin, 47, 1527-1536. SIDDANS, A. W. 1976. Deformed rocks and their textures. Philosophical Transactions of the Royal Society of London, 283, 43- 54. SlEGESMUND, S., UlXEMEYER, K. & DAHMS, M.

1995.

Control of rock fabrics by mica preferred orientation: a quantitative approach. Journal of Structural Geology, 17, 1601-1613. STACEY, F. D., JOPLIN, G. & LINDSAY, J. 1960. Magnetic anisotropy and fabric of some foliated rocks from S.E. Australia. Geofisica pura e applicata, 47, 30-40. STAMATAKOS, J., HIRT, A. M. & LOWRIE, W. 1996. The age and timing of folding in the central Appalachians from paleomagnetic results. Geological Society of America Bulletin, 108, 815-829. WINTSCH, R. P., KVALE, C. M. & KISCH, H. J. 1991. Open-system, constant-volume development of slaty cleavage and strain-induced replacement reactions in the Martinsburg Formation, Lehigh Gap, Pennsylvania. Geological Society of America Bulletin, 103, 916-927. WRIGHT, T. O. & PLATT, L. B. 1982. Pressure dissolution and cleavage in the Martinsburg shale. American Journal of Science, 282, 122-134. WRIGHT, T. O., STEPHENS, G. C. & WRIGHT, E. K. 1979. A revised stratigraphy of the Martinsburg Formation of eastern Pennsylvania and paleogeographic consequence. American Journal of Science, 279, 1176-1186.

An integrated AMS, structural, palaeo- and rock-magnetic study of Eocene marine marls from the Jaca-Pamplona basin (Pyrenees, N Spain); new insights into the timing of magnetic fabric acquisition in weakly deformed mudrocks JUAN C. LARRASOANA1 2 3 , EMILIO L. PUEYO24 & JOSEP M. PARES5 1

Paleomagnetic Laboratory, Institute of Earth Sciences 'Jaume Aimer a', CSIC, c/ Sole i Sabaris s/n, Barcelona 28080, Spain 2 Department of Earth Sciences, University of Zaragoza, c/ Pedro Cerbuna 12, Zaragoza 50009, Spain 3 Corresponding author, present address: Southampton Oceanography Centre, European Way, Southampton S014 3ZH, UK (e-mail: [email protected]) Laboratoire des Mecanismes de Transfert en Geologic, Universite Paul Sabatier, Toulouse, France 5 Department of Geological Sciences, University of Michigan, 2534 C. C. Little Building, Ann Arbor, MI 48109, USA Abstract: In this paper, we present a combined magnetic anisotropy (AMS), structural, palaeo- and rock-magnetic study of Eocene marine mudrocks from the western sector of the Jaca-Pamplona Basin (southwestern Pyrenees, north Spain). Comparison of structural, AMS and palaeomagnetic data reveals a subtle, but evident tectonic overprint affecting the phyllosilicate matrix of the mudrocks and the remanence carriers. The particular structural setting of the studied rocks has allowed us to combine palaeomagnetic and structural data with the AMS results in order to establish a relative chronology between sedimentation, blocking of magnetic fabrics, acquisition of magnetic remanence and deformation. Our data suggest that the blocking of the magnetic fabrics and the lock-in of the remanence occurred simultaneously during the early stages of gentle warping that affected the Jaca-Pamplona basin throughout the Mid-Late Eocene. According to the origin of the remanence carriers and the synsedimentary nature of the Mid-Late Eocene warping, the blocking of the magnetic fabrics and the lock-in of the remanence can be bracketed to a very short time span, of a few (<15?) ka, after sediment deposition. Our findings confirm previous interpretations claiming a very early origin of the magnetic fabric blocking in mudrocks undergoing weak deformation. They therefore validate the use of magnetic fabrics as palaeostress indicators and suggest that AMS data might also provide a useful tool for detecting preferred paths for the migration of geofluids during the early stages of deformation.

The study of the fabric in deep-sea sediments has deserved a special attention in the last few years because it has great potential for furthering our understanding of initial structural processes in environments such as accretionary prisms and fold and thrust belts (e.g. Kanamatsu et al. 2001). One of the most useful techniques used to determine sediment fabrics of mudrocks is the anisotropy of magnetic susceptibility (AMS), because it is generally an effective tool to infer the deformation suffered by rocks that do not show any evidence of deformation at the mesoscopic scale (see reviews by Tarling & Hrouda 1993; and Borradaille & Henry 1997). According to several studies (Borradaille & Tarling 1981; Kissel et al. 1986; Averbuch et al. 1992; Pares & Dinares-Turell 1993; Sagnotti &

Speranza 1993; Mattei et al. 1995; Pares et al. 1999), the magnetic fabric of a mudrock undergoes a series of changes during weak deformation that can be described by four main types of magnetic ellipsoids. Type 1 ellipsoids correspond to the initial, undeformed state and are characterized by a cluster of Kmin parallel to the bedding pole and a dispersion of Kmax and K;nt within the bedding plane. This fabric usually results from the collapse of clay aggregates and their deposition parallel to the water-sediment interface and randomly oriented throughout this surface, as well as from subsequent compaction driven by overburden pressure (Sintubin 1994). Type 2 ellipsoids are defined by an initial alignment of Kmax perpendicular to the shortening direction, with Kmin still normal to bedding.

From\ MARTIN-HERNANDEZ, F., LUNEBURG, C. M., AUBOURG, C. & JACKSON, M. (eds) 2004. Magnetic Fabric: Methods and Applications. Geological Society, London, Special Publications, 238, 127-143. 0305-8719/04/ $15.00 © The Geological Society of London 2004.

128

J. C. LARRASOANA ET AL.

In Type 3 magnetic ellipsoids, Kmax clearly clusters parallel to the fold axis while Kmin and Kint begin to form pseudo-girdles (60°) perpendicular to Kmax. Finally, Type 4 ellipsoids are characterized by a clear girdle distribution of Kmin and Kjnt perpendicular to Kmax, which remains clustered perpendicular to the shortening direction. Among the vast number of studies dealing with AMS in weakly deformed mudrocks, little attention has been paid to the timing of magnetic fabric blocking processes. In order to explain the similar degree of clay fabric development throughout a sequence of shales, regardless of their position within the stratigraphic pile, Sintubin (1994) has suggested that blocking of sediment fabric occurs during the early stages of diagenesis. A similar conclusion can be drawn fro the the work done by Mattei et al. (1995), Sagnotti et al. (1998) and Coutand et al. (2001). In those cases, the magnetic lineation defined by Kmax shows a degree of rotation comparable to the vertical-axis rotations determined palaeomagnetically, suggesting that the blocking of the magnetic fabrics is not much older than the acquisition of the magnetic remanence. An early blocking of the magnetic fabrics can be also derived from the study by Sagnotti et al. (1999), where AMS, borehole breakout and focal mechanism data gave consistent results in terms of palaeostress determinations. Based on the study of Eocene mudrocks from the eastern Pyrenean foreland basin, Pares et al. (1999) have proposed that magnetic fabric blocking occurs when the sediments are only partially lithified and thus relatively soft, provided that the presence of a fluid in the sediment permits particulate flow and grain rotation (Borradaille 1981) and that crystal plasticity can be ruled out considering the P-T conditions in the basin. Despite the growing evidence pointing to an early origin for sediment fabrics in weakly deformed mudrocks, the precise timing of magnetic fabric acquisition is still to be determined and it may have important consequences for better understanding the implications and applications of AMS data. In this paper, we use a combination of structural and palaeomagnetic data from Eocene marine mudrocks from the central-western Pyrenees to bracket the timing of fabric blocking as derived from AMS data. The rationale of our approach is that, if any assemblage of deformed rocks displays coherent primary magnetization directions, it must indicate that any sediment reorganization must have taken place essentially before the lock-in of the magnetization, otherwise the remanence directions should be

significantly affected by the reorientation of magnetic grains according to tectonic deformation. AMS data have been used to establish the fabric of the marls, whereas structural data have been used to assess their tectonic origin and the combination of previous palaeomagnetic and rock-magnetic results has been used to establish the timing of the remanence acquisition processes. Geological setting The Pyrenees constitute a double vergence linear mountain belt formed at the boundary between the Iberian and the European plates. From Late Cretaceous to Miocene, the oblique convergence, collision and subduction of Iberia underneath Europe caused the end of sea-floor spreading in the Bay of Biscay and the initiation of compression along the northern Iberian margin, leading to inversion of Mesozoic extensional basins and ultimately to uplift of the Pyrenees (Choukroune 1992). The Jaca-Pamplona basin represents the external, westernmost unit of the mountain belt in its east-central sector (Fig. 1). In the studied area, which corresponds to the western half of the basin, its geometry can be described by a synclinorium with a small, steep northern limb and a long, subhorizontal southern flank. The later corresponds to a monoclinal flexure that is related to the culmination limb associated with the south Pyrenean frontal system. This thrust system involved a SW tectonic transport of the basin of about 15.6km (Larrasoana et al. 2003a). Neither metamorphism nor high-pressure penetrative deformation has been reported affecting the Jaca-Pamplona basin and other equivalent units in the southern Pyrenees (Choukroune 1992). The Middle-Upper Eocene marls of the Erize (Payros 1997) and Pamplona-Arguis (Puigdefabregas 1975) Formations were deposited in the westernmost sector of the Jaca-Pamplona basin when it constituted a foreland basin. Both formations comprise a monotonous sequence of bluish marine marls that were deposited, respectively, in the upper slope and the distal part of a deltaic system that opened toward the west (Puigdefabregas 1975; Payros 1997). Relatively deep, marly facies are typical throughout the stratigraphic pile although some turbiditic levels are also present, especially in the Erize marls and the base of the PamplonaArguis Formation (Puigdefabregas 1975; Payros 1997). In the studied area, the total thickness of the marls varies between 2000 m on the

Fig. 1. Geological sketch map of the western sector of the Jaca-Pamplona basin with the location of the studied sites. Kmax (black arrows) indicates the mean orientation of the maximum susceptibility axes at every site whereas Ps (white arrows) indicates the site-mean orientation of pencil structures. To clarify the plot, the easterly and westerly directions of Kmax and Ps have been arbitrarily considered, respectively. Fabric types ranging from 1 to 4 reflect an increased tectonic overprint on the magnetic fabrics (see text). Small numbers refer to sites and large numbers to structural units: 1: Lumbier syncline; 2: Izaga syncline; 3: Pamplona syncline; 4: Liedena anticline; 5: Ezkaba syncline; 6: Anezkar syncline; 7: Ibero syncline; 8: Senosiain syncline; 9: Goni syncline; 10: Alaiz unit; 11: Guembe block.

130

J. C. LARRASOANA ET AL.

northern margin of the basin and less than 500 m on its southern margin. From Palaeocene to Late Eocene, the JacaPamplona basin constituted a foreland basin that was formed in response to lithospheric flexure due to tectonic loading in the internal parts of the Pyrenean belt (Teixell 1996). During the Mid and Late Eocene, the studied area underwent a gentle synsedimentary warping along NW-SE striking structures. Such gentle synsedimentary warping explains the differences in the thickness of the Eocene marls and the overlaying coastal deposits of the Gendulain Formation, as well as the presence of tectonically driven slumps and soft-sedimentary structures in both formations (Leon-Chirinos 1985; Baceta et al. 2000). The basin was incorporated into younger thrust sheets during the latest Eocene, and was folded and transported toward the south during the main Mid-Oligocene-Early Miocene southern Pyrenean thrusting event (Teixell 1996). Methods A total of 46 sites distributed throughout the western sector of the Jaca-Pamplona basin were sampled for AMS and palaeomagnetic purposes. At every site, 8 to 15 standard oriented cores spanning a stratigraphic section of up to 15m were collected with a portable gas-powered drill. The AMS was measured for 6 to 19 (typically 10 to 15) standard specimens at every site using two KLY-2 susceptibility meters at the IES Jaume Almera (CSIC) in Barcelona (Spain) and the Department of Geological Sciences in Ann Arbor (Michigan, USA). The anisotropy of the magnetic susceptibility is a second-rank tensor that can be graphically displayed by a three-axis ellipsoid with its own orientation, shape and degree of anisotropy. The orientation of the magnetic ellipsoids has been expressed by means of the directions of their three main axes (Kmax > Kint > Kmin). The shape and degree of anisotropy of the ellipsoids are described using the shape parameter (T) and the corrected anisotropy degree (P') (Jelinek 1981). It has been shown in several studies that the Kmax axes are easily reoriented perpendicular to the shortening direction under very low degrees of deformation, and that they remain parallel to the fold axis under progressive deformation (Borradaille & Tarling 1981; Pares & Dinares-Turell 1993; Sagnotti & Speranza 1993). As a consequence, most of the magnetic ellipsoid evolution path takes place within the same region of the prolate field and therefore the classical P'-T (Hrouda-Jelinek)

plots based on bulk properties fail to distinguish the different fabric types (Pares et al. 1999). To tackle this shortcoming, Pares et al. (1999) have envisaged a method to discern the magnetic fabrics that relies on the distribution of the Kmin axes, which are more sensitive to weak tectonic deformation processes. Following these authors, the shape and strength of the distribution of the Kmin axes are evaluated by plotting the eigenvalues that characterize their distribution in a modified Woodcock diagram (Woodcock 1977). The magnetic fabrics of the studied Eocene marls have been additionally described using this method, which has been applied to evaluate the distribution of both the Kmax and Kmin axes. In order to establish the relative contribution of paramagnetic minerals to the mean susceptibility, hysteresis and low-temperature measurements were performed at the Institute for Rock Magnetism (Minneapolis, USA) with a subset of representative samples using a Micromag-VSM and a Lakeshore susceptibility meter, respectively. After measuring the AMS, the same samples were used to conduct a palaeomagnetic and rock-magnetic study aiming at quantifying the amount of vertical-axis rotations in the area and the origin of the magnetic carriers. These results have been already published (Larrasoana et al. 2003a, b) and only those aspects relevant to the present study will be considered. Results Structural data The studied area has been divided in 11 structural units. Three of them (i.e. the Lumbier, Izaga and Pamplona synclines) correspond to three different sectors of the main synclinorium that constitutes the westernmost sector of the Jaca-Pamplona basin (Fig. 1), and include about half (25) of the studied sites. Another six units correspond to minor folds located especially in the westernmost sector of the basin, and include 20 sites. These 9 structural units correspond to cylindrical folds with gently plunging axes trending parallel to the ENEWSW main Pyrenean direction (Fig. 2 & Table 1). The last two units considered in this study are the Alaiz unit and the Guembe block, and include 7 and 1 sites respectively. The former unit corresponds to the hanging wall of a lateral ramp of the South Pyrenean frontal system (the Alaiz thrust) and has a monoclinal geometry slightly curved at both terminations. The latter unit corresponds to a SW dipping, fault-bounded

Fig. 2. Equal area, lower hemisphere projections showing a comparison between the fold axes determined from bedding poles and the mean orientation of site-mean Kmax axes and pencil structures at every structural unit. N indicates the number of AMS sites at every structural unit. The circles around the mean direction of pencil structures and Kmax site-mean directions represent the a95 angle around the mean values. Due to the low number of data in the Anezkar syncline, the a95 angle around the mean orientation of site-mean Kmax axes has not been plotted. To clarify the plot, the easterly and westerly directions of Kmax and pencil structures have been arbitrarily considered, respectively. Filled (open) symbols indicate projections in the lower (upper) hemisphere.

J. C. LARRASOANA ET AL.

132

Table 1. Comparison of fold axis direction, mean direction of site-mean Kmax axes and mean direction of pencil structures at every structural unit Fold axis

Mean of Kmax

Pencil Structures

Unit

Dec

Plunge

N

Dec

Inc

N

Dec

Inc

N

Lumbier (1) Izaga (2) Pamplona (3) Liedena (4) Ezkaba (5) Anezkar (6) Ibero (7) Senosiain (8) Com (9)

303 294 109 278 293 105 105 121 90

8 6 3 8 15 20 6 0 1

22 26 33 16 7 14 15 6 16

291 300 120 108 304 121 119 118 108

8 11 4 2 16 9 3 9 1

5 7 9 6 2 3 3 2 4

291 305

11 10

12 8

81 310

7 8

8 6

267

2

7

N indicates the number of bedding poles, site-mean Kmax directions and pencil structures measured, respectively.

block located at the westernmost tip of the studied area. Structural lineations ranging from coarse to well-developed pencil structures appear in 12 out of the 46 sites studied. Pencil structures form when the bedding fissility intersects with a weakly developed cleavage and indicate a 9% to 26% of tectonic shortening occurred mainly during layer-parallel shortening and low-amplitude folding (Reks & Gray 1982). As expected, the mean direction of the pencil structures measured at the different structural units matches very closely the orientation of the fold axes (Fig. 2 & Table 1).

AMS data Low temperature measurements of a low-field magnetic susceptibility and hysteresis experi-

ments demonstrate that the magnetic susceptibility in the mudrocks is mainly controlled by the paramagnetic minerals that constitute the matrix of the rock (Fig. 3). X-ray diffraction data indicate that these paramagnetic minerals are chlorite, illite-muscovite and, in much lower amounts, pyrite and ankerite (Larrasoana 2000). According to these results, the magnetic fabric of the studied marls can be unambiguously interpreted in terms of preferred orientation of phyllosilicate grains. The magnetic fabric of the studied marls can be grouped into the four types of fabrics that have been classically described for weakly deformed mudrocks undergoing progressive deformation (e.g. Kissel et al. 1986; Averbuch et al 1992; Pares & Dinares-Turell 1993; Sagnotti & Speranza 1993; Mattei et al. 1995; Pares et al. 1999) (Fig. 4). Most of the studied marls show Type 2 ellipsoids, with only 6, 3

Fig. 3. Rock-magnetic data for representative samples of Eocene marine marls, (a) Low-temperature analysis of a low-field magnetic susceptibility measured at 400 Hz. The curve is well fitted by a hyperbola, indicating a negligible contribution of ferromagnetic (sensu laid] minerals, (b) Hysteresis loop. Notice the very weak hysteresis behaviour, indicative of a very subtle contribution of the ferromagnetic minerals to the room temperature magnetic susceptibility.

MAGNETIC FABRIC ACQUISITION IN MUDROCKS

133

Fig. 4. Comparison of the four different types of magnetic fabrics with a P'-T and a Woodcock-type diagram describing the distribution of the Kmin axes. One representative example of AMS directional data is shown (in equal area, lower hemisphere projection) for each of the four types of fabrics described (see text). Squares indicate Kmax axis whereas circles (triangles) represent Kmax (Kint) axis. The stars correspond to the bedding poles. In the Woodcock-type diagram, C is a measure of the strength of the preferred orientations whereas K gives the shape of the fabrics. Deformation stages after Pares et al. (1999): ED: earliest deformation; PS: pencil structure; WC: weak cleavage. E 1? E2 and E3 are the eigenvalues (normalized to 1) that describe the site-mean distribution of the kmin susceptibility axes.

and 2 sites belonging to Types 1, 3 and 4, respectively (Table 2). We notice that, regardless of the type of fabric shown by the marls, the mean orientation of Kmax at each site matches very closely the direction of the fold axis and the mean orientation of the pencil structures, when present (Fig. 3 & Table 1). The correspondence between the direction of K max , the orientation of pencil structures (if present) and the orientation of the fold axes is also evident at the scale of individual sites but with one exception (site 20; Fig. 1 & Table 2). Such correspondence attests the tectonic origin of the fabrics in the studied marls regardless of their overall low degree of tectonic

overprint. We notice that the site-mean susceptibility values tend to decrease as we move from E to W and from lower to upper stratigraphic levels in the studied marls. This trend might indicate variations in the carbonate content of the marls, and it could be argued that such variations impose a lithological control on the type of fabric and in the formation of pencil structures, regardless of deformation conditions. Sites 18 and 35 are the only two sites showing Type 4 fabrics, yet their site-mean susceptibility values are similar to those of sites 22, 28, 37, 38, 39 and 43 and these sites do not show a Type 4 fabric. In addition, pencil structures are observed only

Table 2. List of palaeomagnetic, AMS and structural data at every site

AMS

ChRM Unit

Site

Dec

Inc

«95

V

max

Type

K mean (SD)

N P' 19 1.0738 15 1.0855 17 1.0658 11 1.0803 15 1.0857

T

Pencil Str

Kmin

Jv

Dec

Inc

ln(El/E2) ln(E2/E3) In (E1/E2) In (E2/E3)

1 1 1 1 1

1 (LU01) 2 (LU02) 3 (LU03) 4 (LU05) 5 (LU04)

9 360 4 7 359

42 49 40 52 47

1.9 4.5 6.7 8.2 3.3

2 2 3 2 2

137.81 (4.75) 135(6.15) 168.62 (4.69) 164.22(8.57) 157.92(7.97)

0.095 0.64 0.374 0.371 0.537

296 291 293 284 293

3 8 11 4 15

4.59 4.33 2.44 3.62 5.52

2.30 2.56 4.38 1.65 1.39

4.92 6.18 5.00 3.68 6.49

1.58 0.83 1.57 1.94 0.91

2 2 2 2 2 2 2 3 3 3 3 3 3 3 3

6 (IZ01) 7 (IZ03) 8 (IZ06) 9 (IZ07) 10 (IZ09) 11 (AJ01) 12 (RI01) 13(IR02) 14 (IR03) 15(IR04) 16 (IR08) 17 (IR09) 18 (BA01) 19 (AZ01) 20 (PT01)

357 9 37 357 358 10 4 8 353 20 1 348 15 16 7

60 55 46 76 31 48 50 -70 -71 -34 -59 -57 61 -56 -58

4.7 19.4 3.6 3.3 17.7 3.8 6.5 10.5 12.5 6.5 9.9 7.3 13.6 4.5 13.3

2 2 2 2 2 2 2 2 2 1 2 2 4 3 1

177(8) 15 .0812 0.598 134.53(12.93) 14 .0352 0.395 155.63(3.9) 10 .0647 0.17 g .0728 0.437 153.4(5.65) 142.73 (2.37) 10 .0489 0.493 7 .0807 0.399 128.78 (6.05) 150.15(11.52) 11 .0867 0.5 138.31 (3.84) ..0613 0.763 131.19(14.05) 11 .0558 0.679 7 .054 0.618 147.08(11.01) 11 .0548 0.647 125.09 (20) 8 .0543 0.585 100.63 (9.17) 88.74 (20.57) 19 .0175 -0.241 109.61 (10.28) 16 .0293 0.433 0.634 15 .027 91.85(8.26)

296 121 309 295 294 308 295 109 112 131 131 116 125 301 186

11 6 13 13 13 19 14 2 2 5 4 5 1 4 14

3.84 3.19 4.06 3.89 4.70 4.82 4.70 5.51 3.41 1.85 3.75 3.28 3.43 2.74 0.69

1.95 1.18 2.83 2.30 1.10 1.39 2.20 0.29 2.08 3.11 2.44 1.79 1.04 2.15 2.19

4.49 3.84 4.25 5.95 6.00 5.96 5.36 5.37 5.11 5.01 5.73 4.67 0.59 3.43 2.25

1.17 0.94 2.44 1.80 1.41 1.61 0.57 0.97 1.24 0.41 0.75 0.82 2.91 1.83 0.79

4 4 4 4 4 4

21 (LI01) 22 (LI02) 23 (LI03) 24 (LI04) 25 (LI05) 26 (LI06)

352 2 8 4 13

5 5

27 (EZ01) 28 (EZ02)

353 13

61 -41

6 6 6

29 (IR05) 30 (IR06) 31 (IR07)

38

-53

19

57

Dec Inc 298 285 288 297

11 13 8 11

309

9

304

14

143.71 (3.37) 88.12(3.98) 135.76 (13.96) 100.83(11.99) 150.89(6.78) 148.56(13.84)

9 10 14 15 14 11

.0916 .0709 .0536 .0401 .078 .0727

0.491 0.612 0.587 0.36 0.729 0.453

77 298 298 290 103 299

16 4 3 6 20 5

5.52 2.34 1.48 3.79 3.70 3.17

1.39 2.67 2.81 0.89 1.57 1.29

4.80 5.32 3.55 4.24 5.35 4.35

2.75 1.85 1.50 0.37 0.81 2.26

77

8

304

4

10.6 2 10.4 3

97.67 (5.71) 73.3 (9.42)

14 14

1.0271 1.0253

0.054 0.436

304 303

18 13

3.20 3.17

1.47 2.08

3.39 3.06

1.35 2.23

308 313

7 9

14.5 2 2 9.3 2

144.9 (20.2) 118.27(5.02) 138.6(7.77)

9 1.0511 7 1.0653 7 1.046

0.644 0.674 0.467

154 119 274

17 16 7

2.28 5.29 5.11

2.73 0.51 1.79

4.86 5.23 5.55

1.42 0.57 0.83

2 44 17.9 2 41 23.6 1 45 10.6 2 59 3.9 2 60 4.5 2

7 7 7

32(IB01) 33(OR01) 34(OY01)

8 8

35(SE01) 36(SE03)

9 9 9 9

37(GO01) 38(GO02) 39(GO03) 40(GO04)

10 10 10 10 10

41 (YA01) 42(YA02) 43(OT01) 44(GU01) 45(TI01)

11

46(GE01)

349 349 343

40 -40 48

17.5 1 9.6 1 7.1 2

99.39(17.32) 91.59(6.36) 112.37(7.5)

8 1.0764 15 1.0275 15 1.0495

120 124 299

14 16 26

0.93 1.65 3.46

3.44 2.88 1.00

3.96 4.35 3.94

1.63 1.89 1.59

9

-32

4 7.1 2

74.4(19.21) 137.02(9.55)

14 1.0224 -0.115 137 15 1.0741 0.617 99

5 13

3.17 3.94

2.59 1.56

1.20 4.99

3.67 1.27

15

45

2 2 2 8.3 1

82.2(11.49) 75.32(6.66) 87.05(8.84) 106.22(8.74)

15 14 13 13

1.027 1.0436 1.0326 1.0343

0.265 0.767 0.224 0.541

111 118 290 271

1 3 0 0

3.79 2.90 2.75 2.24

1.15 2.16 3.40 2.07

3.71 5.00 3.83 3.64

1.04 0.37 0.73 0.58

20 17 355 14 360

48 48 45 48 55

2 2 2 2 2

167.6(5.84) 159.58(5.36) 88.75(11.73) 141.78(4.29) 96.51(5.3)

7 8 15 7 15

1.0686 1.0675 1.0626 1.0753 1.0362

0.688 0.631 0.619 0.536 0.599

118 98 81 278 264

6 11 11 7 14

4.12 3.10 4.05 2.94 3.66

1.67 2.66 0.75 3.22 1.83

4.82 5.72 4.37 5.67 4.79

1.23 0.53 2.54 0.86 0.65

1

66.03(7.17)

15

1.0412

0.695

129

21

1.14

3.17

4.27

1.06

8.7 9.8 8.7 14.4 8.3

0.858 0.599 0.625

266

2

255

19

Structural units are given as in Fig. 1. The declination (Dec), inclination (Inc) arid confidence angle (a95) of the palaeomagnetic directions are given after tilt correction. For the bulk AMS data, the type of fabric (Type), the mean susceptibility value in 10~6 SI units and its standard deviation (Kmean (SD)), the number of samples (N), the corrected anisotropy degreee (P7) and shape parameter (T) are given. For the distribution of Kmax, the declination (Dec) and inclination (Inc) of the mean direction are given together with the logarithm of the eigenvalues that characterize their distribution (ln(El/E2) and ln(E2/E3)). For the distribution of K min , the logarithm of the eigenvalues that characterize their distribution (ln(El/E2) and ln(E2/E3)) are also given.

136

J. C. LARRASOANA ET AL.

in the Erize Formation and in the basal levels of the Pamplona-Arguis marls, without regard to the site-mean susceptibility values (Table 2). Thus, our data rule out a possible lithological control on the AMS results and the petrofabric of the marls, and demonstrate that the type of magnetic fabrics shown by the marls can be interpreted in terms of increasing deformation conditions. In Figure 4, the P'-T diagram is compared with a modified Woodcock plot describing the distribution of the Kmin axes. In the P'-T diagram, most of the sites show oblate fabrics with variable anisotropy degrees and, with the exception of Type 4 sites, no clear correlation can be made between the different types of fabrics and their distribution in the plot. Thus, sites with fabrics that correspond to increasing tectonic overprint do not clearly follow the classic evolution from oblate (sedimentary) to prolate (superimposition of tectonics and sedimentary) and back to oblate (tectonic) fabrics (Pares et al. 1999; Kanamatsu et al. 2001). This lack of correlation is due to the fact that most of the studied sites have a low tectonic overprint and therefore the P'-T diagram fails to recognize potential differences among them. In contrast, the modified Woodcock diagram shows a better correlation between the fabric type and the expected degree of tectonic overprint. All sites showing Type 1 ellipsoids fall in the 'early deformation' field of the diagram. Most of the sites showing Type 2 ellipsoids also fall in the 'early deformation' field. However, the tendency toward a stronger distribution of the Kmin axes (C < 6 for Type 1 sites; C > 6 for most Type 2 sites) and the presence of some sites in the 'pencil structure' field of the diagram point to an increased effect of deformation affecting Type 2 ellipsoids. Sites with Type 3 ellipsoids seem to record a more girdle-like behaviour and a slight decrease in the strength of the Kmin axes distribution (C is usually less than 6 for Type 3 fabrics). Finally, Type 4 ellipsoids show a very well-defined girdle distribution of their Kmin axes that attests the increased effect of deformation on the magnetic fabric. The evolution of the magnetic fabrics can be also examined by plotting in a modified Woodcock diagram the distribution of the Kmax axes (Fig. 5). The transition from Type 1 to Type 2 ellipsoids, which represents the first effect of deformation, is marked by an enhanced clustering of the Kmax axes and an increase in the strength of their distribution (C < 5 and K < 1 for most Type 1 sites; C > 5 and K > 2 for most Type 2 sites). The further increase in deformation that characterizes Types 3 and 4 ellipsoids seems to be accompanied

by a slight loss of Kmax clustering (K < 2 for most Type 3 and 4 sites). The four types of magnetic fabrics observed have been related by Pares et al. (1999) to the four stages of petrofabric development reported in mudrocks undergoing incipient deformation (Ramsay & Huber 1983). The first type of magnetic ellipsoid represents an 'undeformed condition' where the tectonic overprint cannot overcome the primary sedimentary fabric. As tectonic strain increases, the progressive alignment of Kmax along the tectonic Y axis (Type 2) and the incipient dispersion of Kmin in a girdle normal to Y (Type 3) reflect, respectively, the 'earliest deformation' and the 'pencil structure' stages of petrofabric development. Further increase in tectonic strain causes the girdle formed by Kmin to enhance, corresponding to the 'weak cleavage' stage. We notice that the different fabrics of the studied marls do not exactly follow the correspondence found by Pares et al. (1999) between the magnetic ellipsoids and the correspondent petrofabrics (Fig. 4). Weak cleavage has not been observed in the studied area. Moreover, not all sites displaying pencil structures lie on the prolate field of the diagram and several sites falling in the 'early deformation' field of the diagram develop pencil structures.

Palaeo- and rock-magnetic data Palaeo- and rock magnetic results from the Eocene marls of the Jaca-Pamplona basin indicate the presence of a dual polarity, characteristic remanence (ChRM) that is carried by magnetite and, to lesser extent, by magnetic iron sulphides (most probably pyrrhotite) (Larrasoana et al. 2003a, b). The number of samples for which the remanence carried by magnetic iron sulphides can be accurately determined is low (<10% of the samples) and does not allow the application of field tests for checking the time of magnetic iron sulphide formation. However, the directions of the remanence carried by magnetic iron sulphides are reliable and consistent with those carried by magnetite when both minerals appear together, which indicates that magnetic iron sulphides and magnetite acquired the remanence roughly at the same time (Larrasoana et al. 20036). Incremental fold tests results indicate that the ChRM was acquired before folding at the 95% of confidence level. However, a closer examination of the fold tests indicates that the optimum grouping of directions systematically occurs before complete unfolding of the beds (Fig. 6). It is well known that, during the

MAGNETIC FABRIC ACQUISITION IN MUDROCKS

137

Fig. 5. Comparison of the different types of magnetic fabrics with a Woodcock-type diagram describing the distribution of the Kmax axes. The examples shown and the symbols used are as in Fig. 4.

application of the incremental fold test, a prefolding remanence might appear to be synfolding in origin if mineral grains, including the magnetic carriers, rotate as rigid markers in response to deformation (Van der Pluijm 1987; Kodama 1988; Stamatakos & Kodama 1991). Magnetic grains rotate in opposite directions at each limb of a fold in response to different senses of shear, in such a way that the remanence becomes shallower and steeper depending on the structural position (Fig. 7a). The ChRM of the Eocene marls in the studied area has a mean declination of 5-10° and it is therefore broadly perpendicular to the main WNW-ESE structural trend of the basin (Larrasoana et al. 20030). Thus, the studied marls offer an unusual natural opportunity to evaluate the effect of deformation upon the directional properties of a ChRM. In Figure 7b, the site-mean inclinations of the ChRM from the Eocene marls are plotted as a function of dipping. It is apparent that sites displaying SSW and NNW dipping tend to have remanence directions steeper and shallower than the expected reference, respectively. In order to increase the statistical significance of the data, we have

calculated the mean inclination for 30°-wide windows centred at -60, -30, 0, 30, 60, 90 and 120 (overturned) degrees of bedding tilt. The new trend obtained is statistically robust and very similar to that shown in Figure 7b, and is therefore consistent with the expected sense of shear deformation undergone at each limb of the basin. Thus, our data suggest that the apparent synfolding origin of the ChRM in the Eocene marls is caused by the effect of deformation upon the orientation of the remanence directions. This effect must be very subtle, in any case, because the rotation of the ChRM is not large enough to give statistically significant synfolding results in the incremental fold tests. Discussion Mechanisms of magnetic fabric acquisition and remanence rotation The orientation of the susceptibility axes in clayrich weakly deformed mudrocks is produced by the spatial arrangement of the basal plane of

138

J. C. LARRASOANA ET AL.

Fig. 6. Incremental fold tests performed for different structural units. The Pamplona syncline includes data from the Ezkaba and Anezkar units, where the low number of palaeomagnetic sites prevents the application of individual tests. The fold test compares the precision parameter (k) upon folding (k2) with the initial value (kj) (McElhinny 1964).

platy phyllosilicates that constitute the matrix of the rock. According to Pares et al (1999), the underlying mechanical model is a rigid-body rotation in which the platy elements become systematically reoriented during layer parallel shortening when the sediments are soft and only partially lithified. Lateral shortening decreases the pore space, which results in an increase of pore-fluid pressure that reduces the normal stress between phyllosilicate grains and therefore favours particulate flow. The lack of penetrative deformation and metamorphism in the JacaPamplona basin indicates that the pressuretemperature conditions in the basin were limited to diagenesis. Thus, we consider that the mechanism proposed by Pares et al. (1999) to explain the evolution of magnetic fabrics in the mudrocks of the Southern Pyrenean Basin can be applied to our study area. A comparison between the

distribution of the Kmin and Kmax axes (Figs 4 & 5) allows an accurate determination of the evolution of the spatial arrangement of the phyllosilicate grains upon progressive deformation. The transition from Type 1 to Type 2 ellipsoids is characterized by the enhanced clustering of the Kmax axes perpendicular to the shortening direction and the stability of the distribution of the Kmin axes, which remain clustered around the bedding poles. This indicates that the phyllosilicate grains remain included in the bedding plane and rotate around the X tectonic axis. The slight increase in the strength of the distribution of the Kmin axes observed for Type 2 ellipsoids might reflect the effect of increased sedimentary compaction favoured by the rearrangement of the phyllosilicate grains under lateral shortening. This compaction might result in an enhanced bedding fissility. The transition from

MAGNETIC FABRIC ACQUISITION IN MUDROCKS

139

Fig. 7. (a) Schematic representation of a remanence affected by shear at opposite limbs of a fold. The observed remanence (white arrows) is steeper and shallower than the expected reference direction (black arrow) in the S and N dipping limbs, respectively. The best grouping of the directions upon stepwise application of the fold test is attained before the beds are completely restored. The 100% unfolding of the beds result in an overcorrection of the remanence. (b) Site-mean inclinations of the ChRM plotted as a function of dipping. Positive (negative) values indicate dipping to the south (north), and dipping values exceeding 90° indicate overturned beds. Reversed magnetizations have been converted to their antipodal direction and are indicated by open circles. Sites with a95 values larger than 15° have been not considered for constructing the plot. Vertical error bars correspond to the a95 angle. The expected Eocene inclination (calculated from Taberner et al. 1999) has been plotted for comparison with our directional results. Sites dipping SW (NW) tend to have inclinations steeper (shallower) than the expected reference. This trend is indicated by the dashed line and is accompanied by the statistics of the best fit. (c) Inclinations of the ChRM directions shown before, averaged for 30°-wide windows and plotted as a function of dipping. Vertical error bars correspond to the standard error of the mean inclination value obtained for each window, whereas horizontal bars indicate the width of the windows. The new trend fitted for the data is very similar to that obtained before but is statistically more robust.

140

J. C. LARRASOANA ET AL.

Type 2 to Types 3 and 4 ellipsoids is characterized by the girdling of the Kmin axes around the Y tectonic axis and by a simultaneous loss in the clustering of the Kmax axes. At this stage, the relative proportion of phyllosilicate grains rotated around the Y tectonic axis determines the appearance of a second fissility, which results in the formation of pencil structures and, eventually, of weak cleavage (Reks & Gray 1982). We recall that weak cleavage is not observed in the studied marls, even in those sites showing Type 4 fabric where such petrofabric could be expected. Moreover, not all sites displaying pencil structures lie on the prolate field of the modified Woodcock diagram and vice versa (Fig. 4). The formation of pencil structures, and likely of weak cleavage, depends on several factors such as initial rock composition, grain size, weathering and exposure conditions (Reks & Gray 1982). We consider that, apart from these factors, the close interaction between tectonic (e.g. lateral shortening) and early diagenetic processes (e.g. compaction) might also be responsible for the slight discrepancies observed between our results and the expected petrofabrics reported by Pares et al. (1999). Concerning the mechanisms responsible for the rotation of the ChRM, the incremental fold test results and the directional properties of the ChRM suggest that the remanence of the Eocene marls has been slightly affected by simple shear during incipient limb rotation as a result of flexural flow. Keeping in mind the tectonic history of the area, such incipient limb rotation would correspond to the Mid-Late Eocene synsedimentary warping reported by Leon-Chirinos (1985) and (Baceta et al, 2000). If the beds are restored back to the position in which the grouping of the ChRM directions becomes largest, that is, at the optimum percentage of unfolding determined from the incremental fold tests, the tilt of the beds at which the ChRM was effectively locked in can be determined. Following this reasoning, it can be estimated that in most of the sites the ChRM was locked in when the beds were tilted less than 5° (mean value of 4.5° ± 1° standard error), e.g. at the early stages of Mid-Late Eocene warping. This value is consistent with the limb rotation, that is, the shear deformation, required to rotate the ChRM from the expected direction according to the trend depicted in Figure 7 (e.g. <5° for near vertical sites). Given the very low concentration of magnetic grains in the marls, we interpret that their rotation was necessarily caused by intergranular slip conditioned by the impingement of the much larger phyllosilicate grains. This requires that

the phyllosilicate grains must also have been affected by shear deformation during initial Mid-Late Eocene warping.

Timing the blocking of magnetic fabric acquisition Based on the available data, a relative dating between the blocking of the magnetic fabric, the lock-in of the remanence and deformation might be attempted (Fig. 8a). We have seen that the blocking of the magnetic fabrics occurred mainly when the sediments were still soft and subjected to layer parallel shortening. In addition, some degree of phyllosilicate grain rotation must have occurred also during initial Mid-Late Eocene warping in order to allow rotation of magnetic grains and deflection of the ChRM from the expected reference inclination. At this point, two possible interpretations arise: (1) the deflection of the ChRM is very subtle because the magnetic grains were reoriented with the ambient field after being affected by rotation and impingement of the phyllosilicate grains. This would imply that the lock-in of the remanence post-dates the blocking of the magnetic fabric, and also that the fabric of the phyllosilicates records a higher degree of tectonic overprint than the fabric recorded by the ferromagnetic (sensu lato) grains; or (2) the deflection of the ChRM is very subtle because deformation was very weak and only conditioned minor reorientation of phyllosilicate grains. This would imply that rearrangement of phyllosilicate grains and deflection of the ChRM occurred simultaneously, and also that the fabric of the phyllosilicates and the ferromagnetic grains are recording a similar degree of tectonic overprint. Although we speculate that the first possibility might apply to the studied marls, we have no experimental data to constrain the fabric of the ferromagnetic (sensu lato) grains (e.g. the anisotropy of the ARM) in order to compare it with the fabric of the phyllosilicates. In the absence of such data, we consider that the second possibility is more likely because the imprint of tectonic deformation in the magnetic fabrics is very subtle, e.g., the very weak girdling of kmin axes for most sites indicates only minor reorientation of the phyllosilicate grains around the fold axes, and therefore a limited rotation of the magnetic grains. Based on the above, we propose that the lockin of the ChRM and the blocking of the magnetic fabrics occurred simultaneously in the time span going from sediment deposition to initial sedimentary warping. It appears that the

MAGNETIC FABRIC ACQUISITION IN MUDROCKS

141

Fig. 8. (a) Relative chronology between deposition, blocking of the magnetic fabrics, remanence lock-in and deformation, (b) Estimated absolute time constraint for the processes involved.

blocking of the magnetic fabrics can be potentially dated in absolute terms if some independent time constraint is available for the other two processes mentioned (Fig. 8b). Concerning the age of initial warping in the basin, its 'synsedimentary' nature (Baceta et al. 2000) would bracket the blocking of the magnetic fabrics to an undetermined, but certainly short time span after sediment deposition. Concerning the lockin time of the ChRM, we need to recall the discussion about its origin in the studied marls. Diagenetic pyrrhotite, and in general diagenetic magnetic iron sulphides, can be formed at different times ranging from the early (Snowball & Thomson 1988; Roberts & Turner 1993) to the late (Florindo & Sagnotti 1995; Weaver et al. 2002) diagenesis. In our case, the absence of field tests prevents the determination of the timing of magnetic iron sulphide formation. To do so, we have to rely qualitatively on the available palaeomagnetic data, and these data suggest that magnetic iron sulphides are recording reliable palaeomagnetic directions comparable to that carried by detrital magnetite (Larrasoana et al. 20036). Thus, magnetic iron sulphides seem to have formed shortly after deposition, which is consistent with the early diagenetic origin of the pyrite in the marls according to its texture and mode of occurrence (Larrasoana et al. 2003&). Formation of early diagenetic pyrite is restricted

to the surface metre of marine sediments (Canfield & Berner 1987). It is therefore likely that magnetic iron sulphides, which grow as precursors in the pyritization process, will form during early diagenesis within a few (< 3) thousand years after deposition for sediments accumulated at rates of 0.3m/ka (see Roberts & Turner, 1993). The mean accumulation rate of the Eocene marls in the eastern sector of the Jaca-Pamplona basin ranges between 0.1 and 0.5m/ka at different positions of the stratigraphic pile (Pueyo et al. 2002). If we accept that magnetite and magnetic iron sulphides acquired their remanence simultaneously and that formation of early diagenetic magnetic iron sulphides occurred within the upper metre of the sediments, the delay between sedimentation and remanence lock-in might be estimated at between 2 and lOka. Although these two estimates based on the age of initial warping and the lock-in time of the ChRM only represent a rough approximation, they are mutually consistent and certainly bracket the blocking of the magnetic fabrics within few thousand (<15?) years after deposition. Once the AMS and the ChRM were effectively blocked in the marls, subsequent Mid-Oligocene-Early Miocene folding would have resulted in their passive tilting as the beds reached their final structural configuration.

142

J. C. LARRASOANA ET AL.

Conclusions

References

A combination of AMS, structural, palaeo- and rock-magnetic data has allowed us to establish a relative chronology between sediment deposition, the blocking of magnetic fabrics, the acquisition of the ChRM and deformation in the weakly deformed Eocene mudrocks of the JacaPamplona basin. The very weak tectonic overprint that is interpreted as having affected both the AMS and the directional properties of the ChRM suggests that the blocking of the magnetic fabrics and the lock-in of the ChRM occurred simultaneously during the initial, MidLate Eocene synsedimentary warping undergone by the basin. Knowledge of the approximate time when the ChRM was locked in the studied sediments provides an absolute time constraint to bracket the blocking of the magnetic fabrics, which has been estimated to be of the order of few (<15?) thousand years. These results fully confirm previous interpretations invoking an early origin for the magnetic fabrics (Sintubin 1994; Mattei et al 1995; Sagnotti et al 1998 and 1999; Pares et al 1999; Coutand et al 2001). Based on our findings, we propose that combined AMS and palaeomagnetic studies in weakly deformed mudrocks might have important applications in tectonic studies. The short time span during which magnetic fabrics are blocked under tectonic deformation validates their use for palaeostress determinations. Although this application of AMS data needs to be first tested against other methods, the widespread distribution of mudrocks might considerably improve our knowledge of the tectonic processes involved in the evolution of accretionary prisms and fold and thrust belts, especially if we consider that mudrocks are not prone to be extensively studied by standard methods such as micro-fault populations or calcite twins. Additionally, the sensitivity of AMS data to detect preferred orientation of phyllosilicates might provide an excellent tool for tracking preferential paths for the migration of geofluids (e.g. hydrocarbons, orogenic fluids) within sedimentary basins at the early stages of deformation.

AVERBUCH, O., FRIZON DE LAMOTTE, D. & KISSEL, C. 1992. Magnetic fabrics as a structural indicator of the deformation path within a fold-thrust structure: a test case from the Corbieres (NE Pyrenees, France). Journal of Structural Geology, 14, 461— 474. BACETA, J. L, ASTIBIA, H., CEARRETA, A., PEREDA-

We wish to thank the staff of the Institute for Rock Magnetism for their technical assistance, their help in the interpretation of results and their hospitality during the realization of the rock-magnetic measurements. We also thank M. Mattei and L. Sagnotti for their helpful review of the manuscript. The study was supported by two predoctoral grants (Navarra Government, J.C.L.; Ministerio de Educacion y Ciencia, E.L.M.), the DOES (Project BTE2002-04168) and the National Science Foundation (IRM Visiting Fellowship, J.C.L).

SUBERBIOLA, X., MURELAGA, X. & BADIOLA, A.

2000. The Upper Eocene south Pyrenean coastal deposits (Liedena Sandstone, Navarre): sedimentary facies, benthic foraminifera and avian ichnology. Facies, 42, 107-132. BORRADAILLE, G. J. 1981. Particulate flow of rocks and the formation of cleavage. Tectonophysics, 72, 305-321. BORRADAILE, G. J. & HENRY, B. 1997. Tectonic applications of magnetic susceptibility and its anisotropy. Earth Science Reviews, 42, 49-93. BORRADAILLE, G. J. & TARLING, D. H. 1981. The influence of deformation mechanisms on magnetic fabrics in weakly deformed rocks. Tectonophysics, 77, 151-168. CANFIELD, D. E. & BERNER, R. A. 1987. Dissolution and pyritization of magnetite in anoxic marine sediment. Geochimica et Cosmochimica Acta, 51, 645-659. CHOUKROUNE, P. 1992. Tectonic evolution of the Pyrenees. Annual Review of Earth and Planetary Sciences, 20, 143-158. COUTAND, L, COBBOLD, P. R., ET AL., 2001. Style and history of Andean deformation, Puna plateau, northwestern Argentina. Tectonics, 20, 210-234. FLORINDO, F. & SAGNOTTI, L. 1995. Paleomagnetism and rockmagnetism in the upper Pliocene Valle Rica (Rome, Italy) section. Geophysical Journal International, 123, 340-354. JEL!NEK. V. 1981. Characterization of the magnetic fabric of rocks. Tectonophysics, 79, 63-70. KANAMATSU, T., HERRERO-HERVERA, E. & TAIRA, A. 2001. Magnetic fabrics of soft-sediment folded strata within a neogene accretionaty complex, the Miura group, central Japan. Earth and Planetary Science Letters, 187, 333-343. KISSEL, C., BARRIER, E., LAJ, C. & LEE, T. Q. 1986. Magnetic fabric of 'undeformed' marine clays from compressional zones. Tectonophysics, 40, 287-308. KODAMA, K. P. 1988. Remanence rotation due to rock strain during folding and the stepwise application of the fold test. Journal of Geophysical Research, 93, 3357-3371. LARRASOANA, J. C. 2000. Estudio magnetotectonico de la transicion entre el Pirineo central y occidental; implicaciones estructurales y geodinamicas. Ph.D. thesis, University of Zaragoza, Spain. LARRASOANA, J. C., PARES, J. M., MILLAN, H., DEL VALLE, J. & PUEYO, E. L. 20030. Paleomagnetic, structural, and stratigraphic constraints on transverse fault kinematics during basin inversion: the Pamplona fault (Pyrenees, north Spain). Tectonics, 22 (6), 1071.

MAGNETIC FABRIC ACQUISITION IN MUDROCKS LARRASOANA, J. C., PARES, J. M. & PUEYO, E. L. 20036. Stable Eocene magnetization carried by magnetite and magnetic iron sulphides in marine marls (Pamplona-Arguis Formation, southern Pyrenees, N Spain). Studia Geophysica et Geodaetica, 47, 237-254. LEON-CHIRINOS, I. 1985. Etude sedimentologique et reconstitution de cadre geodinamique de la sedimentation detritique fini-Eocene/'Oligocene dans le bassin sud-Pyreneen entre Sangilesa et Pamplona. Ph.D. thesis, University of Pau, France. MATTEI, M., FUNICIELLO, R. & KISSEL, C. 1995. Paleomagnetic and structural evidence for Neogene block rotations in the Central Apennines, Italy. Journal of Geophysical Research, 100, 1786317883. MCELHINNY, M. V. 1964. Statistical significance of the fold test in paleomagnetism. Geophysical Journal of the Royal Astronomical Society, 8, 338-340. PARES, J. M. & DINARES-TURELL, J. 1993. Magnetic fabric in two sedimentary rock types from the Southern Pyrenees. Journal of Geomagnetism and Geoelectricity, 45, 193-205. PARES, J. M., VAN DER PLUIJM, B. A. & DINARESTURELL, J. 1999. Evolution of magnetic fabrics during incipient deformation of mudrocks (Pyrenees, Northern Spain). Tectonophysics, 307, 1-14. PAYROS, A. 1997. El Eoceno de la cuenca de Pamplona: estratigrafia, fades y evolucion paleogeogrdfica. Ph.D. thesis, University of. Pais Vasco, Spain. PUIGDEFABREGAS, C. 1975. La sedimentation molasica de la cuenca de Jaca. Pirineos, 104, 1-118. PUEYO, E. L., MILLAN, H. & Pocovi, A. 2002. Rotation velocity of a thrust: a paleomagnetic study in the External Sierras (Southern Pyrenees). Sedimentary Geology, 146, 191-208. RAMSAY, J. G. & HUBER, M. I. 1°*3. The Techniques of Modern Structural Geology, Volume 1: Strain Analysis. Academic Press, London. RECKS I. J. & GRAY, D. R. 1982. Pencil structures and strain in weakly deformed mudstone and siltstone. Journal of Structural Geology, 4, 161-176. ROBERTS, A. & TURNER, G. M. 1993. Diagenetic formation of ferrimagnetic iron sulfide minerals in rapidly deposited marine sediments, South Island, New Zealand. Earth and Planetary Science Letters, 115, 257-273.

143

SAGNOTTI, L. & SPERANZA, F. 1993. Magnetic fabric analysis of the Plio-Pleistocene clayey units of the Sant'Arcangelo basin, southern Italy. Physics of the Earth and Planetary Interiors, 77, 165-176. SAGNOTTI. L., SPERANZA, F., WINKLER, A., MATTEI, M. & FUNICIELLO, R. 1998. Magnetic fabric of clay sediments from the external northern Apennines (Italy). Physics of the Earth and Planetary Interiors, 105, 73-93. SAGNOTTI. L., WINKLER, A., ET AL., 1999. Magnetic anisotropy of Plio-Pleistocene sediments from the Adriatic margin of the northern Apennines (Italy): implications for the time-space evolution of the stress field. Tectonophysics, 311, 139-153. SINTUBIN. M. 1994. Clay fabrics in relation to the burial history of shales. Sedimentology, 41, 1161-1169. SNOWBALL, I. F. & THOMPSON, R. 1988. The occurrence of grigite in sediments from Loch Lomond. Journal of Quaternary Science, 3, 121-125. STAMATAKOS, J. & KODAMA, K. P. 1991. Flexural flow folding and the paleomagnetic fold test: an example of strain reorientation of remanence in the Mauch Chunk Formation. Tectonics, 10, 807-819. TABERNER, C., DINARES-TURELL, J., GIMENEZ, J. & DOCHERTY, C. 1999. Basin infill architecture and evolution from magnetoestratigraphic cross-basin correlations in the southeastern Pyrenean foreland basin. Geological Society of America Bulletin, 111, 2-21. TARLING, D. H. & HROUDA, F. 1993. The Magnetic Anisotropy of Rocks. Chapman & Hall, London. TEIXELL, A. 1996. The Anso transect of the southern Pyrenees: Basement and cover thrust geometries. Journal of the Geological Society, London, 153, 301-310. VAN DER PLUIJM, B. A. 1987. Grain-scale deformation and the fold test-evaluation of synfolding magnetization. Geophysical Research Letters, 14, 155-157. WEAVER, R., ROBERTS, A. P. & BARKER, A. J. 2002. A late diagenetic (syn-folding) magnetization carried by pyrrhotite: implications for paleomagnetic studies from magnetic ironsulphide-bearing sediments. Earth and Planetary Science Letters, 200, 371-386. WOODCOCK, N. H. 1977. Specification of fabric shapes using an eigenvalue method. Geological Society of America Bulletin, 88, 1231-1236.

This page intentionally left blank

Magnetic properties and magnetic fabrics of Pleistocene loess/palaeosol deposits along west-central Siberian transect and their palaeoclimatic implications GALINA G. MATASOVA & ALEXEY YU. KAZANSKY Institute of Geology of the Siberian Branch of the Russian Academy of Sciences, ac. Koptyug ave., 3, Novosibirsk, 630090, Russia (e-mail: [email protected]) Abstract: Rock-magnetic properties, in particular anisotropy of magnetic susceptibility (AMS) were investigated in detail for eight sections transecting southwestern and central parts of Siberia. The results obtained indicate that magnetic properties and magnetic fabrics of loess/palaeosol deposits in western and southwestern Siberia depend on superposition of two different mechanisms namely a pedogenic mechanism proposed for Chinese loess and a wind-vigour mechanism for Alaskan loess. The wind-vigour mechanism is predominant in loess deposits and this allows palaeowind directions during glacial periods to be determined. In palaeosols, the balance between both models strongly depends on the geographical position of the section and thus reflects the palaeoclimate. In western Siberia, palaeosols corresponding to OIS 3 have sedimentary magnetic fabric, while the magnetic fabric of palaeosols corresponding to OIS 5 is completely reworked by pedogenesis. Such differences indicate a warmer climate during OIS 5. In central Siberia, separated from the west by the Kuznetsk Ala-Tau mountain ridge, the magnetic properties and AMS of loess/palaeosol sequences agree with 'Alaskan' type of loess, suggesting a colder and windier climate during the Late Pleistocene. Therefore, the Siberian subaerial realm may be subdivided into two provinces based on the palaeoclimate conditions prevailing during the Late Pleistocene. These climatic provinces remain in the modern climate.

Anisotropy of magnetic susceptibility (AMS) and magnetic fabric of loess/palaeosol sequences have profitably been applied to solve a wide range of problems in Quaternary geology. Such problems deal for instance with determinations of the sediment source, distance and means of sediment transport, depositional conditions, prevailing palaeowind direction and degree and character of secondary reworking processes of primary aeolian deposits (Liu et al. 1988; Thistlewood & Sun 1991; Lagroix & Banerjee 2002). Shape and degree of anisotropy as well as other magnetic properties of aeolian deposits strongly depend on the palaeogeographical features of sedimentation and subsequent reworking. Thus the magnetic fabric itself cannot be completely understood in isolation from other magnetic properties (such as different types of magnetic susceptibility and remanence) and other processes affecting the loess since its initial deposition. According to the relative variation of magnetic susceptibility (MS) magnitudes between loess and palaeosol, most deposits can be subdivided into two basic types (or models) (Heller & Evans 1995; Maher 1998). The first so-called 'Chinese' or 'pedogenic' type demonstrates 3-4 times higher magnetic susceptibility values in palaeosol than in loess horizons due to the enhancement of magnetic minerals

(secondary pedogenic magnetite/maghemite formation) during warm and wet interglacial periods. The second 'Alaskan' or 'wind-vigour' type shows the opposite trend. Here, higher MS values are generally explained by the increased input of magnetic particles due to stronger winds during glacial periods. An additional criterion to distinguish these two types appears to be the frequency-dependence of magnetic susceptibility (FD). In 'Chinese' loess/palaeosol sections, the FD values in loess horizons are relatively low (1-3%, sometimes up to 5%) while in buried soils it increases up to 10-15% (Maher & Thompson 1991; Liu et al, 1994; Dearing et al. 1996) due to formation of ultra-fine grained biogenie and/or abiotic magnetic minerals during pedogenesis. In contrast, the FD-susceptibility does not exceed 2-3% at all in 'Alaskan' type loess and palaeosols, regardless of the lithology (Chlachula et al. 1998; Vlag et al. 1999; Matasova et al. 2001). The two mechanisms (wind-vigour and pedogenesis) are also responsible for the diversity of magnetic fabrics of loess/palaeosol deposits. In order to reveal the differences, we compiled data from different authors (Liu et al. 1988; Thistlewood & Sun 1991; Reinders & Hambach 1995; Jordanova et al. 1996; Zhu et al. 1998; Lu et al. 1998; Zhu et al. 1999; Pan et al. 2001; Lagroix

From: MARTIN-HERNANDEZ, F., LUNEBURG, C M . , AUBOURG, C. & JACKSON, M. (eds) 2004. Magnetic Fabric: Methods and Applications. Geological Society, London, Special Publications, 238, 145-173. 0305-8719/04/ $15.00 © The Geological Society of London 2004.

146

G. G. MATASOVA & A. YU. KAZANSKY

Table 1. AMS characteristics of some loess jpalaeosol deposits Region

Deposit

AMS degree AMS ellipsoid Kmax preferred K^n % shape orientation orientation

Reference

'Chinese'-type loess/palaeosol deposits Europe Belgium 2 < P' < 5 oblate loess palaeosol P ' < 2 oblate P' < 1.5 Ukraine loess oblate less oblate palaeosol P ' < 1 Germany loess P<3 oblate Czechia loess oblate P'<2 weakly oblate Bulgaria loess palaeosol P' < 1.5 chaotic

yes yes yes no no no no no

vertical vertical vertical vertical vertical vertical vertical chaotic

Asia China

vertical or Thistlewood et al. 1991; near vertical Guo et al. 2002 vertical Zhu et al 1999;

Hus 2003 Hus 2003 Hus 2003 Hus 2003 Reinders & Hambach 1995 Zhu et al. 2001 Jordanova et al. 1996 Jordanova et al. 1996

loess

h<2

oblate

yes

loess

P'<3

oblate

no

palaeosol h < 1.5

oblate

yes

palaeosol P < 3

oblate

no

vertical or Thistlewood et al. 1991 near vertical vertical Liu et al 1988;

'Alaskan'-type loess/palaeosol deposits North America Alaska loess 3 < P' < 7 oblate palaeosol 3 < P' < 7 oblate

yes yes

vertical vertical

Lagroix & Banerjee 2002 Lagroix & Banerjee 2002

Asia Central Siberia

yes yes

vertical vertical

Hus 2003 Hus 2003

China

loess 2 < P' < 5 oblate palaeosol 2 < P' < 4.5 oblate

Panetal. 2001; Liuetal. 1988

Zhuetal. 1999

Reworked loess/palaeosol deposits water lain redeposited China loess 3< P< 6

oblate

yes

near vertical Liu et al., 1988

reworked by slope process Alaska loess P' > 3 Alaska palaeosol P' > 3

oblate oblate

yes yes

inclined inclined

Lagroix & Banerjee 2002 Lagroix & Banerjee 2002

P = ^max/^min (Nagata 1961), h = (Km.dx - Kmin)/Kini (Rees 1966), for P' see Jelinek 1981.

& Banerjee 2002; Guo et al. 2002; Hus 2003), which are represented in Table 1. Mostly both 'Chinese' and 'Alaskan' types of deposits are characterized by typical sedimentary fabrics with vertical or near vertical Kmin orientation. In general, 'Chinese' type loess and palaeosols appear to be less anisotropic than the 'Alaskan' type ones. In Europe, the AMS degree decreases with the distance from the boundary of Pleistocene ice cover (see maps in Mangerud et al. 1999; Naslund et al. 2003) and magnetic fabrics become less expressed: oblateness is weaker and the predominant KmSLX orientation is lost. In Bulgarian palaeosols, the AMS axes are chaotically distributed due to pedogenic processes. The picture is more obscured on the Chinese Loess Plateau. The AMS degree is relatively low and

a preferred Kmax axis can present or absent both in loess and palaeosol. 'Alaskan' type of loess/ palaeosol deposits in Alaska and Central Siberia demonstrate a well-defined oblateness and a distinct orientation of maximal AMS axes in both loess and palaeosol deposits. The nature of the magnetic fabric of loess deposits is a widely debated topic. Some authors are of the opinion that AMS in loess/palaeosol deposits can reflect a primary magnetic fabric, being formed at the moment of deposition among and in association with other sedimentary material (Rolph et al. 1989; Lagroix & Banerjee 2002). Other authors explain the AMS fabric only in terms of secondary reworking processes, i.e. redeposition of loess material by water flow (alluvial, slope and proluvial environments), for

PROPERTIES AND FABRICS OF PLEISTOCENE LOESS/PALAEOSOL DEPOSITS

example in Liu et al. (1988), Derbyshire et al. (1988). It is documented that in the case of a primary magnetic fabric in loess, the preferred orientation of the AMS maximum axes (i.e. magnetic lineation) may correspond to the predominant wind direction (Lagroix & Banerjee 2002), and hence to the general transport direction of the sedimentary material (Thistlewood & Sun 1991). It has been concluded from numerous loess/ palaeosol data sets (magnetic susceptibility, MS; natural remanent magnetization, NRM; saturated isothermal remanent magnetization, SIRM; anhysteretic remanent magnetization, ARM) worldwide, that magnetic proxy data from loess/palaeosol sequences reflect palaeoclimatic changes and hence can be used for the characterization of palaeoenvironmental conditions on regional and global scales (Kukla et al. 1988; Beget et al 1990; Maher & Thompson 1992; Heller & Evans 1995). The interrelation between the magnetic signature of loess/palaeosol sections and climatic fluctuations is supported by a strong correlation between magnetic susceptibility records and marine oxygen-isotope (<518O) records. High £18O values during even-numbered oxygen isotope stages (OIS) (i.e. glacial periods) correlate with low MS values in 'Chinese' type and high values in 'Alaskan' type (Heller & Liu 1986; Beget et al. 1990). The main purpose of this study is to investigate the behaviour of magnetic properties and magnetic fabrics of loess/palaeosol sequences across Siberia in order to make regional palaeoenvironmental reconstructions of the Middle and Late Pleistocene. The studied sections are spread along a sub-latitude transect from the southern part of West Siberia to the eastern part of central Siberia. Study area Siberia, due to its vast area and geographical position, remains a poorly investigated region of the northern hemisphere in terms of Quaternary geology. This territory is characterized by a highly diverse relief with vast plains in the western and north-central regions and broad highlands and high mountain ridges in the south and east. The cover of subaerial deposits with interbedded buried soils is widespread over the non-glacial zone of western, southwestern and central Siberia and is described hereafter as the Siberian subaerial realm (Volkov 1971). Stratigraphy, colour, carbonate content, grain-size distribution and other

147

lithological parameters suggest a predominantly aeolian origin of the sediment sequences in the area. Subaerial deposits are genetically connected to specific landforms like crests, linear depressions, ridges, valley flats and steppe depressions. The subaerial cover in Siberia was formed under a dry and cold climate when valleys of second order rivers and both small and large lakes were dry. The accumulative relief is composed of loads transported from local sources. The global climatic changes during the Late Pleistocene with alternating dry or wet and cold or warm epochs are well recorded in Siberia, because the alternation of glacial and interglacial periods has resulted in sharp changes of the sedimentation environment. Hence, depositional and relief-forming processes have climatically dependent, cyclic characters. The Near-Ob' crest plain is located in the southwestern forest-steppe part of the Siberian subaerial realm. It is composed of large linear landforms (crests) and their origin is still hotly debated. The linear crests, separated by depressions, are aligned in a WSW-ENE orientation coinciding with the predominant modern and palaeowind directions (Moskvitin 1940; Baranov & Blinova 1969) in the region. The height of some crests can exceed 150m. The lower parts of the crests are comprised of alluvial and lake deposits, while the middle and top parts consist of loess and loess-like deposits with interbedded buried soils, partly reworked by bio-/cryoturbation and solifluction processes. The most southwestern section studied Belovo (52.6° N, 83.6° E) - is located on the central part of a crest cut by the Ob' river valley (Fig. 1). Further northward, the clear dissection of linear landforms is less pronounced. However, local river valleys retain an orientation parallel to the crests. Within this marginal part of the Near-Ob' crest plain we studied two sections: Lozhok (54.6° N, 83.3° E) and Mramorny (54.7° N, 83.4° E) (see Fig. 1). Both are located within the watersheds of second order rivers. Going northwards, we studied Toguchin section (55.2° N, 84.4° E) located on the foothills of the Salair mountain ridge, where lowlands are filled by loess and loess -like deposits (Fig. 1). The Kuznetsk depression is located east of the Salair mountain ridge and represents a slightly uplifted crested steppe and forest-steppe plain, which is surrounded by the Salair mountain ridge to the west and by the Kuznetsk Ala-Tau mountain ridge to the east and south. In the depression, we studied two sections: Bachat (54.5° N, 87.1° E) located within the river watershed in the central part of the depression and Novokuznetsk (53.8° N, 87.1°E) located on

148

G. G. MATASOVA & A. YU. KAZANSKY

Fig. 1. Geographic map showing the location of the sections studied (black dots), the transect (black line) and the climatic sub-regions of the study area according to Davitaya & Pastukh (1960). The climatic division is given by white lines, zone names by white signs. Studied sections: 1 - Belovo, 2 - Lozhok, 3 - Mramorny, 4 Toguchin, 5 - Bachat, 6 - Novokuznetsk, 7 - Kurtak, 8 - Tatyshev.

the right bank of the Tom' river in the southern part of the depression (Fig. 1). The Minusa depression is located further NE, across the Kuznetsk Ala-Tau mountain ridge. It is completely surrounded by mountain chains. The studied Kurtak section (55.1° N, 91.4° E) is located in the northern part of the Minusa depression along the left bank of the Yenisey river (Fig. 1). More northward, downstream of the Yenisey river, there is the Rybinsk depression. The most northeasterly section studied, Tatyshev (56.1°N, 92.9° E) is located along the left bank of the Yenisey river within the Rybinsk depression (Fig. 1). Present-day climate According to the regional climate classification of the former USSR (Davitaya & Pastukh 1960) the southwestern sections (Belovo, Lozhok, Mramorny, Toguchin, Bachat and Novokuznetsk) belong to climatic zone IILic,

which is characterized by low humidity (evaporation/precipitation ratio 1.0-3.0), moderately cold (from -13 to -32°C) winters and warm summers: ET° (the annual sum of the surface ground temperatures above +10°C) exceeds 2200 °C and even 2400 °C (in Belovo). The Kurtak section belongs to zone IIIsc, which is also characterized by low humidity and moderately cold winters. The summer here, however, is moderately warm with £2"° less than 2200 °C. Tatyshev falls on the boundary between zones IIIsD and II3. The difference between III3C and 11X30 i§ the thickness of the snow cover (Table 2). Zone II3 is characterized by normal humidity (evaporation/precipitation ratio 0.451.0), moderately cold winters and moderately warm summers. According to modern soil types, Dobrovolsky (1976) suggests that the locations of Belovo, Lozhok, Mramorny, Toguchin, Bachat and Novokuznetsk sections fall into a sharp continental climate region, while the locations of the Kurtak and Tatyshev sections fall into an extra-continental climate region.

Table 2. Some present-day climatic characteristics in localities of studied sections Section

Modern soil type

Environment

* Winter temperature at surface [°C] DJF

* Summer Average temperature T[°C\ at surface [°C] annual JJA

Precipitation [mm/yr] annual

* Potential evaporation [W/m3] annual

* Relative humidity at surface [%] annual

Snow cover Distance to [cm] annual permafrost [km]

Number of days with T > 20 °C

Belovo Mramorny, Lozhok Toguchin Bachat Novokuznetsk

chernozem leached chernozem meadow chernozem leached chernozem grey forest soil (greyzem) grey forest soil (greyzem) sod-podzolic soil

steppe forest steppe meadow steppe steppe meadow steppe

-13.7 -14.9 -15.0 -14.9 -15.0

16.5 16.7 16.4 15.3 14.9

+2.0 -0.3 120 +0.5 +0.4

220-320 450-500 450 460-500 500-550

111 120 113 94 92

88 82 80 88 90

25-30 50 >50 60 90

375 425 450 250 225

35^0 30 20-30 30 20

forest

-16.0

14.5

-1.3.

400

96

87

<50

75

10-20

forest

-17.4

14.4

-1.3

400-500

104

86

>50

25

10-20

Kurtak Tatyshev

*From monthly long-term mean values (1968-1996) Compiled from Baranov & Blinova (1969), Davitaya & Pastukh (1960), Gusthina (1979), Protsuk (1978), NOAA-CIRES Climate Diagnostics Centre internet database: http://www..cdc.noaa.gov/ index.html).

150

G. G. MATASOVA & A. YU. KAZANSKY

Some features of the present-day climate for the sections studied are given in Table 2. The climatic divisions are shown in Figure 1. Sampling and methodology Oriented samples were collected from all eight sections. The thickness of the investigated sections varies from 7 to 24m. All studied sections were cross-correlated using several stratigraphic methods: palaeopedology (morphological descriptions, micromorphological analysis, detailed study of organic matter) and biostratigraphy (small and large mammals) (Arkhipov et al 1997; Zykina 1999; Foronova 1999). Stratigraphic correlations and corresponding age estimations are strongly supported by absolute dating, which is available for all studied sections except Toguchin and Novokuznetsk. Belovo has been dated by radiocarbon (C14) (Zykina et al. 2000) and thermoluminescence (TL) methods (Arkhipov et al. 1997). Lozhok and Mramorny have been dated by C14 (Volkov & Zykina 1984). The M. primingenius bones in the Bachat section have been dated by C14 (Foronova 1999). Kurtak, being a famous archaeological site, has the most comprehensive dating with C14 (Svezhentsev et al. 1992) and TL methods (Zander et al. 2003), while Tatyshev has only a few radiocarbon dates (Yamskikh 1992) and TL dates (Frechen & Yamskikh 1999). Detailed geological descriptions of the studied sections and the correlation of loess/palaeosol horizons with marine oxygen-isotope stages (OIS) can be found in Volkov (1971), Zykina & Kim (1989), Yamskikh (1992); Arkhipov et al. (1997), Chlachula et al. (1997), Chlachula et al. (1998), Chlachula (1999); Zykina (1999), Zykina et al. (2000), Matasova et al. (2001); Matasova et al. (2002). The loess and palaeosol units have different local names assigned; however, the following designation is introduced for all sections: LI to L5 (for loess units corresponding to glacial OIS 2, 4, 6, 8, 10, respectively) and PCI to PC4 (for palaeosols (pedocomplex) corresponding to the interglacial OIS 3, 5, 7, 9, respectively). PCO denotes the modern soil. Although our nomenclature is similar to that of Chinese loess/palaeosol sections, ours should not be correlated with the nomenclature of the Chinese Loess Plateau. During our field studies, we observed the rudiments of another palaeosol, hereafter called LI PC, within loess LI at Mramorny (Fig. 2). LI PC was not observed at all the other sections. Rock-magnetic data (Fig. 2), however, indicate its presence at Belovo, Toguchin, Bachat,

Novokuznetsk and Tatyshev. LI PC is mostly known from other sections located to the northwest of the Siberian realm. According to Volkov & Zykina (1984), LI PC has an absolute date of 14±2ka. Most samples were taken as oriented blocks 5 x 5 x 10cm. The blocks were cut in four to six oriented standard samples (2 x 2 x 2) at the laboratory. The sampling intervals range between 3 and 5cm. More than 2000 samples have been analysed in total. At some sections the loess was partly too loose and coarse grained (Kurtak section from 15.55 to 17.7m and Tatyshev section from 12.5 to 12.9m and from 14.2 to 15.5m) to be sampled in blocks. A piston sampler and plastic boxes have been used in this case. As this sampling technique may induce a secondary magnetic fabric (Jordanova et al. 1996), the samples from the above mentioned intervals were not used for AMS studies. Bulk magnetic low-field susceptibility was measured using a dual-frequency sensor (0.47 and 4.7kHz) from Bartington Instruments. The frequency dependence of magnetic susceptibility (FD), determined as the difference between lowand high-frequency susceptibility (KLP — KHP) and expressed as a percentage of the low-frequency susceptibility, has been calculated. The anisotropy of magnetic susceptibility was measured on a Geofyzika KLY-3 Kappabridge. The three principal axes defining the AMS ellipsoid were determined from the 15position orientation scheme suggested by Jelinek (1977). All anisotropy parameters (P' - corrected anisotropy degree; L - magnetic lineation; F magnetic foliation; T - shape factor) were calculated according to Jelinek (1981) and Tarling & Hrouda (1993). Various techniques were used to characterize the magnetic mineralogy of representative samples and magnetic extractions. The temperature dependence of magnetic susceptibility from room temperature to 700 °C was determined with the KLY-3 and CS-3 furnace in air. The thermal demagnetization of SIRM, acquired in a 1.4T direct current (d.c.) field, was also used to further the magnetic mineral characterization. The samples were heated from room temperature to 700 °C using a hand-made spinner magnetometer with a built-in shielded furnace (Burakov 1977). In order to assess mineralogical changes during the heating process, a second measurement run (SIRM acquisition and its thermal demagnetization) was conducted on the same sample. The NRM behaviour during step wise AF demagnetization up to lOOmT and thermal demagnetization up to 700 °C was also analysed.

Fig. 2. Lithologies and variations of magnetic susceptibility (K) and frequency-dependent susceptibility (FD) for studied sections. Lithology subdivision for Belovo, Lozhok, Mramorny, Bachat, Novokuznetsk and Kurtak sections is given according to Zykina et al. (1981), Zykina et al. (2000), Matasova et al. (2001), Matasova et al (2002) and Matasova et al. (2003). Lithology subdivision for Toguchin section is given according to field description by V. S. Zykina, for Tatyshev section - according to field description by D. Koz'min.

Fig. 2. (cont.)

PROPERTIES AND FABRICS OF PLEISTOCENE LOESS/PALAEOSOL DEPOSITS

153

Table 3. Mean magnetic susceptibilities and corresponding standard deviations for loess (LI and L2) and palaeosol (PCI and PC2) units of studies sections Section Belovo Lozhok Mramorny Toguchin Bachat Novokuznetsk Kurtak Tatyshev Belovo Lozhok Mramorny Toguchin Bachat Novokuznetsk Kurtak Tatyshev

Interval

LI 1.8-4.13 0.8-2.3 0.8-2.0 1.4-2.0 0.9-1.7 0.8-2.1 1.0-7.8 1.7-3.8 L2 6.7-7.7 4.0-4.7 5.3-6.8

2.4-3.4

4.0-5.8 3.9-6.6 11.0-18.3 7.5-8.7

N

•*vmean

if

SD

137 102 42 33 50 86 282 110

159 74 84 53 91 62 303 299

29 14 12 13 12 24 73 36

66 42 82 53 98 56 257 80

166 123 79 65 54 106 322 222

27 24 8 14 7 42 78 30

Interval

PCI 4.2-5.3 3.0-3.5 3.2-3.7 1.9-2.1 2.6-3.6 2.9-3.4 8.0-10.2 5.9-7.0 PC2 7.9-8.9 5.6-6.2 3.4-4.0 6.0-6.8 6.7-7.0 21.1-21.9 9.8-10.8

N

i^ -^mean

SD

34 30 41 10 60 32 176 72

51 73 40 29 55 71 152 164

9 11 4 5 8 11 31 15

66 36 38 46 21 42 58

40 42 34 41 55 193 125

10 4 3 10 7 28 17

Interval - the stratigraphic interval in metres within which susceptibility means are calculated; N - number of samples within the interval; Kmean - mean value of magnetic susceptibility in 10~5 SI units; SD - standard deviation in 10~5 SI units. Each interval for susceptibility means calculation is chosen on the basis of the following principles: (1) Deposits within the interval must be the typical loess (palaeosol) with the pristine fabric, homogenous in colour and granulometry; (2) Sandy layers, gleyed layers and layers of coarse material are excluded; (3) Illuvial horizons of buried soils are excluded; (4) Intervals of expected rudiments of LI PC (according to rock-magnetic data) are excluded.

The S-ratio (IRM_ 03r : SIRM! AT ) has been used for a first-order characterization of the sample's coercivity of remanence (i.e. relative amounts of low-coercivity remanence to highcoercivity remanence) within different stratigraphic units. The magnetic measurements were performed at three different laboratories - the Institute of Geology of the Siberian Branch of the Russian Academy of Sciences (Novosibirsk), the Geophysical Institute of the Czech Academy of Sciences (Prague) and the Institute of Geology and Geophysics of the Chinese Academy of Sciences (Beijing). Results

Magnetic susceptibility (K), frequencydependent susceptibility (FD) The variations of the volume magnetic susceptibility (K) and frequency-dependent susceptibility (FD) for all sections, as a function of depth, are plotted on Figure 2. Near Ob' crest plain The most characteristic picture is found in the section Belovo in the southwestern part of the

plain where the loess units are characterized by higher values of K and palaeosols demonstrate low susceptibility values (Fig. 2). The differences in average K values for LI, L2 PCI and PC2 units at Belovo are significant compared to their standard deviation (Table 3). Typical FDsusceptibility variations are observed. Increased FD values (up to 9%) correspond to palaeosol horizons while in loess units they do not exceed 3% (except in the upper part of L2 and in the upper part of L4). A more complicated picture is found in the central part of the studied area. Loess and palaeosol horizons demonstrate both low and high K values (Fig. 2) in the northeastern part of the Near Ob' crest plain (sections Mramorny, Lozhok and Toguchin). The FD values, in contrast, indicate a more distinct regularity: enhanced FD values in palaeosols (up to 8%) and low values in loess (<3%). Kuznetsk depression The susceptibility behaviour at Bachat and Novokuznetsk is similar to Mramorny, Lozhok and Toguchin, except PCI, which has almost constant values over the whole depth range. Enhanced FD values up to 5% correspond to palaeosol horizons and LI PC rudiments in both sections, while loess units are characterized

154

G. G. MATASOVA & A. YU. KAZANSKY

by FD values less than 3%, except the middle part of L2 loess in Novokuznetsk section (Fig. 2). Central Siberia The susceptibility variations at Kurtak and Tatyshev are similar to that in Belovo: loess units are characterized by higher values of K and palaeosols demonstrate low susceptibility values. The differences in K means of loess and palaeosol units in those sections are significant compared

to the standard deviation (Table 3). The variations of the FD-susceptibility are much smaller (<2%) in Kurtak and Tatyshev than in all other sections and do not depend on lithology (Fig. 2). Magnetic susceptibility variations along the transect Variations of K averages for different stratigraphic units of the Late Pleistocene (LI, L2,

Fig. 3. Generalized relief profiles and changes in average values of magnetic susceptibility ma maximal FD values for LI, L2, PCI and PC2 units along the transect. Numbers of sections correspond to Figure 1. Abbreviations: ESR - East Sayan mountain ridge, RD - Rybinsk depression.

PROPERTIES AND FABRICS OF PLEISTOCENE LOESS/PALAEOSOL DEPOSITS

PCI, PC2) along the transect are plotted in Figure 3. It is evident that the magnetic susceptibility of loess/palaeosol deposits exhibits some trend: gradual decrease in K means of stratigraphic units downwind in a N-NE direction along the Near-Ob' crest plain up to Salair mountain ridge (sections Belovo, Mramorny, Lozhok and Toguchin). This tendency is more pronounced in loess than in palaeosols (Fig. 3). Behind the ridge, in Kuznetsk depression the K means increase slightly in almost all units. A sharp increase of K means (about 3 times) both in loess and palaeosols is observed behind Kuznetsk Ala-Tau mountain ridge in Central Siberia (Kurtak). Further northeastwards (Tatyshev), the K means decrease in the upper units (LI, PCI) and drop slightly or remain nearly constant in lower units (L2, PC2). It should be mentioned that palaeosol susceptibilities in central Siberia are relatively high. These values are comparable to loess susceptibilities from Belovo, only. The FD value is about 2% in loess LI and L2 and does not show much variation along the transect. The FD value of the palaeosols, in contrast, is variable. The highest values of about 9% are observed in the Near Ob' crest plane and intermediate values of about 5% in the Kuznetsk depression. Very low values (~2%) - comparable to loess - are found in the Minusa and Rubynsk depression, behind the Kuznetsk Ala-Tau mountain ridge.

Magnetic mineralogy More than 500 samples of Siberian loess/palaeosol deposits were studied to analyse magnetic mineral composition. Some results have already been published in Zhu et al. (2000), Matasova et al. (2001), Zhu et al (2003) and Matasova et al. (2003), so we shall give the common features with greater attention on unpublished data. Near Ob' crest plain At Belovo, the temperature dependence of magnetic susceptibility (TDS) for loess and palaeosol (Fig. 4a) are similar in appearance. The heating branch exhibits a small peak at about 300 °C and the distinct decrease of susceptibility near 580 °C (loess) and 620 °C (palaeosol). Some palaeosol samples display an additional susceptibility increase between 400 °C and 525 °C. The cooling and heating branches are equal between 575 and 700 °C. Below 575 °C, a considerable susceptibility increase is observed during cooling. The maximal susceptibility values on the cooling branch are observed at 500 °C for palaeosols and between 400 and 500 °C for loess samples.

155

Further cooling causes a gradual decrease of magnetic susceptibility and the final susceptibilities are 2-4 times higher than initial ones. TDS curves for samples from Mramorny (Fig. 4b) and Lozhok (not shown) sections are not significantly different from those from Belovo section. The discrepancy in TDS behaviour between samples of the Belovo, Mramorny and Lozhok sections in general is in the position of the maximal susceptibility peak on the cooling branch. The thermal SIRM demagnetization curves of loess and palaeosol samples from Belovo are quite similar in shape, but vary in the SIRM0 intensity: loess values are 2 times higher than palaeosol values. There is significant decrease of SIRM between 180 and 350 °C, which is not pronounced during the second heating (see below). About 3-5% of the initial SIRM is retained above 580 °C. The complete demagnetization of SIRM is observed at 630 and 680 °C for loess and palaeosols, respectively. In order to assess mineralogical changes during the heating process, a second heating was performed. The sample was remagnetized and again thermally demagnetized (dashed in Fig. 4a). The initial SIRM values of the second heating curve are 30 to 40% lower than the initial values of the first heating. The second heating curve differs considerably from the first one and does not exhibit the significant SIRM decrease between 180 and 350 °C. Above 580 °C both curves are similar and the complete demagnetization of SIRM still remains at 630/680 °C. The thermal demagnetization of NRM of samples from Belovo shows a strong intensity loss between 200 and 300 °C, which is more pronounced in loess than in palaeosols. However, the NRM decay is different. In loess, 75% of NRM is demagnetized by 250-300 °C, while in palaeosols 25% of NRM remains at 620 °C. The complete demagnetization of NRM is observed at 680 °C in both cases. Alternating field (AF) demagnetization of NRM exhibits different behaviour in sections over the Near Ob' crest plain (Fig. 4a-c). In all cases, AF peak fields of 100 mT are not sufficient to demagnetize the NRM completely. Loess samples from Belovo have low NRM 0 intensities. The NRM decays rapidly with median destructive fields (MDF) between 10 and 20 mT. Ten per cent of the NRM remains at lOOmT. Palaeosol samples from the Belovo have MDFs of more than 65 mT (in individual soil samples more than 100 mT). About 40% of initial NRM remains at lOOmT. Loess samples from Mramorny (Fig. 4b) show a moderate NRM decay with MDFs about 30 mT and about 20% of residual NRM at lOOmT. NRM decay in palaeosol samples from

156

G. G. MATASOVA & A. YU. KAZANSKY

Fig. 4. Temperature dependence of magnetic low-field susceptibility (K), thermal demagnetization of SIRM, thermal and AF demagnetization of NRM of representative Siberian loess and palaeosol samples from the sections at: (a) Belovo, (b) Mramorny, (c) Toguchin, (d) Bachat, (f) Novokuznetsk, (g) Kurtak and (i) Tatyshev. XRD spectra are given of magnetic extracts from Bachat (e) and Kurtak (h). The solid (dashed) lines in K vs. T graphs represent heating (cooling) branch. Solid and dashed lines in SIRM-graphs represent first and second heating run, respectively.

the Mramorny section has a linear character with MDFs more than 70 mT. About 45% of initial NRM remains after 100 mT. Loess and palaeosol samples from Toguchin (Fig. 4c) have similar MDFs being 22 and 19mT, respectively. The shape of the decay curves is similar and about 25% of the initial NRM remains at lOOmT. Variations of S ratio versus depth for all sections studied within the crest plain are represented in Figure 5. In loess horizons from

Belovo, the S ratio varies between 0.9 and 1.0 and falls slightly below 0.9 only for some samples from LI, L4 and L5. Sharp decreases of S ratio up to 0.74-0.85 are attributed to humus horizons of palaeosols PCI, PC2 and PC4. A similar behaviour is observed at Mramorny. At Toguchin, the S ratio scatters between 0.8 and 1 in LI, PCI and L2. Below L2 the values are less scattered, being around 0.9. Distinct decreases down to 0.8 and 0.77 are observed in the

PROPERTIES AND FABRICS OF PLEISTOCENE LOESS/PALAEOSOL DEPOSITS

157

Fig. 4. (cont.)

humus horizons of PC2 and PC3, respectively. At Lozhok, the S ratio is about 0.95 over the entire section, except PC2 where it falls slightly below 0.9. X-ray diffraction (XRD) analysis on the whole rock material from the Belovo section (not shown) indicates the presence of magnetite, maghemite and hematite for loess and for palaeosols.

Kuznetsk depression The TDS behaviour of loess and palaeosol samples from Bachat (Fig. 4d) is quite similar to those from Belovo. The heating branch exhibits a small peak at about 300 °C and a distinct decrease of susceptibility near 580 °C. Palaeosol samples demonstrate an additional susceptibility increase between 400 °C and 525 °C. The cooling

158

Fig. 4. (cont.)

G. G. MATASOVA & A. YU. KAZANSKY

PROPERTIES AND FABRICS OF PLEISTOCENE LOESS/PALAEOSOL DEPOSITS

Fig. 5. Variations of the S-ratio versus depth of the studied sections.

159

160

G. G. MATASOVA & A. YU. KAZANSKY

branch follows the heating one between 600 and 700 °C and then increases dramatically. Maximal susceptibility values on the cooling branch are observed at 400-500 QC in palaeosol and between 200 and 500 °C for loess samples. The final susceptibilities are 5-10 times higher than initial ones. The highest ratios between initial and final susceptibility values correspond to palaeosol horizons. The TDS behaviour for samples from the Novokuznetsk section (Fig. 4f) is not significantly different from those from Bachat. The same peak at 300 °C and the peak above 450 °C (in palaeosols) are observed here. The final susceptibility values are 7-8 times higher than initial ones. The thermal SIRM demagnetization curves for loess and palaeosol samples from Bachat are similar, but differ in their initial SIRM (lower in loess and high in palaeosol). Both loess and palaeosols are comparable in quality to the samples from Belovo (see above). However, the initial SIRM values at Bachat are 3-5 times lower than at Belovo. The SIRM of loess and palaeosols is demagnetized at 620/680 °C, respectively. The NRM intensity decay during thermal demagnetization is different for loess and for palaeosol samples from Bachat. In loess, the NRM drops quickly. Less than 25% of the initial NRM remains at temperatures between 250 and 300 °C. The demagnetization curve of the palaeosol samples has a nearly linear character. Approximately 10 to 15% of the initial NRM remains above 580 °C. The demagnetization of the NRM is completed at 680 °C. There is almost no difference in the AF demagnetization curve of the NRM between loess and palaeosols from Bachat. In both cases, the MDF is about 20-25 mT; 15% and 20% of the initial NRM remain at lOOmT in loess and palaeosol, respectively. Samples from the Novokuznetsk section are magnetically harder than in Bachat. The MDF of loess samples generally exceeds 30 mT and the residual NRM after 70mT is about 35% of its initial value. The MDF of the palaeosols is about 50-55 mT and 45% of NRM still remains after 70 mT AF demagnetization. The S ratios at Bachat are relatively low and generally less than 0.9 (Fig. 5). They vary between 0.85 and 0.95 regardless of the lithology of the section. Although lower S values (<0.83) are attributed to humus palaeosol horizons in the previously discussed sections, such low values are observed in a thin sandy layer in the middle part of LI and in some levels of L2. In contrast, small variations between 0.9 and 1.0 are observed at Novokuznetsk. Nevertheless,

here the lowest values (<0.93) are found in humus palaeosol horizons. XRD analysis of the magnetic extraction from Bachat (Fig. 4e) indicates the presence of magnetite and hematite in palaeosols, while in loess horizons maghemite is also present, in addition to magnetite and hematite. Central Siberia The TDS curves for loess samples from the Kurtak section (Fig. 4g) are characterized by high values compared to the sections discussed above. However, the distinct decrease of susceptibility near 580 °C remains a common feature for all samples. Two types of TDS behaviour are found for loess units. The first type (sample from depth 6.9m) shows a peak at about 300°C on the heating curve while after heating to 700 °C the susceptibility decreases and is about half of its initial value at room temperature. The second type (sample from depth 12.4m) exhibits a peak between 400 °C and 550 °C on the heating branch. While cooling down from 700 °C, the susceptibility increases greatly below 600 °C, reaching maximal values between 300 and 400 °C. Further cooling causes a gradual decrease and the final susceptibility at room temperature is 1.5-2.0 times higher than the initial values. The palaeosol samples from Kurtak have similar TDS characteristics to the other palaeosols described before. As in Bachat, the thermal demagnetization of SIRM in Kurtak does not show any significant differences between loess and palaeosol. Both loess and palaeosol differ in their initial SIRM values only. The SIRM vanishes around 680 °C. After remagnetization the SIRM is about 25% lower. There is no difference in the SIRM between different loess. The thermal demagnetization of NRM samples from the Kurtak section shows a considerable intensity loss between 180 and 300 °C, which is more pronounced in loess than in palaeosols. Above 580 °C, 1% of the NRM remains in the loess samples whereas in the palaeosol samples 5% remain. The NRM is completely demagnetized at 680 °C in both. Loess and palaeosols from Tatyshev (Fig. 4i) have a similar NRM intensity loss at 300 °C. At this temperature, the loess are demagnetized to 85% and the palaeosols to 75%. In the loess, the nearly complete NRM decay is observed at 600 °C. In the palaeosols, however, a small portion (<1%) of initial NRM is retained up to 700 °C, but such values are at the noise level of the instrument. AF demagnetization of NRM of samples from Kurtak and Tatyshev may reflect the presence of

PROPERTIES AND FABRICS OF PLEISTOCENE LOESS/PALAEOSOL DEPOSITS

low coercivity minerals. Loess samples demonstrate a rapid NRM decrease with MDFs less than 5mT for Kurtak and lOmT for Tatyshev. Between 5 and 10% of initial NRM remains at lOOmT. NRM decay in palaeosol samples from Kurtak has a nearly linear character with MDFs more than 60mT. About 40% of initial NRM remains at lOOmT AF. In palaeosol samples from Tatyshev, 65% of the NRM decays below lOmT (MDFs are less than 8mT). Above lOmT, the curve demonstrates a nearly linear character. About 15% of initial NRM remains at lOOmT. Variations of S ratio versus depth for the sections Kurtak and Tatyshev (Fig. 5) show a decrease in palaeosol horizons. In general, the S ratio varies between 0.9 and 1.0 in loess horizons of both sections. However, the ratio is rather constant at Kurtak than at Tatyshev. All palaeosols (except PCI in Tatyshev) demonstrate a distinct decrease of the S ratio below 0.9. The minimal S ratio values at Kurtak are 0.75 and at Tatyshev 0.85. XRD analysis of the magnetic extraction from the Kurtak section (Fig. 4h) indicates the presence of magnetite and hematite in both loess and palaeosols. However the relative peak intensities may indicate a higher magnetite concentration compared to Bachat. Anisotropy of magnetic susceptibility (AMS) Near Ob' crest plain The degree of anisotropy P' for samples from the Belovo section (0.6) is relatively high compared to the other sections of the transect. P' of individual loess samples can exceed 10%. In general, P' falls within the range of 1.02-1.08 for loess horizons, while palaeosols predominantly have P' values between 1.00 and 1.03, except for the middle part of PC2 palaeosol, where P' exceeds 1.06. The depth variation of the AMS degree correlates with the magnetic susceptibility (Figs 2 & 6), in contrast to all other sections studied (see below). The AMS ellipsoid shape in loess horizons (LI and L2) is predominantly oblate, more than 80% of samples have a shape factor of T > 0.6 (Fig. 6). The distribution of the principle susceptibility axes in loess demonstrates the typical sedimentary fabric, which is characterized by a minimum susceptibility axis in a nearly vertical direction and by a maximal axis quite close to the bedding plane. A slight magnetic lineation is also expressed. The maximal axes are well grouped in a SW-NE direction (Fig. 7). The mean

161

directions of the minimal and maximal susceptibility axes, including their corresponding confidence limits, are given in Table 4. The magnetic fabric of palaeosol horizons at Belovo shows a different behaviour. Palaeosol unit PCI has a similar magnetic fabric like the loess units, but the Kmax axes cluster along the SSE-NNW direction. Palaeosol unit PC2 shows a rather random distribution of principal AMS axes. The shape factor (T) varies between —1.0 and 1.0 for both palaeosols. The AMS degree in Lozhok (Fig. 6), Mramorny and Toguchin sections (latter two not shown) is lower than at Belovo and weakly correlates with susceptibility (Figs 2 & 6). P' ranges between 1.01 and 1.06 in loess horizons, while palaeosols demonstrate P' values between 1.00 and 1.03. Inclination variations of ATmax and Kmin show the same trend as in Belovo. The oblate shape of AMS ellipsoid is predominant in loess horizons, while palaeosol horizons demonstrate also nearly spherical and partly prolate shapes (Fig. 6). The maximal AMS axes are well grouped in loess horizons (Fig. 7). Unlike Belovo, the maximal AMS axes in LI at Lozhok and Mramorny demonstrate a bimodal distribution: NW-SE and NNE-SSW (Fig. 7). At Toguchin, in contrast, a unimodal distribution is observed. Loess L2 exhibits a bimodal character only at Lozhok, with preferred Kmax direction similar to LI. The preferred AMS Kmsix directions of L2 at Mramorny and Lozhok are similar to the corresponding mode in LI. Palaeosol unit PCI demonstrates a SSW-NNE direction at Mramorny and Toguchin as the preferred orientation of KmSLX, while at Lozhok no directional preference is seen (Fig. 7). Palaeosol unit PC2 has no preferred ^max orientation in all sections. Kuznetsk depression In contrast to the previously discussed sections, the AMS degree is not related to the susceptibility variations in both sections of Kuznetsk depression (Figs 2 & 6). Bachat demonstrates the lowest AMS degrees in the studied area. P' ranges between 1.01 and 1.04 in loess horizons, while in palaeosols P' does not exceed 1.02 (Fig. 6). The Km[n inclinations are quite scattered throughout the sections; however, nearly vertical inclinations, correspond to loess layers and horizontal inclinations to palaeosol units. The shape factors vary between -1.0 and +1.0 in both loess and palaeosols, and the oblateness of the magnetic fabric is less pronounced in loess units. The palaeosols at Bachat show a random distribution of minimal and maximal AMS axes (Fig. 7).

Fig. 6. Variations of AMS degree P', inclinations of minimal and maximal AMS axes versus depth and plots of shape parameter T and AMS degree P' of the studied sections.

Fig. 6. (cont.)

Fig. 6. (cont.)

165

PROPERTIES AND FABRICS OF PLEISTOCENE LOESS/PALAEOSOL DEPOSITS Table 4. AMS data for loess (LI and L2) and palaeosol (PCI and PC2) units of studied sections Horizon

N

Dmax(°)

Belovo section LI 22 242 22 PCI 168 235 L2 18 no preferred orientation PC2 Lozhok section LI 18 13 LI* 22 302 no preferred orientation PCI L2 14 24 L2* 15 263 PC2 no preferred orientation Mramorny section LI 19 223 LI* 8 300 11 29 PCI L2 25 296 Toguchin section 11 LI 56 PCI 6 26 L2 6 17 PC2 no preferred orientation Bachat section LI 20 235 PCI no preferred orientation 232 L2 17 PC2 no preferred orientation Novokuznetsk section 20 285 LI 19 185 LI* no preferred orientation PCI 24 L2 91 12 L2* 181 PC2 no preferred orientation Kurtak section 24 270 LI 272 11 PCI 11 182 PCI* 32 L2 237 PC2 7 25 17 PC2* 98 Tatyshev section LI 213 10 28 239 PCI L2 232 27 224 PC2 17

maxO

«95 (°)

6 2 5

29 8 21

3 6

DminO

A

mmv / T (°}

«95 (°)

58 80 8

86 85 83

4 6 12

9 9

81 100

82 84

6 5

3 2

12 20

94 116

85 87

4 5

5 2 2 1

11 15 11 12

52 39 148 12

88 86 88 80

6 6 3 4

1 1 5

21 24 29

274 165 58

89 82 87

9 9 26

6

22

358

85

7

5

20

27

85

15

2 1

10 3

111 80

84 85

4 5

2 1

12 15

246 155

85 88

4 4

5 7 3 5 1 2

15 21 23 9 13 22

88 76 52 66 348 304

87 86 85 84 88 86

3 6 5 5 5 5

3 4 6 13

38 10 11 15

131 114 29 60

88 88 84 83

12 3 4 5

I

N- number of samples; D and I - mean declination and inclination of the AMS axes, respectively; a95 - semi-angle of cone of confidence (P = 0.05); min and min indexes denote maximal and minimal axes. In case of two differently oriented assemblages of AMS direction in one unit, the second direction is given in the row just below, with the unit name marked by *. The loess/palaeosol deposits at Novokuznetsk are characterized by enhanced AMS degrees compared to those from Bachat. Especially in L2, P7 exceeds 1.09, but ranges between 1.01 and 1.04 in palaeosol units (Fig. 6). The scatter of Kmin and Kmax inclinations are much less than at Bachat and follow a similar trend as in

sections of the Near Ob' crest plain. The shape of the AMS ellipsoid is essentially oblate for loess units. However, values between -0.5 and 0 are also observed. The AMS ellipsoids in palaeosols have rather neutral shapes, only a few samples demonstrate oblateness with T values close to 1. Palaeosols PCI and PC2 in

Fig. 7. Stereoplots of Kmax (squares) and ATmin (dots) principal AMS axes for loess (LI, L2) units and palaeosols (PCI and PC2) for all studied sections. In sections Lozhok Mramorny, Novoreznetsk and Kurtak two different directions were found (marked by different colours). Mean directions are shown by stars and crosses and given in Table 4 with corresponding confidence limits.

PROPERTIES AND FABRICS OF PLEISTOCENE LOESS/PALAEOSOL DEPOSITS

both sections have no preferred Kmax orientations. The maximal susceptibility axes in LI and L2 at Bachat are grouped in SE-NW direction. The means are not statistically different from those of Belovo and Toguchin (Fig. 7). The loess units at Novokuznetsk have a bimodal distribution of preferred Kmax axes (Fig. 7) W-E and S-N, being similar to Lozhok. Central Siberia In the Minusa depression just behind the Kuznetsk Ala-Tau ridge, the AMS degree increases in comparison with the Kuznetsk depression. P' ranges between 1.02 and 1.10 and between 1.00 and 1.08 for loess at Kurtak and Tatyshev, respectively. Palaeosols demonstrate P' values between 1.01 and 1.08 in both sections. The stratigraphic distribution of Kmax and Kmin inclinations at Kurtak and Tatyshev is generally similar to Belovo. Steep Kmax inclinations, however, are quite rare (Fig. 6). Both loess and palaeosol horizons demonstrate a predominant oblate magnetic fabric with mean T values >0.5. The ATmax axes are mainly distributed along an E-W orientation in loess LI at Kurtak; L2 shows a NE-SW orientation. Both loess layers at Tatyshev show NE-SW orientation of the AMS maximum axes. The Kmax axes of the palaeosols units PCI and PC2 at Tatyshev have the same orientation as the loess. At Kurtak, in contrast, a bimodal distribution is observed. The two observed predominant directions (E-W and S-N) correspond to different parts of the palaeosol units. All mean Kmax directions of humus horizons coincide with those of the overlying loess units, while the means of Kmsix directions of illuvial horizons are in good agreement with those of the underlying loess. A similar observation has been made by Matasova et al. (2001). Discussion Magnetic mineralogy Magnetite appears to be the predominant magnetic mineral of Siberian loess/palaeosol deposits. It has clearly been identified by the temperature dependence of magnetic low-field susceptibility. Most of the loess and palaeosol samples demonstrate a considerable decrease around 580 °C (Fig. 4d,f,g,i). Magnetite has also been identified by X-ray diffraction (Fig. 4e, h). Due to its high magnetization (Dunlop & Ozdemir 1997), magnetite is believed to cause higher susceptibilities in loess layers, especially at Kurtak and Tatyshev (Fig. 2). Highest NRM

167

values were also observed at Kurtak and Tatyshev (Fig. 4g, i). The low FD-values in loess layers may indicate the absence of superparamagnetic grains, and the susceptibility of the loess is mainly caused by multidomain grains. Equidimensional single domain grains of magnetite, which are also responsible for high susceptibilities (Dunlop & Ozdemir 1997), are not expected to occur in loess. In palaeosols, however, superparamagnetic grains are present. Concerning the temperature behaviour, the susceptibility drops down above 580 °C in some samples (Belovo, Mramorny, Fig. 4a, b). The curves show a small bulge around 250-350 °C and thermal demagnetization of SIRM demonstrates a considerable loss of intensity at the same temperature range. The second heating run of the SIRM does not show such a feature. This behaviour indicates the presence of maghemite in the samples, which converts to hematite during heating (Dunlop & Ozdemir 1997). The presence of maghemite in samples from Belovo and Kurtak is also supported by (1) the reversible character of partial TDS heating/cooling paths below 300 °C (Zhu et al 2003; Matasova et al. 2003) before maghemite-hematite conversion, and their irreversible character after the conversion above 400 °C (Zhu et al 2003; Matasova et al 2003), and (2) the fact that TDS behaviour of magnetic extracts indicates the same susceptibility peak at 350 °C on the heating branch and a slight decrease in total susceptibility after cooling (Matasova et al 2003). Maghemite may be present as oxidation cover of magnetite grains in loess horizons and magnetite probably can be completely oxidized to maghemite in palaeosol units as the result of pedogenesis. Matasova et al (2001) and Zhu et al (2003) observed a suppression of the Verwey transition in samples from Kurtak and Bachat, being a indicator for weathering of magnetite grains (Dunlop & Ozdemir 1997). Alternating field demagnetization of NRM indicates the presence of high coercivity remanence-carrying minerals. Loess/palaeosol samples from the Near Ob' crest plain and the Kuznetsk depression still retain 10-50% of the NRM at 100 mT (Fig. 4a-d,f). Palaeosols from Kurtak and Tatyshev retain 25% and 10%, respectively (Fig. 4g,i). The S ratio represents the proportion of low-coercivity minerals to high-coercivity minerals. Figure 5 shows that lower values (about 0.8) are related to palaeosol units, whereas loess layers have rather constant values around 1. Due to the field applied of 1.4T, lower S ratios may be related to hematite. Thermal demagnetization of NRM and SIRM (Belovo, Fig. 4a) also suggests the presence of

168

G. G. MATASOVA & A. YU. KAZANSKY

hematite above the superparamagnetic grain-size limit. Hematite is clearly identified by XRD spectra (Fig. 4e, h). The enhanced occurrence of hematite in soils may be indicative of a rather dry climate (Maher 1986, 1998). The relative peak intensities of hematite are much smaller than for magnetite, suggesting a lower percentage of hematite being present. Evidence for remanence-carrying goethite has not been found. It may occur as superparamagnetic particles. As hematite is formed in competition with goethite (Schwertmann 1988), it may have existed at the time when the soil was formed but dehydrated later. As discussed, loess/palaeosol deposits of the Siberian aeolian realm contain three different magnetic minerals and have a similar qualitative magnetic composition. Susceptibility, FD- value and S ratio profiles, however, are different and show variations in dependence of the geographical position of the sections. Hence, the magnetic properties of the sediments are determined by relative concentration differences. It is very important that similar composition of magnetic minerals (magnetite, hematite and maghemite) and also found in the other loess/ palaeosol provinces such as Chinese loess plateau and Alaska (Florindo et al 1999; Guo et al. 2002; Heller & Evans 1995; Lagroix & Banerjee 2002). NRM decay curves during thermal demagnetization of Chinese palaeosols (Heller & Evans 1995; Pan et al. 2001; Guo et al. 2002) are very similar to Siberian palaeosols. Chinese palaeosols during AF demagnetization show MDF < lOmT, while for Siberian palaeosols MDF can exceed 50 and even 80 mT, indicating the greater concentration of high coercivity minerals (probably, hematite). This higher hematite concentration is supported by the difference in Her values of Chinese and Siberian deposits. The first are characterized by Her values from 20 to 50 mT (Florindo et al. 1999; Pan et al. 2001; Guo et al. 2002) and the second by Her of 60-80 mT (Zhu et al. 2003). It may appear that the enhanced hematite concentration in Siberian loess and palaeosols in comparison with Chinese ones is responsible for their higher magnetic anisotropy. However this is probably not the case, because loess/palaeosol deposits in Alaska are characterized by Her values 4055mT (Lagroix & Banerjee 2002) within the range of Chinese Her, while the AMS degree for loess/palaeosol deposits in Alaska more closely resembles the AMS degree of Siberian deposits. The difference in magnetic anisotropy of Chinese loess/palaeosol deposits on the one hand and Siberian and Alaskan deposits on the

other results not from a difference in magnetic composition or the proportion of hematite concentration but is most likely determined by different climatic and environmental conditions in those regions.

Palaeoclimatic reconstruction The magnetic signature of loess/palaeosol units in the Siberian subaerial realm is controlled by the superposition of two mechanisms: the windvigour mechanism and the pedogenic mechanism. Both depend strongly on local and regional climate. Near Ob' crest plain The most distinct picture showing this superposition is observed in the southwestern part of the Near Ob' crest plain. Both mechanisms are most pronounced at Belovo. During glacials, the wind strength is strongest and the windvigour mechanism causes higher susceptibility values with a preferred orientation of magnetic grains. During interglacials and interstadials, high temperatures and moderate humidity force pedogenesis, and much less material is deposited due to lower wind strengths. Pedogenesis may not be strong enough to form larger quantities of single domain magnetite/maghemite particles, because palaeosols susceptibilities are not enhanced, compared to those of loess. Pedogenesis causes enhanced FD values (Figs 2 & 3) and lower S ratios (Fig. 5) and, along with weathering, cryo- and bioturbation processes cause a destruction of magnetic fabric in palaeosols (Fig. 7). The superposition of both processes is seen in a direct correlation between magnetic susceptibility and degree of anisotropy (Figs 2 & 6). Towards the north and NE, the superposition of the two mechanisms still holds true but is less pronounced resulting in an obscured picture due to slight differentiation in magnetic properties between loess and palaeosols. This process is governed by an equilibrium between humidity, temperature conditions and distance from the detrital source. Wind strengths are weaker and the correlation between AMS degree and preferred orientation of magnetic minerals is less pronounced (Fig. 7). The decrease of magnetic susceptibility and slight increase of FD values in loess horizons (Figs 2 & 3) observed towards the Salair mountain ridge may be caused by a relative increase in humidity of the palaeoclimate, as is observed in the modern climate (Table 2). The susceptibility remains nearly constant between different palaeosol units, while FD values and S ratios demonstrate large scattering.

PROPERTIES AND FABRICS OF PLEISTOCENE LOESS/PALAEOSOL DEPOSITS

However, the FD values are still higher, and the 5 ratio is lower than in loess horizons (Fig. 2). This means that pedogenesis in palaeosols from the Near Ob' crest plain is controlled mainly by precipitation and less by the temperature, similar to China (Maher & Thompson 1995). Kuznetsk depression The slight increase of magnetic susceptibility in palaeosol units in the Kuznetsk depression (sections Bachat and Novokuznetsk) results most probably from a little stronger pedogenesis. It may result from higher relative humidity and snow cover, moderate temperatures and lower evaporation (at Novokuznetsk) as indicated by modern climatic conditions (Table 2). The FD values of loess horizons are also slightly increased, suggesting that climatic conditions favour pedogenic activity. However, the magnetic susceptibility and the S ratio both demonstrate irregular variations. The AMS degree is low and not related to the susceptibility. The magnetic fabric of the palaeosol unit PCI and PC2 is completely random, which also favours the argument that stronger pedogenesis has taken place. The AMS degree in palaeosol units shows an inverse correlation with pedogenic intensity determined from FD values (Jordanova 6 Jordanova 1999). Increased FD values correspond to lower preservation of the initial parent loess deposit and the magnetic fabric is distorted or lost by pedogenic reworking. The annual temperature in the Kuznetsk depression is much lower relative to Belovo, while humidity is higher (Table 2). Apparently, increased humidity influences the strength of pedogenic activity rather than temperature. Minusa and Rybinsk depression Further northeastward, behind the Kuznetsk Ala-Tau mountain ridge, the difference between loess and palaeosol is clearly expressed in the magnetic properties. Most likely, the palaeoclimate was rather dry and cold with strongest wind intensity due to the closer position to the Siberian High. Thus, much more loess was deposited and the sections are thicker (Kurtak vs. Belovo). Deposits from Kurtak and Tatyshev sections evidently correspond to the pure 'Alaskan' model without significant pedogenesis. However, lower S ratios in palaeosol units at Kurtak may indicate the formation of pedogenic hematite (see below). The S ratios indicate that high coercivity minerals, namely hematite, occur mainly in palaeosol units. This occurrence seems to be also dependent upon the climatic conditions, for example S ratios PCI vs. PC2 units for all

169

sections (Fig. 5). Hematite forms via dehydration and structural rearrangements from poorly crystalline ferrihydrite in competition with goethite (Schwertmann 1988). This reaction favours hematite in low humidity and nearly neutral pH values. Such conditions are met in the present day soils of the Siberian realm (especially in the southwestern part), which are mostly chernozems (Table 2). We argue that pedogenic hematite formation in dry climates with low humidity is an important factor. Novokuznetsk has the highest relative humidity (90%), highest precipitation and lowest evaporation. Therefore, it demonstrates the smallest variation of the S ratio (Fig. 5). It seems that temperature and humidity are affecting the neoformation of pedogenic magnetite/maghemite rather than the formation of pedogenic hematite. On the Chinese loess plateau, remanence carried by hematite seems to be of detrital origin, whereas pedogenic hematite is superparamagnetic (Spassov et al. 2003). We propose the same situation for Siberian loess/palaeosols units, where hematite probably is of both detrital and pedogenic origin for southwestern sections and mainly of pedogenic origin in palaeosol units from central Siberia. Thus, pedogenic enhancement of magnetic minerals seems be a non-negligible factor in loess/palaeosol sections of the Siberian subaerial realm. Palaeowind direction The unaltered loess has retained its primary magnetic fabric with orientation of magnetic grains (W-E or SW-NE) according to the predominant palaeowind direction. High AMS degree, preferred orientation of AMS axes and essentially oblate magnetic fabric testify that these loess layers have undergone compaction only, without any secondary reworking. The palaeowind intensity probably is pronounced in the AMS degree. Palaeowind directions for LI loess units are more or less uniform (from WSW-NNE to W-E, Table 4, Fig. 7) and close to the modern wind direction all over the studied area (Fig. 7). At Lozhok and Mramorny, two wind directions are observed. This is in agreement with modern winds during spring and autumn at these localities which have two predominant orientations (SW-NE and S-N, respectively, Gusthina... 1979). The same situation is expected for Toguchin where the modern wind direction also has bimodal character, but the number of samples is not enough to draw such a conclusion (Fig. 7). The difference in palaeowind directions

170

G. G. MATASOVA & A. YU. KAZANSKY

for adjacently located sections Lozhok and Mramorny is probably connected with their different landscape position. Lozhok is located on a leeward slope of the river valley, while Mramorny is located on the windward slope of the watershed and is 30m higher in altitude. A similar situation holds in the Kuznetsk depression where differences in the palaeowind direction are observed. Bachat is located in the middle of the Kuznetsk depression and is partially screened by the Salair ridge, so the wind intensity must be weakened. Novokuznetsk, in contrast, is located on the southwestern windward slope of the depression in the foothills of the Kuznetsk Ala-Tau mountain ridge where the atmospheric circulation is essentially affected by mountain relief. Tatyshev has almost constant AMS patterns. The Kmax AMS axes are similar to LI and L2 at Belovo/Bachat and L2 at Kurtak. Since palaeosol unit PC2 derived its magnetic fabric from the underlying loess L3, it can be concluded that the main wind directions have generally not changed along the transect over the last 180000 years respectively since OIS 6. The AMS pattern of palaeosol PC2 appears to be more distorted than in PCI (Fig. 7, Belovo, Lozhok, Toguchin). The S ratio at Kurtak, Belovo and Lozhok is a little smaller in PCI than in PC2. There is no difference at all between PCI and the over- and underlying loess horizons at Tatyshev, in contrast to PC2. It is therefore suggested that pedogenesis was stronger during OIS 5 than during OIS 3. This is in agreement with observations from the central Chinese Loess Plateau. In China, the susceptibility of palaeosol LiSi, which is correlated with OIS 3 and PCI, is two times lower than that of palaeosol SI (OIS5/ PC2) (Sartori 2000). Conclusions The magnetic composition of Siberian loess/ palaeosol deposits consists of three main minerals: magnetite, maghemite and hematite. This qualitative composition is similar to other loess/palaeosol provinces in China and Alaska. Their variable concentration depends on climate and determines the magnetic properties of loess/palaeosol deposits and the mechanism of magnetic response on climatic changes ('Chinese', 'Alaskan' or 'Siberian'). Although the 'Siberian' mechanism is the superposition of the 'Chinese' and 'Alaskan' ones, from our viewpoint, it may be considered as a new individual mechanism due to its own peculiarities and advantages.

In general, the 'Alaskan' wind-vigour mechanism predominates the magnetic enhancement in loess of the Siberian subaerial realm. 'Chinese' pedogenic mechanism plays a minor, but non-negligible role. Superparamagnetic minerals are formed and their presence is mainly indicated by high FD values in palaeosols. Both mechanisms are active and their relative contribution depend on local climate conditions. The Kuznetsk Ala-Tau mountain ridge is the geographical barrier between rather humid climate in the SW and a rather dry/cold climate in the NE. Correspondingly, the wind-vigour enhancement dominates in the NE, whereas in the SW pedogenic magnetic mineral enhancement gains in importance. According to present-day climate parameters, the Siberian subaerial realm is divided into different sub-zones. The magnetic parameters reflect this subdivision and suggest that the climate of the past 180000 years has not changed to a greater or lesser extent. However, the amplitude of past global climate changes is also reflected in Siberian palaeosols. The marine oxygen isotope stage 3 is less pronounced in Siberia than stage 5, which is in agreement with observations on the Chinese Loess Plateau and other sedimentary record worldwide. The distribution of Kmax axes reflects palaeowind directions in unaltered loess with a pristine sedimentary fabric. In palaeosols, however, the distribution of Kmin and Kmax axes is a sensitive indicator for the degree of pedogenesis. Humidity seems to be an important climate factor for the degree of pedogenesis. During the Late Pleistocene, the Kuznetsk depression was a region with its own local mild/humid microclimate due to its specific isolated geographical position. The climate was characterized by the optimal balance between humidification and temperature over the studied area, which is still retained in modern climate. The authors are grateful to R. Zhu for allowing the use of the laboratory equipment of the Institute of Geology and Geophysics of the Chinese Academy of Sciences (Beijing) and useful discussion of the results; to E. Petrovsky and N. Jordanova for the help in experiments at the Geophysical Institute (Prague) and constructive comments on the data interpretation; to E. Solotchina and N. Paltchik for performance and interpretation of XRD analysis; to V. Zykina and D. Koz'min for guidance and assistance during fieldwork. We also greatly appreciate the hard work of reviewer S. Spassov in order to improve the manuscript and F. Lagroix for useful comments. This work was supported

PROPERTIES AND FABRICS OF PLEISTOCENE LOESS/PALAEOSOL DEPOSITS by joint Russian Basic Research Foundation and China National Foundation of Natural Sciences through Grant NO.99-05-39077.

References ARKHIPOV, S. A., ZYKINA, V. S., KRUKOVER, A. A., GNIBIDENKO, Z. N. & SHELKOPLYAS, V. N. 1997. Stratigraphy and palaeomagnetism of glacial and loess-soil deposits on the West-Siberian plain. Russian Geology and Geophysics, 38(6), 1065-1985. BARANOV, A. N. & BLINOVA, A. N. (eds). 1969. Atlas of the USSR, GUGK, Moscow. (In Russian) BEGET, J. E., STONE, D. B. & HAWKINS, D. B. 1990. Paleoclimatic forcing of magnetic susceptibility variations in Alaskan loess during the late Quaternary. Geology, 18, 40-43. BURAKOV, K. S. 1977. Thermomagnetometer. Izvestia AN SSSR. Fizika Zemli, 5, 92-96. (In Russian). CHLACHULA, J. 1999. Loess-palaeosol stratigraphy in the Yenisey basin, Southern Siberia. Sbornik Geologickych Ved,. Journal of geological sciences. Anthropozoic, 23, 55-70. CHLACHULA, J., EVANS, M. E. & RUTTER, N. W. 1998. A magnetic investigation of a Late Quaternary loess/palaeosol record in Siberia. Geophysical Journal International, 132, 128-132. CHLACHULA, J., RUTTER, N. W. & EVANS, M. E. 1997. A late Quaternary loess-palaeosol record at Kurtak, southern Siberia. Canadian Journal of Earth Sciences, 34(5), 679-686. DA VITAYA, F. F. & PASTUKH, V. P. (eds) 1960. Climatic Atlas of the USSR. V. 1. GUGK, Moscow. (In Russian). DEARING, J., LIVINGSTONE, I. & ZHOU, L. P. 1996. A Late Quaternary magnetic record of Tunisian loess and its climatic significance. Geophysical Research Letters, 23, 2, 189-192. DERBYSHIRE, E., BILLARD, A., VLIET-LANOE, B. V., LAUTRIDOU, J.-P. & CREMASCHI, M. 1988. Loess and palaeoenvironment: some results of European joint program of research. Journal of Quaternary Science, 3, 147-169. DOBROVOLSKY, V. V. 1976. Soil Geography and Basic Pedology. Prosveshenie Publishing, Moscow. (In Russian). DUNLOP, D. J. & OZDEMIR, O. 1997. Rock Magnetism: Fundamentals and Frontiers, Cambridge University Press, Cambridge. FORONOVA, I. V. 1999. Quaternary mammals and stratigraphy of the Kuznetsk Basin (South-Western Siberia). Sbornik Geologickych Ved. Anthropozoic. 23, 71-99. FLORINDO, F., ZHU, R., Guo, B., YUE, L., PAN, Y. & SPERANZA, F. 1999. Magnetic proxy climate results from the Duanjiapo loess section, southernmost extremity of the Chinese loess plateau. Journal of Geological Society, London, 104, 3, 645-659. FRECHEN, M. & YAMSKIKH, A. F. 1999. Upper Pleistocene loess stratigraphy in the southern Yenisei Siberia area. Journal of Geophysical Research, 156,3,515-525.

171

Guo, B., ZHU, R. X., FLORINDO, F., DING, Z. L. & SUN, J. M. 2002. A short, reverse polarity interval within the Jaramillo subchron: Evidence from the Jingbian section, northern Chinese Loess Plateau. Journal of Geophysical Research, 107, 2124-2124. GUSTHINA, A. G. (ed). 1979. Atlas of the Novosibirsk Region. GUGK, Novosibirsk, (in Russian). HELLER, F. & EVANS, M. E. 1995. Loess magnetism. Reviews of Geophysics and Space Physics, 33, 211-240. HELLER, F. & Liu, T. S. 1986. Paleoclimatic and sedimentary history from magnetic susceptibility of loess in China. Geophysical Research Letters, 13, 1169-1172. Hus, J. J. 2003. The magnetic fabric of some loess/ palaeosol deposits. Physics and Chemistry of the Earth, 28, iss. 16-19, 689-699. JELINEK, V. 1977. The Statistical Theory of Measuring Anisotropy of Magnetic Susceptibility of Rocks and Its Application. Brno, Geofyzika. JELINEK, V. 1981. Characterisation of magnetic fabric of rocks. Tectonophysics, 79, 63-67. JORDANOVA, D. & JORDANOVA, N. 1999. Magnetic characteristics of different soil types from Bulgaria. Studia Geophysica et Geodaetica, 43, 303-318. JORDANOVA, N., JORDANOVA, D. & KARLOUKOVSKI, V. 1996. Magnetic fabric of Bulgarian loess sediments derived by using various sampling techniques. Studia Geophysica et Geodaetica, 40, 36-49. KUKLA, G. F., HELLER, F., Liu, X. M., Xu, T. S. & AN, Z. S. 1988. Pleistocene climates in China dating by magnetic susceptibility. Geology, 16, 811-814. LAGROIX, F. & BANERJEE, S. K. 2002. Paleowind directions from the magnetic fabric of loess profiles in central Alaska. Earth and Planetary Science Letters, 195,99-112. Liu, X., Xu, T. & Liu, T. 1988. The Chinese loess in Xifeng. II: A study of anisotropy of magnetic susceptibility of loess from Xifeng. Geophysical Journal of the Royal Astronomical Society, 92, 349-353. Liu, X. M., J. SHAW, T. S. Liu, F. HELLER & M. Y. CHENG. 1994. Rock Magnetic Properties and Palaeoclimate of Chinese Loess. J. Geomagnetism and Geoelectricity, 45, 117-124. Lu, H. Y., YANG, W. F., Liu, X. D. & AN, Z. S. 1998. Orbital tuned time scale for loess deposit in Luochuan of Central Chinese Loess Plateau. Acta Geophysica Sinica, 41, 6, 810-816. MAKER, B. A. 1986. Characterization of soils by mineral magnetic measurements. Physics of the Earth and Planetary Interiors, 42, 76-92. MAHER, B. A. 1998. Magnetic properties of modern soils and Quaternary loessic palaeosols: paleoclimatic implications. Palaeogeography, Palaeoclimatology, Palaeoecology, 137, 25-54. MAHER B. A. & THOMPSON R. 1991. Mineral magnetic record of the Chinese loess and palaeosols. Geology, 19, 3-6. MAHER, B. A. & THOMPSON, R. 1992. Paleoclimatic significance of the mineral magnetic record of Chinese loess and palaeosols. Quaternary Research, 37, 155-170.

172

G. G. MATASOVA & A. YU. KAZANSKY

MAKER B. A. & THOMPSON R. 1995. Paleorainfall reconstructions from pedogenic magnetic susceptibility variations in the Chinese loess and palaeosols. Quaternary Research, 44, 383-391. MANGERUD, J., SVENDSEN, J. I. & ASTAKHOV, V. 1.1999. Age and extent of the Barents and Kara ice sheets in Northern Russia. Boreas, 28, 46-80. MATASOVA, G., PETROVSKY, E., JORDANOVA, N., ZYKINA, V. & KAPICHKA, A. 2001. Magnetic study of Late Pleistocene loess/palaeosol sections from Siberia: palaeoenvironmental implications. Geophysical Journal International, 147, 367-380. MATASOVA, G. G., KAZANSKY, A. Yu. & ZYKINA, V. V. 2002. Magnetic grain size in loess/soil deposits of Novokuznetsk section (Kuznetsk depression) as an indicator of climatic oscillations. In: VAGANOV, E. A., DEREVYANKO, A. P., GRACHEV, M. A. (eds) Main Legitimates of Global and Regional Changes of Climate and Environment in Late Cainozoic of Siberia. IAIE SO RAN, Novosibirsk, 323-337. (in Russian). MATASOVA, G. G., KAZANSKY, A. Yu. & ZYKINA, V. S. 2003. Superposition of 'Alaskan' and 'Chinese' proxy climatic signals recorded in magnetic properties of Upper and Middle Neopleistocene deposits in Southern West Siberia. Russian Geology and Geophysics, 44, 607-619. MOSKVITIN, A. I. 1940. Loess and loess-like deposits from Siberia. Publishing of Geological Institute. Geological series. 14(4), 1-83 (in Russian). NAGATA T. 1961. Rock Magnetism. Tokyo, Maruzen. NASLUND, J. O., RODHE, L., FASTOOK, J. L. & HOLMLUND, P. 2003. New ways of studing ice sheet flow directions and glacial erosion by computer modelling - examples from Fennoscandia. Quaternary Science Reviews, 22, 245-258. PAN Y. X., ZHU, R. X., SHAW, J., Liu, Q. S. & Guo, B. 2001. Can relative paleointensities be determined from the normalized magnetization of the windblown loess of China? Journal of Geophysical Research, 106, B9, 19221- 19232. PROTSUK, I. S. (ed.) 1978. Altai Region Atlas. GUGK, Moscow-Barnaul, (in Russian). REES, A. I. 1966.The effect of depositional slopes on the anisotropy of magnetic susceptibility of laboratory deposited sands. The Journal of Geology, 74, 6, 856-867. REINDERS, J. & HAMBACH, U. 1995. A geomagnetic event recorded in loess deposits of the Tonchesberg (Germany): identification of the Blake magnetic polarity episode. Geophysical Journal International, 122, 407-418. ROLPH, T. C, SHAW, J., DERBYSHIRE, E. & WANG, J. T. 1989. A detailed geomagnetic record from Chinese Loess. Physics of the Earth and Planetary Interiors. 56, 151-164. SARTORI, M. 2000. The Quaternary climate in loess sediments: Evidence from rock and mineral magnetic and geochemical analysis, Doctoral Thesis, Diss. ETHNo. 13570, Zurich. SCHWERTMANN, U. 1988. Occurrence and formation of iron oxides in various pedoenvironments. In: STUCKI, J. W., GOODMAN, B.A. & SCHWERTMANN, U. (eds) Iron in Soils and Clay Minerals. D. Reidel

Publishing, Dordrecht, Boston, Lancaster, Tokyo, 267-308. SPASSOV, S., HELLER, F., KRETZSCHMAR, R., EVANS, M. E., YUE, L. P. & NOURGALIEV, D. K. 2003. Detrital and pedogenic magnetic mineral phases in the loess/palaeosol sequence at Lingtai (Central Chinese Loess Plateau). Physics of the Earth and Planetary Interiors, 140, 4, 255-275. SVEZHENTSEV Yu. S., LlSITSIN N. F. & VASIL'EV S. A.

1992. Radiocarbon chronology of the Yenisey Palaeolithic. In: Chronostratigraphy of Palaeolithic of Northern, Central and Eastern Asia and America. Publications of International Sympothium. RAN, Siberian Branch. Institute of History, Philosophy and Philology , Novosibirsk, 57-64 (in Russian). TARLING, D. H. & HROUDA, F. 1993. The Magnetic Anisotropy of Rocks. Chapman & Hall, London. THISTLEWOOD, L. & SUN, J. 1991. A palaeomagnetic and mineral magnetic study of the loess sequence at Liujiapo, Xian, China. Journal of Quaternary Science, 6, 13-26. VLAG, P. A., OCHES, E. A., BANERJEE, S. K. & SOLHEID, P. A. 1999. The paleoenvironmental-magnetic record of the Gold Hill Steps loess section in central Alaska. Physics and Chemistry of the Earth, 24 (9), 779-783. VOLKOV, I. A. 1971. Late Quaternary subaerial sedimentary association. Nauka Press, Moscow. (In Russian). VOLKOV, L A. & ZYKINA, V. S. 1984. Loess stratigraphy in the south-western Siberia. In: VELICHKO A. A. (ed.) Late Quaternary Environments of the Soviet Union. University of Minneapolis Press, Minneapolis, 119-124. YAMSKIKH, A. F. 1992. Polycyclic terrace formation and stratigraphic subdivision of Quaternary deposits of river valleys from non-glacial zone of Near-Yenisey region in Siberia. Zodiac Publishing, Krasnoyarsk. (In Russian). ZANDER, A., FRECHEN, M., ZYKINA, V. & BOENIGK, W. 2003. Luminescence chronology of the Upper Pleistocene loess record at Kurtak in Middle Siberia. Quaternary Science Review. 22, 9991010. ZHU, R. X., COE, R. S. & ZHAO, X. X. 1998. Sedimentary record of two geomagnetic excursions within the last 15,000 years in Beijing, China. Journal of Geophysical Research, 103, B12, 30323-30333. ZHU, R. X., PAN, Y. X. & Liu, Q. S. 1999. Geomagnetic excursions recorded in Chinese loess in the last 70,000 years. Geophysical Research Letters, 26(4), 505-508. ZHU, R. X., KAZANSKY, A., MATASOVA, G., Guo, B., ZYKINA, V., PETROVSKY, E. & JORDANOVA, N. 2000. Rock-magnetic investigation of Siberia loess and its implication. Chinese Science Bulletin, 45, 23, 2190-2197. ZHU, R. X., MATASOVA, G., KAZANSKY, A., ZYKINA, V. & SUN, J. M. 2003. Rock magnetic record of the last glacial-interglacial cycle from Kurtak loess section, southern Siberia. Geophysical Journal International, 152, 335-343.

PROPERTIES AND FABRICS OF PLEISTOCENE LOESS/PALAEOSOL DEPOSITS ZYKINA, V. S. 1999. Pleistocene pedogenesis and climatic history of Western Siberia. Sbornik Geologickych Ved,. Journal of geological sciences. Anthropozoic, 23, 49-53. ZYKINA, V. S. & KIM, Y. V. 1989. Pedogenesis and loess sedimentation in Pleistocene in south-eastern part of West Siberia In: ARKHIPOV, S. A. & VOLKOVA, V. S. (eds) Pleistocene of Siberia. Stratigraphy and interregional correlations. SB RAS Publishing, Novosibirsk, 657, 81-86. (In Russian).

173

ZYKINA, V. S., VOLKOV, I. A. & DERGACHEVA, M. I. 1981. The Upper Quaternary Deposits and Fossil Soils of Novosibirsk PreObje. Nauka, Moscow (in Russian). ZYKINA, V. S., VOLKOV, I. A. & SEMENOV, V. V. 2000. Neopleistocene climatic reconstruction in West Siberia based on study of Belovo key section. In: VAGANOV, E. A. & DEREVYANKO, A. P. (eds.) Problems of Reconstruction of Holocene and Pleistocene Climate and Environment in Siberia, SB RAS Press, Novosibirsk, 229-249. (In Russian).

This page intentionally left blank

The puzzle of axis-normal magnetic lineations in folded low-grade sediments (Bude Formation, SW England) MARK W. ANDERSON & ANTONY MORRIS School of Earth, Ocean and Environmental Sciences, University of Plymouth, Drake Circus, Plymouth, PL4 8AA, UK Abstract: A single upright, open anticline from sub-greenschist fades sedimentary rocks of the Bude Formation (Cornwall, UK) was sampled in order to investigate the kinematic relationships between fold development and anisotropy of magnetic susceptibility (AMS). The mean magnetic susceptibility of these samples is 0.25 x 10~ SI, suggesting low concentrations of ferromagnetic phases. AMS ellipsoids have a mean corrected anisotropy degree of 1.03 and a mean shape parameter of -0.54 (prolate). Kmin and Kini define a girdle distribution striking sub-parallel to the fold axial plane, with Kmin tending to cluster around the fold axis. ^rnax axes from both limbs of the fold define a cluster with a mean azimuth perpendicular to the fold axis. This arrangement of Kmax and Km^n could represent an inverse magnetic fabric of composite primary/tectonic origin. This is discounted, however, on the basis of broad correlation between the orientation of AMS and AIRM (anisotropy of isothermal remanence) ellipsoids. The prolate shapes and axis-normal orientation of ^max axes contrast markedly with the widely observed relationship of AMS ellipsoids in folds, which are typically oblate and have Kmax parallel to the fold axis. This relationship is interpreted to represent progressive overprinting of primary depositional/compactional fabrics (Kmin perpendicular to bedding) by a tectonic fabric (A^min perpendicular to cleavage). Consistency of KmaK orientations irrespective of position within the fold clearly points to a fabric of tectonic origin. Prolate ellipsoids with long axes perpendicular to the fold hinge line are indicative of superimposed sub-horizontal stretching at a late stage or post-dating fold formation. Such a situation is not inconsistent with superimposed southward-directed thrusting simple shear that has been suggested in this area to account for variations in fold attitude on a regional scale. It is more likely, however, that the fabric reflects post-orogenic extension, with the fold occupying a position in the immediate hanging-wall of a major northward dipping normal fault. In either case, the AMS fabrics around the fold record only the last increments of deformation in this area, with earlier primary and fold-related fabrics being entirely obliterated.

A frequently observed relationship between the orientation of anisotropy of magnetic susceptibility (AMS) ellipsoids in folded low-grade metamorphic rocks is an alignment of the maximum principal susceptibility axes, KmSLX9 along the intersection of an axial planar cleavage and bedding (e.g. Borradaile & Tarling 1981; Hrouda et al. 2000). This relationship is interpreted to represent progressive overprinting of primary depositional/compactional fabrics (characterized by alignment of the minimum principal susceptibility axes, Kmin, perpendicular to bedding) by a tectonic fabric (characterized by Km[n axes arranged perpendicular to cleavage) (e.g. Housen et al. 1993). Such composite magnetic fabrics are usually dominantly oblate in shape with Kmax axes typically aligned perpendicular to the direction of maximum tectonic shortening (e.g. Averbuch et al. 1995; Hirt et al. 2000; Pares & van der Pluijm 2003). The main aim of this study is to investigate further these relationships in order to assess the

general applicability of using AMS fabrics as a proxy for petrofabrics in low-grade fold and thrust belts. Very low-grade meta-sedimentary rocks of the Upper Carboniferous Bude Formation have been sampled around a single upright, open anticline at Widemouth Bay, North Cornwall (Fig. 1), which lies within the Variscan belt of SWEngland. Additional samples were collected away from the target fold in order to constrain the 'background' patterns of AMS fabrics in these rocks. The area was chosen because the geometrical and kinematic development of the folds have been studied extensively (e.g. Ramsay 1974; Sanderson 1974, 1979; Lloyd & Whalley 1986, 1997; Tanner 1989, 1992). Furthermore, the rocks have complex pre- and post-fold deformational histories (e.g. Lloyd & Chinnery 2002, 2003; Treagus 2003) potentially allowing the progressive development of finite AMS fabrics from depositional, fold-related and post-folding stages to be evaluated.

From: MARTIN-HERNANDEZ, F., LUNEBURG, C. M., AUBOURG, C. & JACKSON, M. (eds) 2004. Magnetic Fabric: Methods and Applications. Geological Society, London, Special Publications, 238, 175-190. 0305-8719/04/ $15.00 © The Geological Society of London 2004.

Fig. 1. (a) Simplified geology of the North Cornwall and Devon, showing the location of area investigated in detail (white star); (b) Generalized cross-section from Widemouth Bay to Boscastle showing interpretation as a regional Variscan fold dismembered by numerous northward dipping normal faults; (c) Reconstruction of major southward facing fold in North Cornwall, showing fan-like arrangement of minor chevron folds and position of later normal faults (prior to displacement), (b) and (c) Modified from Sanderson (1979).

AMS IN LOW-GRADE SEDIMENTS

Geological background and regional structural setting Culm Basin development The spectacular coastal sections between Boscastle and Hartland Quay (Fig. 1) represent the most complete exposure of the Upper Carboniferous Culm Basin succession (Freshney et al. 1972). North of the regionally significant Rusey Fault (Andrews et al. 1988), two formations are identified within this succession, the lower Crackington Formation (Namurian to Westphalian A) and higher Bude Formation (Westphalian A-C). The Crackington Formation comprises organic-rich mudrocks and thin ( 1 m), medium-grained sandstones and shales with sporadic slumped and slurried horizons ('disturbed' beds). The Key Shales (Melvin 1986), dark grey to black marine marker beds, often containing pyrite, occur throughout the succession. Sandstone units may obtain thicknesses up to 9 m, but typically lack internal structure. Planar- and cross-lamination is seen rarely, although thin, laterally discontinuous screens of claystone within the thicker sand bodies suggest amalgamation from thinner beds. The bases of the sandstones against the background shales are typically sharp but very irregular, with widespread development of spectacular load structures and more locally developed scours. Palaeocurrent indicators suggest flow towards the southwest, in contrast to the easterly directed flow suggested for the Crackington Formation. The interpretation of the Bude Formation is a matter of continuing debate. Burne (1995) and Melvin (1986, 1987) favoured deposition as part of a large-scale, southward-sloping submarine fan. Higgs (1987, 1991), however,

177

argued, for deposition in shallow water on a horizontal lake floor, with the Key Shales representing occasional marine incursions. Disturbed beds ('slump beds') occur throughout the upper levels of the Crackington Formation and much of the Bude Formation. They are especially well developed towards the top of the Bude Formation. Hecht (1992) suggests that the earliest slump horizons in the Crackington Formation may reflect tectonic uplift of a northern source area in the late Namurian. Slump horizons within the Bude Formation also show a general southward fold vergence, confirming slump movement on a southward sloping surface. Hartley's (1991) interpretation that both slumps and basin-wide debris flows are represented, implied a significantly sloping floor. However, if they are seismically disturbed soft sediment (Higgs 1987, 1991), then much gentler slopes could have been involved. Slump folds with a northerly sense of vergence have been recognized by Enfield et al. (1985) towards the top of the succession and probably record a reversal of the slope of the Bude Basin floor suggesting tectonic uplift to the south. It is generally accepted that the Culm basin is syn-orogenic with respect to an advancing Variscan deformation front to the south. Hartley & Warr (1990) and Warr (1993) suggest a foreland basin origin whereas Gayer & Jones, (1989) prefer to interpret it as a 'piggy-back' basin developed above a northward translating thrust sheet. A further possibility is a pull-apart basin developed during oblique dextral convergence (Jackson 1991; Andrews 1993). The Rusey Fault marks the southern margin of the Culm Basin, although there is widespread evidence of multiple reactivation along this boundary and its initial geometry remains unclear. It is regarded by many authors as being a major basin-bounding normal fault that was reactivated during later Variscan thrusting (Andrews et al. 1996). Variscan transpression Variscan thrusting and folding produced in excess of 50% shortening of the Culm Basin succession and clearly post-dates deposition of Westphalian C units within the Bude Formation. Early bed-parallel thrusts (Whalley & Lloyd 1986; Mapeo & Andrews 1991) are folded by E-W-trending chevron folds that characterize all parts of the basin. Local thrust duplexes may also have accommodated bed-parallel flexural slip during folding (Tanner 1992). The chevron folds display a regional fan-like geometry (Sanderson & Dearman 1973).

178

M. W. ANDERSON & A. MORRIS

Upwards-facing folds characterize the greater part of the basin between Hartland Point and Widemouth Bay although increasingly recumbent, south-facing folds typify the southern margin, including the well-known Millook Haven locality (Fig. 1). The transition from upright to recumbent chevron folds is considered by Sanderson (1979) and Rattey and Sanderson (1982) to be the result of regional southerly directed simple shear acting at a low angle to an originally subhorizontal stratigraphic template. He invoked a major southerly directed overfold (the 'Millook Nappe') with simple shear strains increasing with depth in the structure (Fig. Ic), The range of observed geometries in the chevrons is produced, in this model, by both regional and local strain variations through the structure. Subsequent normal faulting dismembers the nappelike structure to produce the present distribution of stratigraphic units and associated chevrons. Deepest levels of the original overfold (highest strains) are exposed south of the Rusey Fault Zone towards the Tintagel High Strain Zone. Sanderson's model invokes southerly directed simple shear during both initiation and modification of the chevrons and requires, therefore, a reversal of the main northward directed Variscan deformation in this area. Lloyd & Whalley (1986) suggest an alternative, partitioned simple shear model to explain the variation in observed chevron geometry. In this model initial chevron formation is associated with northerly directed thrusting and imbrication along decollement horizons (Whalley & Lloyd 1986). They argue that southerly directed shear is a mechanically necessary deformation associated with continued overall northward-directed deformation. As such the southerly directed shear that is locally superimposed upon the initial chevron geometry is considered analogous to backthrusts observed in the external portions of many fold-thrust belts. They also demonstrate that southerly directed simple shear requires the synchronous development of both low angle and high angle normal faults during modification of the initially upright chevron fold geometry. A feature of both models is that they invoke distributed southerly simple shear (either regionally distributed or in relatively narrow zones) rather than the development of discrete backthrusts. Indeed the absence of major south-directed thrusts in this area is a paradox of the overall transition in fold style from north to south. The main Variscan deformation in SW England is generally considered as a dextral transpression. Andrews et al (1998) summarize the available evidence, which is based upon the geo-

metries of pre-folding vein arrays, obliquity between the trend of the folds and the margins of the basin, and rare, asymmetrical plunging folds with dextral vergence. NW-SE-trending dextral wrench faults are also common throughout the Culm Basin, and SW England generally, the largest having kilometre-scale displacements (Dearman 1963). These cross-cut earlier fold and thrust structures and, together with bed-parallel dextral strike-slip duplexes, developed along the steep limbs of the chevron folds (Mapeo & Andrews 1991), probably represent an increasingly important strike-slip component to the transpressive regime during the latter phases of basin closure. Very low grades of metamorphism associated with basin development and subsequent thrustrelated crustal thickening characterize the Culm Basin succession (Warr et al. 1991). Illite crystalUnity values of A26> = 0.37 - 0.44 (Warr & Hecht 1993) indicate metamorphic grades no higher than transitional diagenesis to anchizone. Post-transpressional structures Widespread normal faults cross-cut thrusts, folds and strike-slip faults related to dextral transpression throughout SW England (e.g. Shail & Alexander 1997). These structures dip consistently to the north-northeast or south-southwest with angles of dip varying from 30-80°. Local reactivation of pre-existing thrust surfaces at least partly explains normal faults with angles of dip at the low end of this range. In the area immediately south of the sampled fold at Widemouth Bay numerous north-dipping extensional faults dismember the overturned limb of the regional south-facing anticline (Freshney et al. 1972). These represent the latest structures identified in the area and probably developed during a phase of extension in the thickened Variscan crust, either immediately before, or during, protracted emplacement of the Cornubian batholith in the Late Carboniferous and Early Permian (e.g. Alexander & Shail 1996). Local geology of the sampled fold The sampled fold at Widemouth Bay (Fig. 2) is an upright, open chevron fold (Fig. 3) in the immediate hanging wall of a major north dipping normal fault. Although not well exposed, this structure is shown on the maps of Freshney et al. (1972) to cut-out an earlier NW-SE trending dextral strike-slip fault (Widemouth South Fault) that in turn cross-cuts the fold and

AMS IN LOW-GRADE SEDIMENTS

179

Fig. 2. Detailed geological map of the foreshore south of Widemouth Sand (part of Ordnance Survey 100km grid square SS) showing the location of the fold sampled in detail (arrowed) and additional sampling sites to the north. Modified from Freshney et al. (1972). A-B shows line of cross-section in Figure 3.

thrust structures in the area. The Widemouth South Fault juxtaposes correct way-up rocks of the Bude Formation to the north and generally overturned rocks of the Crackington Formation to the south. No stratigraphic contacts are observed between these two units in this area. The folded rocks of the Bude Formation comprise medium-grained sandstone beds of 120mm to 2m thickness, interbedded with grey-black laminated siltstone and organic-rich shale units up to 1 m in thickness. The sandstone beds are typically massive towards their bases, becoming laminated and cross-laminated upwards. Load structures and sole marks are common on the bases. No tectonic fabric is observed petrographically in the sandstone beds,

which are locally fractured and exhibit several generations of quartz-filled vein arrays. A subvertical cleavage is developed in the shales, but only in the hinge region of the sample fold. Steep, northward-dipping reverse faults also locally breach the hinge of the fold. Samples for AMS analysis were collected from each of these three lithologies from both north and south dipping limbs and the hinge region of the fold. Further upright chevrons are exposed in the area north of the sample fold for a distance of 150 m. Beyond this the beds dip more uniformly to the north at angles of 60-75°. Samples for comparative AMS analysis were collected from this area over a distance of 225 m from the fold axis (Fig. 2).

180

M. W. ANDERSON & A. MORRIS

Fig. 3. (a) Detailed N-S cross-section along line A-B (Figure 2) showing fold profiles along sample line, subsequent brittle faulting and the location of the area sample fold (boxed). Symbols as in Figure 2. Modified from Freshney et al. (1972); (b) View of the fold sampled in detail, looking towards the east. Numerous undergraduate students for scale; (c) Sketch of sample fold illustrating minor reverse faulting in core, interepreted to represent local accommodation structures during folding. Arrows indicate three beds sampled around the fold in this study.

AMS IN LOW-GRADE SEDIMENTS

Sampling and laboratory methodologies Core samples of 25mm diameter were drilled in situ following normal palaeomagnetic procedures. Cores were oriented using both magnetic and sun compasses, and subsequently cut into standard 22mm long specimens. In the laboratory, the anisotropy of low field magnetic susceptibility of each specimen was determined using an AGICO Kappabridge KLY-3S. Anisotropy of isothermal remanent magnetization (AIRM) was determined for a sub-set of samples in order to test for inverse magnetic fabrics (involving interchange of the orientation of maximum and minimum principal susceptibility axes due to the presence of single domain magnetite (Rochette 1988)). AIRM ellipsoids were determined using the method of Stephenson et al. (1986) by imparting isothermal remanent magnetizations (IRMs) along three orthogonal specimen axes using an 80 mT direct field generated by a Molspin pulse magnetizer, and alternating field demagnetization at a peak field of lOOmT between remanence determinations. Additional

181

rock magnetic characterization of the sampled lithologies was achieved via: (i) acquisition of IRM experiments, in applied fields up to 800 mT; (ii) thermal demagnetization of threecomponent IRMs (acquired in fields of 800, 300 and 50 mT) using the method of Lowrie (1990); and (iii) determination of the low temperature variation of low field magnetic susceptibility using an AGICO CS-L low temperature cryostat apparatus on the KLY-3S Kappabridge. All IRMs were measured using a Molspin fluxgate spinner magnetometer (on the long spin setting). Results and analysis

Structure and petrofabrics Structural data collected from the sample fold are presented in Figure 4a. Poles to bedding measurements from around the fold describe a N-S girdle distribution with two weak clusters of data from the limb regions, reflecting the chevron style of folding. Poles to cleavage measurements

Fig. 4. Summary of structural (stereograms a and b) and AMS data (stereograms c and d) from the study area. Star in (a) denotes mean orientation of vr-axis associated with best fit 7r-girdle (grey line). The mean orientation of Si cleavage is also shown (dashed line).

182

M. W. ANDERSON & A. MORRIS

cluster around the N-S horizontal. Following the eigenvector approach of Woodcock (1977) a best fit 7r-girdle has been fitted through the bedding data, producing a mean ?r-axis that plunges 7° towards 267°. The maximum eigenvector estimates the mean cluster of cleavage poles and plunges, 2° towards 178°, suggesting a mean cleavage orientation dipping 88° towards 358°, axial planar to the fold. As might be expected, structural data collected in the cliff section and wave-cut platform north of the sample fold show more scatter, although bedding poles also follow a broad N-S girdle (Fig. 4b). Two clusters of bedding poles fall off the general girdle, plunging moderately steeply to the NE and moderately to the SW respectively. These represent data from folds on the wave-cut platform that have axes that trend more NW-SE than the majority of folds, probably caused by proximity to late NW-SE striking dextral strike-slip faults (Fig. 3). Poles to the cleavage developed in the mudrocks are generally sub-parallel to the poles to axial planes of minor folds and again cluster close to the N-S horizontal. Plunges of minor fold axes are generally coaxial with the vr-axis for the sample fold although there is a slight spread from W to NW plunges, again associated with proximity to the late dextral strike-slip faults. Where present, petrofabrics in the sandstone beds are all depositional in origin. These reflect planar and cross-lamination and local scouring with no grain shape preferred orientation apparent in thin section. The sandstones are dominated by quartz, minor amounts of clay minerals including chlorite and very occasional lithic fragments. Mudrocks also display mainly depositional petrofabrics, primarily planar lamination, again with no visible grain shape preferred orientation. Exceptions are samples of black shale from both the hinge region of several folds and locally where bed parallel shear has occurred. A spaced cleavage and poorly defined S-C fabric are developed respectively. In thin section these fabrics are defined by alignment of clay minerals (including chlorite) and organic matter but with little or no visible grain shape preferred orientation of fine grained quartz.

Magnetic mineralogy Since the AMS of a specimen may result from contributions from diamagnetic, paramagnetic, antiferromagnetic and ferrimagnetic minerals (see Tarling & Hrouda 1993 for a review), it is necessary to understand the nature of the minerals carrying the susceptibility signal before

geological interpretation is possible. Low field (bulk) magnetic susceptibilities of the specimens are low with a mean value of 2.5 x 10~ SI, suggesting that paramagnetic minerals dominate the magnetic fabrics in these rocks. This is confirmed by the variation in magnetic susceptibility (K) with temperature, which shows near linear relationships on plots of \/K versus T(K) during warming from liquid nitrogen temperatures (77 K), suggesting that the susceptibility obeys a near Curie-Weiss law (Fig. 5a). The relative contributions of ferromagnetic and paramagnetic phases to the total susceptibility may be determined using the method of Richter and van der Pluijm (1994). This involves subtraction of an assumed ferromagnetic contribution from the total susceptibility, which is increased until a purely paramagnetic straight-line relationship is produced on a plot of K0/K versus temperature (indicated by a maximum correlation coefficient). The paramagnetic contribution is found by subtracting the best-fitting ferromagnetic susceptibility from the initial room temperature low field magnetic susceptibility. Results suggest that the paramagnetic fraction contributes between 64 and 99% of the observed susceptibilities in these specimens. Petrographically, these rocks are dominated by diamagnetic quartz and (minor) calcite. Optical microscopy and XRD observations indicate presence of significant paramagnetic chlorite and mixed illite-smectite clay minerals, at concentrations sufficient to swamp the diamagnetic signal and to account for the observed susceptibilities. The nature of the ferromagnetic phases in these rocks can not be determined easily by direct observation because of the inferred low concentrations. IRM acquisition curves indicate the presence of both high and low coercivity ferromagnetic minerals in these rocks (Fig. 5b). An initial rapid rise in isothermal remanence in fields below 200 mT is consistent with presence of magnetite, whereas a more gradual rise in IRM up to fields of 800 mT suggests presence of either hematite or goethite. Step wise thermal demagnetization of three component IRMs (Lowrie 1990) provides additional constraints on the ferromagnetic mineralogy (Fig. 5c), although interpretation is complicated by chemical changes induced by laboratory heating. Medium to low coercivity grains show distributed unblocking temperature spectra up to 500 °C, confirming presence of magnetite. A low intensity high coercivity component unblocks by 150°C, confirming presence of goethite. Thermal alteration observed above 400 °C (Fig. 5c) is consistent with decomposition of pyrite (the presence of which is confirmed petrographically) to produce

AMS IN LOW-GRADE SEDIMENTS

183

new magnetite. The original IRM signal is swamped at these higher temperatures, making determination of the maximum unblocking temperature of the magnetite grains impossible. In summary, the susceptibility signal in these rocks is dominated by paramagnetic phases with a minor, variable contribution from magnetite and a negligible contribution from goethite.

Anisotropy of magnetic susceptibility (AMS) )

Fig. 5. Typical rock magnetic data: (a) Variation of low field magnetic susceptibility upon warming from liquid nitrogen temperature demonstrating a near Curie-Weiss dependency indicative of a dominantly paramagnetic signal. Verwey transition at c. 100K suggests presence of minor pure magnetite; (b) Example of isothermal remanent magnetization (IRM) acquisition curve indicating a dominance of low coercivity phases within the ferromagnetic fraction; (c) Thermal demagnetization of 3-component composite IRMs (method of Lowrie 1990), confirming presence of magnetite with distributed unblocking temperature spectra and minor high coercivity goethite. Note the effects of thermal alteration above temperatures of 450 °C.

AMS reflects the preferred orientation of grains, grain distributions and/or the crystal lattices of minerals that contribute to the magnetic susceptibility of a rock. It corresponds to a second order tensor and can be represented by an ellipsoid specified by the orientation and magnitude of its principal axes (Xmax, Kini and A:min, being the maximum, intermediate and minimum susceptibility axes respectively). Diamagnetic and antiferromagnetic contributions are minimal in the sampled lithologies, and the AMS signal is carried dominantly by paramagnetic chlorite and clay minerals and by minor amounts of magnetite. The oblateness or prolateness of individual specimen AMS ellipsoids is described by the AMS shape parameter T (Jelinek 1981), with 0 < r < 1 . 0 for oblate ellipsoids and -1.0 < T < 0 for prolate ellipsoids. Specimens from the sampled fold display a dominance of prolate fabrics at the specimen-level (Fig. 6), with a mean value of T — —0.5. Strength of anisotropy is described by the corrected anisotropy degree, Pj (Jelinek 1981), and ranges from 1.01 to 1.08 (Fig. 6), with a mean value of 1.03 (i.e. 3% anisotropy). There is no preferred relationship between Pj and T, and no correlation between Pj and mean susceptibility, indicating that the degree of anisotropy is not dependent on ferrimagnetic concentration. The sampled fold has a well-defined fabric characterized by a cluster of Kmax axes orthogonal to an E-W girdle distribution of £jnt and kmin axes (Fig. 4c), indicating an overall prolate

fabric. Samples collected north of the sample fold show a similar overall strength of anisotropy. The majority of these data have corrected anisotropy degree values in the range 1.01-1.07, with a mean value for Pj of 1.05 (i.e. 5% anisotropy). The mean value for Pj is higher than in the sample fold since 5 samples have much higher values of corrected anisotropy degree, being in the range 1.1-1.3. The shape of the AMS ellipsoids north of the sample fold is noticeably more

184

M. W. ANDERSON & A. MORRIS

Fabrics north of the sample fold are characterized by a much greater variability in the orientation of each of the three principal axes of magnetic susceptibility (Fig. 4d). There is a weak clustering of Kmax axes close to the N-S horizontal, similar to the strong clustering of KmSLX in the sample fold. A sub-ordinate cluster of Kmax axes is also apparent, plunging moderately to the ENE. Kmin axes form a broad E-W girdle, although this is weakly defined and contains a sub-ordinate clustering of Kmin axes that plunge moderately to the SW (Fig. 4d).

Anisotropy of isothermal remanent magnetization (AIRM)

Fig. 6. Jelinek plot of shape parameter against corrected anisotropy degree for AMS ellipsoids throughout the study area.

variable, however, with a much greater spread of data from the prolate field into the oblate field on the Jelinek plot (Fig. 6). The 4 samples displaying the highest values for corrected anisotropy degree are also strongly prolate ellipsoids.

AIRM ellipsoids reflect the preferred orientation of ferromagnetic phases and are not affected by domain state. Comparison of AIRM and AMS fabrics can, therefore, be used to detect the potential presence of inverse AMS fabrics resulting from single-domain effects in magnetite. The mean IRM intensity observed in samples during AIRM experiments was SSmArrT1. In general, AIRM ellipsoids determined for the sample fold have a higher degree of anisotropy (mean PJ = 1.13) than the corresponding AMS ellipsoids (Fig. 7), as observed in other studies (e.g. Stephenson et al. 1986). Again, the majority of samples display prolate fabrics although there is a greater spread of data into the oblate field on the Jelinek plot (Fig. 7). The orientations of Kmin and K-mt principal axes display a diffuse E-W girdle distribution whereas Kmax axes are

Fig. 7. Anisotropy of IRM data from the sample fold. Symbols as for AMS data in Figure 4.

185

AMS IN LOW-GRADE SEDIMENTS

sub-horizontal and cluster weakly around a NNW-SSE azimuth. This arrangement of principal axes is similar to that observed using AMS, and is consistent with an overall prolate ferromagnetic fabric. The low intensities of isothermal remanence observed in these rocks are consistent with very low concentrations of magnetite and probably introduce an element of experimental error in the determination of the AIRM ellipsoids. This may account for the greater dispersion of data observable in Figure 7 compared to the corresponding AMS data. The broad agreement in the distribution of AIRM principal axes compared with those of the AMS ellipsoids indicates that singledomain-related inverse fabric effects are not significant in the sampled lithologies. Given the higher level of analytical uncertainty involved in the determination of AIRM fabrics in these ferromagnetic-poor rocks, these are not considered further, and subsequent kinematic analyses are based on the much better defined AMS fabrics. Comparison of AMS orientation data and structural orientation The degree of clustering or scatter for the principal axes of the AMS ellipsoids may be evaluated following the eigenvector and eigenvalue approach of Woodcock (1977). The distribution of data for each set of axes, Kmax, K-mi and Kmin, have three orthogonal eigenvectors calculated (VI, V2 and V3) using the algorithm of Davis (1973), with VI estimating the mean orientation of a clustered distribution and V3 the pole to a girdle distribution (Table 1). All data pass the statistical test of significance suggested by Woodcock & Naylor (1983) at a 99% confidence level. Cluster and girdle distributions are distinguished by the shape parameter, K (ratio of eigenvalues), with high values of K (>1) indicating clustering

and low values (<1) indicating a girdle distribution. A further shape parameter, C, indicates the strength of the cluster/girdle. Data from the sample fold suggest a strong clustering of ^max about a mean axis that plunges gently to the north. KUX and Km^n are distributed along a strong girdle whose pole plunges parallel to the mean Kmax, gently to the north. Data from north of the fold are less coherent, with generally transitional shape fabrics (K close to 1 and moderate values for C). The exception is Kmax, which tends towards a girdle with a pole plunging moderately to the SW, sub-parallel to a poorly defined cluster for K^ (Table 1). This overall distribution may have little geological significance since the data probably comprise a mix of data clustered close to the N-S horizontal and moderately plunging to the ENE. When comparing the AMS fabrics with structural orientation data from the area some unusual relationships become apparent (Fig. 4). First, it is clear that Kmax axes in the sample fold cluster close to the poles to the axial planar cleavage, perpendicular to the fold axis. No clear relationship exists between ellipsoid orientation and either lithology, degree of petrofabric development or position (either limbs or hinge) within the fold. Second, K-mt and Kmin define an E-W girdle distribution with a mean great circle that lies close to, but is not parallel with, the axial plane/mean cleavage orientation determined for the fold. Neither of these relationships conform to the expected result based on previous studies of AMS fabric development in folds at low metamorphic grades. North of the sample fold the comparison between AMS and structure becomes less obvious. One of the weak clusters of Kmax axes corresponds with those in the sample fold (N-S horizontal) but otherwise no clear relationship exists between the AMS fabrics and the structural data (Fig. 4). Figure 8 plots the azimuthal dispersion of ATmax axes for all samples (away

Table 1. Eigenvectors determined for each of the 3 principal axes of AMS for the sample fold and the area immediately north

Sample fold

-* ^vmax

North of sample fold

^int Jf ^min if ^max ^int ^min

v,

v2

v3

K

c

23/001 8/266 67/161 25/139 48/354 29/250

62/144 62/161 8/272 35/030 41/155 45/125

15/265 27/000 21/005 44/256 9/253 30/359

3.44 0.25 0.15 0.08 0.65 1.02

3.50 2.28 3.01 1.78 1.28 2.33

K represents a shape parameter associated with each eigenvector and indicates the degree of cluster about the Vj axis (K > 1) or girdle about V3 axis (K < 1). The strength of the shape fabric is indicated by C. The shaded values are those that are considered to be both statistically and geologically meaningful (see text for details).

186

M. W. ANDERSON & A. MORRIS

Fig. 8. Plot of dispersion of the azimuths of Kmax principal axes versus distance northwards from the sample fold. White arrows indicate sites where Kmax azimuths are well grouped and consistent with those in the sample fold.

from the mean KmSLX orientation in the sample fold) versus distance northwards from the sample fold. It is clear that the azimuths of Kmax axes become very scattered away from the mean at distances greater than 30m from the sample fold. Well-defined northerly azimuths (consistent at a site level) occur at only three

localities north of this (at approximately 80m, 130m and 225m north of the sample fold). A comparison of the AMS shape parameter (T) with distance north of the fold (Fig. 9) indicates a progressive shape change from oblate to prolate dominated ellipsoids towards the sample fold, which is itself dominated by prolate ellip-

Fig. 9. Variation of AMS shape parameters with distance north of the sample fold. Black star denotes mean shape parameter determined from the sample fold. Black arrows indicate sites beyond 50 metres north of the sample fold where shape fabrics are dominantly prolate.

AMS TN LOW-GRADR SEDTMFNTS

187

soids (mean T = -0.5). Furthermore, anomalous prolate ellipsoids greater than 50m north of the sample fold correspond with those sites that have well-defined northerly azimuths for axes. K m Interpretation and conclusions Dominantly prolate AMS ellipsoids with Kmax axes perpendicular to the fold hinge line are difficult to reconcile with existing models for the development of AMS fabrics in folded rocks at low metamorphic grades (e.g. Hirt et al. 2000; Pares & van der Pluijm 2003). Rather, the fabrics from the sample fold in this study are more indicative of a superimposed sub-horizontal stretching at a late stage or post-dating fold formation (assuming Kmax to be generally sub-parallel to the X-axis of principal strain). Measured fabrics lie consistently out of the plane of both the folded bedding and the axial planar cleavage, suggesting that both earlier depositional and fold-related fabrics have been completely overprinted. For this to be the case, therefore, the AMS fabrics detected must represent only the very latest increments of late-Variscan or post-Variscan deformation in this area. Two geologically realistic possibilities exist for the origin of a superimposed stretching in this area (Fig. 10): (1) southward-directed simple shear associated with modification of the regional chevron folds from upright to recumbent structures in high strain zones, as predicted by the models of Sanderson (1979) and Lloyd and Whalley (1986); or (2) late- or post-orogenic extension of the thickened Variscan crust, associated with late normal faulting (e.g. Freshney et al. 1972), mineral veins (e.g. Moore 1975) and granite emplacement (e.g. Shail & Alexander 1997). Southward directed, simple shear might be expected to produce superimposed strain ellipsoids with moderate (<45°) plunges for the X axis of principal strain. A sub-horizontal geometry is proposed by Lloyd & Whalley (1986) but, if this is to produce the observed cluster of gently plunging axes for Km8LX, superimposed shear strains in excess of 7 = 1 are required. For 7 = 1 , the plunge of the X axis should be approximately 30° with an ellipticity (R) equal to 3.5 (Fig. lOa). Sanderson's (1979) model proposes a low angle, northerly dip to the zones of superimposed shear, which requires still higher shear strains to produce the gentle plunges observed for Kmax. Both models invoke localized south-directed shear as the cause of modification of an initial upright chevron geometry. Sanderson (1979) models the evolution of fold axial

Fig. 10. Geological models for superimposing N-S stretch on pre-existing chevron folds in North Cornwall, (a) Geometry of southward directed simple shear superimposed on initially upright chevron folds (Sanderson 1979; Lloyd & Whalley 1986). A representative orientation and shape of strain ellipse (shear strain, 7 = 1) is shown for comparison with the orientations of mean ATmax and Kmin axes determined in this study; (b) The proposed orientation of conjugate normal faults associated with superimposed N-S extension, assuming that the angle between Kmax and the dip of the fault planes cannot exceed 45°; (c) Modification of the upright chevron folds in the study area by north dipping normal faults during late- or post-orogenic extension (this paper).

planes and interlimb angles with increasing shear strain, predicting axial plane dips of 45° (northwards) and interlimb angles of 45-60° at a shear strain, 7 = 1 . Lloyd & Whalley (1986), however, consider only the limbs of the initial chevrons to behave as material planes and their models result in complex chevron profiles more closely matching the range of observed profiles across the region. Numerous geometrically necessary normal faults on either the limbs or along the axial planes of the modified folds are a feature of their models. At high shear strains (7 = 1), however, each of their models predict north-dipping axial planes and either a tightening of interlimb angles to less than 60°, or else

188

M. W. ANDERSON & A. MORRIS

complex box fold geometries. We conclude that none of the modified fold geometries predicted by south-directed superimposed simple shear models are consistent with those observed in the sample fold which has an upright axial surface and an interlimb angle between 60 and 80°. The gently plunging nature of Kmax is perhaps equally consistent with regional extension. Numerous north dipping normal faults have been documented and these clearly post-date development of the chevrons as they dismember the so-called 'Millook Nappe' structure to the south of Widemouth Bay (Freshney et al 1972; Selwood et al 1985; and Andrews et al 1988). These structures are shown on cross-sections of this area to be dipping at very low angles, typically 30-40° (Freshney et al 1972). At Widemouth Bay, low angle normal faults occur together with more steeply dipping normal faults having dips in excess of 60°. Figure lOb illustrates the limits on fault dip if the gentle plunge of Kmax is indeed sub-parallel to the X axis of principal strain. Assuming that the angle between the X axis and the fault planes never exceeds 45°, then the observed northerly plunge of 23° constrains northerly dipping normal faults to dip between 23 and 68°. Similarly, conjugate southerly dipping normal faults should have dips that do not exceed 22°. The position of the sample fold in the immediate hanging wall of a northerly dipping normal fault (Fig. 3) is consistent with this interpretation (Fig. lOc). This is further supported by the progressive change in the AMS shape parameter from north (oblate) to south (prolate) as this fault is approached (Fig. 9). Other anomalous prolate fabrics north of the sample fold may also correspond with late normal faults inferred independently on the maps of Freshney et al. (1972) (Fig. 3). The geometry proposed in Figure lOc assumes a bulk shear strain in the hanging wall of 7 = 1, mainly for comparison with the superimposed thrust shear models discussed previously. It seems unlikely, however, that such penetrative strains would accompany essentially a brittle faulting process, particularly when there is no post-folding petrofabric development. Lower overall strains would also require steeper fault dips according to the geometrical constraints outlined in Fig. lOb. Remarkably, the AMS fabrics appear to be sensitive to the superimposed stretching even though the principal deformation mechanism is a brittle one. A further possibility is that the AMS fabric does not reflect penetrative deformation of the hanging wall away from the fault, but rather directional fluid flow and secondary mineral growth at a grain scale

during sub-horizontal stretching. If fluid flow is controlled by micro-dilation of grain boundaries in particular orientations then it may be possible to produce an AMS fabric that reflects the regional stresses. On this note, it is also worth commenting that, although not exposed in the study area, E-W trending mineral lodes and elvan dykes are common throughout this area of SW England and are thought to have opened as mode I extension fractures immediately prior to, and during granite emplacement (Halls 1987). The well-defined N-S azimuth obtained for Kmax in the sample fold is consistent with opening directions on these fractures. Summary Analysis of anisotropy of magnetic susceptibility (AMS) in low-grade metamorphic rocks is capable of revealing pervasive tectonic fabrics that remain undetected by standard petrofabric analysis. Remarkably, these fabrics may record only the latest deformation in rocks that have complex depositional and earlier tectonic histories. In this study, AMS fabrics are interpretable in the context of reasonable regional geological models, providing a potential proxy for mapping regional patterns of subtle late- or post-orogenic deformation in the Variscan crust of SW England. We would like to acknowledge the contributions made during sampling and analysis of the following: J. Phillips, B. Dunning and N. Shelford. P. Davis is thanked for help with sample preparation. F. Hrouda and O. Averbuch are both thanked for providing constructive reviews that improved the final version of the manuscript. This work was funded by the University of Plymouth.

References ALEXANDER, A. C. & SHAIL, R. K. 1996. Late- to PostVariscan structures on the coast between Penzance and Pentewan, south Cornwall. Proceedings of the Ussher Society, 9, 72-77. ANDREWS, J. R. 1993. Evidence for Variscan dextral transpression in the Pilton Shales, Croyde Bay, north Devon. Proceedings of the Ussher Society, 8, 198-199. ANDREWS, J. R., BARKER, A. J. & PAMPLIN, C. F. 1988. A reappraisal of the facing confrontation in north Cornwall: fold- or thrust-dominated tectonics? Journal of the Geological Society of London, 147, 777-788. ANDREWS, J. R., DAY, J. & MARSHALL, J. E. A. 1996. A thermal anomaly associated with the Rusey Fault and its implications for fluid movements. Proceedings of the Ussher Society, 9, 68-71.

AMS IN LOW-GRADE SEDIMENTS ANDREWS, J. R., ISAAC, K. P., ROBINSON, D. SELWOOD, E. B. & SHAIL, R. K. 1998. Variscan Structure and Regional Metamorphism. In: SELWOOD, E. B., DURRANCE, E. M. & BRISTOW, C. M. (eds) The Geology of Cornwall. Exeter University Press, 82-119. AVERBUCH, O., MATTEI, M., KISSEL, C., DE LAMOTTE, D. F. & SPERANZA, F. 1995. Kinematics of deformations within a blind thrust-system: The example of the 'Montagna dei Fiori' structure (Central Apennines front, Italy). Bulletin de la Societe Geologique de France, 166 (5), 451-461. BORRADAILE, G. D. & TARLING, D. H. 1981. The influence of deformation mechanisms on magnetic fabrics in weakly deformed rocks. Tectonophysics, 77, 151-168. BURNE, R. V. 1995. The return of the 'fan that never was': Westphalian turbidite systems in the Variscan Culm Basin: Bude Fm (South West England). In\ Plint, A. G. (ed.) Sedimentary Fades Analysis. International Association of Sedimentologists Special Publication, Oxford, 22, 101-135. DAVIS, J. C. 1973. Statistics and data Analysis in Geology. Wiley, New York. DEARMAN, W. R. 1963. Wrench faulting in Conrwall and South Devon. Proceedings of the Geologists' Association, 74, 265-287. EDMONDS, E. A., MCKEOWN, M. C. & WILLIAMS, M. 1975. British Regional Geology: South-West England. Institute of Geological Sciences, HMSO, London. ENFIELD, M. A., GILLCRIST, J. R., PALMER, S. N. & WHALLEY, J. S. 1985. Structural and sedimentary evidence for the early tectonic history of the Bude and Crackington Formations, north Cornwall and Devon. Proceedings of the Ussher Society, 6, 165-172. FRESHNEY, E. C., MCKEOWN, M. C. & WILLIAMS, M. 1972. Geology of the coast between Tintagel and Bude. Memoir of the Geological Survey of Great Britain, HMSO, pp 92. GAYER, R. A. & JONES, J. A. 1989. The Variscan foreland in South Wales. Proceedings of the Ussher Society,!, 177-179. HALLS, C. 1987. A mechanistic approach to the paragenetic interpretation of mineral lodes in Cornwall. Proceedings of the Ussher Society, 6, 548-554. HARTLEY, A. 1991. Debris flow and slump deposits from the Upper Carboniferous Bude Formation of SW England: implications for Bude Formation facies models. Proceedings of the Ussher Society, 1, 424^26. HARTLEY, A. & WARR, L. N. 1990. Upper Carboniferous foreland basin evolution in SW Britain. Proceedings of the Ussher Society, 1, 212-216. HECHT, C. A. 1992. The Variscan evolution of the Culm Basin, south-west England. Proceedings of the Ussher Society, 8, 33-38. HIGGS, R. 1986. 'Lake Bude' (early Westphalian, SW England): storm-dominated siliciclastic shelf sedimentation in an equatorial lake. Proceedings of the Ussher Society, 6, 417-418. HIGGS, R. 1987. The fan that never was? - discussion of 'Upper Carboniferous fine-grained turbiditic

189

sandstones from southwest England: a model for growth in an ancient, delta-fed subsea fan.' Journal of Sedimentary Petrology, 57, 378-382. HIGGS, R. 1991. The Bude Formation (Lower Westphalian), SW England: siliciclastic shelf sedimentation in a large equatorial lake. Sedimentology, 38, 445-469. HIRT, A., JULIVERT, M. & SoLDEViLA, J. 2000. Magnetic fabric and deformation in the Navia-Alto Sil slate belt, northwestern Spain. Tectonophysics, 320, 1-16. HOUSEN, B. A., RlCHTER, C. & VAN DER PLUIJM, B. A.

1993. Composite magnetic anisotropy fabrics: experiments, numerical models, and implications for the quantification of rock fabrics. Tectonophysics, 220, 1-12. HROUDA, F., KREJCI, O. & OTAVA, J. 2000. Magnetic fabric in folds of the easternmost RhenoHercynian Zone. Physics and Chemistry of the Earth (A), 25(5), 505-510. jELiNEK, V. 1981. Characterisation of the magnetic fabric of rocks. Tectonophysics, 79, 63-67. JACKSON, R. R. 1991. Vein-arrays and their relationship to transpression during fold development in the Culm Basin, central south-west England. Proceedings of the Ussher Society, 7, 356-362. LLOYD, G. E. & CHINNERY, N. 2002. The Bude Formation, SW England - a three dimensional, intraformational Variscan imbricate stack? Journal of Structural Geology, 24, 1259-1280. LLOYD, G. E. & CHINNERY, N. 2003. The Bude Formation, SW England - a reply to comments by J. Treagus. Journal of Structural Geology, 25, 21612163. LLOYD, G. E. & WHALLEY, J. S. 1986. The modification of chevron folds by simple shear: examples from North Cornwall and Devon. Journal of the Geological Society of London, 143, 89-94. LLOYD, G. E. & WHALLEY, J. S. 1997. Simple shear modification of chevron folds: implications for facing interpretations, strain analysis and deformation history. In: SENGUPTA, S. (Ed). Evolution of Geologic Structures from Micro to Macro Scales. Chapman and Hall, London, 373-396. LOWRIE, W. 1990. Identification of ferromagnetic minerals in a rock by coercivity and unblocking temperature properties, Geophysical Research Letters, 17, 159-162. MACKINTOSH, D. M. 1964. The sedimentation of the Crackington Measures. Proceedings of the Ussher Society, 1, 88-89. MAPEO, R. B. M. & ANDREWS, J. R. 1991. Pre-folding tectonic contraction and extension of the Bude Formation, North Cornwall. Proceedings of the Ussher Society, 4, 350-355 MELVIN, J. 1986. Upper Carboniferous fine-grained turbiditic sandstones from southwest England: a model for growth in an ancient, delta-fed subsea fan. Journal of Sedimentary Petrology, 56, 19-34. MELVIN, J. 1987. Upper Carboniferous fine-grained turbiditic sandstones from southwest England: a model for growth in an ancient, delta-fed subsea fan - reply. Journal of Sedimentary Petrology, 57, 378-382.

190

M. W. ANDERSON & A. MORRIS

MOORE, J. McM. 1975. A mechanical interpretation of the vein and dyke system of the SW England orefield. Mineralium Deposita, 10, 374—388. PARES, J. M. & VAN DER PLUIJM, B. A. 2003. Magnetic fabric and strain in pencil structures of the Knobs Formation, Valley and Ridge Province, US Appalachians. Journal of Structural Geology, 25, 1349-1358. RAMSAY, J. G. 1974. Development of chevron folds. Geological Society of America Bulletin, 85, 17411754. RATTEY, P. R. & SANDERSON, D. J. 1982. Patterns of folding within nappes and thrust sheets: examples from the Variscan of southwest England. Tectonophysics, 88(3-4), 247-267. RICHTER, C. & VAN DER PLUIJM, B. A. 1994. Separation of paramagnetic and ferrimagnetic susceptibilities using low-temperature magnetic susceptibilities and comparison with high-field methods. Physics of the Earth and Planetary Intereriors, 82, 113123. ROCHETTE, P. 1988. Inverse magnetic fabric in carbonate-bearing rocks. Earth and Planetary Science Letters, 90, 229-237. SANDERSON, D. J. 1974. Chevron folding in the Upper Carboniferous rocks of north Cornwall. Proceedings of the Ussher Society, 3, 96-103. SANDERSON, D. J. 1979. The transition from upright to recumbent folding in the Variscan fold belt of southwest England: a model based on the kinematics of simple shear. Journal of Structural Geology, 1(3), 171-180. SANDERSON, D. J. & DEARMAN, W. R. 1973. Structural styles of the Variscan fold belt in SW England, their location and development. Journal of the Geological Society of London, 129, 527-536. SELWOOD, E. B., STEWART, I. J. & THOMAS, J. M. 1985. Upper Palaeozoic sediments and structure in north Cornwall - a reinterpretation. Proceedings of the Geologists' Association, 96, 129-141. SHAIL, R. K. & ALEXANDER, A. C. 1997. Late Carboniferous to Triassic reactivation of Variscan basement in the western English Channel: evidence from onshore exposures in south Cornwall.

Journal of the Geological Society of London, 154, 163-168. STEPHENSON, A., SADIKUN, S. & POTTER, D. K. 1986. A theoretical and experimental comparison of the anisotropies of magnetic susceptibility and remanence in rocks and minerals. Geophysical Journal of the Royal Astronomical Society, 84, 185-200. TANNER, P. W. G. 1989. The flexural-slip mechanism. Journal of Structural Geology, 11, 635-655. TANNER, P. W. G. 1992. Morphology and geometry of duplexes formed during flexural slip folding. Journal of Structural Geology, 14, 1173-1192. TARLING, D. H. & HROUDA, F. 1993. The Magnetic Anisotropy of Rocks. Chapman & Hall, London. TREAGUS, J. 2003. The Bude Formation, SW England: discussion. Journal of Structural Geology, 25, 2159-2160. WARR, L. N. 1993. Basin development and foreland basin development in the Rhenohercynian of south-west England. In: GAYER, R. A., GREILING, R. & VOGEL, A. (eds) The Rhenohercynian and Sub- Variscan Fold Belts. Earth Evolution Series, Vieweg & Sohn, Braunschweig/Wiesbaden, 197224. WARR, L. N. & HECHT, C. A. 1993. A clay mineral investigation of the Upper Carboniferous Culm Basin of south-west England. Proceedings of the Ussher Society, 8(2), 94-98. WARR, L. N. PRIMMER, T. J. & ROBINSON, D. 1991. Variscan very low-grade metamorphism in SW England - a diastathermal and thrust-related origin. Journal of Metamorphic Geology, 9, 751— 764. WHALLEY, J. S. & LLOYD, G. E. 1986. Tectonics of the Bude Formation, north Cornwall - the recognition of northerly directed decollement. Journal of the Geological Society of London, 143, 83-88. WOODCOCK, N. H. 1977. Specification of fabric shapes using an eigenvalue method. Geological Society of America Bulletin, 88, 1231 - 1236. WOODCOCK, N. H. & NAYLOR, M. A. 1983. Randomness testing in three-dimensional orientation data. Journal of Structural Geology, 5, 539-548.

How deformed are weakly deformed mudrocks? Insights from magnetic anisotropy JOSEP M. PARES Department of Geological Sciences, University of Michigan, 2534 C.C. Little Building, Ann Arbor, MI-48109, USA (e-mail: [email protected]) Abstract: For over thirty years the anisotropy of magnetic susceptibility (AMS) has been exploited in rock fabric studies. Our knowledge of the mechanisms leading to magnetic fabric development in deformed rocks has grown substantially, even though some details are still under debate. This paper reviews recent developments in AMS studies applied to the characterization of early deformation stages in mudrocks. From the current data set of AMS in rocks deformed at low temperature and low pressure it is possible to revisit the concept of weakly deformed rocks. We contend that weak deformation in mudrocks and the appearance of magnetic cryptofabric are concatenated. Furthermore, magnetic anisotropy studies in weakly deformed rocks suggest that cleavage fabric, the most common planar fabric in deformed mudrocks, builds up gradually from the earliest stages of deformation rather than suddenly at some strain threshold.

The mechanisms leading to the earliest deformation stages in rocks have intrigued structural geologists for almost a century. Specifically, great effort has been made in addressing the question of when and how mudrocks become weakly deformed and develop cleavage fabric. Mudrocks are one of the most abundant sedimentary rocks and they are very common in areas that are of special interest in structural geology studies such as accretionary prisms and fold-and-thrust belts. Hence it is important to understand the mechanisms by which these rocks deform. By weakly deformed mudrocks is meant flat-lying or gently tilted strata with very incipient deformation that might include a very crude cleavage, perhaps only detectable because of the presence of fissility or a very incipient lineation. In general, early deformation of mudrocks results in a rearrangement of their constituents, or what we know as rock fabric (=the complete spatial and geometrical configuration of all those components that make up a rock, which should be penetrative and repeatedly developed throughout a volume of rock; Hobbs et al. 1976). Evidence for such an early deformation in pelitic rocks is the appearance of secondary foliation or cleavage fabric. Cleavage is without a doubt the most common planar fabric produced during tectonic deformation. Identifying cleavage and hence deformation in rocks is scale dependent though. Since the seminal study by Sorby (1853), the scale and resolution of our observations have broadened, ranging from characteristics that are observable in the field to electron microscopy. Characterization of physical properties such as magnetic

anisotropy offers an approach of determining rock fabric and, under certain circumstances, of finite rock strain on even seemingly undeformed rocks. As a result, we are currently pushing the frontiers of concepts such as 'weak deformation', which is the scope of this paper. The questions that we specifically address include (1) what is the meaning of weak deformation in mudrocks? (2) how weak is the deformation in what we commonly call 'weakly deformed mudrocks'? and (3) how soon in a deformational process, does cleavage fabric develop? Here we present recent results from a number of published studies on AMS, focusing on those obtained in mudrocks, based on which we discuss the potential and limitations of magnetic fabric studies ultimately to define weak deformation in mudrocks.

Scope This paper focuses on the anisotropy of magnetic susceptibility (AMS) properties in mudrocks and therefore, it is necessary to define the meaning of such a lithology. There has been a plethora of terms and qualifiers to describe mudstones (including clay, shale, claystone, siltstone, argillite and so on). Defining clay as particles finer than 1/256 mm diameter and silt as particles with diameters between 1/256 and 1/16 mm, Blatt el al. (1980) provided a useful and simplified terminology based purely on grain size. According to their classification scheme, mud is sediment predominantly (>75%) composed of clay and silt. Mudstone, the indurated equivalent of mud is a blocky, non-fissile rock, whereas

From: MARTIN-HERNANDEZ, F., LUNEBURG, C. M., AUBOURG, C. & JACKSON, M. (eds) 2004. Magnetic Fabric: Methods and Applications. Geological Society, London, Special Publications, 238, 191-203. 0305-8719/04/ $15.00 © The Geological Society of London 2004.

192

J. M. PARES

shale is usually laminated and fissile (fissility being the property of splitting). Argillite is used for a more indurated mudrock and slate for one that possesses cleavage. A sedimentary rock of clay-grade material only is called a claystone, and one that contains more silt-grade particles than clay is called a siltstone. A classification centred on these parameters is useful as grain size (below ~I/16mm) is closely linked to mineralogy and the physical properties of muds. As a general rule, we will follow Tucker's (1985) recommendation and will use the term mudrock for all these rock-types (mudstones, claystone and siltstone). Beyond this, there is no generally accepted classification of mudrocks although several schemes have been put forward (see references in Tucker 1985). From the mineralogical standpoint, mudrocks are composed primarily of clay minerals (~42%) and fine-grained quartz (~37%) (Fig. 1). The four common clay mineral constituents of mudrocks include illite (the commonest of the clay minerals), kaolinite, smectite and chlorite. In addition, the occurrence of mixed-layer clays is quite common. These consist of an interleaving of sheets of the common clays, in particular illitesmectite and chlorite-smectite. The bulk of these paramagnetic clay minerals will thus typically control the total low-field magnetic susceptibility in mudrocks. Quartz (diamagnetic) in mudrocks is mostly of silt-grade although coarser, sandsize, grains may occur. Other constituents may include feldspars, muscovite and chamosite. Organic matter and with it pyrite is common in black shales.

Fig. 1. Histogram showing the average percentage mineral composition of mudrocks of different ages (data from O'Brien & Slatt 1990).

What is 'weak deformation' really? In their seminal work on strain analysis, Ramsay & Huber (1983) described six main stages of fabric development arising from layer parallel tectonic shortening in sediments. The main features of such stages may be summarized as follows: (1) An 'undeformed condition' in sedimentary rocks is characterized by a fabric that is parallel to the bedding plane and develops mostly from the effect of diagenetic compaction on platy minerals present in the initial mud by pressure and water expulsion. In part, the original fabric is also inherited from the hydrodynamic orientation of platy minerals parallel to bedding. (2) The 'earliest deformation stage' is usually accompanied by considerable volume loss as a result of closure of pore spaces and expulsion of pore water. The elongation in the X direction is small compared to the contraction in the Z direction. The rotation of platy grains around the Y direction makes the plane bedding fabric less pronounced. It is important to remark that the rock at this stage has a very weak linear fabric. (3) With increasing tectonic strain the ellipsoid takes on a more prolate form with the longest axis being parallel to the Y direction. Mineral rotations are sufficient at this stage to produce a fabric strong enough to impart the well-known 'pencil structure'. (4) Further increasing strain leads to the progressive movement of the ellipsoid into the flattening field. Mineral rotations lead to the production of an imperfectly formed cleavage crossing bedding, called embryonic cleavage. (5) A cleavage stage is achieved when the tectonic strain imprint is sufficiently strong to form a dominant planar cleavage fabric, where platy and acicular minerals are oriented in this cleavage plane. (6) With increasing strain, the planar cleavage and lineation are intensified, this last one very close to or parallel to the strain X direction. We would like to point out that in this evolutionary strain path for mudrocks, the so-called 'earliest deformation stage' involves usually a weak linear fabric (Ramsay & Huber 1983). As we will show later, the advantage of AMS is precisely its ability to reveal very weak tectonic fabric that often predates this stage called 'earliest deformation'. Interestingly, this deformation in mudrocks, as we shall see later, is

AMS AND DEFORMATION IN MUDROCKS

revealed by a (weak) magnetic foliation rather than lineation. Such a realm of incipient deformation in mudrocks is the main focus of the present study. Cleavage, AMS and Strain The term cleavage repeatedly appears in AMS literature and has led to some confusion, particularly as far as the strain magnitude that a given fabric conveys. In order to preserve objectivity

193

in the usage of the term, it is better to give the term 'cleavage' descriptive and morphological attributes rather than genetic (e.g. Dennis 1967). Thus the definition of Dennis (1967) for 'rock cleavage' as a 'secondary planar fabric elements which impart mechanical anisotropy to the rock without apparent loss of cohesion' is very useful. Cleavage results from a coherent rearrangement and realignment of minerals and therefore is essentially ductile rather than brittle (see also Durney & Kisch 1994). In Figure 2 I represent the main cleavage types observable in

Fig. 2. Cleavage types observable in mudrocks and relationship to the magnetic ellipsoid. Clay and slaty cleavage types are penetrative, with mineral alignment down to the dominant grain size, compraising nondomainal (clay t.) and domainal (slaty t.) phyllosilicate rich types. Spaced type cleavage is non-penetrative and has domainal mineral alignment spaced wider than the dominant grain size. Scaly type is a non-penetrative structure with smooth and highly polished or slickensided discrete surfaces that intersect or merge with one another to yield lensoid flakes (Durney & Kisch 1994). The width and spacing between bands of greater and lesser concentration of cleavage domains and immobile minerals will determine, to a large extent, the magnetic fabric ellipsoid. Open circles on the cleavage type diagrams represent the sample volume measured relative to the cleavage spacing. For each cleavage type, the orientation of the corresponding Km[n (dots) and Kmax (squares) principal axes are shown on stereonet and the shape (7) and anisotropy degree (Pr) of the ellipsoid are represented on a Pf-T diagram.

194

J. M. PARES

mudrocks (based on the nomenclature of Durney & Kisch 1994) in relationship to their corresponding magnetic ellipsoid. Such a classification is field-based and hence a very convenient way to categorize cleavage fabric in mudrocks. Clay and slaty cleavage are penetrative cleavage types, with mineral alignment down to the dominant grain size. Clay type is non-domainal and has also been called clay foliation. Slaty type is by far the most common and is domainal and phyllosilicate rich. Spaced-type cleavage is non-penetrative and has domainal mineral alignment spaced wider than the dominant grain size. Scaly type is a non-penetrative structure with smooth and highly polished or slickensided discrete surfaces that intersect or merge with one another to yield lensoid flakes. Notice that the AMS ellipsoid properties will depend in part on the ratio between the sample volume used and the cleavage domain size (Fig. 2). A number of mechanisms can contribute to development of such cleavage fabric types in mudrocks. The processes leading to cleavage formation and hence to magnetic anisotropy in mudrocks include: (a) rotation of platy and acicular minerals (similarly to passive markers); (b) grain growth of micas parallel to (001); (c) restricted growth parallel to platy minerals (Passchier & Trouw 1998); and (d) pressuresolution. Although this last mechanism is not a direct process for phyllosilicate preferred orientation, it results in passive reorientation of grains into cleavage domains as matrix minerals are progressively removed (e.g. Borradaile & Tarling 1981). A number of studies show that below the anchimetamorphic field (<200°C) process (a) will largely control cleavage development in mudrocks (e.g. Kanagawa 1991). Rotation of platy minerals will lead to cleavage development and therefore anisotropy of magnetic susceptibility (AMS) in mudrocks provides a bulk measure of rock fabric and deformation intensity. Low field magnetic susceptibility (K) of a rock (the ratio of magnetization [M] to the applied field [H], or K — M/H) is given by the total contribution of its bulk mineralogy, including paramagnetic (e.g. phyllosilicates, ironbearing silicates), diamagnetic (e.g. quartz, calcite) and ferromagnetic sensu la to (e.g. magnetite, goethite, hematite) grains. An intrinsic property of such rock-forming minerals is that their magnetic susceptibility is anisotropic (Nye 1957); i.e. K fj = M//H,-. Consequently, the the magnetic susceptibility of virtually all rock types is anisotropic. The anisotropy of magnetic susceptibility (AMS) defines a symmetric, second-rank tensor that has six independent

matrix elements. These elements trace an ellipsoid that is called the magnitude ellipsoid (Nye 1957), whose semi-axes are the three principal susceptibilities (maximum, intermediate and minimum susceptibility axes, or ^max > Ain, > *min) (Hrouda 1982). Although some details of the mechanisms controlling the development of cleavage and hence AMS are still under debate, there is a general consensus that the magnetic ellipsoid is related to the finite strain state. The magnetic foliation of the AMS ellipsoid, i.e. the plane containing KmSLX and j£int axes, forms perpendicular to the shortest axis (Z direction) of the finite strain ellipsoid and increases in intensity with the strain ratio Rxz (strain ratio X: Z), just like cleavage does. Since the 1800s (e.g. Sorby 1853) all work on cleavage that employed strain measurement techniques has supported such results (see Ramsay & Huber 1983). Therefore, AMS has been used to characterize finite strain in a variety of tectonic settings, ranging from low to very high strain rocks (e.g. Goldstein 1980; Rathore 1980; Kligfield et al. 1981; Rathore et al 1983; Siddans et al 1984; Cogne & Perroud 1988; Goldstein & Brown 1988; Ruf et al 1988; Aubourg et al 1991; Housen & van der Pluijm 1991; Hirt et al 1993; Sagnotti & Speranza 1993; Averbuch et al 1995; Housen et al 1995; Aranguren et al 1996; Borradaile & Henry 1997; Saint-Blanquat & Tikoff 1997; Luneburg et al 1999; Pares & van der Pluijm 20020). In most deformation environments, the AMS axes show a good correlation with the orientation of the principal strain directions (see reviews by Hrouda 1982; Borradaile 1987, 1988, 1991; Tarling & Hrouda 1993; Borradaile & Henry 1997). When cleavage is present, the principal magnetic susceptibility directions parallel the flattening plane of the finite-strain ellipsoid, with the minimum susceptibility perpendicular to cleavage whereas the maximum susceptibility or magnetic lineation typically parallels either the tectonic extension direction or the intersection of bedding and cleavage (Singh et al 1975; Coward & Whalley 1979; Kligfield et al 1981; Hrouda 1982; Borradaile 1987; Borradaile & Henry 1997; Pares et al 2001; Pares & van der Pluijm 20020, 2003). A unique advantage of AMS over other techniques is that the magnetic ellipsoid orientation also reflects strain in very weakly deformed rocks, where penetrative fabric markers are often absent in hand specimens or at outcrop scale (Kissel et al 1986; Pares & Dinares-Turell 1993; Sagnotti & Speranza 1993; Aubourg et al 1995; Sagnotti et al 1994; Pares et al 1999). This aspect will be discussed in the following section.

AMS AND DEFORMATION IN MUDROCKS Although qualitative correlations between strain and magnetic ellipsoids have been widely explored, there has been considerable debate about the quantitative relationship between AMS and strain, as the magnitudes of magnetic fabrics and strain are more complexly related. Among the studies focusing on AMS-strain correlations, there are two main approaches. One is numerical modelling that simulates magnetic fabric development (Owens 1974; Richter 1992; Benn 1994). The second includes empirical correlations between strain markers and susceptibility ellipsoids (e.g. Rathore 1980; Kligfield et al 1981, 1982; Borradaile & Tarling 1984; Hirt et al. 1988; Borradaile 1991; Pares & van der Pluijm 2003). Difficulties in correlating AMS and strain arise from: (a) an incomplete understanding of the relationship between the sources of magnetic susceptibility and the AMS ellipsoid; (b) recrystallization of AMS-carriers; and (c) lack of a standard method to correct for compositional variations. What follows from the existing results is that AMS-strain correlations require (a) a knowledge of the initial predeformational fabric ellipsoid (magnitude and orientation); (b) understanding of the rockmagnetic mineralogy; and (c) absence of recrystallization. Most AMS-strain correlations are based on the predictions of the March model for the relationship between linear-, linear-planar- and planar-fabrics and the strain ellipsoid (e.g. Borradaile 1991; Richter 1992). The strain response model of March (1932) considers platy or acicular elements that are passively reoriented in response to progressive deformation. Grains in this model are assumed to be mechanically indistinguishable from the rock matrix in which they are embedded. Even though both conditions are unlikely to occur and that the March model does not account for all the features of fabric development, it can be used to calculate magnetic ellipsoids with realistic properties. According to March's model, the application of strain to an ensemble of marker planes increases their pole density. The formula that relates the final pole densities in the principal directions of the strain to an original unitary and uniform density is given by £,• = p^ where pt is a principal pole density, normalized by dividing it by the average pole density for all orientations, and et is a principal strain, expressed as the change of length divided by the original length. The behaviour of such rigid particles in a deforming system is very complex, mostly because of the degree of packing of large and small components, grain size and strain sequence (Ramsay & Huber 1983). In addition, and crucial for AMS analysis,

195

processes such as pressure-solution and crystals that kink, bend or that interfere with one another, lead to deviations from the March model. For example, Sintubin (1994a, b) notes that the degree of clay particle preferred orientation based on the March model reflects not only strain but also the presence of non-platy particles and the relative concentration of the different clay minerals. In short, despite great efforts that have been made in numerous studies to use AMS as a strain gauge (e.g. Singh et al. 1975; Coward & Whalley 1979; Goldstein 1980; Kligfield et al. 1981; Hrouda 1982; Rathore et al. 1983; Siddans et al. 1984; Kissel et al. 1986; Borradaile 1987, 1988, 1991; Cogne & Perroud 1988; Aubourg et al. 1991; Housen & van der Pluijm 1990a,6, 1991; Hirt et al. 1993; Tarling & Hrouda 1993; Sagnotti & Speranza 1993; Pares & DinaresTurell 1993; Sagnotti et al. 1994; Averbuch et al. 1995; Housen et al. 1995; Aranguren et al. 1996; Saint-Blanquat & Tikoff 1997; Borradaile & Henry 1997; Liineburg et al. 1999; Pares & van der Pluijm 2003) the quantitative relationship between AMS and strain remains rather poorly understood. AMS in 'weakly deformed rocks' The current proliferation of magnetic fabric literature is to a large extent due to both the development of highly sensitive and rapid measuring devices and a better understanding of the AMS tensor, including methods for separation of subfabrics (e.g. Rochette & Pillion, 1988; Richter & van der Pluijm 1994; Martin-Hernandez & Hirt 2001; Pares & van der Pluijm 20026; Kelso et al. 2002). Specifically, an increasing number of studies are revealing that apparently undeformed mudrocks contain a subtle, yet identifiable tectonic magnetic fabric. Because AMS enables the measurement of weak although not macroscopically visible tectonic fabrics, many rocks previously thought to be undeformed actually indicate weak deformation fabrics (e.g. Kissel et al. 1986; Sagnotti & Speranza 1993; Pares & Dinares-Turell; 1993; Sagnotti et al. 1994; Pares et al. 1999). Notably, AMS reveals evidence that mudrocks that are flat-lying and without mesoscopic cleavage yield magnetic fabrics that is tectonic in origin, as revealed by mesoscopic or regional structures (trend of fold axes and thrust faults). From a number of studies (Kissel et al. 1986; Sagnotti & Speranza 1993; Pares & Dinares-Turell; 1993; Sagnotti et al. 1994; Pares et al. 1999) it follows that the earliest stage of deformation in mudrocks from such settings is

196

J. M. PARES

revealed by a weak planar magnetic fabric, although it has no macroscopic morphological expression. For example, in the Pyrenean Foreland Basin, Pares et al. (1999) reported development in seemingly undeformed mudrocks of magnetic fabric that is tectonic in origin. Even though mudstones are not cleaved, regional structures (folding) in the area show a shortening direction that conforms the magnetic fabric ellipsoid orientation that is observed. Larrasoana et al. (this volume) describe analogous results obtained in mudstones from the western Pyrenees. Similar conclusions were also reached by Kissel et al. (1986) in the Taiwan accretionary prism. Overall, naturally deformed rocks conform experiments and mathematical modelling (e.g. Housen et al. 1993). Because cleavage is typically established on the basis of micro- or mesoscopic scales of observation, early stages of deformation that did not lead to a planar fabric development that is recognizable at those scales (like the studies mentioned above) have obviously remained unnoticed in the field in many cases. So far, embryonic cleavage or pre-pencil cleavage has been advocated by Durney & Kisch (1994) to be the earliest type of incipient cleavage, but no other planar or linear fabric has been suggested to develop before that deformation stage. It is worth recalling that the early stage of deformation of Ramsay & Huber (1983) is commonly associated with a lineation. AMS analysis described above reveals that there is indeed evidence for deformation even before the appearance of embryonic cleavage or the 'early deformation stage' of Ramsay & Huber (1983) and that the corresponding magnetic ellipsoid is oblate, as opposed to the tectonic lineation noticed by these authors. Magnetic anisotropy obtained in mudrocks that appear undeformed shows an incipient planar fabric that results from very subtle realignment of grains at very early stages of deformation, perhaps even before the total lithification of the mudstones (e.g. Knipe 1986; Pares et al. 1999; Larrasoana et al., this volume). Such a nascent tectonic fabric or cryptofabric, precursor to incipient cleavage, is characterized by relatively low values of anisotropy degree (typically P' < 1.05). Particulate flow, mostly by intergranular slip and kinking (Borradaile 1981; Oertel 1983; Hounslow 1990; Bhatia & Soliman 1991; Owens 1993; van der Pluijm et al. 1998) is most likely the dominant mechanism for reorienting phyllosilicate grains at this stage of deformation. With increasing strain, changes in the degree of preferred grain orientation often produce pencil structure. At higher

Fig. 3. Model for anisotropy of magnetic susceptibility ellipsoid distribution as a function of magnetite content. AMS ellipsoid for magnetite is represented as a white square (Pf = 1.18, T — —0.30; Tarling & Hrouda 1993). A, B, C and D show data from four different hypothetical sites, where the AMS tensor is mostly due to phyllosilicates. Samples form site A show an elongated distribution, whose best fitting line (or long axis of the distribution) goes through the AMS ellipsoid of magnetite. Such a stretched AMS distribution towards magnetite suggests that variations in concentration of this mineral control the site AMS distribution. Site B shows a more homogeneous distribution of data points and no visible effect of magnetite variations are observed. Site C has elongated distribution, but the best fitting line does not include the values of Pf and T of magnetite, suggesting that the AMS variations at site level reflect changes in preferred grain orientation. Site D has elongated distribution, and the best fitting line falls near magnetite, indicating a possible effect of magnetite content.

strain bedding is eventually obliterated and cleavage becomes the main planar tectonic fabric. Such a cleavage fabric development in mudrocks results from progressive grain preferred orientation, and its relative intensity can to a first approximation be quantified by the anisotropy degree of the magnetic anisotropy, or the ratio ^max/^min (or its more sophisticated equivalent P'see Appendix), as discussed later. Mudrocks composition and AMS We have shown earlier that clay minerals and quartz constitute ~90% of the composition of mudrocks. Quartz is diamagnetic and hence has a negative and very low susceptibility, negligible for most AMS applications on deformation intensity in mudrocks. On the other hand, clay minerals are paramagnetic and typically dominate the magnetic anisotropy tensor in most mudrocks. Most mudrocks reviewed in this

AMS AND DEFORMATION IN MUDROCKS

paper contain magnetite as the main ferromagnetic phase. Magnetite has much higher magnetic susceptibility than phyllosilicates (magnetite, Kf ~ 0.07 to 20 SI; phyllosilicates, Kp = 122 to 1500 x 10~6 SI) and hence variations in concentration can produce changes in the magnetic ellipsoid (e.g. Borradaile 1987; Rochette et al.9 1992). In our view the concentration of magnetite in mudrocks is usually so low that it will barely affect the bulk anisotropy. Our rationale is as follows: The average value for the saturation magnetic moment in mudrocks is 30 to 300 x 10~8 Am2, suggesting a magnetite concentration of 0.2 to 2 x 10~3%, assuming a magnetite grain size around 1 um (O'Reilly 1984). The corresponding volume susceptibility for this percentage of magnetite is ~10 x 10~6 SI, which is less than 3% of the bulk susceptibility in most mudrocks. In practice, the effects of mineralogical changes on the AMS ellipsoid can be checked by inspecting the dispersion of the T and Pf parameters (see appendix for definitions). On the so-called Jelinek plot (P1 versus 7), a group of data points for a given site will have a dispersion that will reflect both the intrinsic magnetic fabric (i.e. the geometric arrangement of the particles contributing to the AMS tensor) and changes in the relative contribution of ferromagnetic and paramagnetic minerals. Figure 3

197

illustrates four possible scenarios of changes in the AMS ellipsoid, expressed in terms of the Pf and T parameters, due to the different amounts of ferromagnetic minerals. The possible effects of magnetite concentration changes on the magnetic ellipsoid can be evaluated by inspecting the distribution or dispersion of Pf and T values. Data points (P;, T) will tend to align towards the magnetite ellipsoid if such a mineral contributes significantly to the total fabric. An elongated distribution of Pf—T data towards magnetite will most likely reflect a variable amount of magnetite rather than changes in fabric intensity. Therefore, when comparing different sites among the same rock formation (e.g. determining deformation intensity across a transect) or the distribution at site level, particular attention is needed in identifying elongated distributions of P'-T towards the magnetite ellipsoid (Fig. 4). Changes in bulk magnetic susceptibility are an additional indicator for possible content of a significant ferromagnetic contribution. Ideally, at site level, rock composition should be rather uniform and so a fairly constant Kf/Kp ratio is expected. In such a case, one would expect that changes in the AMS ellipsoid (Pf, T) respond to preferred grain orientation and not to compositional variations. The usage of a plot comparing

Fig. 4. Single-mineral model of rock magnetic anisotropy versus degree of mineral alignment (modified from Pares & van der Pluijm 20020). For a rock whose anisotropy is carried by a mineral with 1.33 < P1 < 1.40, the degree of crystalline alignment can be directly determined from the degree of magnetic anisotropy Pf as indicated by the envelope (biotite: 1.33; chlorite: 1.36; muscovite: 1.40). Case studies include only pelitic rocks, where phyllosilicates have been reported to be the main carriers of AMS and those where the anisotropy degree P1 has been calculated (or the necessary data to calculate is provided): 1 - Aubourg et al. (1991); 2 - Pyrenean Foreland Basin mudrocks (Pares et al 1999); 3 - Sagnotti et al (1999); 4 - Sagnotti & Speranza (1993); 5 - Mattei et al (1999); 6 - Lamarche & Rochette (1987); 7 - Paleozoic slates with stretching lineation (Pares & van der Pluijm 20020); 8 - Hirt et al (2000); 9 - Knobs Formation pencils (Pares & van der Pluijm 2003); 10 - Cap de Creus mylonites (Pares & van der Pluijm 20020); 11 - Liineburg et al (1999).

198

J. M. PARES

low field bulk susceptibility with saturation magnetization to test whether rock composition is controlling changes in the AMS ellipsoid is highly recommended. Weak deformation realm in the models for AMS development To better illustrate the relationship between AMS and intensity of grain preferred orientation, Tarling & Hrouda (1993) estimated the degree of anisotropy (P1) with the degree of grain alignment (Fig. 4) for a group of common phyllosilicate minerals. Progressive grain reorientation and its effect on the magnetic anisotropy are simulated by the anisotropic properties of a single crystalline grain, derived from a summation of the various orientations of individual grains (see Owens 1974, and Hrouda & Schulmann 1990 for details). Therefore the degree of magnetic anisotropy of a rock whose anisotropy is carried by a single mineral can theoretically provide an estimate of the degree of mineral alignment (Hrouda & Schulmann 1990; Richter 1992; Tarling & Hrouda 1993). To first order, as the degree of strain increases, the intensity of preferred orientation increases and hence a plot of P7 versus degree of alignment offers a basis for comparing relative deformation intensity (Fig. 4). Note that the degree of alignment in a given rock cannot exceed the magnetic value when all grains of that mineral are perfectly aligned. For example, a rock whose magnetic fabric is solely due to muscovite will attain a maximum anisotropy degree Pf of 1.40 when all muscovite grains are perfectly aligned. Although the model is based on a firm assumption, it presumes that the original rock is almost isotropic, i.e. P' = 1, which is very unlikely. AMS studies in mudrocks (i.e. phyllosilicates are the main susceptibility contributors) show that initially magnetically isotropic rocks occur but are very unlikely in nature (Hounslow 1985; Sintubin 19940; Bouchez 1997). Although flocculation, which is a common pattern in clays, depends on the load and the water chemistry, mechanical working during emplacement will tend to convert any flocculated microfabric towards a more remoulded distribution. As burial and consolidation proceed, increasingly better-aligned clay domains will develop and the increase of vertical stress (load) will produce an increasing horizontal alignment of platy clay grains. Even so, the overall particle alignment does not typically reach very high levels of packing (Bennett et al 1991). In consolidation experiments on pure

clays, an increasing load produces higher clay particle orientation, but most of it occurs in the very early stages of loading, while the material is still a slurry. Other studies of particle orientation of natural sediments as a function of burial depth show little increase in orientation beyond a few tens of metres of burial (e.g. Bennett et al. 1991). Once a remoulded distribution is achieved, well-aligned mineral particles will slip relatively easily. The presence of such an initial anisotropy that is pre-deformational poses an additional challenge to the AMS-strain models and has barely been explored (Benn 1994; Pares & van der Pluijm 20020; Housen et al. 1993). In this sense, our plot of deformation intensity as revealed by progressive grain reorientation (Fig. 4) is an oversimplification as it assumes an original isotropic distribution (Pf — 1). Rather than using the anisotropy degree only for quantifying strain our approach is to consider both shape parameter T and anisotropy degree Pf. A large number of AMS studies from naturally deformed rocks result in a gross grouping of both anisotropy degree Pr and shape parameter T values as a function of the dominant fabric (Fig. 5a). For the vast majority of deformed rocks, notice that the anisotropy degree ranges from 1.05 to 2.50 (Fig. 5a). Overall, both shape (T parameter) and degree of anisotropy (P parameter) crudely correspond to strain, i.e. low, intermediate and high values of P correspond respectively to cleavage domains of clay/spaced type, slate and mylonitic foliation. From such a distribution of P' and 7" data, it is possible to infer a general path of the anisotropy parameters with strain. Figure 5b shows a synopsis of the development of magnetic anisotropy with increasing deformation intensity. As an example, we show some of our own studies that cover a particular region of the Pf-T strain evolution diagram. Also, the observed trend is compared to a numerical model developed by Benn (1994) (Fig. 5c). Note that for axially symmetric shortening (X = Y > Z), the trend of the shape parameter conforms observations from naturally deformed mudrocks (Fig. 5c). We notice that two factors will determine the details of the magnetic ellipsoid evolution during deformation. First, the degree of anisotropy that is present before the onset of deformation (typically a sedimentary magnetic foliation). Second the angle between the original magnetic ellipsoid relative to the strain ellipsoid. Numerous mudrocks with a degree of anisotropy of 1.05 or lower (Fig. 4) show magnetic ellipsoids whose axes of orientation reflect the magnetic cryptofabric described earlier. In this

AMS AND DEFORMATION IN MUDROCKS

199

Fig. 5. (a) Plot of the shape parameter (J) versus degree of anisotropy (P7) (Jelinek 1981) for various cleavage types. Shaded regions include AMS studies where both T and Pf are provided or else the necessary variables to compute them. Those include Rathore et al (1983); Borradaile & Mothersill (1984); Rochette & Vialon (1984); Lamarche & Rochette (1987); Goldstein & Brown (1988); Housen & van der Pluijm (1990); Aubourg et al. (1991); Sagnotti & Speranza (1993); Sagnotti et al. (1994, 1999); Averbuch et al. (1995); Housen et al. (1995); Aranguren et al. (1996); Alfonsi & Sagnotti (1996); Bakhtari et al. (1998); Pares et al. (1999); Hirt et al. (2000). (b) Synopsis of the evolution of magnetic parameters T and Pf with strain in mudrocks, from compaction to well-developed tectonite. Encircled areas show regions of the P'-T path that we have studied and that provide good examples of the magnetic ellipsoid and its relation to strain: (1) Pares et al. (1999); (2) Joseph et al. 1998; (3) Pares & van der Pluijm (2003); (4) Pares et al. (2001). (c) Evolution of the shape parameter T of the AMS ellipsoid under progressive strain for axially symmetric shortening (Benn 1994). The shortening axis is near-perpendicular to the initial magnetic lineation (see discussion in text). deformation realm, K^m axes are normal to bedding and Kmax axes cluster perpendicular shortening direction. Overall, the magnetic ellipsoid is oblate (0 < T < 4-1) although the magnetic foliation is weak. Most mudrocks that fall in this category do not display cleavage micro- or macroscopically at this stage. Such a magnetic cryptofabric is due to an incipient preferred grain orientation, perhaps by reorientation of grains in discrete zones or stripes (e.g. Durney & Kisch 1994), enough to develop an intersection

axis, which is ultimately the origin of the magnetic lineation or cluster of Km^ axes.

Final remarks AMS studies on mudrocks reveal a magnetic cryptofabric that predates incipient cleavage formation (and the 'early deformation stage') even though it does not lead to fissility or any other rock anisotropy that has macro- or microscopic

200

J. M. PARES

recognition. If we use the traditional definition of cleavage (as a secondary meso- or microscopically observable planar structure, e.g. Dennis 1967), then magnetic cryptofabric cannot be considered as cleavage fabric. Nevertheless, given that the magnetic ellipsoid senses rock anisotropy that is tectonic in origin, we should certainly consider the observed magnetic cryptofabric as cleavage. Such a cryptofabric (characterized by an oblate magnetic ellipsoid and a cluster of Kmax axes) is the earliest evidence for weak deformation in mudrocks, predating the embryonic cleavage type of Durney & Kisch (1994) and the earliest deformation stage of Ramsay & Huber (1983). AMS ellipsoid and strain data show a crude correlation in mudrocks deformed under a variety of tectonic environments. With increasing strain, magnetic ellipsoids grade from the oblate field (sedimentary fabric) towards the prolate field (where both flattening and bedding plane compete) while the anisotropy degree slightly decreases (Fig. 5b). At higher strains, the eccentricity of the magnetic ellipsoid increases. Under certain circumstances, a quantitative correlation between AMS and strain is possible (e.g. Borradaile 1991; Hirt et al 1993; Pares & van der Pluijm 2003). Notably such circumstances include homogeneous lithology of the rock formation, and absence of recrystallization or intracrystalline plasticity. Durney & Kisch (1994) pointed out that even though cleavage appears to be closely related to strain, it only seems to appear above some finite strain threshold (e.g. -35%, Wood 1974). Magnetic anisotropy of a variety of seemingly undeformed mudrocks suggests that cleavage fabric builds up gradually from the earliest stages of deformation (e.g. Kligfield et al. 1983) rather than suddenly at some strain threshold. This has important implications in our understanding of deformation processes and strain in tectonic regions where deformation occurs shortly after deposition, before the final lithification of mudrocks (e.g. accretionary prisms). A number of colleagues have enhanced our understanding by their generous discussions, especially B. Housen, J. Stamatakos and B. van der Pluijm. Reviews by C. Richter and M. Sintubin and the editor C. Liineburg led to improvements in the manuscript. Research supported by NSF Grant EAR9814343.

Appendix We summarize as follows the most commonly used magnetic parameters in this paper:

(a) Corrected anisotropy degree The so-called corrected anisotropy degree (Pf) was suggested by Jelinek (1980) to include the value of the axis of intermediate susceptibility in the calculation of the anisotropy degree (P = Kmax/Kmin). Instead of the principal susceptibilities, Jelinek (1980) considers their logarithms T?}, rj2, 773 and the scatter of the latter. As a measure of the scatter of the logarithms of the principal susceptibilities it is suggested the factor P>:

where 17= fa +% + ^3)/3. The corrected anisotropy degree has values that typically range from 1 (for a sphere) to ~1.4 (micas) (b) Shape parameter The so-called shape parameter T provides a quantitative measure of the susceptibility ellipsoid shape. Jelinek (1980) proposed the T parameter inspired on the strain ellipsoid shape of Ramsay (1967) (R = ln(A 2 /A 3 )/ln(A 1 /A 2 ). The T parameter has an important advantage, namely that it ranges from — 1 to +1:

where foliation F = K-mi/K^m

and lineation

L = ^max/^int-

Therefore, the values of the shape parameter allow separating prolate (-1 < T < 0) from oblate (0 < T < 4-1) magnetic ellipsoids. References ALFONSI, L. & SAGNOTTI, L. 1996. Magnetic fabric of Plio-Pleistocene clayey sediments from the footwall of the Avezzano earthquake fault (Central Apennines, Italy). // Quaternario, 9, 145-154. ARANGUREN, A., CUEVAS, J. & TUBIA, J. M. 1996. Composite magnetic fabrics from S-C mylonites. Journal of Structural Geology, 18, 863-869. AUBOURG, C., ROCHETTE, P. & VIALON, P. 1991. Subtle stretching lineation revealed by magnetic fabric of Callovian-Oxfordian black shales (French Alps). Tectonophysics, 185, 211-223. AUBOURG, C., ROCHETTE, P. & BERGMULLER, F. 1995. Composite magnetic fabric in weakly deformed black shales. Tectonophysics, 87, 267-278. AVERBUCH, O., MATTEI, M., KISSEL, C., FRIZON DE LAMOTTE, D. & SPERANZA, F. 1995. Cinematique des deformations au sein d'un systeme chevauchant aveugle: Fexemple de la 'Montagna dei Fiori' (front des Apenins centraux, Italic). Bulletin de la Societe Geologique de France, 166(5), 451— 461.

AMS AND DEFORMATION IN MUDROCKS BAKHTARI, H. R., FRIZON DE LAMOTTE, D., AUBOURG, C. & HASSANZADEH, J. 1998. Magnetic fabrics of Tertiary sandstones from the Arc of Pars (Eastern Zagros, Iran). Tectonophysics, 284, 299316. BENN, K. 1994. Overprinting of magnetic fabrics in granites by small strains: numerical modeling. Tectonophysics, 233, 153-162. BENNETT, R. H., BRYANT, W. R. & HULBERT, M. H. 1991. Microstructures of Fine Grained Sediments from Mud to Shale. Springer, New York. BHATIA, S. K. & SOLIMAN, A. 1991. The application of image analysis techniques to microstructure studies in geotechnical engineering. In: BENNETT, R. H., BRYANT, W. R. & HULBERT, M. H. (eds). Microstructure of Fine-Grained Sediments. Springer-Verlag, New York, 367-378. BLATT, H., MIDDLETON, G. V. & MURRAY, R. C. 1980. Origin of Sedimentary Rocks, Prentice-Hall, New Jersey. BORRADAILE, G. J. 1981. Particulate flow of rock and the formation of cleavage. Tectonophysics, 72, 305-321. BORRADAILE, G. 1987. Anisotropy of magnetic susceptibility: rock composition versus strain. Tectonophysics, 138, 327-329. BORRADAILE, G. J. 1988. Magnetic susceptibility, petrofabrics and strain. Tectonophysics, 156, 1-20. BORRADAILE, G. J. 1991. Correlation of strain with anisotropy of magnetic susceptibility (AMS). Pure and Applied Geophysics, 135, 15-29. BORRADAILE, G. J. & HENRY, B. 1997. Tectonic applications of magnetic susceptibility and its anisotropy. Earth Sci. Rev. 42, 49-93. BORRADAILE, G. J. & MOTHERSILL, J. S. 1984. Coaxial deformed and magnetic fabrics without simply correlated magnitudes of principal values. Physics of the Earth and Planetary Interiors 35, 294-300. BORRADAILE, G. J. & TARLING, D. H. 1981. The influence of deformation mechanisms on magnetic fabrics in weakly deformed rocks. Tectonophysics, 77, 151-168. BORRADAILE, G. J. & TARLING, D. H. 1984. Strain partioning and magnetic fabrics in particulate flow. Canadian Journal of Earth Sciences, 21, 694-697. BOUCHEZ, J. L. 1997. Granite is never isotropic: an introduction to AMS studies of granitic rocks. In: BOUCHEZ, J. L., HUTTON, D. H. W. & STEPHENS, W. E. (eds). Granite: From Segregation of Melt to Emplacement Fabrics. Kluwer Academic Publishers, Dordrecht, 95-112. COGNE, J. P. & PERROUD, H. 1988. Anisotropy of magnetic susceptibility as a strain gauge in the Flamanville granite, NW France. Physics of the Earth and Planetary Interiors, 51, 264-270. COWARD, M. P. & WHALLEY, J. S. 1979. Texture and fabric studies across the Kishorn Nappe, near Kyle of Lochalsh, Western Scotland. Journal of Structural Geology, 1, 259-273. DENNIS, J. G. 1967. International Tectonic Dictionary. American Association of Petroleum Geologists Memoir, 7, Tulsa. DURNEY, D. W. & KISCH, H. 1994. A field classification and intensity scale for first-generation cleavages.

201

Journal of Australian Geology and Geophysics, 15, 257-295. FERRE, E. C., THILL, J. W., RAINEY, E. S. G. & TEYSSIER, C. 2000. Ductile flow in migmatites deduced from combined low- and high-field AMS measurements. Eos Transactions, AGU, 81, 48: F367. GOLDSTEIN, A. G. 1980, Magnetic susceptibility anisotropy of mylonites from the Lake Char mylonite zone, southeastern New England. Tectonophysics, 66, 197-211. GOLDSTEIN, A. G. & BROWN, L. L. 1988. Magnetic susceptibility anisotropy of mylonites from the Brevard Zone, North Carolina, USA. Physics of the Earth and Planetary Interiors, 51, 290-300. HIRT, A., LOWRIE, W., CLENDENEN, W. S. & KLIGFIELD, R. 1988. The correlation of magnetic anisotropy with strain in the Chelmsford Formation of the Sudbury Basin, Ontario. Tectonophysics, 145, 177-189. HIRT, A., LOWRIE, W., CLENDENEN, W. S. & KLIGFIELD, R. 1993. Correlation of strain and the anisotropy of magnetic susceptibility in the Onaping Formation: evidence for a near-circular origin of the Sudbury basin. Tectonophysics, 225, 231-254. HIRT, A., JULIVERT, M. & SOLDEVILA, J. 2000. Magnetic fabric and deformation in the Navia-Alto Sil slate belt, northwestern Spain. Tectonophysics, 320, 1-16. HOBBS, B. E., MEANS, W. D. & WILLIAMS, P. F. 1976. An Outline of Structural Geology, Wiley International, New York. HOUNSLOW, M. W. 1985. Magnetic fabric arising from paramagnetic phyllosilicate minerals in mudrocks. Journal of the Geological Society of London, 142, 995-1006. HOUNSLOW, M. W. 1990. Grain fabric measured using magnetic susceptibility anisotropy in deformed sediments of the Barbados accretionary prism: Leg 110. Proceedings of the Ocean Drilling Program, Scientific Results, 1001, Ocean Drilling Program, 257-275. HOUSEN, B. A. & VAN DER PLUUM, B. A. 1990fl. Slaty cleavage development and magnetic Anisotropy fabrics. Journal of Geophysical Research, 96, 9937-9946. HOUSEN, B. A. & VAN DER PLUUM, B. A. 19905. Chlorite control of correlations between strain and anisotropy of magnetic susceptibilty. Physics of the Earth and Planetary Interiors, 61, 315-323. HOUSEN, B. A. & VAN DER PLUUM, B. A. 1991. Slaty cleavage development and magnetic anisotropy fabrics (AMS and ARM A). Journal of Geophysical Research, 96, 9937-9946. HOUSEN, B. A., RICHTER, C. & VAN DER PLUUM, B. A. 1993. Composite magnetic anisotropy fabrics: experiments, numerical models, and implications for the quantification of rock fabrics. Tectonophyscis, 200, 1-12. HOUSEN, B. A., VAN DER PLUUM, B. A. & ESSENE, E. J. 1995. Plastic behaviour of magnetite and high strains obtained from magnetic fabrics in the Parry Sound shear zone, Ontario, Grenville Province. Journal of Structural Geology, 17, 265278.

202

J. M. PARES

HROUDA, F. 1982. Magnetic anisotropy of rocks and its application in geology and geophysics. Geophysical Surveys, 5, 37-82. HROUDA, F. & SCHULMANN, K.1990. Conversion of the magnetic susceptibility tensor into the orientation tensor in some rocks. Physics of the Earth and Planetary Interiors, 63, 11-11. jELiNEK, V. 1981. Characterization of the magnetic fabric of rocks. Tectonophysics, 79, T63-T67. JOSEPH, L, H., REA, D. K. & VAN DER PLUIJM, B. A. 1998. Use of grain size and magnetic fabric analyses to distinguish among depositional environments. Paleoceanography, 13, 491-501. KANAGAWA, K. 1991. Change in dominant mechanisms for phyllosilicate preferred orientation during cleavage development in the Kitakami slates of NE Japan. Journal of Structural Geology, 13, 927-945. KELSO, P. R., TIKOFF, B., JACKSON, M. & SUN, W. 2002. A new method for the separation of paramagnetic and ferromagnetic susceptibility anisotropy using low field and high field methods. Geophysical Journal International, 151, 345-359. KISSEL, C., BARRIER, E., LAJ, C. & LEE, T. Q. 1986. Magnetic fabric in 'undeformed' marine clays from compressional zones. Tectonics, 5, 769-781. KLIGFIELD, R., LOWRIE, W. & DALZIEL, I. W. D. 1977. Magnetic susceptibility anisotropy as a strain indicator in the Sudbury basin, Ontario. Tectonophysics, 40, 287-308. KLIGFIELD, R., OWENS, W. H. & LOWRIE, W. 1981. Magnetic susceptibility anisotropy, strain and progressive deformation in Permian sediments from the Maritime Alps (France). Earth and Planetary Science Letters, 55, 181-189. KLIGFIELD, R., LOWRIE, W. & PFIFFNER, O. A. 1982. Magnetic properties of deformed oolitic limestones from the Swiss Alps: the correlation of magnetic anisotropy and strain. Eclogae Geologica HelvQtiaQ, 75, 127-157. KLIGFIELD, R., LOWRIE, W., HIRT, A. & SIDDANS, A. W. B. 1983. Effects of progressive deformation on remanent magnetization of Permian redbeds from the Alpes Maritimes (France). Tectonophysics, 97, 59-85. KNIPE, R. J. 1986. Microstructural evolution of vein arrays preserved in deep sea drilling project cores from the Japan Trench, Leg 57. In: CASEY MOORE, J. (eds). Structural Fabrics in Deep Sea Drilling Project Cores from Forearcs. Geological Society of America Memoir, 166, 75-88. LAMARCHE, G. & ROCHETTE, P. 1987. Microstructural analysis and origin of lineations in the magnetic fabric of some Alpine slates. Tectonophysics, 139, 285-293. LARRASOANA, J. C., PUEYO-MORER, E. & PARES, J. M. 2004. An integrated AMS, structural, palaeo- and rock-magnetic study of Eocene marine marls from the Jaca-Pamplona Basin (Pyrenees, N Spain): New insights into the timing of magnetic fabric acquisition in weakly deformed mudrocks. In: MARTIN-HERNANDEZ, F., LUNEBURG, C. M., AUBOURG, C. & JACKSON, M. (eds). Magnetic Fabric: Methods and Applications. Geological Society, London, Special Publications, 238,127-143.

LUNEBURG, C. M., LAMPERT, S. A., LEBIT, H. D., HIRT, A. M., CASEY, M. & LOWRIE, W. 1999. Magnetic anisotropy, rock fabrics and finite strain in deformed sediments of SW Sardinia (Italy). Tectonophysics, 307, 51-74. MARCH, A. 1932. Mathematische Theorie der Regelung nach der Korngestalt bei affiner deformation. Zeitschrift fur Kristallographie, 81, 285-297. MARTIN-HERNANDEZ, F. & HIRT, A. M. 2001. Separation of ferrimagnetic and paramagnetic anisotropies using a high-field torsion magnetometer. Tectonophysics, 337, 209-221. MATTEI, M., SPERANZA, F., ARGENTIERI, A., ROSSETTI, F., SAGNOTTI, L. & FUNICIELLO, R. 1999. Extensional tectonics in the Amantea basin (Calabria, Italy): a comparison between structural and magnetic anisotropy data. Tectonophysics, 307, 33^9. NYE, J. F. 1957. The Physical Properties of Crystals. Clarendon Press, Oxford, 322 pp. O'BRIEN, N. R. & SLATT, R. M. 1990. Argillaceous Rock Atlas, Springer-Verlag, New York. OERTEL, G. 1983. The relationship of strain and preferred orientation of phyllosilicate grains in rocks - a review. Tectonophysics, 100, 413-447. O'REILLY, W. 1984. Rock and Mineral Magnetism, Blackie & Son, Glasgow. OWENS, W. H. 1974, Mathematical model studies on factors affecting the magnetic anisotropy of deformed rocks. Tectonophysics, 24, 115-131. OWENS, W. H. 1993. Magnetic fabric studies of samples from Hole 808C, Nankai Trough. Proceedings of the Ocean Drilling Program, Scientific Results, 131, Ocean Drilling Program, 301-310. PARES, J. M. & DINARES-TURELL, J. 1993. Magnetic fabric in two sedimentary rock types from the Southern Pyrenees. Journal of Geomagnetism and Geoelectricity, 45, 193-205. PARES, J. M. & VAN DER PLUIJM, B. 2002<2. Evaluating magnetic lineations (AMS) in deformed rocks. Tectonophysics, 350, 283-298. PARES, J. M. & VAN DER PLUIJM 20026. Phyllosilicate fabric characterization by Low-Temperature Anisotropy of Magnetic Susceptibility (LT-AMS). Geophys. Res. Lett., 10. 1029/2002GLO15459. PARES, J. M. & VAN DER PLUIJM, B. 2003. Magnetic fabrics and strain in pencil structures of the Knobs Formation, Valley and Ridge Province, US Appalachians. Journal of Structural Geology, 25, 1349-1358. PARES, J. M., VAN DER PLUIJM, B & CIRES, J. 2001. Tectonic lineation and magnetic fabrics in deformed rocks. GSA Annual Meeting, Boston, MA. PARES, J. M., VAN DER PLUIJM, B. & DINARES-TURELL, J. 1999. Evolution of magnetic fabrics during incipient deformation of mudrocks (Pyrenees, northern Spain). Tectonophysics, 307, 1-14. PASSCHIER, C. W. & TROUW, R. A. J. 1996. Microtectonics, Springer-Verlag, New York. RAMSAY, J. G. & HUBER, M. I. 1983. The techniques of modern structural geology, Volume 1: Strain Analysis. Academic Press, London. RATHORE, J. S. 1980. The magnetic fabric of some slates from the Borrowdale Volcanic Group in the

AMS AND DEFORMATION IN MUDROCKS English Lake District and their correlations with strain. Tectonophysics, 67, 207-220. RATHORE, J. S., COURRIOUX, G. & CHOUKROUNE, P. 1983. Study of ductile shear zones (Galicia, Spain) using texture goniometry and magnetic fabric methods. Tectonophysics, 98, 87-109. RICHTER, C. 1992. Particle motion and the modelling of strain response in magnetic fabrics. Geophysical Journal International, 110, 451-464. RICHTER, C. & VAN DER PLUIJM, B. A. 1994. Separation of paramagnetic and ferrimagnetic susceptiblities using low temperature magnetic susceptibilities and comparison with high field methods. Physics of the Earth and Planetary Interiors, 82, 113— 123. ROCHETTE, P. & PILLION, G. 1988. Identification of multicomponent anisotropies in rocks using various field and temperature values in a cryogenic magnetometer. Physics of the Earth and Planetary Interiors, 51, 379-386. ROCHETTE, P. & VIALON, P. 1984. Development of planar and linear fabrics in Dauphinois shales and slates (French Alps) studied by magnetic anisotropy and its mineralogical control. Journal of Structural Geology, 6, 33-38. ROCHETTE, P., JACKSON, M. & AUBOURG, C. 1992. Rock magnetism and the interpretation of anisotropy of magnetic susceptibility. Reviews of Geophysics, 30, 209-226. RUF, A. S., NARUK, S. J., BUTLER, R. F. & CALDERONE, G. J. 1988. Strain and magnetic fabric in the Santa Catalina and Pinaleno mountains metamorphic core complex mylonite zones, Arizona. Tectonics, 1, 235-248. SAGNOTTI, L. & SPERANZA, S. 1993. Magnetic fabrics analysis of the Plio-Pleistocene clayey units of the Sant'Arcangelo basin, Southern Italy. Physics of the Earth and Planetary Interiors, 77, 165-176. SAGNOTTI, L., FACCENNA, C., FUNICIELLO, R. & MATTEI, M. 1994. Magnetic fabric and structural setting of Plio-Pleistocene clayey units in an extensional regime: the Tyrrhenian margin of central Italy. Journal of Structural Geology, 16, 12431257.

203

SAGNOTTI, L., WINKLER, A., MONTONE, P., Di BELLA, L. D., FLORINDO, F., MARIUCCI, M. T., MARRA, F., ALFONSI, L. & FREPOLI, A. 1999. Magnetic anisotropy of Plio-Pleistocene sediments from the Adriatic margin of the northern Apennines (Italy): implications for the time-space evolution of the stress field. Tectonophysics, 311, 139-153. SAINT-BLANQUAT, M. & TIKOFF, B. 1997. Development of magmatic to solid-state fabrics during syntectonic emplacement of the Mono Creek granite, Sierra Nevada Batholith. In: BOUCHEZ, J. L., HUTTON, D. H. W. & STEPHENS, W. E. (eds). Granite: From Segregation of Melt to Emplacement Fabrics. Kluwer Academic Publishers, Dordrecht, 231-252. SIDDANS, A. W., HENRY, B., KLIGFIELD, R., LOWRIE, W., HIRT, A. & PERCEVAULT, M. N. 1984. Finite strain patterns and their significance in Permian rocks of the Alpes Maritimes (France). Journal of Structural Geology, 6, 339-368. SINGH, J., SANDERSON, D. J. & TARLING, D. H. 1975. The magnetic susceptibility anisotropy of deformed rocks from north Cornwall, England. Tectonophysics, 27, 141-153. SINTUBIN, M. 1994a. Clay fabrics in relation to the burial history of shales. Sedimentology, 41, 1161— 1169. SINTUBIN, M. 19946. Phyllosilicate preferred orientation in relation to strain path determination in the lower Paleozoic Stavelot-Venn Massif (Ardennes, Belgium). Tectonophysics, 237, 215231. SORBY, H. C. 1853. On the origin of slaty cleavage. Edinburgh New Philosophical Journal, 55, 137-148. TARLING, D. H. & HROUDA, F. 1993. The Magnetic Anisotropy of Rocks. Chapman & Hall, London. TUCKER, M. E. 1985. Sedimentary Petrology. An Introduction. Vol. 3. Blackwell Scientific, Oxford. VAN DER PLUIJM, B. A., Ho, N. -C., PEACOR, D. R. & MERRIMAN, R. J. 1998. Contradictions of slate formation resolved? Nature, 392, 348. WOOD, D. S. 1974. Current views of the development of slaty cleavage. Annual Review of Earth and Planetary Sciences, 2, 369-401.

This page intentionally left blank

Anisotropy of magnetic susceptibility of lava flows and dykes: A historical account EDGARDO CANON-TAPIA CICESE, Department of Geology, P.O. Box 434843, San Diego CA, 92143, USA (e-mail: [email protected]) Abstract: The basic assumptions made in the interpretation of the anisotropy of magnetic susceptibility (AMS) of lava flows, dykes and other tabular intrusive rocks have changed with time. This paper presents a historical account of those changes. Because several aspects of the AMS of these types of rocks are still open to debate, the historical perspective helps us to appreciate better the limitations that had to be faced at different times, therefore providing some clues that can be used to solve such controversies. Also, by benefiting from hindsight it is possible to devise alternative approaches that can be employed to interpret the AMS of these types of rocks. Although some adjustments to current models will be made as new results become available, it seems that at present the main components of the basic model of AMS in lava flows and tabular intrusive rocks have been finally reached, leaving behind most of the apparent contradictions found in earlier works. This progress is undoubtedly an important advance in our understanding of the nature of AMS in general.

Low-field anisotropy of magnetic susceptibility (AMS) is an increasingly used petrofabric tool that allows us to solve a large number of problems of geological interest. Although the sources of the AMS of any rock type ultimately reside in their mineral components, whether associated with their crystalline structure, the shape of individual mineral grains or their distribution within a more or less isotropic matrix (e.g. Hrouda 1982; Rochette et al. 199la; Tarling & Hrouda 1993), it is often necessary to make some basic assumptions concerning how the rock fabric relates to the magnetic measurements. Additionally, to make a geological interpretation of results, it is often necessary to make some assumptions concerning the processes controlling the acquisition of the rock fabric itself. In contrast with those made in sedimentary or tectonic realms, the basic assumptions made in the interpretation of the AMS of lava flows and dykes have undergone a drastic shift over the years following the seminal paper by Khan (1962). To some extent, this is due to the large diversity of AMS results obtained in lava flows and dykes, which has resulted in many anomalous cases that required some explanation beyond the accepted models at the time. In retrospect, such diversity of the magnetic fabric of lava flows and dykes might signal a greater diversity of emplacement processes relative to those accounting for most sedimentary or metamorphic rocks. This paper presents a historical account of how the basic assumptions used to interpret AMS measurements in lava flows and dykes have evolved in the last four decades. Although this review concentrates on lava flows

and dykes, some of the lessons learned from these rocks are directly applicable to other scenarios such as granitic rocks. However, the role played by minerals other than titanomagnetites, or the occurrence of deformation under conditions different from those occurring in lava flows or dykes can become dominant in other rock types, therefore precluding the direct application of the same criteria for AMS interpretation as those summarized here. Sources of AMS in igneous rocks The AMS of any rock is the result of the combined contributions of all its constituent minerals (and glassy material, should this exist). From the earlier works in the 1960s to the present, however, there has been a shift in the importance given to various sources of the measured AMS of lava flows and small tabular intrusive rocks. To some extent, the relative importance of the three main sources that have been identified over time has influenced the models of fabric acquisition in lava flows and dykes. Therefore, it is convenient to start by establishing the evolution of the ideas concerning sources of AMS. The evolution of the concepts concerning fabric acquisition models can then be examined within this framework. Due to the large differences in the magnitude of bulk magnetic susceptibility (fcm) of different mineral species, it is sometimes possible to identify a dominant phase controlling the measured AMS. For example, Borradaile (1988) and Hrouda & Kahan (1991) have shown that a

From. MARTIN-HERNANDEZ, F., LUNEBURG, C. M., AUBOURG, C. & JACKSON, M. (eds) 2004. Magnetic Fabric: Methods and Applications. Geological Society, London, Special Publications, 238, 205-225. 0305-8719/04/ $15.00 © The Geological Society of London 2004.

206

E. CANON-TAPIA

small amount of ferromagnetic minerals (~lvol%) is enough to dominate the magnetic properties of a rock including its characteristic km. Consequently, a very useful rule of thumb that can be used to discriminate between two main sources of the rock anisotropy (i.e. whether it is controlled by the ferromagnetic or the paramagnetic fraction) can be devised (Tarling & Hrouda 1993). If km > 5 x 10~ 3 SI, then it is very likely that the contributions from all the paramagnetic and diamagnetic minerals is negligible, and that the measured anisotropy is controlled by the ferromagnetic minerals alone (commonly titanomagnetites in lava flows or tabular intrusive rocks). At the other extreme, if km < 10~4SI the ferromagnetic contribution can be safely neglected and the interpretation of the anisotropy is only relevant to the paramagnetic minerals (and probably some contributions from the diamagnetic fraction). For rocks with a bulk susceptibility intermediate between these two extremes, a more detailed examination of the mineral content might be required before drawing a valid petrofabric interpretation of the measured AMS. Because most lava flows and dykes yield km ~ 10~2, the source of their AMS is mainly related to the fabric of the titanomagnetites, and contributions from all the other minerals are generally negligible. This implies that the measured AMS may not provide a simple proxy of an optically determined mineral fabric (like that defined by plagioclase, for example), and it can occur that the orientation of each fabric is different from the other (e.g. Archanjo & Araujo 2002) even when both fabrics could be flow-related. On the other hand, if more than one generation of titanomagnetites exists within the rock (e.g. one generation formed directly from the melt and the other as the

result of alteration of a ferrosilicate mineral), the measured AMS will provide an average of both generations of grains, which may be difficult to reconcile with a particular optically observed flow-related structure. Recent studies have shown, however, that even when a recrystallized ferromagnetic phase might dominate the bulk susceptibility, the measured anisotropy might still be determined by the fabric of the mafic silicates (Borradaile & Gauthier 2003). In any case, it is noted that there might be more than one possible interpretation of the AMS in terms of mineral fabric, even when attention is focused exclusively on the titanomagnetites. Historically, it is important to note that the relative importance of the three main sources of anisotropy illustrated in Figure 1 has changed over time. The first to be recognized, and most appealing because of its simplicity, associates the physical dimensions of a mineral grain with the size of the principal susceptibilities. It is based on the tensorial description of the anisotropic properties of crystals made by Nye (1960) that allows a geometrical representation of the susceptibility tensor; a representation that is sometimes confused with the actual shape of a mineral particle. In the strictest sense, the tensorial approach used by Nye (1960) applies only to paramagnetic and diamagnetic susceptibilities, and it is only in the limit of weak magnetic fields that the anisotropy arising from the unequant shape of a ferromagnetic mineral (especially titanomagnetites) can be approximated by the same tensor description (see also Uyeda et al. 1963). Based on this approximation, Khan (1962) established that the AMS of a group of elongated grains of magnetite is directly related to their degree of preferred orientation, and therefore implied

Fig. 1. Diagrams illustrating the three sources of the AMS in igneous rocks, (a) The AMS of each mineral is such that the axis of maximum susceptibility (large arrow in each particle) coincides with the largest dimension of the particle and the minimum axis of susceptibility (short arrow in each particle) coincides with its shortest dimension. In this case the AMS of the whole assemble of grains (indicated on the margin of the figure) is a direct consequence of the shape preferred orientation of the large ferromagnetic grains, (b) The single domain (SD) effect results in a switching of susceptibility axes relative to the case shown in (a). In this case, the resulting AMS is also a consequence of the preferred orientation of the grains, but the resulting £min points along the direction of orientation of the petrofabric. (c) When the magnetic grains are close enough, magnetic interaction effects result in AMS axes that reflect the shape of the cluster rather than the shape of individual grains, (d) Rose diagram showing the orientation of particle long axis (petrofabric) measured in each of the previous diagrams.

HISTORICAL ACCOUNT OF AMS OF LAVAS & DYKES

that measuring the AMS would be equivalent to knowing the average orientation of the minerals in the rock (i.e. its petrofabric). From that moment, and for nearly two decades, the AMS of igneous rocks was almost exclusively interpreted in terms of the preferred orientation of elongated ferromagnetic minerals within the rock. Subjacent in the 'orientation' model advanced by Khan (1962) is the relation that exists between the physical dimensions of a single mineral grain and its principal susceptibility axes. The assumed relation is such that the directions of higher and smaller susceptibilities of a single, elongated grain of magnetite correspond to the longest and shortest dimensions of the grain respectively (Fig. la). Although this assumption is granted for large enough grains, it was established in the late 1980s that on very small, or singledomain (SD) particles the relationship of proportionality between magnitude of susceptibility and physical dimensions of the grain might be reversed (Potter & Stephenson 1988). The size threshold dividing the large enough from the too small ferromagnetic grains depends on the mineral species and is a function of the exact chemical composition, geometry and state of stress of the grain (Butler & Banerjee 1975; Dunlop & Ozdemir 1997). This implies that even when some measurements of magnetic granulometry are made, there will always be some uncertainty as to the extent to which the SD effect is really controlling the AMS of a rock specimen. Nonetheless, the SD effect (Fig. Ib) became increasingly invoked to explain some 'anomalous' AMS measurements in dykes in the absence of another viable explanation (e.g. Rochette 1988; Rochette et al. 19916, 1999). In the earlier part of the 1990s, another source of anisotropy gained in importance in the interpretation of the AMS of igneous rocks. Wolff et al. (1989) and Margraves et al. (1991) argued that the magnetite contained in many igneous rocks is not elongated, but is rather almost perfectly isometric, and that consequently the shape effect might not be as good an explanation as it had been assumed until then. Both of these studies suggested that a viable alternative was the magnetic interaction that should exist between neighbouring grains. Stephenson (1994) showed that indeed magnetic interaction can result in large degrees of anisotropy even in cases where all the particles are completely isotropic. If grain separation is very small (less than 2 grain diameters), it was later shown (Canon-Tapia 1995) that magnetic interaction can even modify the AMS of unequant particles so that the measured AMS will actually reflect

207

the shape of the cluster of grains rather than their preferred orientation (Fig. Ic). In extreme cases the direction of maximum susceptibility can become perpendicular to the direction of alignment of the longest dimensions of a set of grains due to interaction effects (Canon-Tapia 1995; Gregoire et al. 1995), therefore overriding the shape effect. Consequently, in order to explain some of the variability observed in lava flows that could not be explained at that time by using the single particle orientation effect (whether multi- or single domain), interaction effects were proposed as another important source of the AMS of igneous rocks (e.g. Canon-Tapia et al. 1996, 1997). Of these three sources of anisotropy, the first two have received most attention because both the original shape effect and the SD effect rely on the same preferred orientation of minerals to produce the observed AMS. Due to this common characteristic of both source models it is relatively easy to link the AMS with a conceptual model explaining the production of a preferred orientation of minerals as the result of flow of magma or lava. The only difference between these two source models of anisotropy is the predicted orientation of the £max and A:min axes relative to the observed mineral fabric. In contrast, the acquisition of a mineral fabric as the result of mechanical forces aiming to explain the AMS through magnetic interaction effects is difficult to model qualitatively. Consequently, this source model of AMS has not received much attention in the interpretation of AMS measurements made in dykes or lava flows. Despite this, Canon-Tapia (2001) has shown that it is possible to evaluate theoretically the extent to which magnetic interactions can be important in controlling the resulting AMS by using information from optical measurements that constrain the proportions of interactive and non-interactive magnetic grains. Use of this model can be helpful in establishing how widespread these deviations from the simpler singleshape mechanism are. Models of fabric acquisition in igneous rocks In the interpretation of AMS measurements, the first step is to decide the most probable source of the observed anisotropy, as already discussed. A second, independent step, is to adopt a model capable of explaining the mineral fabric characteristic of the rock so that it is possible to relaic it to the measured AMS. These models ^oui monly concentrate on the uuvhaiiical forces acting during rock formation, although other

208

E. CANON-TAPIA

aspects concerning mineral changes can be important in some particular instances. Despite the fact that alteration of lava flows or dykes can result in the creation of new mineral phases, more attention has been given to the mechanical aspects of fabric acquisition within the context of AMS studies. This not only reflects the historical perspective, but is also justified by the results of the few works that have shown that even very old rocks may have preserved the original AMS signature (e.g. Perroud et al. 1991; Walderhaug 1993; Bates & Mushayandebvu 1995). This section summarizes the evolution of the mechanical models that have been adopted to interpret AMS measurements in lava flows and dykes.

Lava flows and dykes: why are they special? By definition, 'igneous rocks' are the result of the cooling of an originally melted material. Presumptions about the amount of melt that can be accumulated at depth, about the relative proportion of liquid and other phases that it may contain in suspension, and about the details of its ascent to the surface, have all changed over time and may still evolve in the future (Marsh 2000; Canon-Tapia & Walker 2004). For present purposes, however, it suffices to note that not all the igneous rocks share a common original melt, nor are they formed by the same processes or in the same environment. Different histories of various igneous rocks can lead to significant differences in their mineral composition and in the

relative proportion of minerals that is sustained in suspension at a given time. In general, the differences in mineral compositions might not be important in controlling the AMS because of the overall predominance of the Fe-Ti oxides. Indeed, in most igneous rocks the ferromagnetic minerals are generally Fe-Ti oxides that constitute at least 1 vol %, and therefore the bulk susceptibility of the whole rock commonly exceeds 10~2 SI. Consequently, contributions from other minerals to the AMS of these rocks can be neglected in most cases. The most common exceptions to this rule include some granites, highly silicic lava flows or domes and some pyroclastic flows. In these rocks with a low bulk susceptibility, the interpretation of AMS must take into consideration the contribution of all minerals present, and in this sense it might be more akin to the interpretation of AMS of most metamorphic and sedimentary rocks. More importantly, and despite the preponderance of Fe-Ti oxides, not all igneous rocks share a common history responsible for the acquisition of a mineral fabric. For present purposes it is convenient to introduce three separate categories attending more to their mechanism of formation than to the environment where they were formed or to their chemical (mineralogical) composition (Fig. 2). The first group is defined by intrusive bodies of irregular form and large dimensions, possibly forming the remnants of ancient magma chambers (Fig. 2a). From the AMS point-ofview, this category includes almost exclusively granitic rocks, as mafic intrusive rocks with

Fig. 2. Diagrams showing three types of igneous rocks that can be defined attending to their rheological characteristics during emplacement. The left column shows the environment (intrusive = outline of volcano is dashed, extrusive = outline of the volcano is a solid line), and typical flow regimes (arrows) within the rock. The right size column displays a typical array of particles resulting from the localized (cm scale) flow regime, and indicates the phases present in the rock during AMS acquisition, (a) Large size plutons. (b) Pyroclastic deposits, (c) Tabular intrusions. Note that in (c) only a dyke is shown for clarity, but this group includes sills and lava flows as discussed in the text.

HISTORICAL ACCOUNT OF AMS OF LAVAS & DYKES

these characteristics have not been extensively studied through AMS. The second category of igneous rock as defined in this work is formed by volcaniclastic rocks. These rocks are characterized by preserving a fragmental nature resulting from the disruption of an originally continuous liquid phase rich in volatiles. Although the various types of fall and flow (including surge) deposits as well as the deposits formed by remobilization of the fragments immediately after their first settling belong to this group, flow deposits have been almost exclusively the subject of study through AMS (Fig. 2b). The third category introduced in this work includes lava flows, lava domes, lava tubes and small tabular intrusive rocks (dykes or cone sheets), therefore encompassing both extrusive and intrusive igneous rocks (Fig. 2c). Despite the difference in their environment of formation, all the rocks considered in the third group are formed as the result of the flowing of a continuous liquid phase that has solid and gaseous particles in suspension. Although the rocks included in the first group might be Theologically similar to those included in the third group, important differences in their fluid history distinguish them from each other. As illustrated in Figure 2, most of these differences are related to the dimensions and geometry of the 'channel' over which flow takes place. Whereas in the rocks of Figure 2a the reservoir containing the fluid rock is relatively large in all its three dimensions, the channel in Figure 2c is relatively small in at least one dimension (from a few centimetres to a few metres most of the time) so that boundary effects cannot be disregarded. In addition, flow within the larger reservoirs may not be laminar and could include changes in flow direction associated with convective cells (Fig. 2a), whereas in the tabular intrusive rocks and many lava flows, movement of the fluid is almost always laminar and has a well-determined direction for long distances (Fig. 2c). Finally, cooling times are much longer in the rocks of the first group, which might favour relaxation of any flowrelated fabric, whereas the mineral fabric of rocks in the third group has a larger probability of being preserved by a rapid cooling of the melted rock. It is noted that both historically and physically some coincidences can be established between different types of rock. Therefore, some reference will be made to studies of rocks other than lava flows or dykes throughout this paper. Similarly, differences can be found between dykes and lava flows (or even between two different types of lava flow) that can lead to important differences in the interpretation of lava flows between

209

specific examples. For example magma flow in dykes is driven by a dynamic pressure gradient and can be considered an essentially two-dimensional problem likely to be closer to simple shear; whereas in lava flows it is driven mainly by gravity, three-dimensional effects can be very important and both simple and pure shear are very likely to occur. Nevertheless, the acquisition of a mineral fabric in lava flows and dykes can be visualized as the result of the movement of rigid particles immersed in a very viscous fluid, contrasting with the fundamental processes of fabric acquisition characteristic of the other two groups considered here (e.g. turbulence and particulate flow are likely to be important in the other two scenarios, not to mention deposition and post-flow events). In addition, lateral variations in the mineral content may not be as common in lava flows or small tabular intrusive rocks as in some of the rocks included in the other two groups defined above. Consequently, the differences between specific examples of lava flows and other small tabular intrusive rocks seem to be smaller than those existing between the various groups of igneous rocks defined above. For these reasons, lava flows and dykes seem to belong to a unique type of rocks whose AMS can be studied independently from the other two types.

AMS in lava flows and tabular intrusions: the earlier confusion Two main assumptions were made by the pioneer studies of the AMS irrespective of the rock type of interest: (1) the AMS of the whole rock can be regarded as the result of the superposition of the anisotropies of individual grains within the rock; and (2) these individual anisotropies are represented by an ellipsoid whose axes are proportional to the axes of the best fitting ellipsoid of each grain. For lava flows and tabular intrusions, Khan (1962) supplemented these assumptions with the adoption of a model predicting the movement of ellipsoidal particles immersed in a viscous fluid (Jeffery 1922). Unfortunately, the interpretation of the mechanical model made by Khan (1962) was limited to some special cases described at the end of the original paper by Jeffery (1922), and later explored in more detail by Bretherton (1962). This limited interpretation explains why Khan (1962) postulated that in both lava flows and tabular intrusions the intermediate susceptibility direction would be in the direction of flow 'as expected from the theory' (Fig. 3a). It is important to mention that Taylor (1923) reported the

210

E. CANON-TAPIA

Fig. 3. Examples of the earlier AMS results obtained in dykes and lava flows. The measured orientation of /cmax axes (unornamented lines) is shown relative to the flow direction as indicated by the arrows in (a) dykes (Khan 1962), (b) the Aiyansh lava flow (Symons 1975) and (c) the Chfisby les lava flow (Kolofikova 1976).

results of an experimental confirmation of the limiting cases mentioned by JefFery's model (although there was no mention that these were limiting cases), therefore lending some support to the adoption made by Khan (1962). However, it became apparent very soon that the 'theoretical prediction' was in apparent contradiction with the experimental results reported by WingFatt and Stacey (1966) in which kmax was found to point along flow direction. Nevertheless, little effort was made at the time to reconcile both findings. Almost a decade later, Symons (1975) implicitly associated lava flow direction and maximum susceptibility axes without providing any justification of why he had decided to dismiss Khan's model. Another important difference of Symons's study relative to Khan's is that flow direction was well constrained in the Aiyansh flow (Fig. 3b), whereas it was only reasonably assumed for all the lava flows and tabular intrusive rocks studied by Khan (1962). Unfortunately, due to the young age of the Aiyansh flow, Symons (1975) was forced to collect samples from the flow crust, therefore introducing some noise in the geological interpretation of results. Although Symons was aware of such a possibility, as shown by the reservations he expressed in discussing his results, the text of both the abstract and the conclusions of his paper is categorical in affirming that 'the

anisotropy of magnetic susceptibility is not a reliable indicator of the flow direction in subaerial lavas'. This assertion marked much of the subsequent studies concerning the AMS of lava flows. Three years after Symons's study, Ellwood (1978) simply overlooked the difference in the fundamental association assumed to exist between flow direction and either k-mi or kmax axes by Khan (1962) and Symons (1975), respectively. Instead of calling attention to such a discrepancy in this fundamental assumption, Ellwood (1978) decided to follow Khan's interpretation of the theoretical model of ellipsoidal movement, highlighting the negative results reported in Symons's work. Therefore, by analogy with Khan (1962), the flow directions reported by Ellwood (1978) both in lava flows and dykes were a priori assumptions that apparently did not benefit from any clear field evidence supporting them. A more conservative interpretation of AMS measurements in the absence of well-established field evidence was made by Halvorsen (1974) who restricted his interpretation to the orientation of the km^n axes as perpendicular to the flow plane of magma. Adding to this confusion, Kolofikova (1976) reported changes in the orientation of the principal susceptibilities of one lava flow (Fig. 3c). The conclusions of this work pointed out that reliable flow directions could be inferred from the orien-

HISTORICAL ACCOUNT OF AMS OF LAVAS & DYKES

tation of £max axes only in the medial parts of the flow, whereas the frontal end was likely to yield unreliable results. Probably because of the uncertainties in the interpretation of measurements, and due to the lack of a sound model that could be used as a guiding principle, the interest in the AMS of lava flows declined at this date, and no other studies were produced for some time.

The AMS of columnar basalts: a dead-end detour During the 1960s and 1970s a few studies of the AMS of columnar basalts were made, searching for defined patterns that could be used to identify the occurrence of convective cells during the formation of such structures. Results from these studies showed from the beginning (Brown et al. 1964) that convection cells were not a likely explanation for the observed orientation of principal susceptibility axes (Fig. 4). Probably for this reason, only a few studies of this type were completed, and interest in the subject waned very soon. Nevertheless, all these studies (Brown et al. 1964; Symons 1967; Ellwood & Fisk 1977; Ellwood 1979) contributed, at least in part, to augmenting the confusion concerning the exact nature of the factors controlling the orientation of the principal susceptibility axes in lava flows and tabular intrusions. For example, Brown et al (1964) followed the model prediction of Khan (1962) concerning the alignment of k-mi axes with flow. As they found that all the k^ rvmin axes were vertical, it was easy for them to conclude that no evidence of convective flow was recorded at the columns they had studied. Although the £max axes in the horizontal plane seemed to yield a relatively good grouping of directions (Fig. 4a), they

211

considered that the scatter was excessive, preventing them from attempting any interpretation on the direction of flow of lava. Similar conclusions were reached by Symons (1967), even when he found that the plane of maximum susceptibility was nearly vertical, which could have been interpreted as evidence for a vertical flow (Fig. 4b). The relative independence of the orientation of this plane from the shape of the column, however, was a key observation interpreted as evidence against the occurrence of convective cells within a single column, although the possibility of larger-size cells was not discounted. Ellwood & Fisk (1977) also found a good grouping of &max axes from a single basalt column but, unlike the two previous works, they interpreted this clustering of susceptibility axes as related to the direction of thermal stress responsible for the column formation. Surprisingly, these authors dismissed the fact that thermal stresses must change in orientation within the column (hence the formation of the cracks with different orientation). Two years later, Ellwood (1979) presented evidence that made him conclude that thermal contraction was not the primary control in the orientation of the susceptibility axes, but he continued to invoke the effect of stress build-up during solidification, contrary to stresses related to flow, to explain the large dispersion of susceptibility axes found in the two dykes sampled. Although the three sites from one of his dykes have truly dispersed susceptibility axes, all three sites of the other dyke (Fig. 4c) could be identified as either the intermediate or the inverse fabric later defined by Rochette et al. (1991&) (see below). In any case, the possible influence of post-solidification stresses in the AMS of lava flows and dykes became entrenched in the literature and, despite good evidence pointing to the contrary (Perroud et al. 1991), it is still invoked from time to time.

Fig. 4. Examples of AMS results obtained from columnar basalts, (a) Orientation of &max axes across a single column (Brown et al. 1964), (b) Kmax axes from a single column (Symons 1967). (c) Kmax (squares), /rmin (triangles), and kmin (circles) from one column in a dyke. The plane of the dyke is shown as the great circle in the projection. Note the relatively good grouping of kmax axes in all three examples shown, irrespective of column-wall orientation. Also note that kmax axes are almost normal to the plane of the dyke in most of the samples shown in (c).

212

E. CANON-TAPIA

The AMS of Hawaiian dykes: a general consensus seems to appear Parallel to the studies made in lava flows some works were made focusing exclusively on the AMS of dykes or other tabular intrusive rocks. Stacey (1960) was among the first workers to report results of AMS measurements made in a tabular intrusion (sill). Although he was unable to present the geographic orientation of the principal susceptibility axes because the samples were taken from a drill core, several aspects of his work constitute guiding principles for many subsequent investigations. For example, he clearly stated that 'flow in partially solidified magma and deformation of solid rocks commonly results in grain alignment which may be evident as foliation or lineation'. Further, he anticipated some of the later discussions concerning the time of formation of the magnetic minerals in a flowing magma by saying that 'even when the magnetic minerals themselves have not been formed at the time of the alignment process, growth of magnetic grains may be controlled by the alignment of the other minerals.' Therefore, it is not strange to find in his work an overall optimism in the use of AMS measurements as an extension of the traditional petrofabric methods when he affirmed that 'where grain elongations are evident as magnetic anisotropy but are too small to be apparent visually, they are still indications

of stress or flow history'. The basic implications of these ideas were subsequently expanded by Knight & Walker (1988) in their study of the AMS of Hawaiian dykes. Despite some internal contradictions that can be found within this paper (perhaps the most notable of which is found when comparing the two first paragraphs of the section that starts on page 4305 in the original publication), their study presented enough evidence allowing them to infer magma flow direction independently from the AMS measurements. Their report of the AMS of Hawaiian dykes, therefore, constitutes the first study that used empirical findings to establish that the ^max axis is commonly parallel to magma flow direction. The evidence presented by Knight & Walker (1988) was so compelling that almost immediately the controversy found in earlier works was considered as completely solved and, whether explicitly influenced by Knight & Walker's results or not, almost every subsequent study of AMS in dykes and other intrusive rocks (e.g. Park et al. 1988; Ernst 1990; Cadman et al 1992; Ernst & Baragar 1992; Puranen et al 1992; Staudigel et al 1992; Lanza 1994; Raposo & Ernesto 1995; Raposo 1997; Varga et al 1998; Herrero-Bervera et al 2001; Ferre et al 2002) has considered that &max must be parallel to flow direction without further questioning (Fig. 5).

Fig. 5. (a) Diagram showing the imbrication angle of both elongated and flattened minerals in dykes, as proposed by Knight & Walker (1988). (b) Equal area projection (lower hemisphere) of the dyke shown in (a), illustrating the orientation of the /cmax axes that would be obtained from the east (open squares) and west (solid squares) walls of the dyke, respectively. The plane of the dyke is shown by the dashed line.

HISTORICAL ACCOUNT OF AMS OF LAVAS & DYKES

The AMS of Omanian dykes: new problems in the interpretation of results Although nearly 84 % of the 61 dykes and lava tubes studied by Knight & Walker (1988) were characterized by having fcmax axes parallel to flow direction, in 4 of the remaining 10 dykes the geological flow direction was more consistent with the orientation of fcint axes despite having the A:max axes contained in the plane of the intrusion, and in 6 other dykes there was a random distribution of susceptibility axes. Probably because of the small number of these cases (16% of the total), Knight & Walker (1988) did not provide a clear explanation of the mechanism producing such a shift in the relative orientation of susceptibility axes and magma flow direction, or at best they explained them as the result of turbulent flow. Almost simultaneous with the work by Knight & Walker (1988), Park et al (1988) provided a more detailed explanation for some of these 'anomalous' cases. The explanation offered by the latter authors was based on the influence of post-emplacement stresses in the domain structure of magnetite grains. Therefore, a change in the domain structure of magnetite was considered to be responsible for the observed shift in the positions of /qnt and km^n axes relative to the plane of intrusion. A few years later, in their study of the Oman ophiolite, Rochette et al. (1991Z?) associated all the observed anomalous fabrics with the influence of hydrothermal alteration, which, according to these authors, selectively affected the thicker (>1.5m) dykes of MORE affinity, and few dykes of calc-alkaline affinity. An alternative explanation for the special case of an inverse fabric (i.e. one in which the &max axes are normal to the plane of the intrusion) was also proposed by Rochette et al. (199la,/?) based on the relatively recent establishment of the effect of single domain magnetite. Magnetic granulometry tests carried out in dykes with a reverse fabric (i.e. £max normal to the plane of intrusion, and for which /rmin was interpreted as parallel to flow direction), however, did not support such an interpretation as categorically as would have been expected from the theory. Even so, the single domain (SD) effect became established as an alternative to explain the occurrence of abnormal cases in which fcmax axes were normal to the plane of the intrusion. As the number of studies of AMS in dykes increased, it became apparent that departures from the normal case were more common that could be guessed at from the evidence provided by the Hawaiian dykes. For example, anomalous

213

fabrics have been reported in different proportions for dykes in Brazil (Raposo & Ernesto 1995; Raposo 1997; Archanjo et al. 2000; Raposo & D'Agrella-Filho 2000; Archanjo & Araujo 2002), Canada (Ernst 1990; Cadman et al. 1992; Ernst & Baragar 1992), Finland (Puranen et al. 1992), French Guyana (Nomade et al. 2000), Greenland (Callot et al. 2001), India (Prasad et al. 1999), Israel (Baer 1995), Scotland (Herrero-Bervera et al. 2001), and Troodos (Borradaile & Gauthier 2001, 2003). In all these cases, the anomalous fabrics have been explained as the result of a superimposition of a post-emplacement-acquired fabric, as the result of SD effects, or simply have been disregarded (Fig. 6). The spectrum of post-emplacement processes that has been used to explain the presence of anomalous fabrics, as first invoked by Rochette et al. (19916), includes mineral changes associated with hydrothermal alteration, metamorphism or tectonic activity, although such explanations have proved not to be satisfactory for every instance. For example, many cases where magnetic mineralogy tests have been made on the samples presenting an inverse fabric (presumably due to the SD effect) are at best compatible with mixtures of MDSD magnetite. Actually, in some of the better documented cases (e.g. Raposo & Ernesto 1995; Raposo 1997; Raposo & D'Agrella-Filho 2000; Walderhaug 1993) this hypothesis has proved to be of no use for explaining the exact nature of the inverse fabric. All of the other possible explanations, including hydrothermalism or tectonic activity, have also been of limited success in explaining departures from the normal case, or require many specific considerations that can only be satisfied by a limited number of very special cases. Nevertheless, following the study of the Omanian dykes, departures from the 'normal case' have been systematically associated with causes unrelated to magma flow, even when the effect of a changing particle orientation during flow may have been considered explicitly at some stage during the interpretation (e.g. Bates & Mushayandebvu 1995; Archanjo et al. 2000; Archanjo & Araujo 2002). The AMS of lava flows revisited: order among the chaos Following the negative results reported by Symons (1975), the AMS of lava flows received almost no attention for a period of nearly 15 years. Additionally, it is noteworthy that the reported AMS measurements in works published during the earlier 1990s were only of secondary

214

E. CANON-TAPIA

Fig. 6. Diagrams showing the various relations among crystal orientation, magma flow direction and principal susceptibilities that have been proposed to explain abnormal magnetic fabrics, (a) Single domain effects result in &max axes perpendicular to flow direction. Note that the mineral fabric is parallel to flow direction, as in the normal fabric of Figure 5. (b) Late growth of ferromagnetic minerals in a direction perpendicular to the dyke walls result in kmax perpendicular to dyke walls and fcmin in the plane of the dyke, (c) Rolling effects on large grains orient &max normal to flow direction, but within the plane of the dyke. In this case, hmi is parallel to flow direction, (d) Turbulence during flow results in a disperse mineral fabric and consequently in a random orientation of principal susceptibility axes.

HISTORICAL ACCOUNT OF AMS OF LAVAS & DYKES

interest, as the main objective of those studies centred on palaeomagnetic observations (Perroud et al. 1991; Macdonald et al 1992; Canon-Tapia et al 1994, 1995). In all those cases, the influence of the results obtained in dykes by Knight & Walker (1988) can be identified, as it was assumed without further question that /cmax should be parallel to flow direction. Nevertheless, each of those studies made important contributions that paved the way for future developments. For example, based on the absence of optical evidence suggesting internal deformation after lava flow emplacement, Perroud et al (1991) concluded that the AMS measured on the Siluric lavas of Almaden probably originated during the emplacement of the flows, and that the regional folding had only been responsible for rigid rotations of whole sections of the lavas without affecting them internally. Most importantly, they pointed out that the mineral content, and the characteristics of the AMS fabric of the lava, could be divided in two main groups corresponding to the middle and basal parts of the flow respectively. This division also linked the observed shift in the orientation of the &max axes from the middle section (becoming parallel to the kmin axes of the basal

215

section) to changes in the size of the ferromagnetic minerals as observed in thin sections. In a different venue, MacDonald et al (1992) reported an agreement between the direction of ^max and field measured foliations and lineations, presumably acquired during the flow of lava, in 80% of their lava flow sites. In some of their sites, where departures from this agreement were more evident, they pointed out that the lack of coincidence between average susceptibilities and field structures was the result of the presence of interspersed populations of susceptibility axes, although no clear explanation for the production of such dispersion was advanced. This variability in the orientation of the susceptibility axes of individual specimens was also documented in the works by Canon-Tapia et al (1994, 1995) although, unlike previous cases, these authors presumed that such variability reflected systematic variations of the flow regime rather than random fluctuations around a single direction (Fig. 7). Therefore, by the mid-1990s there was enough evidence suggesting that (1) post-solidification stresses could not be responsible for the systematic reorientation of AMS if there was no evidence for a change in the original fabric of the rock; (2) the AMS of

Fig. 7. Three examples of the variability in the orientation of principal susceptibilities (lines centred in the solid symbols) relative to flow direction found in a single lava flow unit. Note the different vertical scales in each diagram. The solid symbols denote whether &max (squares), k-mi (triangles) or the plane &max-A;int (pentagons) was projected in the plane of the figure, (a) Thick unit from Xitle volcano (Canon-Tapia et al. 1996). (b) and (c) Vertical cross sections of toothpaste lava from Hawaii (Canon-Tapia et al. 1997).

216

E. CANON-TAPIA

a single unit could present different orientations of principal susceptibility axes that were functions of position within the unit; and (3) AMS measurements are closely related to other petrofabric indicators, which can also change orientation as a function of position within the unit. Canon-Tapia et al. (1996) explored in more detail the AMS of lava flows, integrating the most recent findings, and in particular paying attention to the possible effects of systematic variations of the flow regime on the measured AMS. That study used the deformation patterns defined by flattened vesicles in two lava flows for which the flow direction was unambiguously determined to constrain the deformation regimes experienced during the emplacement of the lava flows. Due to the small dimensions of the flows, to the relatively steep slope over which they had been emplaced, and to their overall similarities despite differences in the chemical (and hence mineral) compositions, those two flows provided a unique opportunity to calibrate the AMS method as a petrofabric indicator. As a result of that study it was confirmed that the AMS of lava flows bears a close relationship with the three-dimensional deformation experienced by flowing lava, but it was also suggested that such a relationship can be used to identify regions likely to have experienced more intense shear within the same flow. This idea was further explored by Canon-Tapia et al. (1997), who suggested that differences in the degree of anisotropy can be related to different modes of lava emplacement, and that AMS measurements could be used to infer not only the direction of lava movement, but also to explore in more detail the deformation history of the whole rock. Both of these studies also showed that the scatter of the orientations of principal susceptibilities, including some of the exchanges between two types of susceptibility axes previously documented by Perroud et al (1991) and MacDonald et al. (1992), could be due to changes in the direction and/or intensity of shear rather than to statistical fluctuations around a mean value that had been assumed as valid without a real physical base. All of these results were difficult to explain using the model of movement of a single particle as adopted by the AMS community until that time. For that reason, an alternative source of AMS was adopted as a more viable alternative to explain the observed exchanges of principal susceptibility axes. This source of AMS was the recently (at the time) advanced hypothesis of magnetic interaction effects, which could be related to AMS through the formation of particle clusters during magma movement. Although physically reasonable, this model of AMS was

received with scepticism by some authors (e.g. Rochette et al. 1999) mainly because of the lack of a numerical model of interactions that use initially random assemblages of magnetite. The AMS of lava flows and of tabular intrusions: the deterministic and the systematic-variation approaches The most recent studies of the AMS of lava flows or dykes can be divided in two main groups defined as a function of the main approach followed in the interpretation. To the first group belong studies based in a deterministic approach in which a given type of susceptibility axes is considered to point along flow direction anywhere within the same unit. The works in the second group base the interpretation of AMS results in the identification of systematic variations of the orientation of susceptibility axes within the same rock unit. Following the results by Knight & Walker (1988), most of the studies of the first group assume that flow direction will be represented by the orientation of £max axes. In an extreme situation, this approach has been followed to infer a regional flow direction from a group of lava flows on the assumption that all the lavas on that group were issued from the same vent and moved in the same direction. For instance, Morris (2000) reported AMS results obtained from groups of lava flows that could be found at the same locality in the Aegina island, Greece. The number of samples collected from individual lava flows in this study was small, and little or no attention seemed to be given to the position of each sample relative to the boundaries of the flow unit from which it had been collected. By using this approach, the occasional switch between £max and /qnt axes, or the large scatter of these two axes found within a plane, remained features of the AMS that do not seem to have any relation with flow direction. Glen et al. (1997), Plenier et al. (2002) and Henry et al. (2003) also showed section means interpreted to represent regional flow directions. In these three cases, however, attention was given to the relative position of the samples within the flow unit and various degrees of clustering of susceptibility axes could be identified either from the vertical distribution of samples within the same unit (Glen et al. 1997) or by the use of density contour plots (Plenier et al. 2002; Henry et al. 2003). Although the approach followed by Plenier et al. (2002) and Henry et al. (2003) is essentially deterministic, because at the end the interpretation is made for all the flows in

HISTORICAL ACCOUNT OF AMS OF LAVAS & DYKES

the same conditions, these authors were aware of the possible effects of changing emplacement conditions in the resulting AMS signal. For example, Henry et al. (2003) associated a larger scatter of £max axis as the probable result of lava flow emplacement in gentle slopes, and a tighter clustering as the result of emplacement in steeper slopes, which is similar to the difference suggested by Canon-Tapia et al. (1995) on some flows of the Xitle volcano. A similarly deterministic approach has commonly been adopted to interpret the AMS of dykes or tabular intrusions. For example, Aifa et al. (1999), used AMS measurements to investigate the flow direction in the St Malo dyke swarm, France. In that study, the flow direction was inferred from contour plots of fcmax axes. These plots showed that 40% of their dykes yield at least two clusters of fcmax orientations, some of which were normal to each other, and in at least 30% of the dykes some of the fcmax axes were closer to the normal to the plane of the dyke. Despite such variability in fcmax orientations and the discrepancy observed between the ^max inclinations and the measured dyke dips identified by the authors, they suggested that flow direction was compatible with a fanning pattern apparently based on the orientation of the mean clusters of &max axes. Although transverse sections were recovered from each dyke in that study, Ai'fa et al. (1999) seemed to take no advantage of the relative sample position within the dyke, which should result in an opposite imbrication according to Knight & Walker (1988). In most other recent studies of the AMS of dykes, the role played by an imbrication angle of principal susceptibility axes has received more attention, mainly within a deterministic approach. In a extreme situation, it has been proposed that flow directions only can be determined by using a rigorous statistical approach in which the clusters of fcmax axes obtained from opposite halves of the intrusive unit are completely independent from each other (Tauxe et al. 1998). Although it is certainly possible to find some intrusive rocks that adjust to this model, it should be noted that not all the cases with overlapping regions of confidence (obtained from samples collected at opposite sides of a dyke) are due to an aleatory scatter of susceptibility axes. For example, a decrease in the angle of imbrication as a function of distance from the dyke wall, or even the effect of only one or two specimens with susceptibility axes that have exchanged orientations, can result in overlapping regions of confidence (e.g. Staudigel et al. 1992; Baer 1995; Rapalini & Lopez de Luchi 2000; Borradaile & Gauthier 2001;

217

Herrero-Bervera et al. 2001). In all these studies a meaningful flow direction has been determined based on a deterministic approach despite the observed overlapping of the regions of confidence, therefore showing that a meaningful flow direction can be extracted from the AMS data even in these cases. Nevertheless, the observed changes in orientation of the principal susceptibility axes have remained unexplained features of the AMS, and identification of possible outliers is subjective. The deterministic procedure followed in all these works contrasts with the method followed by the works belonging to the 'systematicvariation' approach. In these works (e.g. Canon-Tapia & Coe 2002; Canon-Tapia 2003) the relative position of each specimen within the unit is extremely important in the analysis of AMS results. The fundamental aspect of this approach is that the orientation of susceptibility axes, and their degree of clustering, can change as a function of position within the unit due to the effects on the particle movement introduced by the geometry of the flow itself, and therefore, it is presumed that information concerning the flow field can be retrieved from the systematic variation of the AMS. Although this relationship between flow regime and AMS was suspected from the observations of lava flows of various morphologies (Canon-Tapia et al. 1996; Canon-Tapia et al. 1997; CanonTapia & Walker 1998), some of the details of such relation were better observed in the experimental results reported by Canon-Tapia & Pinkerton (2000). These experiments showed that in those cases where lava movement stops while still at high temperatures, the magnetic fabric is lost because the low viscosity of the lava prevents it from fixing the preferred orientation of the particles achieved during lava movement. This orientation was inferred to have been achieved by microlites present during the experiments, even at the highest temperatures, and it could be preserved if the lava was chilled instead of slowly cooled. These experiments also showed that, in chilled lavas, the rate of deformation is proportional to the degree of anisotropy, and that a switching of susceptibility axes could be found even in samples separated by a few centimetres. Although the experimental results of CanonTapia & Pinkerton (2000) showed that the principal susceptibility axes may produce a complicated distribution pattern due to the systematic changes in orientation due to flow complexities, the close relationship existing between the flow regime and AMS can turn out to be advantageous for unravelling details of flow that would

218

E. CANON-TAPIA

otherwise be missed. For instance, despite the large scatter observed when all the measurements from one thick lava flow were plotted together, Canon-Tapia & Coe (2002) were able to isolate the occurrence of lava injection pulses within a single lava flow by making a detailed analysis of axial distributions within specific sections in the flow (Fig. 8), in combination with the observed fluctuations in the degree of anisotropy. In tabular intrusive rocks, Staudigel et al. (1999) have illustrated how it is possible to

obtain a well-defined pattern from apparently diffuse AMS data. Although in their work the regional trends of flow directions in the Troodos ophiolite are obtained by considering dyke location and chemical type instead of systematic variations of AMS data alone, the approach followed in that work is completely analogous to that followed by Canon-Tapia & Coe (2002) in their analysis of the AMS results obtained within a single flow unit. Both of those studies show that it is only when each piece of

Fig. 8. Equal area projections (lower hemisphere) showing the principal susceptibilities of the Birkett flow, Columbia River Basalt (Canon-Tapia & Coe 2001). (a) All the specimens collected from the same unit yield a large scatter of directions that can be dissected into coherent groupings as a function of vertical position of the samples. Note the general agreement between the diagrams showing density contour plots (b & c) and both the location of the mean principal susceptibilities and the orientation of the regions of confidence around those means (d & e). The examples shown correspond to the layers 6 (b & d) and 18 (c & e) of the Birkett flow.

HISTORICAL ACCOUNT OF AMS OF LAVAS & DYKES

information (whether flow direction of a dyke or orientation of the principal susceptibilities of each specimen) is inserted in the adequate context that the otherwise diffuse pattern breaks down into a well defined pattern that can be interpreted in geological terms. As an example of how both approaches can influence the interpretation of AMS results it is illustrative to examine the results obtained by Liss et al. (2002). These authors found systematic changes in the orientation of principal susceptibility axes as a function of distance from the contact of the Whin sill, northern England. Because of the assumed parallelism between fcmax and flow direction, the observed variation in fcmax orientation was interpreted as the result of a change in flow direction as the intrusion of the sill progressed. Although reasonable, this interpretation of the observed change from a NE-SW direction near the lower contact of the intrusive to a nearly N-S direction recorded in its centre could also reflect different degrees of shear associated with a common flow direction if a less deterministic approach were used in the interpretation. Geoffrey et al. (2002) also observed a large variability in the orientation of both £max and k-ini axes in dykes from Greenland. Due to the parallelism observed in these dykes between the mean orientation of plagioclase crystals and the imbrication of the fcmax — fcint plane, however, these authors suggested that the &min axes could be used to infer a reliable flow direction. Callot & Guchet (2003) reached a similar conclusion based in the modelling of a mineral fabric composed of two subfabrics with a well-defined orientation relative to flow direction. This approach is reminiscent of that followed by Halvorsen (1974) in lava flows and of the criteria commonly used to infer flow direction of pyroclastic rocks (e.g. Incoronato et al. 1983; Baer et al. 1997; Ort et al. 1999). Although a reliable flow direction may be inferred using this approach in cases in which the AMS is very oblate, and in which the £max axes change orientation randomly within a plane, the methodology remains deterministic in its most basic aspects because it assumes a constant orientation of the principal susceptibilities along the walls of the dyke. Interestingly, the effects of a changing wall orientation can be somewhat examined with the methodology described by Geoffroy et al. (2002), as one of the two techniques of flow-vector determination that they proposed (the 'core flow-vector method') may result in a scatter of directions whereas a well-constrained average can be obtained with the other method (the 'mean flow-vector method').

219

The AMS of lava flows and of tabular intrusions: towards a unified model? Probably one of the main objections that can be made to the systematic-variation approach of AMS interpretation is the lack of a simple model that can be invoked to understand why it is that the orientation of principal susceptibilities should change in a laminar flow. Paradoxically, the same physical model that has been used until now to support the basic principles followed in the deterministic approach can be used to create a model that predicts a systematic change in orientation of the principal susceptibilities. Surprisingly, one of the fundamental characteristics of the model of ellipsoidal movement proposed by Jeffery (1922) has been completely ignored by almost every worker interested in magnetic fabric studies. Contrary to the assumption made in the deterministic approach, this model predicts a cyclic movement of the symmetry axis of any particle represented by a rotational ellipsoid, and only in very special circumstances predicts the acquisition of a stable orientation of the particle relative to the orientation of shear. Although the cyclic motion of an ellipsoidal particle predicted by Jeffery's equations has received much attention in the geological literature (e.g. Gay 1966; Freeman 1985; Fernandez 1987; Jezek et al. 1994; Ildefonse et al. 1997; Manga 1998; Canon-Tapia & Chavez-Alvarez 2004), Dragoni et al. (1997) were the first to consider the implications of this particle cyclic movement in the context of AMS studies (Fig. 9). Unfortunately, the model proposed by Dragoni et al. (1997) to explain the cyclic orientation of the principal susceptibilities as the result of magma flow within a dyke has passed almost unnoticed by the AMS community even when a variability in the distribution of susceptibility axes has been observed in many instances where the same dyke has been sampled at various locations (e.g. Cadman et al. 1992; Walderhaug 1993; Baer 1995; Bates & Mushayandebvu 1995; Callot et al. 2001; Hrouda et al. 2002). The cyclic behaviour in particle orientation predicted by Jeffery's equations introduces an undesired uncertainty into the determination of flow direction from AMS measurements. As shown by Chavez-Alvarez & Canon-Tapia (2003, 2004), part of this uncertainty can be removed by paying attention to aspects such as (1) sample position relative to significant flowrelated regions; (2) predominant particle shape; and (3) amount of shear experienced by the sampled volume. The predictions of their model represent an adequate framework allowing a

220

E. CANON-TAPIA

Fig. 9. Diagrams showing the predicted particle rotation as the result of movement in a viscous fluid. The trajectory of a single particle depends on the initial particle orientation, on the shape of the particle and on the total amount of shear of a given volume, (a) Particle orientation changes systematically along the dyke, (b) Trajectories followed by the symmetry axes of elongated particles (hence of &max). Particles with same elongation ratio but different initial orientation (dashed and solid lines), and particles with the same initial orientation but different elongation ratio (dashed and dotted lines) describe different trajectories, the plane of the dyke (not shown for clarity) is along the N-S direction.

simple explanation of the observed variations of AMS along a single dyke (Cadman et al. 1992; Walderhaug 1993; Baer 1995; Bates & Mushayandebvu 1995; Callot et al. 2001; Hrouda et al. 2002)), as a function of distance from the contact of thick intrusions (Liss et al. 2002), or at different parts of a single lava flow along flow direction (Kolofikova 1976; Canon-Tapia et al. 1997; Canon-Tapia & Pinkerton 2000). Therefore, by adopting the complete model of elliptical movement as originally established by Jeffery (1922), it is possible to have an alternative model to explain the systematic changes in orientation of AMS axes as a function of flow-related forces that does not rely on the presence of magnetic interaction effects. The extent to which this alternative model of fabric acquisition is widespread remains to be established. Conclusion This historical account of the development of ideas concerning the AMS of lava flows and tabular intrusive rocks has shown how some of the basic assumptions behind the interpretation of this type of measurements have passed from a simple deterministic model to a more complicated approach in which aspects of each sampled unit are important in the interpretation of results. Although it is undoubtedly easier to attempt an interpretation of AMS measurements based on a simple rule of thumb that applies to a large number of specimens indiscriminately, than it is to carry out a specimen by specimen analysis, the easy road may prove to be extremely deterministic and too restrictive. Due to this lack of flexibility, the deterministic approach hampers

a clear interpretation of those cases that fall outside the limits artificially imposed by the theory and leaves unexplained many features of the AMS of various rock types. On the other hand, an approach in which equal attention has to be given to each individual specimen can become cumbersome, making the AMS unattractive in the interpretation of large data sets. Based on the lessons learnt from lava flows (CanonTapia et al. 1996, 1997; Canon-Tapia & Pinkerton 2000; Canon-Tapia & Coe 2002) these extremes can be balanced when making the interpretation of results by using two types of diagrams: one including the average principal susceptibilities and their regions of confidence and the other displaying the orientation density of axes without making a distinction between the type of susceptibility involved. Contour plots that focus on individual types of susceptibility axes may not be adequate, however, as these diagrams will not reflect the effect of switching axes in the same manner as those diagrams constructed by considering the three types of axes simultaneously. In any case, adequate use of two types of diagrams should provide enough evidence to test the hypothesis of systematic variations related to differences in the flow regime, even when other more rigorous statistical tests can be ambiguous. Consequently, the interpretation of the AMS will be put on a sounder physical basis, and a quality factor can be attached to each site. For example, cases in which an absolute flow direction can be inferred from the observations will be more reliable than sites in which only a direction is inferred, and these in turn will be more reliable than those in which it is only possible to identify two alternative directions. Nevertheless, by acknowledging

HISTORICAL ACCOUNT OF AMS OF LAVAS & DYKES

the possibility that two (commonly mutually perpendicular) flow directions are compatible with the AMS measurements, some anomalous sites may in fact become part of a coherent interpretation at a regional scale, without relying on ad hoc explanations as shown by Canon-Tapia (2003). The model of fabric acquisition based on the complete interpretation of the equations of a single particle derived by Jeffery (1922), and used here in the context of the AMS of lava flows and tabular intrusive rocks, has direct application in the interpretation of petrofabric of other igneous bodies in which deformation was likely to be of magmatic character. This is justified because the underlying physics remains the same in all those instances. It is remarked, however, that the effects of the boundary conditions, which are likely to change from case to case, have to be clearly stated to avoid an incorrect interpretation. It is in this sense that it becomes impossible to validate generalized statements concerning the orientation of principal susceptibility axes relative to flow direction, but it remains certain that there is a close relation between shear and AMS that can be used as a guiding principle to make a physically reasonable interpretation of almost any individual situation. Application of the basic aspects of fabric acquisition in conditions of magmatic deformation requires the fulfilment of several requirements both during sampling and at the moment of the analysis of results. In particular, the sampling scheme must be designed to take note of the position of the sample relative to physical boundaries of the unit (lateral or vertical ends of a lava flow, or contacts with the country rock in the case of intrusives). Whenever possible, two zones likely to have experienced different deformation histories should be included, and sampling must be as detailed as possible within these zones. In the case of lava flows, these two zones can be the bottom (preferred), central or upper parts of the unit, whereas in tabular intrusions these can be each of the two margins of the intrusion and the central part of wider dykes. Lateral ends of lava flows are not recommended, as these can result in a systematic bias of the orientation of susceptibility axes along a direction not necessarily reflecting the direction of emplacement of the bulk of the flow, nor easily explained by an imbrication effect. Nevertheless, if enough samples are collected from different parts of the flow (including its lateral ends), at least some constraints can be imposed concerning the local flow direction of that particular unit. Clearly, there will be some instances in which implementation of these recommendations is not possible due to difficulties of access in the field, to the

221

size of the rock to be considered, or to the exposure (or lack thereof) of the boundaries of the rock under study. This can be especially true for granitic bodies or other similarly sized plutons. However, it is stressed that even when a much larger number of specimens must be analysed to check for the internal congruency of the AMS data, the time required for these measurements and the analysis of results is likely to be much less than the time required to make any other observation (whether optical or magnetic) attempting to explain the anomalous cases in mineralogical terms. The above statements do not imply that mineral-related effects have to be completely disregarded in any interpretation of AMS measurements, but rather that a simpler and comprehensive explanation can probably be envisaged for most of the anomalous AMS fabrics without the need to invoke ad hoc explanations. Undoubtedly, it will be possible to find cases in which mineral effects remain as the easiest (or only) possible explanation. For instance, cases in which solid-state deformation has occurred, or even in regimes transitional between solidstate and suspension flow where the particle concentration is relatively large (Nicolas 1992) can deviate from the conditions underlying the equations of movement derived by Jeffery (1922). In these cases, mechanical interaction between particles (e.g. Ildefonse et al I992a,b; Arbaret et al. 1996, 1997) can become the most important factor controlling the acquisition of mineral fabrics. Fortunately, in the simplest cases where particle interaction is expected, the fabric acquisition will be much simpler than in the case of deformation without particle interaction, because the development of tiling structures will result in a more or less constant imbrication along the direction of the deformation. When more than one particle shape is to be considered, however, the development of various subfabrics can take place, making necessary a more detailed analysis of both the AMS and optical measurements (e.g. Geoffroy et al. 2002; Callot & Guchet 2003). Different subfabrics can also be formed as the result of post-flow growth of mineral grains. This situation, specially important in places where extensive hydrothermalism may have occurred, can modify the AMS signal related to flow. Under favourable circumstances, the growth of accessory minerals may not be anisotropic enough to completely erase the flow-related fabric, even if the late forming phase has a higher value of bulk susceptibility than the minerals defining the flow-related fabric (Borradaile & Gauthier 2003). In these cases, use of other techniques (e.g. the anisotropy

222

E. CANON-TAPIA

of magnetic remanence) can help to isolate a meaningful orientation of magnetic fabrics. These results also indicate that even in cases with a km ~ 1(T3, the anisotropy of the paramagnetic fraction can be large enough to define a measurable AMS that bears a direct relation to the only mineral fabric observed optically (which may not be a ferromagnetic one) (Cafion-Tapia & Castro 2004). Finally, it should be mentioned that in some instances the interpretation of AMS might remain ambiguous. In these cases, besides using other sources of information concerning the mineral fabric (whether magnetic or not), it seems reasonable to attempt an interpretation that takes into consideration regional trends (e.g. Bates & Mushayandebvu 1995; CanonTapia 2003). Using this approach, it might be possible to devise alternative approaches that can be used later to test the validity of the regional, AMS-suggested model. In the long run all the available evidence (not only AMS based) should be congruent, and a better understanding of the particular geologic problem will be possible. This work has benefited from the comments made by all the reviewers of my papers in the last ten years. Criticisms raised, sometimes anonymously and not in very constructive terms, have always been welcomed as they have served as a strong motivation in my search for alternatives that could provide explanations for the abnormal and hard-to-interpret observations. Although every effort was made to keep this review as unbiased and historically accurate as possible, the comments made by G. Laurent, M. Ort and G. Plenier helped me to identify those places where I had lost that historical perspective. They are also to be thanked for calling my attention to recent works that were inadvertently overlooked in the first version of this paper. I also would like to thank F. Canon for her support and commitment, which helped me to find the time for finishing this work within the required deadlines.

References AIFA, T., LEFORT, J. P. & GUENNOC, P. 1999. Anisotropy of magnetic susceptibility investigations of the St Malo dyke swarm (Brittany, France): emplacement mechanism of doleritic intrusions. Geophysical Journal International, 139, 573-582. ARBARET, L., DIOT, H. & BOUCHEZ, J. L. 1996. Shape fabrics of particles in low concentration suspensions: 2D analogue experiments and application to tiling in magma. Journal of Structural Geology, 18, 941-950. ARBARET, L., DIOT, H., BOUCHEZ, J. L., LESPINASSE, P. & DE SAINT-BLANQUAT, M. 1997. Analogue 3D simple-shear experiments of magmatic biotite subfabrics. In: BOUCHEZ, J. L., MUTTON, D. H. W. &

STEPHENS, W. E. (eds) Granite: From Segregation of Melt to Emplacement Fabrics. Kluwer Academic Publishers, Dordrecht, 129-143. ARCHANJO, C. J. & ARAUJO, M. G. S. 2002. Fabric of the Rio Ceara-Mirim mafic dike swarm (northeastern Brazil) determined by anisotropy of magnetic susceptibility and image analysis. Journal of Geophysical Research, 107(B3). ARCHANJO, C. J., TRINDADE, R. L, WILSON, J. MACEDO, P. & ARAUJO, M. G. 2000. Magnetic fabric of a basaltic dyke swarm associated with Mesozoic rifting in northeastern Brazil. Journal of South American Earth Sciences, 13, 179-189. BAER, E. M., FISHER, R. V., FULLER, M. D. & VALENTINE, G. 1997. Turbulent transport and deposition of the Ito pyroclastic flow: determinations using anisotropy of magnetic susceptibility. Journal of Geophysical Research, 102, 22565-22586. BAER, G. 1995. Fracture propagation and magma flow in segmented dykes: Field evidence and fabric analyses, Makhtesh Ramon, Israel. In: BAER, G. & HEIMANN, A. (eds) Physics and chemistry of dykes. A.A. Balkema, Rotterdam, 125-140. BATES, M. P. & MUSHAYANDEBVU, M. F. 1995. Magnetic fabric in the Umvimeela Dyke, satellite of the Great Dyke, Zimbabwe. Tectonophysics, 242, 241-254. BORRADAILE, G. J. 1988. Magnetic susceptibility, petrofabrics and strain. Tectonophysics, 156, 1-20. BORRADAILE, G. J. & GAUTHIER, D. 2001. AMSdetection of inverse fabrics without AARM, in ophiolite dikes. Geophysical Research Letters, 28, 3517-3520. BORRADAILE, G. J. & GAUTHIER, D. 2003. Interpreting anomalous magnetic fabrics in ophiolite dikes. Journal of Structural Geology, 25, 171-182. BRETHERTON, F. P. 1962. The motion of rigid particles in shear flow at low Reynolds number. Journal of Fluid Mechanics, 14, 284-304. BROWN, H. C., KHAN, M. A. & STAGEY, F. D. 1964. A search for flow structure in columnar basalt using magnetic anisotropy measurements. Pure and Applied Geophysics, 57, 61-65. BUTLER, R. F. & BANERJEE, S. K. 1975. Theoretical single-domain size range in magnetite and titanomagnetite. Journal of Geophysical Research, 80, 4049-4058. CADMAN, A. C., PARK, R. G., TARNEY, J. & HALLS, H. C. 1992. Significance of anisotropy of magnetic susceptibility fabrics in Proterozoic mafic dykes, Hopedale Block, Labrador. Tectonophysics, 207, 303-314. CALLOT, J. P., GEOFFROY, L., AUBOURG, C., Pozzi, J. P. & MEGE, D. 2001. Magma flow directions of shallow dykes from the East Greenland volcanic margin inferred from magnetic fabric studies. Tectonophysics, 335, 313-329. CALLOT, J. P. & GUCHET, X. 2003. Rock texture and magnetic lineation in dykes: A simple analytical model. Tectonophysics, 366, 207-222. CANON-TAPIA, E. 1995. Single-grain versus distribution anisotropy: a simple three-dimensional model. Physics of the Earth and Planetary Interiors, 94, 149-158."

HISTORICAL ACCOUNT OF AMS OF LAVAS & DYKES CANON-TAFIA, E. 2001. Factors affecting the relative importance of shape and distribution anisotropy in rocks: theory and experiments. Tectonophysics, 340, 117-131. CANON-TAPIA, E. 2004. Flow direction and magnetic mineralogy of lava flows from the central parts of the Peninsula of Baja California, Mexico. Bulletin of Volcanology, 66, 431-442. CANON-TAPIA, E. & CASTRO, J. 2004. AMS measurements on obsidian from the Inyo Domes, CA: a comparison of magnetic and mineral preferred orientation fabrics. Journal of Volcanology and Geothermal Research, 134, 169-182. CANON-TAPIA, E. & CHAVEZ-ALVAREZ, M. J. 2004. Rotation of uniaxial ellipsoidal particles during simple shear revisited: the influence of elongation ratio, initial distributionof a multiparticle system and amount of shear in the acquisition of a stable orientation. Journal of Structural Geology, in press. CANON-TAPIA, E. & CHAVEZ-ALVAREZ, M. J. 2004. Theoretical aspects of particle movement in flowing magma: implication for the anistropy of magnetic susceptibility of dykes and lava flows. In: MARTIN HERNANDEZ, F., AUBOURG, C., JACKSON, M. AND LUNEBURG, C. (eds) Magnetic fabric methods and applicatim. The Geological Society, London, in press. CANON-TAPIA, E. & COE, R. 2002. Rock magnetic evidence of inflation of a flood basalt lava flow. Bulletin of Volcanology, 64, 289-302. CANON-TAPIA, E. & PINKERTON, H. 2000. The anisotropy of magnetic susceptibility of lava flows: an experimental approach. Journal of Volcanology and Geothermal Research, 98, 219-233. CANON-TAPIA, E. & WALKER, G. P. L. 1998. Mecanismo de emplazamiento de las mesetas basalticas gigantes continentales estudiado a traves de mediciones de anisotropia de susceptibilidad magnetica. GEOS, 18, 2-10. CANON-TAPIA, E. & WALKER, G. P. L. 2004. Global aspects of volcanism: the perspectives of 'plate tectonics' and 'volcanic systems'. Earth Science Reviews, 66, 163-182. CANON-TAPIA, E., HERRERO-BERVERA, E. & WALKER, G. P. L. 1994. Flow directions and paleomagnetic study of rocks from the Azufre Volcano, Argentina. Journal of Geomagnetism and Geoelectricity, 46, 143-159. CANON-TAPIA, E., WALKER, G. P. L. & HERREROBERVERA, E. 1995. Magnetic fabric and flow direction in basaltic pahoehoe lava of Xitle volcano, Mexico. Journal of Volcanology and Geothermal Research, 65, 249-263. CANON-TAPIA, E., WALKER, G. P, L. & HERREROBERVERA, E. 1995. The internal structure of lava flows - insights from AMS measurements I: near vent aa. Journal of Volcanology and Geothermal Research, 70, 21-36. CANON-TAPIA, E., WALKER, G. P. L. & HERREROBERVERA, E. 1997. The internal structure of lava flows - insights from AMS measurements II: Hawaiian pahoehoe, toothpaste lava and 'a'a. Journal of Volcanology and Geothermal Research, 76, 19-46.

223

CHAVEZ-ALVAREZ, M. J. & CANON-TAPIA, E. 2003. The AMS of dykes: should £max axes be imbricated everywhere? In: Annual Fall Meeting (edited by AGU) 84 Suppl. EOS, F531. DRAGONI, M., LANZA, R. & TALLARICO, A. 1997. Magnetic anisotropy produced by magma flow: theoretical model and experimental data from Ferrar dolerite sills (Antarctica). Geophysical Journal International, 128, 230-240. DUNLOP, D. J. & OZDEMIR, O. 1997. Rock Magnetism. Fundamentals and Frontiers. Cambridge University Press, Cambridge. ELL WOOD, B. B. 1978. Flow and emplacement direction determined for selected basaltic bodies using magnetic susceptibility anisotropy measurements. Earth and Planetary Science Letters, 41, 254264. ELL WOOD, B. B. 1979. Anisotropy of magnetic susceptibility variations in icelandic columnar basalt. Earth and Planetary Science Letters, 42, 209212. ELLWOOD, B. B. & FISK, M. R. 1977. Anisotropy of magnetic susceptibility variations in a single icelandic columnar basalt. Earth and Planetary Science Letters, 35, 116-122. ERNST, R. E. 1990. Magma flow directions in two mafic Proterozoic dyke swarms of the Canadian shield: as estimated using anisotropy of magnetic susceptibility data. In: PARKER, A. J., RICKWOOD, P. C. & TUCKER, D. H. (eds) Mafic dykes and emplacement mechanisms. A.A. Balkema, Rotterdam, 231-235. ERNST, R. E. & BARAGAR, W. R. A. 1992. Evidence from magnetic fabric for the flow pattern of magma in the Mackenzie giant radiating dyke swarm. Nature, 356, 511-513. FERNANDEZ, A. 1987. Preferred orientation developed by rigid markers in two-dimensional simple shear strain: a theoretical and experimental study. Tectonophysics, 136, 151-158. FERRE, E. C., BORDARIER, C. & MARSH, J. S. 2002. Magma flow inferred from AMS fabrics in a layered mafic sill. Insizwa, South Africa. Tectonophysics, 354, 1-23. FREEMAN, B. 1985. The motion of rigid ellipsoidal particles in slow flows. Tectonophysics, 113, 163183. GAY, N. C. 1966. Orientation of mineral lineation along the flow direction in rocks: a discussion. Tectonophysics, 3, 559—564. GEOFFROY, L. CALLOT, J. P., AUBOURG, C. & MOREIRA, M. 2002. Magnetic and plagioclase linear fabric discrepancy in dykes: a new way to define the flow vector using magnetic foliation. Terra Nova, 14, 183-190. GLEN, J. M. G., RENNE, P. R., MILNER, S. C. & COE, R. S. 1997. Magma flow inferred from anisotropy of magnetic susceptibility in the costal ParanaEtendeka igneous province: Evidence for rifting before flood volcanism. Geology, 25, 1131-1134. GREGOIRE, V., DE SAINT-BLANQUAT, M., NEDELEC, A. & BOUCHEZ, J. L. 1995. Shape anisotropy versus magnetic interactions of magnetite grains: experiments and application to AMS in granitic rocks. Geophysical Research Letters, 20, 2765-2768.

224

E. CANON-TAPIA

HALVORSEN, E. 1974. The magnetic fabric of some dolerite intrusions, northeast Spitsbergen; implications for their mode of emplacement. Earth and Planetary Science Letters, 21, 127-133. MARGRAVES, R. B., JOHNSON, D. & CHAN, C. Y. 1991. Distribution anisotropy: The cause of AMS in igneous rocks? Geophysical Research Letters, 18, 2193-2196. HENRY, B., PLENIER, G. & CAMPS, P. 2003. Post-emplacement tilting of lava flows inferred from magnetic fabric study; the example of Oligocene lavas in the Jeanne d'Arc peninsula (Kerguelen Islands). Journal of Volcanology and Geothermal Research, 127, 153-164. HERRERO-BERVERA, E., WALKER, G. P. L., CANONTAPIA, E. & GARCIA, M. O. 2001. Magnetic fabric and inferred flow direction of dikes, conesheets and sill swarms, Isle of Skye, Scotland. Journal of Volcanology and Geothermal Research, 106, 195-210. HROUDA, F. 1982. Magnetic anisotropy of rocks and its application in geology and geophysics. Geophysical Survey, 5, 37-82. HROUDA, F. & KAHAN, S. 1991. The magnetic fabric relationship between sedimentary and basement nappes in the High tatra Mountains, N. Slovakia. Journal of Structural Geology, 13, 431-442. HROUDA, F., CHLUPACOVA, M. & NOVAK, J. K. 2002. Variations in magnetic anisotropy and opaque mineralogy along a kilometer deep profile within a vertical dyke of the syenogranite porphyry at Cinovec (Czech Republic). Journal of Volcanology and Geothermal Research, 113, 37-47. ILDEFONSE, B., ARBARET, L. & DIOT, H. 1997. Rigid particles in simple shear flow: is their preferred orientation periodic or steady-state? In: BOUCHEZ, J. L., HUTTON, D. H. W. & STEPHENS, W. E. (eds) Granite: from segregation of melt to emplacement fabrics. Petrology and structural geology 8. Kluwer academic publishers, Dordrecht, The Netherlands, 177-185. ILDEFONSE, B., LAUNEAU, P., BOUCHEZ, J. L. & FERNANDEZ, A. 1992a. Effect of mechanical interactions on the development of shape preferred orientations: a two-dimensional experimental approach. Journal of Structural Geology, 14, 73-83. ILDEFONSE, B., SOKOUTIS, D. & MANCKTELOW, N. S. 19926. Mechanical interactions between rigid particles in a deforming ductile matrix. Analogue experiments in simple shear flow. Journal of Structural Geology, 14, 1253-1266. INCORONATO, A., ADDISON, F. T., TARLING, D. H., NARDI, G. & PESCATORE, T. 1983. Magnetic fabric investigations of pyroclastic deposits from Phlegrean Fields, southern Italy. Nature, 306, 461^63. JEFFERY, G. B. 1922. The motion of ellipsoidal particles immersed in a viscous fluid. Proceedings of the Royal Society of London, 102, 161-179. JEZEK, J., MELKA, R., SCHULMANN, K. & VENERA, Z. 1994. The behaviour of rigid triaxial ellipsoidal particles in viscous flows - modeling of fabric evolution in a multiparticle system. Tectonophysics,229, 165-180.

KHAN, M. A. 1962. The anisotropy of magnetic susceptibility of some igneous and metamorphic rocks. Journal of Geophysical Research, 67, 2873-2885. KNIGHT, M. D. & WALKER, G. P. L. 1988. Magma flow directions in dikes of the Koolau Complex, Oahu, determined from magnetic fabric studies. Journal of Geophysical Research, 93, 4301-4319. KOLOFIKOVA, O. 1976. Geological interpretation of magnetic properties of basalts. An example of the Chrisby Les lava flow of the Velky Rouny volcano (Nizky Jesenick, Cas). Casopispro Mineralogii a Geologii, 21, 387-396 (in Czech). LANZA, R. 1994. Paleomagnetic investigations in Victoria Land. Terra Antartica, 1, 123-124. Liss, D., HUTTON, D. H. W. & OWENS, W. H. 2002. Ropy flow structures: A neglected indicator of magma flow direction in sills and dikes. Geology, 30, 715-718. MACDONALD, W. D., PALMER, H. C. & HAYATSU, A. 1992. Egan Range Volcanic Complex, Nevada: geochronology, paleomagnetism and magnetic fabrics. Physics of the Earth and Planetary Interiors, 74, 109-126. MANGA, M. 1998. Orientation distribution of microlites in obsidian. Journal of Volcanology and Geothermal Research, 86, 107-115. MARSH, B. D. 2000. Magma Chambers. In: SIGURDSSON, H., HOUGHTON, B. F., McNuTT, S. R., RYMER, H. & STIX, J. (eds) Encyclopedia of Volcanoes. Academic Press, San Diego, 191-206. MORRIS, A. 2000. Magnetic fabric and palaeomagnetic analyses of the Plio-Quaternary calc-alkaline series of Aegina Island, South Aegean volcanic arc, Greece. Earth and Planetary Science Letters 176, 91-105. NICOLAS, A. 1992. Kinematics in magmatic rocks with special reference to gabbros. Journal of Petrology, 33, 891-915. NOMADE, S., THEVENIAUT, H., CHEN, Y., POUCLET, A. & RIGOLLET, C. 2000. Paleomagnetic study of French Guyana early Jurassic dolerites: hypothesis of a multistage magmatic event. Earth and Planetary Science Letters, 184, 155-168. NYE, J. F. 1960. Physical Properties of Crystals. Oxford University Press, Oxford. ORT, M. H., Rosi, M. & ANDERSON, C. D. 1999. Correlation of deposits and vent locations of the proximal Campanian Ignimbrite deposits, Campi Flegrei, Italy, based on natural remanent magnetization and anisotropy of magnetic susceptibility characteristics. Journal of Volcanology and Geothermal Research, 91, 167-178. PARK, J. K., TANCZYK, E. I. & DESBARATS, A. 1988. Magnetic fabric and its significance in the 1400 Ma Mealy diabase dykes of Labrador, Canada. Journal of Geophysical Research, 93, 13689-13704. PERROUD, H., CALZA, F. & KHATTACH, D. 1991. Paleomagnetism of the silurian volcanism at Almaden, southern Spain. Journal of Geophysical Research, 96, 1949-1962. PLENIER, G., CAMPS, P., HENRY, B. & NICOLAYSEN, K. 2002. Paleomagnetic study of Oligocene (2430 Ma) lava flows from the Kerguelen Archipelago (southern Indian Ocean): directional analysis and

HISTORICAL ACCOUNT OF AMS OF LAVAS & DYKES magnetostratigraphy. Physics of the Earth and Planetary Interiors, 133, 127-146. POTTER, D. K. & STEPHENSON, A. 1988. Singledomain particles in rocks and magnetic fabric analysis. Geophysical Research Letters, 15, 10971100. PRASAD, J. N., SATYANARAYANA, K. V. V. & GAWALI, P. B. 1999. Palaeomagnetic and low-field AMS studies of Proterozoic dykes and their basement rocks around Harohalli, South India. Journal of the Geological Society of India, 54, 57-67. PURANEN, R., PEKKARINEN, L. J. & PESONEN, L. J. 1992. Interpretation of magnetic fabrics in the early Proterozoic diabase dikes of Keuruu, central Finland. Physics of the Earth and Planetary Interiors, 72, 68-82. RAPALINI, A. E. & LOPEZ DE LUCHI, M. 2000. Paleomagnetism and magnetic fabric of middle Jurassic dykes from western Patagonia, Argentina. Physics of the Earth and Planetary Interiors, 120, 11-27. RAPOSO, M. I. B. 1997. Magnetic fabric and its significance in the Florianopolis dyke swarm, southern Brazil. Geophysical Journal International, 131, 159-170. RAPOSO, M. I. B. & D'AGRELLA-FILHO, M. S. 2000. Magnetic fabrics of dike swarms from SE Bahia State, Brazil: their significance and implications for Mesoproterozoic basic magmatism in the Sao Francisco Craton. Precambrian Research, 99, 309-325. RAPOSO, M. I. B. & ERNESTO, M. 1995. Anisotropy of magnetic susceptibility in the Ponta Grossa dyke swarm (Brazil) and its relationship with magma flow direction. Physics of the Earth and Planetary Interiors, 87, 183-196. ROCHETTE, P. 1988. Inverse magnetic fabric in carbonate-bearing rocks. Earth and Planetary Science Letters, 90, 229 -237. ROCHETTE, P., AUBOURG, C. & PERRIN, M. 1999. Is this fabric normal? A review and case studies in volcanic formations. Tectonophysics, 307, 219-234. ROCHETTE, P., JACKSON, M. & AUBOURG, C. 1991«. Rock magnetism and the interpretation of anisotropy of magnetic susceptibility. Reviews of Geophysics, 30, 209-226. ROCHETTE, P., JENATTON, L., DUPUY, C., BOUDIER, F. & REUBER, I. 19916. Diabase dikes emplacement in the Oman Ophiolite: a magnetic fabric study with reference to geochemistry. In: PETERS, T., NICOLAS, A. & COLEMAN, R. G. (eds) Ophiolite genesis and evolution of the oceanic lithosphere. Ministry of petroleum and minerals, Sultanate of Oman, 55—82. STACEY, F. D. 1960. Magnetic anisotropy of igneous rocks. Journal of Geophysical Research, 65, 2429— 2442.

225

STAUDIGEL, H., GEE, J., TAUXE, L. & VARGA, R. J. 1992. Shallow intrusive directions of sheeted dikes in the Troodos ophiolite: anisotropy of magnetic susceptibility and structural data. Geology, 20, 841-844. STAUDIGEL, H., TAUXE, L., ET AL. 1999. Geochemistry and intrusive directions in sheeted dikes in the Troodos ophiolite: Implications for mid-ocean ridge spreading centers. Geochemistry Geophysics Geosy stems, 1, 1999. STEPHENSON, A. 1994. Distribution anisotropy: two simple models for magnetic lineation and foliation. Physics of the Earth and Planetary Interiors, 82, 49-53. SYMONS, D. T. A. 1967. The magnetic and petrologic properties of a basalt column. Geophysical Journal of the Royal Astronomical Society, 12, 473-490. SYMONS, D. T. A. 1975. Age and flow direction from magnetic measurements on the historic Aiyansh flow, British Columbia. Journal of Geophysical Research, 80, 2622-2626. TARLING, D. & HROUDA, F. 1993. The Magnetic Anisotropy of Rocks. Chapman and Hall, London. TAUXE, L., GEE, J. & STAUDIGEL, H. 1998. Flow directions in dikes from anisotropy of magnetic susceptibility data: The bootstrap way. Journal of Geophysical Research, 103, 17775-17790. TAYLOR, G. I. 1923. The motion of ellipsoidal particles in a viscous fluid. Proceedings of the Royal Society of London, 103A, 58-61. UYEDA, S., FULLER, M. D., BELSHE, J. C. & GIRDLER, R. W. 1963. Anisotropy of magnetic susceptibility of rocks and minerals. Journal of Geophysical Research, 68, 279-291. VARGA, R. J., GEE, J., STAUDIGEL, H. & TAUXE, L. 1998. Dike surface lineations as magma flow indicators within the sheeted dike complex of the Troodos Ophiolite, Cyprus. Journal of Geophysical Research, 103, 5241-5256. WALDERHAUG, H. 1993. Rock magnetic and magnetic fabric variations across three thin alkaline dykes from Sunnhordland, Western Norway; influence of initial mineralogy and secondary chemical alterations. Geophysical Journal International, 115, 97-108. WING-FATT, L. & STACEY, F. D. 1966. Magnetic anisotropy of laboratory materials in which magma flow is simulated. Pure and Applied Geophysics, 64, 78-80. WOLFF, J. A., ELLWOOD, B. B. & SACHS, S. D. 1989. Anisotropy of magnetic susceptibility in welded tuffs: application to a welded-tuff dyke in the tertiary Trans-pecos Texas volcanic province. Bulletin of Volcanology, 51, 299-310.

This page intentionally left blank

Theoretical aspects of particle movement in flowing magma: implications for the anisotropy of magnetic susceptibility of dykes EDGARDO CANON-TAPIA & M. JAZMIN CHAVEZ-ALVAREZ

CICESE, Department of Geology, P.O. Box 434843, San Diego CA, 92143, USA (e-mail: [email protected]) Abstract: Most studies of the anisotropy of magnetic susceptibility (AMS) of dykes have assumed that the axes of maximum susceptibility (&max) should define an opposed imbrication pointing along the direction of magma flow, and that this orientation should be preserved along the dyke. This assumption is partially based on a model predicting the orientation of ellipsoidal particles floating in a moving liquid although the model actually predicts a cyclic movement of the particles that has been overlooked in most AMS studies without further justification. The consequences of considering the full rotation of the ellipsoidal particles, as actually predicted by the theory, in the expected AMS of dykes are examined in this work. The complete version of the motion of ellipsoidal particles is then incorporated in a model of rnagma movement that takes into consideration the distribution of shear deformation within the dyke as predicted from the velocity gradient of the moving magma. Results of this model show that both particle elongation and the amount of shear that is sampled will affect the quality of the AMS results. By paying attention to the systematic variations of the AMS predicted by the theory, however, it is possible to devise sampling schemes that can be used to add more confidence to the interpretation of the AMS results. Although based on an idealized scenario of magma movement within a dyke, the model developed here explains satisfactorily the sometimes observed variation of AMS along flow direction in one dyke, and provides a simple explanation for many of the 'abnormal' magnetic fabrics that have been reported in dyke swarms around the world.

Whether the maximum (fcmax), intermediate (£int) or minimum (k^n) susceptibility axes are selected to infer flow direction of lava flows and dykes, most AMS studies made to date have assumed that the orientation of the principal susceptibilities bears a unique relation to the direction of flow of magma or lava (see references in Canon-Tapia, this volume). This deterministic approach started to be adopted following the work by Knight & Walker (1988) in the dyke swarm of the Ko'olau volcano. They found that in nearly 84% of the dykes that they studied, the orientation of &max axes was parallel to the flow direction inferred from other indicators, whereas the km^n axes were normal to the dyke walls. The remaining 16% of their dykes yield a different orientation of the magnetic fabric, but received little attention from these authors. The orientation of the principal susceptibilities relative to the plane of intrusion found in most of the dykes studied by Knight & Walker (1988) later became known as a 'normal' magnetic fabric (Rochette et al 1991), and any fabric that was not normal came to be considered, by definition, an 'abnormal' magnetic fabric. Although some controversy can be found in earlier works, following Knight & Walker (1988) and Rochette et al. (1991), most workers to date have assumed that a normal fabric is the result of well-defined

conditions during flow, and that an abnormal fabric has to be produced by some extraordinary processes. Explanations for the abnormal magnetic fabrics in dykes have included specific mineral and physic characteristics of each intrusive or region. For example, based on a combination of two hypothesis (one magnetic and one mechanical), any dyke in which £max is perpendicular to the dyke walls is suspected to have an inverse fabric. The magnetic hypothesis concerns the existence of very small grains in which the ^max and kmin axes are switched relative to the physical axes of the particle (Potter & Stephenson 1988), whereas the mechanical hypothesis concerns the orientation of those particles based on Jeffery's equations (Jeffery 1922). Both of these hypotheses are hard to justify when more details are available, as shown by the results of granulometry tests that do not confirm the presence of SD magnetite in the same samples in which the inverse fabric is suspected (Walderhaug 1993; Raposo & Ernesto 1995; Raposo 1997; Raposo & D'Agrella-Filho 2000). Furthermore, the mechanical model of fabric acquisition adopted in most of these studies is based on a limited interpretation of the predictions made by Jeffery's equations (see below), which indicates that, in at least some cases there must be another mechanism not related to

From: MARTIN-HERNANDEZ, F., LUNEBURG, C. M., AUBOURG, C. & JACKSON, M. (eds) 2004. Magnetic Fabric: Methods and Applications. Geological Society, London, Special Publications, 238, 227-249. 0305-8719/04/ $15.00 © The Geological Society of London 2004.

THEORETICAL AMS OF A DYKE

The AMS of a group of ellipsoidal grains in magma flowing in a dyke or sill The AMS (Kt) of a multi-particle system formed by N magnetic grains can be evaluated theoretically from the expression proposed by Hrouda (1993): (1)

In this equation, A is the orientation matrix of each grain and K is its corresponding susceptibility tensor in diagonal form. K is defined as: (2)

where Kt is the particle intrinsic susceptibility and NJ are the demagnetization factors, which are directly related to the particle shape (Osborn 1945; Stoner 1945). The calculated Kt in equation 1 is a real symmetric matrix with real eigenvalues that correspond to the characteristic maximum, intermediate and minimum susceptibilities of the system formed by the N grains. By paying attention to the mechanical behaviour of each grain in this multi-particle system, it is possible to find the evolution of the susceptibility tensor of the system as a function of deformation. Dragoni et al (1997) followed this approach and solved the equations of movement of a uniaxial particle immersed in a simple shear flow in combination with equation 1. As shown by Dragoni et al. (1997), it is convenient for the calculations to express the solution to the equations of movement in terms of Euler angles (9 and <£) of the particles:

(3) (4)

where r ~ a/b (a and b are the semi-axes of the ellipsoid) is the elongation ratio of the grain, t is time, 7 = 7, is the shear strain resulting from the application of a shear rate during a time t, and $o and 0 are the initial values of 9 and 0 at t = 0. Using equations 3 and 4 it is therefore possible to obtain the orientation of any ellipsoidal particle with given elongation ratio for any value of deformation 7, provided that the initial orientation of the particle at the time t = 0 (when it is assumed that 7 = 0) is known. Due to the fact that all the particles are ellipsoids of revolution, it is noted that the T parameter defined by Jelinek (1981) will be equal to 1 for each of the

229

prolate grains and equal to — 1 for each of the oblate particles, irrespective of the value of the aspect ratio characterizing them. Consequently, the shape of each particle is better described by considering the value of r alone. Computer simulation of the AMS of a group of particles in flowing magma

Program description Based on equations 1 to 4, we wrote a MATLAB program to examine the evolution of the AMS of a multi-particle system as a function of increasing deformation. In Figure 1 we show a flow diagram of the program. At the beginning of each session three variables have to be specified by the user: the number of particles forming the system (A/), the shape of the particles (oblate or prolate) and their elongation ratio (r). In the strictest sense it is not necessary to define independently the shape of the particle because the value of the elongation ratio determines whether the particle is prolate (r < 1) or oblate (r > 1). We decided to modify the original definition of r given by Dragoni et al. (1997), however, to facilitate a qualitative comparison between particles of different shapes but equal aspect ratios. To this end, r was forced to represent the ratio of the shortest/longest dimension of each particle. Consequently, a value of r to be entered in the program could be only in the range 0.050.95 for both prolate and oblate shapes. Distinction between different shapes is therefore obtained by specifying the shape of the particle as an independent variable. Thus, the value of r entered in the program could be used without further manipulation to solve equations 3 and 4 if the grains are prolate, but the value 1/r would be used if the grain shape is defined as oblate. Once the shape and elongation ratio of the particles are entered in the program, the initial orientation of the particles is required. Our program addresses this problem (Fig. 1, step 2) by generating 3N random numbers that are assigned to three variables 'JT, *y and 'Z'. The engine used to create these random numbers is the command 'randn' in the MATLAB platform, which generates random numbers normally distributed with mean 0 and variance 1. The numbers on these variables are then used to form triplets representing the three Cartesian components of the N axes corresponding to the orientations of the symmetry axes of the N grains forming the system. The axial distributions thus obtained were nearly uniform, as revealed by the use of

230

E. CANON-TAPIA & M. J. CHAVEZ-ALVAREZ

Fig. 1. Flow diagram of the program used to calculate the AMS of a multi-particle system as a function of shear deformation.

magnetic susceptibility tensor of a single particle of a given shape using the demagnetization factors given by Osborn (1945) and a fixed value of intrinsic susceptibility (eq. 2; Fig. 1, step 3). The value of kf obtained in step 3 is used in step 4 to calculate the AMS of the N grains randomly distributed utilizing equation 1. Although for multi-particle systems composed of identical uniaxial particles the eigenvectors of the magnetic susceptibility are parallel to those of the orientation tensor (e.g. Jezek & Hrouda 2002), in order to model systems formed by either oblate or prolate particles it becomes necessary to maintain the distinction between the eigenvectors of the orientation matrix and the eigenvectors of the magnetic susceptibility in the program. This is better illustrated by comparing the meaning of the larger eigenvalue of the orientation matrix (SI) in systems formed by either prolate grains or oblate grains. In either case, SI represents the mean orientation of a dominant cluster that may have been formed by the physical axes of the particles (Woodcock 1977). This cluster represents the orientation of the mean &max axis in the case of prolate particles, but it represents the orientation of the mean fcmin axis in the case of oblate particles. Clearly, although the eigenvectors of both tensors are parallel in both cases, the corresponding eigenvalues are not proportional. In step 5 of the program, the principal susceptibilities of the system at the current deformation stage are shown graphically. In step 6, the value of deformation is increased by a prescribed amount (commonly one full unit of gamma), and in step 7 the orientation of the axes corresponding to the new value of deformation is calculated (eqs. 3 and 4). The cycle of calculations from step 4 is then repeated as many times as desired. Although solution of equations 3 and 4 can be achieved without relying on an iterative scheme like that just described, it was found useful to follow this approach to better appreciate the fundamental aspects of the various system behaviours.

density contour plots calculated using the approach described by Fisher et al. (1993). Local clusters could be observed in a number General conditions of the simulations of the initial distributions at values of 7V = 200 probably due to the relatively small number of To assess the influence of the shape of the grains particles used in each system. These clusters we used a range of the elongation ratio between formed at random are in part responsible for 0.05-0.95 for both oblate and prolate shapes. the various system behaviours reported below. In addition, to evaluate whether the AMS The presence of such clusters, however, does depends on the initial distribution of the grains not influence the fundamental aspects of those we studied the behaviour of systems having the same shape and elongation ratios but with differbehaviours. Once the initial conditions of the system have ent initial random orientation of the grains. In been specified, the program calculates the this section we describe results obtained with

THEORETICAL AMS OF A DYKE

systems formed by N = 200 particles and 7 = 0 , 1 , . . . , 20. This value of N was selected as representative of the number of particles commonly included in image analysis of thin sections. To check for the effect of this variable in the simulations we ran the program with various values of N, up to 10000. This revealed that the variability of the system behaviours is to some extent a function of N, as a larger number of particles influences the relative frequencies of behaviours reported below. Nevertheless, the fundamental aspects of the various behaviours seem to be preserved regardless of the value of TV. As for the range selected for 7, although deformation values much larger than 20 are expected in flowing magma in a dyke or a lava flow (de Rosa et al. 1996), the periodicity of the movement of the particles makes unnecessary a larger number of iterations for most systems (only particles with r < 0.30 do not complete one full cycle in the orientation of their axes, and even particles with r —0.15 have completed a turn of 180°). Consequently, the selected range of deformation is enough to illustrate the fundamental aspects of each behaviour described. The total number of systems considered here was 1 200, of which 600 were of prolate grains and 600 were of oblate grains (Tables 1 & 2). In the presentation of results we used an orientation of axes that bears a closer resemblance to geological situations than the orientation of axes used in the paper by Dragoni et al. (1997). The relation between the Cartesian coordinate system used to solve equations 3 and 4, the direction of flow, the orientation of the velocity gradient, the plane of shear (or intrusion), and the stereograms used to represent the results is illustrated in Figure 2. All subsequent stereograms (lower hemisphere) use this orientation.

Results of the simulations of fabric acquisition Due to the cyclic motion of a particle predicted by equations 3 and 4, it is a natural consequence of the model that each of the principal susceptibility axes eventually should describe a closed trajectory as well. Although in principle the mean £max axis of prolate particles will be close to the direction of flow if a complete cycle of rotation is considered, there are some aspects of a multi-particle system that may deviate the location of the mean from this position. For instance, differences in the velocity of rotation of the particles due to their orientation and aspect ratio result in fluctuations of the fabric intensity

231

as a function of deformation (Jezek et al 1996; Canon-Tapia & Chavez-Alvarez 2004). Also, due to the dependence of the period of rotation on the aspect ratio of the particles, the position of the principal axes of susceptibility distributes unevenly along the trajectory that they describe. Both of these factors produce different orientations of mean &max axes, and associated regions of confidence if these are calculated for equal ranges of deformation in systems with different particle shape. We took advantage of such differences to establish an objective classification scheme that could be used to identify patterns of behaviour based on the results of the 1 200 systems modelled. The classification scheme consisted in calculating the average principal susceptibilities and confidence regions of the 21 AMS tensors generated for each run of our program. Each of these 21 tensors represents the AMS of one system at a given deformation stage (i.e. 7 = 0 , 1 , 2 , . . . , 20). It is stressed that the average of the 21 AMS tensors does not have a well-defined physical meaning in terms of the AMS expected at a given place in a dyke, but it is a convenient feature that can be used in classifying system-behaviours that share characteristics of two or more of the mentioned types. A more meaningful form to calculate averages from these results is described in the second half of the paper. Using the average susceptibility of the 21 AMS tensors of each program-run, it was possible to identify five general types of behaviours (Fig. 3). Tables 1 and 2 show the summary of the system behaviours generated for r = 0.05-0.95 by the 600 prolate systems and 600 oblate systems respectively. In these tables it is observed that the two kinds of particle shapes (oblate and prolate) generate almost the same system behaviours in approximately the same proportions (Fig. 4), and that not all the behaviours appear in the same range of r. In most of the cases (87% in prolate systems and 90% in oblate systems), the average &max of the 21 tensors used in the classification remained on the shear plane whereas the average &min remained on (or near) the poles of this plane. Based on the orientation of the average fcmax, however, it is possible to distinguish two cases within this large group. If the average fcmax forms a small angle with the flow direction (i.e. the region of confidence around this axis does not include the real flow direction), the system behaviour was named 'Im', whereas if the average &max is effectively parallel to flow direction (i.e. the region of confidence includes the flow direction) it was called 'M' (Figs 3a & 3b respectively). In a very limited number of cases (6% in prolate systems

Table 1

a

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40

b

c

0.05

0.1

0.15

0.2

0.25

0.3

0.35

0.4

0.45

0.5

0.55

0.6

0.65

0.7

0.75

0.8

0.85

0.9

0.95

Im Im Im Im Im Im Im Im Im Im Im Im Im Im Im Im Im Im Im Im Im Im Im Im Im

Im Im Im Im Im Im Im Im Im Im Im Im Im Im Im Im Im Im Im Im Im Im Im Im Im

m m m m m M m m M m m m m m m M M M m M M m m m m

M M M M M M M M M m M M M m m M M m m M M M M M m

M M m M M M M M m M M M M M M M M M m M M M M M M

M M M M M M M M M M M M M M M M M M M M M M M M M

M M M M M M M M M M M M M M M M M M M M M M M M M

M M M M M M M M M M M M M M M M M M M M M M M M M

M M M M M M M M M M M M M M M M M M M M M M M M M M M M M M M M

M M M M M M M M M M M m M M M M M M M M M M M M M M M M M M M M

M M M M M M M M M M M M M M M M M M M M M M M M M M M M M M M M

M M M M M M M M M M M M M M M M M M M M M M M M M M M M M M M M

M M M M M M M M M M M M M M M M M M M M M M M M M M M M M M M M

M M M M M M M M M M M M M M M M M M M i M M M M M M M M M M M M M M M M M M M M

M M M M M M M M M I M M M M M M M M M M M M M M M M M M M M M M M M M M M j M M

M M M M M M M i M M I M M M M M M i M M M i m M M M M i M M M M i M I i I M I I

I M M M i I M M I i M M I M i M M NO M M M NO I m M M NO I M NO I M I M M i NO M M

i NO M i M I i i NO NO M NO NO i i M i I M i M M I I M NO I I NO NO I I i I M I M i NO m

M NO i NO I NO M I i I I NO i I I NO I I I I NO NO I I I I NO NO NO M M NO I I I NO NO I NO NO

1

Table 2

b

a

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40

c

0.05

0.1

0.15

0.2

0.25

0.3

0.35

0.4

0.45

0.5

0.55

0.6

0.65

0.7

0.75

0.8

0.85

0.9

0.95

Im Im Im Im Im Im Im Im Im Im Im Im Im Im Im Im Im Im Im Im Im Im Im Im Im

Im Im Im Im Im Im Im Im Im Im Im Im Im Im Im Im Im Im Im Im Im Im Im Im Im

m m M m m M m m m M M M m M M m m m M m M M m m M

M M M M m M M M M m m M M M M m M M M m M M M M M

M M M M M M M M M M M M M M M M M M M M M M M M M

M m M M M M M M M M M M M M M M M M M M M M M M M

M M M M M M M M M M M M M M M M M M M M M M M M M

M M M M M M M M M M M M M M M M M M M M M M M M M

M M M M M M M M M M M M M M M M M M M M M M M M M M M M M M M M

M M M M M M M M M M M M M M M M M M M M M M M M M M M M M M M M

M M M M M M M M M M M M M M M M M M M M M M M M M M M M M M M M

M M M M M M M M M M M M M M M M M M M M M M M M M M M M M M M M

M M M M M M M M M M M M M M M M M M M M M M M M M M M M M M M M

M M M M M M M M M M M M M M M i M M M I M M M M M M M M M M M M M M M M M M M M

M M M M M M M M I M M i M M M M M M M M M M I M M M M M M M M M M I M M I M I I

M I I M M I M I M M M M M M M i M M M M M M M M M i M M I M M M I I M M I M M M

M I M M M I I I I I M I M M M M I M M M M NO I M M I M M I M I M I M I M I I M I

M I M I M M I I M I I M I I I I M M NO I I I NO I I I M I I I I I M M NO M M M NO i

NO NO NO NO I NO NO I I I I I I NO I NO NO i I I NO NO I I I NO I I I NO NO M I I NO NO NO M I I

234

E. CANON-TAPIA & M. J. CHAVEZ-ALVAREZ

Fig. 2. Schematic relationship between the plane of intrusion, magma flow direction, orientation of the velocity gradient and coordinate system in 'real world' view (a) and equal area projections (b). The projection in b (and in each of the diagrams showing model results) is a lower hemisphere. The sense of rotation of the axis of symmetry of the particle shown in (a) as a result of the velocity gradient of the right side of the intrusive is illustrated in the projection in (b).

and 4% in oblate systems) the average A;max formed a very small angle with the flow direction but within the shear plane. In this case the system behaviour was named 'm' to distinguish it from the behaviour Im in which the imbrication of the average &max is not in the shear plane. The angle between fcmax and flow direction in this case was very small and it is not observed in Figure 3c where this type of behaviour is

illustrated. However, the numerical results of the calculations confirmed that the regions of confidence of the systems displaying this type of behaviour were small enough to not include the direction of flow. The fourth type of system behaviour was named T, and is characterized by having an average k-mi parallel to flow direction while the average kmax is on the shear plane and the average &min is on its pole

THEORETICAL AMS OF A DYKE

235

Fig. 3. Examples of the five magnetic behaviours obtained by modelling the AMS of a multi-particle system as a function of shear deformation. In each diagram the initial state of the system plus 20 steps of deformation are shown.

(Fig. 3d). Although the orientation of the principal susceptibilities of some of these behaviours seem to be unrelated to the distribution of the individual AMS of the 21 deformation stages, the difference in the fabric intensity between these tensors and the symmetry of the orientations of the principal susceptibilities shown contribute to create the counterintuitive orientation of the mean principal susceptibilities shown in the figure. Finally, the fifth type of magnetic

behaviour was named T. We considered this behaviour as distinct from M because the average /cmin of the type i is on the shear plane and not in its pole as in the case of M, even when both i and M have the average fcmax parallel to flow direction (Fig. 3e). In all these descriptions it was implicit that the confidence regions around the mean principal axes were less than 25°, which is an upper limit to consider the average as significant in a statistical

236

E. CANON-TAPIA & M. J. CHAVEZ-ALVAREZ

Fig. 4. Proportions of magnetic behaviours for systems of either prolate or oblate particles.

sense (Jelinek 1981). In some of our simulations, two of the regions of confidence had one semiaxis larger than this threshold value, although little seemed to be gained by distinguishing these cases from the five types of behaviour described. In some other instances, however, all three mean susceptibilities had confidence regions with both axes bigger than 25°. In these cases the magnetic fabric was labelled as NO in Tables 1 and 2, irrespective of the orientation of the mean axes. The proportion of cases in which this occurred is relatively small (Fig. 4), but together with the I, i and Im types they represent nearly 24% of all the behaviours observed.

Implications of the simulations of fabric acquisition based on Jeffery's equations One of the main features of the theoretical formulation of the equations of movement of isolated particles immersed in a laminar flow is that

the particles will rotate therefore defining a cyclic movement of the mineral fabric (Jeffery 1922; Dragoni et al. 1997). The 'speed' of rotation is a function of the elongation ratio of the particles, and can be expressed in terms of the deformation experienced by the flowing magma. This property of the model explains many of the features of the systems listed in Tables 1 and 2. For instance, it is possible to identify three zones (a, b and c) in these tables, each being characterized by the preponderance of a given magnetic behaviour. Zone a is defined for r < 0.1, and contains exclusively behaviours of the Im type. Zone b has 0.15 < r < 0.65 and is characterized by the content of behaviours of the types M and m. Zone c is defined for r > 0.7. It is in this zone that it is possible to find four of the five behaviours identified. Furthermore, many systems in this zone generated behaviours in which all three regions of confidence were >25°. The overall increase in the size of the confidence regions observed in the systems of this zone reflects the absence of a continuous preferred orientation of the particles that is sustained during a large enough amount of deformation, as revealed when the evolution of the system is studied in detail (Chavez-Alvarez & Canon-Tapia 2002). Although a much more detailed examination of the results presented in this section can be helpful to better appreciate some general aspects of fabric acquisition (Canon-Tapia & ChavezAlvarez 2004), in the present context it is more convenient to focus on the most general result (i.e. that the principal susceptibilities of a multiparticle system will follow a cyclic movement as a function of deformation). By having removed the symmetry constraints in the initial distribution of particles imposed in the original work by Dragoni et al. (1997), it has been shown that even in conditions dominated by laminar flow it is possible to find /rmax axes at almost any angle to the flow direction or to the plane of deformation, depending on the amount of strain that is examined. Consequently, it is clear that by adopting Jeffery's equations as the basis of the interpretation of AMS results in lava flows and dykes, it must be concluded that the principal susceptibilities do not bear a unique relation to flow direction. Although this variety of possible orientations seems to be undesired for the purposes of inferring magma flow direction from AMS measurements, the patterns of behaviour identified in different regions of Tables 1 and 2 suggest some rules that can be used as guiding principles in the interpretation of results. For example, it is clear from the classification shown in these tables that the more elongated particles are more likely to yield an imbricated

THEORETICAL AMS OF A DYKE

AMS. Dykes containing mineral grains that are not extremely elongated (flattened) will still yield fcmax axes parallel to flow direction in many cases, but in some instances they may result in orientations of fcmax axes other than this. Consequently, it might be possible to find some specimens in one dyke that depart from the 'normal' fabric without implying that turbulent flow or post-emplacement processes have altered the flow-acquired preferred orientation of minerals. In the cases of systems formed by more isometric particles, the flow-related fabric may be the most difficult to relate to flow direction. The results of Tables 1 and 2 suggest, however, that a coherent fabric may be developed even in these cases. Another aspect highlighted by the results of this section is that different mineral subfabrics can be present in a sample representing a given stage of deformation. The orientation of each subfabric will depend on the shape characteristic of each mineral type. Identification of each subfabric can be achieved by optical measurements, but the AMS results will yield an integrated view of all of them (including a weighting factor due to the bulk susceptibility of each mineral type (Canon-Tapia 2001)). This result provides a simple explanation for the sometimes-observed lack of correspondence between the apparent orientation of the magnetic fabric and a specific mineral fabric of a nonmagnetic species, such as plagioclases (e.g. Callot & Guchet 2003). All these findings are important for a better understanding of the origin of'abnormal' fabrics in dykes, but cannot be applied directly to examine the AMS of real dykes or lava flows due to some basic aspects of the macroscopic behaviour of flowing magma. To fully appreciate the implications of the mechanical model of ellipsoidal movement in the AMS of this type of rocks, it is therefore necessary to focus our attention on the distribution of shear deformation within a single cooling unit as constrained by magma movement during its emplacement. Additionally, the scale over which Jeffery's equations are applicable and the scale over which the fabric results are to be obtained, has to be considered. All of these problems are addressed in the next section. Model of the AMS of a dyke In this part of the paper we present a model that can be used to predict some important features of the magnetic fabric of dykes acquired as the result of the movement of magma. This model

237

is based on the equations of movement of isolated particles (eqs. 3 & 4), and incorporates the basic aspects of multi-particle systems obtained in the first half of the paper. Although magma flow in a dyke and the development of flow-related mineral textures may present many complexities due to changes in the rheology of the magma associated with a variable content of crystals and/or bubbles, to deviations from a simple tabular shape of the intrusive, or to complex interactions with the host rock (e.g. Nicolas 1992; Smith 1998; Fialko & Rubin 1999), it is of some importance to examine the consequences of a model of fabric formation based on an idealized case as a first step in the understanding of more complex situations. One advantage of focusing on a simplified model of fabric acquisition as a first stage in the modelling process is that such a simplified model suggests tests that can be used to check for the internal congruency of results when implemented in real cases. This can result in strategies that effectively reduce the ambiguity inherited from the original model of ellipsoidal movement. Additionally, it is noted that even a simplified model such as that developed here may be directly applicable in the interpretation of magnetic fabrics obtained from many rocks. For example, mechanical interactions among particles during magma flow can be disregarded safely in many situations because crystallization of magma within a dyke seems to take place mainly as magma approaches the surface (Teasdale et al. 2003). Furthermore, if we consider that it is only near the surface when most of the mineral content characteristic of pahoehoe flows (typically <20% of groundmass microlite formation) is formed (Teasdale et al. 2003), it seems reasonable to assume that the threshold value of mechanical interaction (~15%, Arbaret et al. 1996) may have not been exceeded along much of the magma path during its ascent. Finally, it is noted that the main results of the model developed below are applicable to lava flows in a general context although it was constructed by assuming magma flow within a dyke. A more realistic model of fabric acquisition not only represents the next stage in the modelling process, but must also take into consideration aspects of flow specific to each situation to be modelled. For instance, although flow in dykes can be reasonably well approximated by a twodimensional movement even if other complexities are introduced (like mechanical interaction of particles or the effect of bubbles in the fabric acquisition), flow in lavas is probably better represented by a three-dimensional scenario in most circumstances. Also, mechanical interaction

238

E. CANON-TAPIA & M. J. CHAVEZ-ALVAREZ

between particles might be more common in lava flows than in dykes, as explained above, and changes in the flow regime are also probably more common. Nevertheless, within the scope of a general model as that developed here, the similarities in the rheological behaviour of lava flows and dykes justify the use of the same basic principles in both scenarios. In this context, it is noted that it is possible to find examples of lava flows in which mineral content is extremely small and for which our simple model should apply without much modification (Canon-Tapia & Castro 2004).

measured in lava flows (de Rosa et al. 1996) it seems that two regions of differing shear rate should suffice to model most dykes in the range of < 1 m up to 20 m thickness (see below) resulting in a reliable first-order approximation of the general distribution of shear deformation within a dyke. Thicker dykes might need more than two divisions to include their whole thickness, but the basic aspects of the model are also valid in these cases. For present purposes it is convenient to concentrate our attention on the shear distribution along the three profiles shown in Figure 5. These three profiles include the two commonest

Distribution of shear deformation within a dyke It is possible to visualize magma movement within a dyke as a Newtonian or Binghamian fluid moving within two stationary plates (e.g. Lister & Kerr 1990). Due to friction with the dyke walls the velocity gradient of the flowing magma achieves its maximum value near the walls of the intrusion, decreasing to a minimum value of zero at the centre of the dyke. The region where the velocity gradient is null is very narrow (theoretically it is just a line) in the case of a Newtonian fluid, whereas on a Binghamian fluid this region is wider and may be a large proportion of the total width of the dyke. Even in this idealized scenario, the velocity gradient of either a Binghamian or a Newtonian fluid within a dyke is not linearly related to the distance from its walls, as assumed on the model of ellipsoidal movement (JefTery 1922). Due to the small size of the minerals floating in a moving magma (typically <5mm), however, the velocity gradient of magma in a small volume of fluid (~ cm3) can be assumed to be constant, therefore satisfying the conditions required by Jeffery's equations at least locally. Although it seems reasonable to extend this approximation to a small volume of fluid that contains many particles, it is not possible to justify its validity on the whole extent of the dyke. Consequently, even a simplified model has to be adapted to reflect changes in the shear rate as a function of distance from the walls of the intrusive. One way of doing this is simply dividing the dyke in regions of constant shear rate as a function of distance from its walls. The number of divisions required to model one dyke will depend on its total width and on the presumed velocity at its centre. Although undoubtedly departures from this idealized model of magma movement are likely to occur in nature, based on the deformation

Fig. 5. (a) Scheme of a vertical dyke showing a typical velocity profile and magma flow direction. The expected amounts of deformation sampled by each of the three profiles A A', BB' and CC' are shown in (b), (c) and (d), respectively.

THEORETICAL AMS OF A DYKE

sampling schemes used when collecting specimens for the measurements of AMS. The velocity gradient of the magma along the line AA' shown on Figure 5a implies that the difference of velocities in the fluid is higher near to the walls than it is near to the centre of the dyke. Consequently, the regions of higher shear stress along this profile will be near to the dyke walls, whereas the regions of smaller shear stress will be near to the centre of the intrusion (Lister & Kerr 1990; Dragoni et al 1997). The difference between the more deformed and less deformed states depicted in Figure 5b will depend on the total thickness of the dyke and on the actual value of the velocity gradient (which may be related to the bulk velocity of the magma, or to the discharge rate). Despite the different orientation relative to the flow direction, a qualitatively similar increase of the deformation is expected along the line BB7 of Figure 5a. Fluid elements along this line are located at the same distance from the dyke walls and consequently they all experience the same shear rate during magma movement. A fluid element that is farther from the magma source, however, accumulates larger amounts of total shear than fluid elements closer to the source because it has experienced the same shear rate for a longer period of time. The stages of deformation in B7 are therefore higher than in B as is illustrated in Figure 5c. The difference between the more deformed and less deformed states depicted in this figure will depend on the total length of the profile and on the actual value of the velocity gradient. The qualitative equivalence between profiles AA7 and BB7 is reinforced in this figure, however, by assigning identical values of deformation at the extremes of the profile, although it is noted that in most real cases the exact ranges of deformation included in these profiles would be different for each case. A slightly different situation is found along the profile CCf of Figure 5a. At first sight, this profile seems equivalent to BB7 (both lines are parallel to the dyke walls), but as CC' is oriented perpendicular to the flow direction the deformation traversed by each profile is qualitatively different. Unlike BB7, along the line CC7 the velocity of magma does not change (i.e. the shear rate is the same for all the fluid elements within the profile and all the fluid elements in the profile will have experienced the same shear rate for equal periods of time). Consequently, all the fluid elements along this line will have the same total shear (Fig. 5d). The deformation of individual fluid elements within the profile increases equally for all of them as the fluid moves farther

239

from the source. Therefore, two profiles of type CC7 will not, in general, have sampled the same states of deformation. The deformation along the three lines A A7, BB7 and CC7 in a dyke can be roughly assessed quantitatively using measurements of deformation actually made in lava flows (de Rosa et al. 1996). Typical ranges of deformation at a distance of 0.5-1.Om measured from the ground, and parallel to the plane of flow, are from 55 to 76. Smaller values of deformation, between 2 and 23, were found in the central parts of the flow 2.0-2.5 m from the ground. Such a contrast in the value of shear deformation as measured far from and near to the ground in lava flows is consistent with the scheme of deformation along AA7 as proposed here for dykes. In addition, the increase on the deformation measured by de Rosa et al. (1996) at a constant distance from the ground along a line parallel to flow is also compatible with the profile of deformation along BB7 of Figure 5c. Although there is no direct measurement of the shear associated with the profile CC7, this can be assumed as valid due to the geometrical constraints imposed by the dyke itself. AMS variation along a dyke: general considerations A general model of the AMS of a dyke can then be formed by combining the general distribution of shear deformation within a dyke presented in the previous section with the theoretical results of the first part of the paper. Because the profiles AA7 and BB7 are qualitatively equivalent, it is sufficient to concentrate on modelling only a profile like A A7; the profile CC7 needs to be treated separately. We start by noticing that one of the systems formed by 200 grains modelled in the first part of this paper can be envisaged as a small volume of magma that is deformed as the result of magma movement. In this form the AMS generated by each of the systems included in Tables 1 and 2 at a given stage of deformation represents the AMS expected at a specific place in the dyke. Because more than one specimen has to be collected to infer flow direction from AMS measurements, and considering that it is impossible to collect the same system of particles twice in a dyke, it becomes necessary to combine the results of more than one system (each system at a known stage of deformation) in the same AMS diagram. It is recalled that the sequence of orientations of the principal susceptibilities portrayed in Figure 3 represents

240

E. CANON-TAPIA & M. J. CHAVEZ-ALVAREZ

the time-evolution of one particular fluid element, and it does not correspond to the AMS of 21 independent systems. Therefore, in order to calculate an average AMS that has a better defined physical meaning than the averages shown in Figure 3, it is necessary to select individual stages of each diagram, keeping track of the deformation stage that they represent, to combine them in one single AMS plot. The plot thus obtained is then more representative of the AMS results that would be found under the physical conditions used in its construction. The values of shear deformation of each system to be included in the modelled AMS can be selected using the constraints imposed by the ranges of deformation discussed above. Clearly, the number of possible combinations of shear values and specific systems is extremely large even when restricting the modelling to the relatively small number of systems shown in Tables 1 and 2. For this reason, instead of attempting to reproduce a large number of dykes it is more illustrative to concentrate our attention on some examples that represent extreme situations that can be used as limiting cases. Additionally, it is possible to simplify the modelling process by taking advantage of the cyclic behaviour of the systems. Such repetition in the fabric makes necessary only to consider a limited range of shear for each system of particles. This range 7(rot) corresponds to the deformation required to complete one full cycle of its movement, and can be either obtained visually from the AMS plots of a particular system (hence the advantage of using the iterative scheme in our program) or calculated more precisely from the equations of motion of a single particle or from the position of the mean cluster (Jezek et al. 1996; Canon-Tapia & ChavezAlvarez 2004). Finally, it must be noted that the effect of different mineral shapes and the exact number of particles forming the system are variables that increase the number of possible examples exponentially. We can select illustrative examples, however, by restraining the models to just four types of shapes, each dominated by mineral grains with elongation ratios of 0.1, 0.3, 0.65 and 0.95 respectively. These values of r were chosen because they represent the characteristic form of some abundant phenocrystals common in dykes, such as olivines, plagioclases, pyroxenes and magnetites (Blanchard et al. 1979; Nicolas 1992). Also, because the basic features of the system behaviours are found in systems formed for 200 and up to 10000 particles, little seems to be gained in a general model by using a model formulation that uses a large number of

particles without having more specific information about the mineral content of the dyke under examination. AMS variation along a dyke: profiles including fluid elements with various amounts of total shear One of the characteristics of profiles AA' of Figure 5 is that they include fluid elements that have experienced at least two different amounts of deformation (higher deformation near the walls and lower deformation near the centre of the dyke, Fig. 5b). It is possible to envisage two extreme situations to encompass the results of AMS for this type of profile. One of these situations is represented by systems that have a small range of deformations along the profile. Physically, this case can be associated with the AMS expected in profiles of type AA' located close to the magma source, where no much deformation has occurred yet (XI in Fig. 6a). Alternatively, the AA7 profile might be farther from the source in a position where the cyclic movement of the particles has resulted in returning to the initial 'undisturbed' state. In the case of profiles BB' this case would represent a very short profile that does not include large variations of deformation. The other extreme situation useful to encompass the results of AMS likely to be encountered in profiles including fluid elements with various amounts of total shear corresponds to cases in which the ranges of deformations represented in the diagram includes systems that are near 7rot, and systems that are as far from this stage of deformation as possible. These cases would be represented by most AA' profiles that are far from the magma source or by BB; profiles that are very long. The examples shown in Figures 6 to 9 were constructed by combining ten systems randomly selected from one column of Table 1. Due to the similar behaviours observed for both prolate and oblate particles, the examples are also representative of the AMS that would be found if systems from Table 2 had been considered. Skeletal and almost equant particles, like many oxides, will be represented in these examples by systems formed by r values close to 1. Although these minerals are commonly the phase that dominates the AMS signature of the rock, their formation may be associated with other minerals with lower values of r, such as clynopyroxenes and olivines. For this reason, the examples examined here include the whole range of r values of the tables.

THEORETICAL AMS OF A DYKE

241

7rot characteristic of the given r. On the contrary, on the profiles associated with X2 (Figs 6f, 6g, 6h & 6i), the deformation stages used in the calculations encompassed a wider range, although it was always limited by 7rot. Note that unlike the regions of confidence shown in Figure 3, the regions of confidence shown in Figures 6 to 8 have a well-defined physical meaning in all of these cases. When r = 0.1, (Figs 6b & 6f) it can be observed that regardless of the total range of shear considered in the profile there is a tendency of the mean A;max to define an imbrication angle, in agreement with the profile of velocities prevailing in the dyke. The region of confidence around the average £max is very small in both cases and it does not include the actual flow direction (or plane of the dyke). The correct imbrication of the average fcmax also is observed when r = 0.3 and r = 0.65 regardless of the total range of shear considered in the profile, although in these cases it is possible to find individual systems with an imbrication angle contrary to that expected from the profile of velocities. Such dispersion in the imbrication of kmax axes of individual systems naturally results in the increase of the size of confidence regions shown in Figures 6c, 6d, 6g and 6h, and on the consequent inclusion of the flow direction within these regions of confidence. When r = 0.95 the modelled AMS remains independent of the distance from the magma source, although in both examples shown there is a bad grouping of the mean susceptibility axes (Figs 6e & 6i). Consequently, the region of confidence around the average kmax is extremely large in these two latter cases, and the relation between AMS and magma flow direction is obscured for particles of this r. AMS variation along a dyke: profiles including fluid elements with the same shear

Fig. 6. Examples of the AMS obtained by collecting 10 samples along a line AA'. See text for a detailed explanation, and Fig. 2 for the relation between flow direction, plane of intrusion, etc.

In the profiles corresponding to XI (Figs 6b, 6c, 6d & 6e), the stages of deformation for each system included in the calculations were very similar to each other, and always less than the

Unlike the previous cases in which the deformation stage of each of the systems included in the model was selected independently from each other, to model a profile more akin to the shear expected along CC' (Fig. 5d) it is necessary to assign the same stage of deformation to all the systems used to calculate the resulting AMS. Figures 7, 8 and 9 show examples of the AMS obtained along profiles of this type, but with different shear values in each figure. Physically, the situation illustrated in Figure 7 (7 = 0) can be envisaged as representing the initial stage of the flow, as for example very near to the magma chamber. Alternatively, the examples of

242

E. CANON-TAPIA & M. J. CHAVEZ-ALVAREZ

Fig. 7. Examples of the AMS obtained by collecting 10 samples along a line CC in Fig. 5. All the systems included in each diagram have the same amount of deformation, as indicated. See Fig. 2 for the relation between flow direction, plane of intrusion, etc.

Fig. 8. Examples of the AMS obtained by collecting 10 samples along a line CC' in Fig. 5. The systems included in each diagram are the same that those used in the diagrams of Fig. 7, but the amount of deformation of each system is equal to 7 in this case. See Fig. 2 for the relation between flow direction, plane of intrusion, etc.

THEORETICAL AMS OF A DYKE

243

Fig. 9. Examples of the AMS obtained by collecting 10 samples along a line CC; in Fig. 5. The systems included in each diagram are the same that those used in the diagrams of Figs 7 and 8, but the amount of deformation of each system is equal to 11 in this case. See Fig. 2 for the relation between flow direction, plane of intrusion, etc.

this figure could represent profiles of samples collected near the centre of the dyke or at a level in the dyke in which the deformation has reached a value that is closer to the 7rot of the system (i.e. just after one full cycle of individual particle rotation had been completed). If the profile is collected at a different level in the dyke, or closer to the dyke walls where the expected deformation is slightly larger, the AMS results of the same systems would not be those of Figure 7 but rather those of Figure 8. Even slightly larger amounts of shear would result in a third set of AMS results, as shown in Figure 9. Although clearly arbitrary, the difference between the examples shown in Figures 7, 8, and 9 is representative of the influence that can be exerted by the exact amount of shear that is represented by the individual specimens of actual profiles; influence that should be taken into consideration when planning a strategy for sample collection as will be discussed below. Another interesting feature of these figures is that AMS distributions with big confidence regions can be found irrespective of the elongation ratio of the particles depending on the total amount of deformation that is being sampled. If the level of exposure of the dyke

allows access to a part of the dyke that is characterized by small amounts of deformation, or at a level where the cyclic movement of the particles had just restarted (even when the total deformation might be large) the AMS results can yield large dispersion of susceptibility axes even if the particles floating in the magma were all very elongated (flattened), as shown in Figure 7a. On the contrary, it is observed that particles with intermediate values of r can yield small groupings of principal susceptibility axes provided that the 'correct' amount of deformation is sampled (e.g. Fig. 9c). It is only the dykes formed by nearly spherical particles that would never yield good groupings of susceptibility axes (Figs 7d, 8d & 9d). Yet another feature of all of these examples is that despite the good grouping of kmax axes observed in dykes with unequant grains, only the dykes with the lower value of r (r = 0.1) will have the correct sense of imbrication of ^max f°r the range of deformations shown. When grouping of susceptibilities is tighter in the dyke formed by particles with r = 0.3, the AMS is such that all of the systems have an imbrication that is opposite to the orientation expected from the direction of flow (Fig. 8b). This result is a direct consequence of the cyclic

244

E. CANON-TAPIA & M. J. CHAVEZ-ALVAREZ

movement of the particles, in which the rotation of the system produces the wrong imbrication for half of the time that it takes to complete one full cycle. The incorrect imbrication of the mean kmax is also observed both in the dykes with r = 0.3 and 0.65 at a slightly large value of shear (Figs 9c & 9d), although in this case the regions of confidence are also slightly larger. Discussion Effect of mineral shape and ^rot on the acquisition of a magnetic fabric As already pointed out in the first half of the paper, one of the consequences of adopting the full model of movement of ellipsoidal particles (Jeffrey 1922) is that the principal susceptibilities do not have a unique relationship to the flow direction or with the plane of flow. By incorporating these results in a model of magma movement that makes explicit the role played by shear distribution within a dyke, the nonuniqueness of orientation of the principal susceptibilities along flow is confirmed, and a large variety of physically possible distributions of susceptibility axes can be produced under conditions of normal flow. These results suggest that many of the occurrences of 'anomalous' magnetic fabrics in dykes (e.g. Ernst 1990; Cadman et al 1992; Ernst & Baragar 1992; Puranen et al 1992; Walderhaug 1993; Baer 1995; Raposo & Ernesto 1995; Raposo 1997; Archanjo et al. 2000; Raposo & D'AgrellaFilho 2000; Borradaile & Gauthier 2001; Callot et al. 2001; Herrero-Bervera et al. 2001; Archanjo et al. 2002) may not be related to the occurrence of post-emplacement processes, but may be entirely due to some of the characteristics of the mineral content of the magma during emplacement. Furthermore, the non-uniqueness of orientation of magnetic fabric with flow direction characteristic of the model, and some of the characteristic fluctuations of the fabrics obtained provide a simple explanation for the observed fluctuations in magnetic fabric along a single dyke (Cadman et al. 1992; Baer 1995; Callot et al. 2001; Hrouda et al. 2002). The model developed in the second half of the paper not only confirms the predictions made by a careful analysis of the model of elliptical movement, but also provides important clues concerning the diversity of orientations of the principal susceptibilities predicted by the fundamental theory but that are not evident from the results of the first half of the paper. For instance, it is observed in the examples shown in Figures 7c,

8c and 9c that systems containing grains sufficiently elongated to generate a good grouping of their susceptibility axes at some stages of deformation will generate a large dispersion at other times as a consequence of the fundamental aspects of the movement of each particle within the system. This finding makes clearer the fact that both particle shape and the value of 7rot must be considered simultaneously in the interpretation of results. The value of shear deformation of Figure 8c is approximately equal to the 7rot characteristic of particles with r = 0.65, and for this reason the AMS is comparable to that of Figure 7c where the deformation was assigned a value of zero. For values of deformation intermediate between the multiples of 7rot, like that shown in Figure 9c, the grouping of axis can be reduced noticeably (although in this particular example the imbrication of the systems is in the wrong direction). Relevance of sampling scheme The influence of 7rot also can be observed to some extent in the examples of profiles AA7 shown in Figure 6. The fact that at least some of the systems included in each the diagrams of Figure 6 are not at a deformation stage equal to 7rot, however, results in less dramatic changes between diagrams involving particles of identical aspect ratios than those observed when comparing the AMS of profiles CCf. We can therefore take advantage of the influence exerted by 7rot in the acquisition of a magnetic fabric to devise a sampling scheme that can help us to maximize the information obtained from AMS measurements. In general, when making an AMS study attempting to infer a flow direction we have to collect the samples in the field while ignoring the magma flow direction within the intrusive. The three profiles AA', BB7 and CC7 discussed earlier, encompass the two most commonly followed schemes of sample collection, and their probable orientation relative to the magma flow direction. In practice, by collecting samples in a profile parallel to the exposed walls of the intrusive there will be an inherent uncertainty concerning the exact orientation of the profile relative to magma flow direction, but the two extremes are represented by profiles BB7 (parallel to magma flow direction) and CC7 (normal to magma flow direction). In the extreme situation represented by CC7, the obtained AMS results may or may not contain some useful information concerning the direction of magma flow, as shown by comparing Figures 7, 8 and 9, irrespective of the shape of the

THEORETICAL AMS OF A DYKE

particles that were floating in the magma during the emplacement of the intrusive. The rate of success in this case depends on the level of exposure that is sampled in one dyke. In the extreme cases represented by the BB7 profile, the rate of success depends on the length of the profile. Larger profiles, likely to include contrastingly different stages of shear are more likely to average out the effects of 7rot than shorter profiles, in which the effect of 7rot will be more evident. In the cases of profiles of the type AA7 the rate of success in inferring the flow direction from AMS measurements is somewhat independent of the factors affecting the profiles parallel to the walls of the intrusive. Such independency arises because these profiles will always be perpendicular to flow direction and are more likely to include various stages of deformation than the BB7 or CC7 profiles independently of the orientation of the outcrop relative to the real flow direction. Consequently, the profiles of the type AA7 seem to be more advantageous than the profiles collected parallel to the walls of the dyke. The advantage of profiles A A', however, can be diluted by the cyclic movement of the particles predicted by the theory. Such cyclic behaviour can result in an imbrication in the 'wrong' direction depending on the stage of deformation that is sampled (unknown in most cases) that also can affect the profiles A A7. Fortunately, the model results suggest one way to test whether the imbrication obtained at one place of the dyke is pointing downflow or upflow. As observed in some of the examples shown in Figure 5 (5g, 5h) it is possible to obtain a mixture of imbrication in samples collected from the same side of the dyke along the profile AA7. This type of mixed distribution is not uncommon in real dykes as observed for example in the study of the Ko'olau dykes made by Knight & Walker (1988), so far considered as the classic example of opposite imbrication as a reliable flow direction indicator in dykes. The systematic changes defined by any of the principal susceptibility axes in a data set of AMS results obtained from two or more profiles of the AA7 type collected from the same dyke might, however, provide enough clues allowing us to infer the correct flow direction. In the absence of such a complete data set examination of other microscopic textures to identify the local sense of shear would be required. In any case, the obtained results highlight the fact that the occurrence of isolated specimens with the wrong imbrication may arise as a natural consequence of the cyclic movement of the particles floating in the magma, and it does not imply the occurrence of any post-emplacement alteration, or of

245

a change in flow direction in time as has been suggested occasionally (Philpotts & Asher 1994). Mineral content of magma during emplacement: towards a more realistic AMS model One of the main limitations of the general model of the AMS of a dyke developed here is that it assumes that all the minerals defining the AMS have the same elongation ratio and contribute equally to the bulk susceptibility of the rock. In more realistic cases this assumption may not be fully satisfied as the mineral content of a flowing magma depends on many variables, some of which can change even during the emplacement of a single dyke. For example, changes in the amount of volatiles dissolved in the magma can promote crystallization of a different mineral species than that already floating in the magma, so that more than one mineral species is likely to be present during magma emplacement. Assuming that the mineral content of the magma during flow remains small enough to prevent mechanical interactions between mineral grains, our model can be used to investigate the resulting magnetic fabric of a mixed population of particles. As each mineral species can be characterized by its own aspect ratio, Jeffery's equations predict that each of these mineral species will rotate at a different velocity. Consequently, the final mineral fabric will be composed of the superposition of various mineral subfabrics, as shown by Fernandez et al. (1983). In the context of magma emplacement within a dyke, the more elongated (flattened) minerals commonly are diamagnetic (e.g. plagioclases) or paramagnetic (e.g. olivine and pyroxenes), whereas the ferromagnetic minerals are commonly nearly equant (Blanchard et al 1979; Nicolas 1992). This implies that the purely ferromagnetic fabric will be unlikely to have a unique, well-defined relation to flow direction in a dyke during its emplacement even if these minerals were present during magma flow. If there is a reason to suspect, however, that the ferromagnetic grains are of a different shape, it could be possible to have a closer relationship between the magnetic fabric and the flow direction. For instance, if the magnetic grains are assumed to be of a rather oblate shape, the orientation of these particles predicted by our simple model is such that the £min axes will remain close to the normal of the dyke (with the ^max'^int plane forming an imbrication to the flow direction), whereas other particles of a rather elongated shape will be oriented subparallel to the flow direction. This situation

246

E. CANON-TAPIA & M. J. CHAVEZ-ALVAREZ

actually corresponds to the relation of the magnetic and mineral fabric described by Geoffroy et al. (2002) and Callot & Guichet (2003) who postulated the presence of oblate ferromagnetic minerals to explain the discrepancy observed between the orientation of the elongated minerals (plagioclases) and the AMS results. Another example of dykes in which very elongated grains of magnetite have been reported is the case of the Ceara-Mirim dyke complex (Archanjo et al. 2002). In some of these dykes the elongated grains were found normal to the dyke walls, whereas in other cases the £max axes were parallel to the plane of the dyke. An interpretation of these results, based on our model, is that both of these particular fabrics are flow related but reveal different stages of magma deformation. A similar situation has been found when examining the orientation distribution of microlites in the rhyolitic lavas of Obsidian Dome (Canon-Tapia & Castro 2004). In this case, the mineral fabric defined by the orientation of diamagnetic microlites and with the orientation of trains of equant magnetite grains is generally in good agreement with the AMS of the same sample. In some cases where a bimodal distribution of microlites was observed, the AMS could be found to yield an average of the two observed fabrics, but the ferromagnetic fraction seemed to be parallel to one of the orientation maxima. Based on all these examples, the role played by the presence of the non-magnetic minerals must be re-evaluated despite their lower value of bulk susceptibility. Stacey (1960) and Margraves et al. (1991) suggested that the more unequant minerals, although magnetically weak, could control the AMS of the dyke by controlling the direction of growth of the late forming ferromagnetic minerals. The role played by the non-magnetic minerals in the context of AMS measurements can be extended beyond this template model if combined with some observations concerning the specific mineralogy of a group of dykes. For example, a generally valid consequence of having a different chemical composition is that the mineral content will also be different. Based on the results of our model, a direct consequence of this difference in the mineral content of two magmas with different chemical composition is that their AMS might bear a different signature for each case. Although the published mineral content of dykes used to make AMS studies precludes a more stringent test of this assertion at this time, there are some key observations that can be made now. For example, it is notable that the proportion of dykes with anomalous magnetic fabrics found by Knight & Walker (1988) is much smaller

than the proportion of anomalous fabrics found in other swarms, such as the ophiolite of Oman (Rochette et al. 1991). Leaving aside the differences in the tectonic scenarios of these two places, a basic difference between the Hawaiian and the Omanian dykes is the different proportion of phenocrysts arising from the reported difference in composition. Furthermore, the postulated difference in mineral content in these two examples is also consistent with the observed differences in the magnetic fabric of dykes of different chemical affinity in the same tectonic scenario, as suggested by the same results of the dykes from Oman where the inverse fabric was more commonly observed in dykes compositionally akin to MORB whereas it was relatively uncommon in the calk-alkaline dykes (Rochette et al. 1991). Similar conclusions could be reached by examining the results obtained in other dykes around the world, where a compositional control of the magnetic fabric has been suggested (e.g. Staudigel et al. 1999; Rapalini & Lopez de Luchi 2000). Therefore, the results of our simplified model suggest that the mineral content present during magma emplacement may play a more active role in determining the quality of AMS than simply acting as a template controlling the formation of the ferromagnetic fraction. This interpretation can not be fully explored at this time, but might prove to be a reasonable alternative that can help to explain the variety of magnetic fabrics found in dykes by processes other than a selective (and not clearly understood) set of post-emplacement alterations. The success of our model in explaining a large variety of magnetic fabrics as the result of the expected variations in orientation of an ellipsoidal particle during movement in a fluid does not imply that we deny the possibility of finding situations in which mineral changes related to alteration of the magnetic fabric of dykes are the most likely explanation of an abnormal fabric. However, we think that probably too much attention has been given to this type of processes, whereas no attention has been given to a simpler alternative. For this reason, it is stressed that by paying attention to the systematic changes predicted to take place within a dyke as the result of the movement of ellipsoidal particles, it might be possible to discriminate between situations in which alteration has become dominant from those in which the flow-related fabric is still preserved. Conclusions By focusing our attention on a complete model of ellipsoidal movement in a moving viscous

THEORETICAL AMS OF A DYKE fluid it has been shown that many of the features commonly observed in the magnetic fabrics of dykes around the world can be reproduced. This suggests that the role played by post-emplacement processes in the production of abnormal fabrics in dykes may not be as important as has been assumed. On the contrary, at least some of these cases may reflect completely normal circumstances of flow emplacement, differing only from the 'normal' fabrics in the shape of the minerals present during the emplacement of the dyke. Also, the amount of deformation represented by the sampled section might be important in determining the type of magnetic fabric that is found at a given location. Although the theoretical considerations presented in this work explain almost any magnetic fabric in terms of the primary flow of magma, apparently reducing the potential of this method to infer flow direction in dykes reliably, the same theory provides guidelines that can be used to test the internal consistency of the AMS measurements. In this sense, the scheme followed when collecting samples for AMS measurements in dykes has been shown to be extremely important. Whenever possible, it is advised that more than one profile is collected within the same dyke to ensure that enough information is available to check the results in terms of the patterns predicted by theory. Undoubtedly, this approach results in a larger number of measurements and a more detailed analysis of the fabric present in each intrusive than is currently made in most AMS studies of dykes. Nevertheless, the time and effort required to do the additional work is much less than the time that would be required to get equivalent amounts of information with other, more traditional, petrofabric techniques. Besides, many of the problems concerning the acquisition of a mineral fabric raised by the complete model of ellipsoidal movement examined in this paper are also found when using those techniques. Consequently, it is concluded that despite a more complex relation between magma kinematics and magnetic fabric than assumed 30 years ago, the measurement of the AMS remains a very powerful method to unravel the processes of emplacement of many igneous rocks. This paper would have not been conceived without the questions asked by J. Glen during the 2002 Fall Meeting of the AGU, where the results of the first half were originally presented. We also are indebted to L. Delgado and the Direction General de Posgrado de CICESE for providing the financial support for MJ Chavez-Alvarez that allowed her to continue working on this manuscript beyond her regular scholarship. E Canon-Tapia also acknowledges the support of CONACYT through grant 39535. Comments made

247

on the previous version of this paper by J. Jezek, J.P. Callot and G. Plenier helped to delineate its scope more clearly, and are greatly appreciated. References ARBARET, L., DIOT, H. & BOUCHEZ, J. L. 1996. Shape fabrics of particles in low concentration suspensions: 2D analogue experiments and application to tiling in magma. Journal of Structural Geology, 18, 941-950. ARBARET, L., DIOT, H., BOUCHEZ, J. L., LESPINASSE, P. & DE SAINT-BLANQUAT, M. 1997. Analogue 3D simple-shear experiments of magmatic biotite subfabrics. In: BOUCHEZ, J. L., HUTTON, D. H. W. & STEPHENS, W. E. (eds) Granite: From Segregation of Melt to Emplacement Fabrics. Kluwer Academic Publishers, Dordrecht, 129-143. ARCHANJO, C. J., TRINDADE, R. I., WILSON, J., MACEDO, P. & ARAUJO, M. G. 2000. Magnetic fabric of a basaltic dyke swarm associated with Mesozoic rifting in northeastern Brazil. Journal of South American Earth Sciences, 13, 179-189. ARCHANJO, C. J., ARAUJO, M. G., S. & LAUNEAU, P. 2002. Fabric of the Rio Ceara-Mirim mafic dike swarm (northeastern Brazil) determined by anisotropy of magnetic susceptibility and image analysis. Journal of Geophysical Research, 107(B3), 10. BAER, G. 1995. Fracture propagation and magma flow in segmented dykes: Field evidence and fabric analyses, Makhtesh Ramon, Israel. In: BAER, G. & HEIMANN, A. (eds) Physics and Chemistry of Dykes. A.A. Balkema, Rotterdam, 125-140. BLANCHARD, J. P., BOYER, P. & GAGNY, C. 1979. Un nouveau critere des sens de mise en place dans une caisse filonienne: le 'pincement' des mineraux aux epontes. Tectonophysics, 53, 1-25. BORRADAILE, G. J. & GAUTHIER, D. 2001. AMS-

detection of inverse fabrics without AARM, in ophiolite dikes. Geophysical Research Letters, 28, 3517-3520. CADMAN, A. C., PARK, R. G., TARNEY, J. & HALLS, H. C. 1992. Significance of anisotropy of magnetic susceptibility fabrics in Proterozoic mafic dykes, Hopedale Block, Labrador. Tectonophysics, 207, 303-314. CALLOT, J. P. & GUCHET, X. 2003. Rock texture and magnetic lineation in dykes: A simple analytical model. Tectonophysics, 366, 207-222. CALLOT, J. P., GEOFFROY, L., AUBOURG, C., Pozzi, J. P. & MEGE, D. 2001. Magma flow directions of shallow dykes from the East Greenland volcanic margin inferred from magnetic fabric studies. Tectonophysics, 335, 313-329. CANON-TAPIA, E. 2001. Factors affecting the relative importance of shape and distribution anisotropy in rocks: theory and experiments. Tectonophysics, 340, 117-131. CANON-TAPIA, E. & CASTRO, J. 2004. AMS measurements on obsidian from the Inyo Domes, CA: a comparison of magnetic and mineral preferred orientation fabrics. Journal of Volcanology and Geothermal Research, 134, 169-182.

248

E. CANON-TAPIA & M. J. CHAVEZ-ALVAREZ

CANON-TAPIA, E. & CHAVEZ-ALVAREZ, M. J. 2004. Rotation of uniaxial ellipsoidal particles during simple shear revisited: the influence of elongation ratio, initial distribution of a multiparticle system and amount of shear in the acquisition of a stable orientation. Journal of Structural Geology, 26,2073-2087. CHAVEZ-ALVAREZ, M. J. & CANON-TAPIA, E. 2002. Mechanical model of grain alignment revisited: implications for anisotropy of magnetic susceptibility measurements. In: AGU Fall meeting 83. EOS, San Francisco, F418. DE ROSA, R., MAZZUOLI, R. & VENTURA, G. 1996. Relationships between deformation and mixing process in lava flows: a case study from Salina (Aeolian Islands, Tyrrenian Sea). Bulletin of Volcanology, 58, 286-297. DRAGONI, M., LANZA, R. & TALLARICO, A. 1997. Magnetic anisotropy produced by magma flow: theoretical model and experimental data from Ferrar dolerite sills (Antartica). Geophysical Journal International, 128, 230-240. ERNST, R. E. 1990. Magma flow directions in two mafic Proterozoic dyke swarms of the Canadian shield: as estimated using anisotropy of magnetic susceptibility data. In: PARKER, A. J., RICKWOOD, P. C. & TUCKER, D. H. (eds) Mafic dykes and emplacement mechanisms. A.A. Balkema, Rotterdam, 231-235. ERNST, R. E. & BARAGAR, W. R. A. 1992. Evidence from magnetic fabric for the flow pattern of magma in the Mackenzie giant radiating dyke swarm. Nature, 356, 511-513. FERNANDEZ, A. 1987. Preferred orientation developed by rigid markers in two-dimensional simple shear strain: a theoretical and experimental study. Tectonophysics, 136, 151-158. FERNANDEZ, A., FEYBESSE, J. L. & MEZURE, J. F. 1983. Theoretical and experimental study of fabrics developed by different shaped markers in twodimensional simple shear. Bulletin de la Societe geologique de France, 25, 319-326. FIALKO, Y. A. & RUBIN, A. M. 1999. Thermal and mechanical aspects of magma emplacement in giant dike swarms. Journal of Geophysical Research, 104, 23033-23049. FISHER, N. I., LEWIS, T. & EMBLETON, B. J. J. 1993. Statistical Analysis of Spherical Data. Cambridge University Press, Cambridge. FREEMAN, B. 1985. The motion of rigid ellipsoidal particles in slow flows. Tectonophysics, 113, 163-183. GAY, N. C. 1966. Orientation of mineral lineation along the flow direction in rocks: a discussion. Tectonophysics, 3, 559-564. GEOFFROY, L., C ALLOT, J. P., AUBOURG, C. & MOREIRA, M. 2002. Magnetic and plagioclase linear fabric discrepancy in dykes: a new way to define the flow vector using magnetic foliation. Terra Nova, 14, 183-190. HARGRAVES, R. B., JOHNSON, D. & CHAN, C. Y. 1991. Distribution anisotropy: The cause of AMS in igneous rocks? Geophysical Research Letters, 18, 2193-2196. HERRERO-BERVERA, E., WALKER, G. P. L., CANON-TAPIA, E. & GARCIA, M. O. 2001. Magnetic fabric and

inferred flow direction of dikes, conesheets and sill swarms, Isle of Skye, Scotland. Journal of Volcanology and Geothermal Research, 106, 195-210. HROUDA, F. 1993. Theoretical models of magnetic anisotropy to strain relationship revisited. Physics of the Earth and Planetary Interiors, 77, 237-249. HROUDA, F., CHLUPACOVA, M. & NOVAK, J. K. 2002. Variations in magnetic anisotropy and opaque mineralogy along a kilometer deep profile within a vertical dyke of the syenogranite porphyry at Cinovec (Czech Republic). Journal of Volcanology and Geothermal Research, 113, 37-47. IEZZI, G. & Ventura, G. 2002. Crystal fabric evolution in lava flows: results from numerical simulations. Earth and Planetary Science Letters, 200, 33^6. ILDEFONSE, B., ARBARET, L. & DIOT, H. 1997. Rigid particles in simple shear flow: is their preferred orientation periodic or steady-state? In: BOUCHEZ, J. L., HUTTON, D. H. W. & STEPHENS, W. E. (eds) Granite: from segregation of melt to emplacement fabrics. Petrology and structural geology 8. Kluwer academic publishers, Dordrecht, The Netherlands, 177-185. JEFFERY, G. B. 1922. The motion of ellipsoidal particles immersed in a viscous fluid. Proceedings of the Royal Society of London, 102, 161-179. JELINEK, V. 1981. Characterization of the magnetic fabric of rocks. Tectonophysics, 79, T63-T67. JEZEK, J. & HROUDA, F. 2002. Software for modelling the magnetic anisotropy of strained rocks. Computers and Geosciences, 28, 1061-1068. JEZEK, J., MELKA, R., SCHULMANN, K. & VENERA, Z. 1994. The behaviour of rigid triaxial ellipsoidal particles in viscousflows- modelling of fabric evolution in a multiparticle system. Tectonophysics, 229, 165-180. JEZEK, J., SCHULMANN, K. & SEGETH, K. 1996. Fabric evolution of rigid inclusions during mixed coaxial and simple shear flows. Tectonophysics, 257, 203221. KHAN, M. A. 1962. The anisotropy of magnetic susceptibility of some igneous and metamorphic rocks. Journal of Geophysical Research, 67, 28732885. KNIGHT, M. D. & WALKER, G. P. L. 1988. Magma flow directions in dikes of the Koolau Complex, Oahu, determined from magnetic fabric studies. Journal of Geophysical Research, 93, 4301-4319. LISTER, J. R. & KERR, R. C. 1990. Fluid-mechanical models of dyke propagation and magma transport. In: PARKER, A. J., RICKWOOD, P. C. & TUCKER, D. H. (eds) Mafic dykes and emplacement mechanisms. A.A. Balkema, Rotterdam, 69-80. MANGA, M. 1998. Orientation distribution of microlites in obsidian. Journal of Volcanology and Geothermal Research, 86, 107-115. NICOLAS, A. 1992. Kinematics in magmatic rocks with special reference to gabbros. Journal of Petrology, 33, 891-915. NOMADE, S., THEVENIAUT, H., CHEN, Y., POUCLET, A. & RIGOLLET, C. 2000. Paleomagnetic study of French Guyana early Jurassic dolerites: hypothesis of a multistage magmatic event. Earth and Planetary Science Letters, 184, 155-168.

THEORETICAL AMS OF A DYKE OSBORN, J. A. 1945. Demagnetizing factors of the general ellipsoid. Physical Review 67, 351-357. PARK, J. K., TANCZYK, E. I. & DESBARATS, A. 1988. Magnetic fabric and its significance in the 1400 Ma Mealy diabase dykes of Labrador, Canada. Journal of Geophysical Research, 93, 13689-13704. PHILPOTTS, A. R. & ASHER, P. M. 1994. Magmatic flowdirection indicators in a giant diabase feeder dike, Connecticut. Geology, 22, 363-366. POTTER, D. K. & STEPHENSON, A. 1988. Single-domain particles in rocks and magnetic fabric analysis. Geophysical Research Letters, 15, 1097-1100. PURANEN, R., PEKKARINEN, L. J. & PESONEN, L. J. 1992. Interpretation of magnetic fabrics in the early Proterozoic diabase dikes of Keuruu, central Finland. Physics of the Earth and Planetary Interiors, 72, 68-82. RAPALINI, A. E. & LOPEZ DE LUCHI, M. 2000. Paleomagnetism and magnetic fabric of middle Jurassic dykes from western Patagonia, Argentina. Physics of the Earth and Planetary Interiors, 120, 11-27. RAPOSO, M. I. B. 1997. Magnetic fabric and its significance in the Florianopolis dyke swarm, southern Brazil. Geophysical Journal International, 131, 159-170. RAPOSO, M. I. B. & D'AGRELLA-FiLHO, M. S. 2000. Magnetic fabrics of dike swarms from SE Bahia State, Brazil: their significance and implications for Mesoproterozoic basic magmatism in the Sao Francisco Craton. Precambrian Research, 99, 309-325. RAPOSO, M. I. B. & ERNESTO, M. 1995. Anisotropy of magnetic susceptibility in the Ponta Grossa dyke swarm (Brazil) and its relationship with magma flow direction. Physics of the Earth and Planetary Interiors, 87, 183-196.

249

ROCHETTE, P., JENATTON, L., DUPUY, C., BOUDIER, F. & REUBER, I. 1991. Diabase dikes emplacement in the Oman Ophiolite: a magnetic fabric study with reference to geochemistry. In: PETERS, T., NICOLAS, A. & COLEMAN, R. G. (eds) Ophiolite genesis and evolution of the oceanic lithosphere. Ministry of petroleum and minerals, Sultanate of Oman, 55-82. SMITH, J. V. 1998. Interpretation of domainal groundmass textures in basalt lavas of the southern Lamington Volcanics, eastern Australia. Journal of Geophysical Research, 103, 27383-27391. STAGEY, F. D. 1960. Magnetic anisotropy of igneous rocks. Journal of Geophysical Research, 65, 24292442. STAUDIGEL, H., TAUXE, L., ET AL. 1999. Geochemistry and intrusive directions in sheeted dikes in the Troodos ophiolite: Implications for mid-ocean ridge spreading centers. Geochemistry Geophysics Geosys terns, 1, 1999. STONER, E. C. 1945. The demagnetizing factors for ellipsoids. Philosophical Magazine, 36, 803-821. TEASDALE, R., BROOKER, R. A. & SPARKS, S. 2003. Devolatilization induced crystallization experiments of hydrous basaltic magma, implications for lava flow emplacement. In: Annual Fall meeting (edited by AGU) 84. Eos, F 1615. WALDERHAUG, H. 1993. Rock magnetic and magnetic fabric variations across three thin alkaline dykes from Sunnhordland, Western Norway; influence of initial mineralogy and secondary chemical alterations. Geophysical Journal International, 115, 97-108. WOODCOCK, N. H. 1977. Specification of fabric shapes using an eigenvalue method. Bulletin of the Geological Society of America, 88, 1231-1236.

This page intentionally left blank

Magmatic flow paths and palaeomagnetism of the Miocene Stoddard Mountain laccolith, Iron Axis region, Southwestern Utah, USA M. S. PETRONIS1, D. B. HACKER2, D. K. HOLM2, J. W. GEISSMAN1 & S. S. HARLAN3 1

Department of Earth and Planetary Sciences, University of New Mexico, Albuquerque, NM, 87131, USA (e-mail: [email protected]) 2 Department of Geology, Kent State University, Kent, Ohio 44242, USA ^Environmental Sciences and Policy, George Mason University, Fairfax, VA 22030, USA Abstract: The Stoddard Mountain laccolith is part of a complex of Early Miocene laccoliths intruded along the western edge of the Colorado Plateau in the Iron Axis region of Southwestern Utah. Most Colorado Plateau laccoliths (e.g. Henry and La Sal Mountains) are considered to be fed by a central axial feeder system. However, detailed mapping in the Iron Axis region suggests that the Stoddard Mountain laccolith was fed laterally from the west. Structural and field data suggest the quartz monzonitic magma initially migrated laterally eastward at ~1 km depth as a sill before spreading laterally north-south where it inflated to ~ 1-1.5 km thickness. To test the model of a lateral feeder system, data were collected from 32 palaeomagnetic sites and 76 AMS stations (763 accepted specimens) sampled over the ~54km 2 exposed part of the N-S oval-shaped laccolith. The in situ AMS fabrics, inferred to correlate with magmatic fabrics, typically show NE trending lineations in the north and S—SE trending lineations in the south part of the intrusion. The palaeomagnetic data are interpreted to indicate a very minimal amount of post-emplacement deformation of the intrusion. The overall lack of westerly-directed and steep magnetic lineations argues against emplacement via a central axial feeder system.

Laccoliths have been well documented and studied for over a century (Gilbert 1877) and their mode of emplacement and magma dynamics are generally understood (Corry 1988; Pollard & Johnson 1973; Johnson & Pollard 1973). Additional field based studies, however, are needed to test competing models for specific laccoliths because a single emplacement model for all laccoliths does not seem warranted given the complexities associated with their emplacement dynamics. Conventionally, structural studies of laccoliths and of plutons in general involve field and laboratory measurement of their mineral phase arrangement. The shape, orientation and spatial distribution of the mineral phases define a fabric, and the kinematic condition of its formation can be inferred from their arrangement. During magmatic flow, the early-formed crystals rotate and undergo relative translations, and they may provide indirect evidence of magmatic flow path. In this study, we use the orientation of the magnetic fabric to infer magma flow patterns during emplacement of a shallow level laccolith, Foliation and lineation orientations determined in this manner can then be used to understand better the emplacement path of intrusions with different geometries (Hutton 1988; Pitcher

1993; Bouchez 1997). Magmatic fabrics are thought to form after ascent and before final crystallization of the magma (Cruden 1990; John & Blundy 1993; Clemens et al. 1997). As early ascent structures are likely obliterated during magmatic flow, only the final stages of magma flow can be inferred from petrofabric studies (Hutton 1992; Bouchez 1997; Vigneresse & Clemens 2000). Although important information can be obtained from structural studies, they have typically been limited in their effectiveness in that magmatic fabrics tend to be only weakly developed and difficult to identify and measure in the field. Recently, studies aimed at describing and interpreting the internal fabrics of plutons have been greatly advanced by the application of the low-field anisotropy of magnetic susceptibility (AMS) technique (Balsey & Buddington 1960; King 1966; Heller 1973; Borradaile 1980; Hrouda 1982; Guillet et al. 1983; Bouchez et al. 1990; Archanjo et al. 1992; Benn et al. 1993; Bouillin et al. 1993). The AMS method can detect anisotropies of only a few per cent and, therefore, the technique has proved to be a powerful method to complement field studies of plutons (Bouchez 1997). In this contribution, we report AMS and palaeomagnetic data from the Stoddard

From\ MARTIN-HERNANDEZ, F., LUNEBURG, C. M., AUBOURG, C. & JACKSON, M. (eds) 2004. Magnetic Fabric: Methods and Applications. Geological Society, London, Special Publications, 238, 251-283. 0305-8719/04/ $15.00 © The Geological Society of London 2004.

252

M. S. PETRONIS ET AL.

Fig. 1. (a) Selected igneous features of the Colorado Plateau. Black circles represent Colorado Plateau laccolithic centres; MVG - Marysvale volcanic field, and SJVF - San Juan volcanic field. Modified from Mutschler et al. (1998). (b) Index map of southwest Utah showing intrusions of the Iron Axis: B - Big Mountain; BV - Bull Valley; D - The Dairy; G - Granite Mountain; H - Hardscrabble Hollow; I - Iron Peak; IM - Iron'Mountain; LP - Lookout Point; MM - Mineral Mountain; PP - Pinto Peak; PV - Pine Valley; SM - Stoddard Mountain: T - Three Peaks. Towns: CC - Cedar City; SG - St. George.

Mountain laccolith, Iron Axis magmatic province, Southwestern Utah. The Iron Axis laccoliths are part of a Late Oligocene-Early Miocene hypabyssal laccolith swarm that lies within and borders the Colorado Plateau (Fig. la). In the Iron Axis region, mineralization associated with the emplacement of over a dozen laccoliths has resulted in significant iron-ore production (Mackin 1968). The Stoddard Mountain intrusion, one of the largest and best exposed of these laccoliths, is an example of a granitoid in which the AMS fabrics are well defined. Its relation to the surrounding country rock and to other smaller nearby laccoliths is well understood through detailed mapping (Hacker 1998; Rowley et al 2004 in press). The magnetic fabrics in the Stoddard Mountain

laccolith were systematically mapped on a relatively tight sampling grid (76 sites distributed over the 54km2 map area) using low-field anisotropy of magnetic susceptibility (AMS). The magnetic susceptibility, degree of magnetic anisotropy, and the shapes of the AMS ellipsoids are homogeneous at the outcrop scale (10m2) and, on the laccolith scale, they define a pattern that is interpreted to reflect magma flow during emplacement. The remanence directions recorded in the main body of the laccolith are well grouped (D = 351.1°, / = 57.8°, c*95 = 3.7°, k = 82.2) and only slightly discordant to a Late Miocene expected field direction (358°, 58°). Sites in chill zone rocks collected throughout the perimeter of the laccolith, however, yield a magnetization of south declination and very shallow inclination.

MAGMATIC FLOW PATHS AND PALAEOMAGNETISM

AMS studies on other Colorado Plateau laccoliths are beginning to provide important information about magma flow patterns and magma dynamics (Horsman & Tikoff 2002). Whereas many of the Colorado Plateau laccoliths appear to have been fed from below by a central axial feeder system (Jackson & Pollard 1988; Ross 1998), our results from the Stoddard Mountain laccolith are more consistent with magma emplacement from a shallow level sill that spread laterally from the west to east. These results, used together with recent mapping, allow us to document the mode of laccolith emplacement and to assess the influence of surrounding geologic features on its development. This study may have important implications, specifically, for testing hypotheses regarding the formation of laccolith related ore-deposits in the Iron Axis region (Mackin & Ingerson 1960), as well as for better understanding the kinematics of shallow level laccolith emplacement in general. Geological setting The Iron Axis laccolith swarm consists of a series of Early Miocene calc-alkaline hypabyssal laccoliths and associated volcanic rocks located just west of the present Colorado Plateau in Southwestern Utah (Fig. Ib). The Iron Axis igneous rocks are part of the general Mid-Cenozoic calc-alkaline igneous sequence that spans much of the western United States. This magmatism is generally considered coincident with oblique convergence during subduction of oceanic lithosphere beneath western North America that produced large fluxes of mantle-derived mafic magma injected into the overlying continental lithosphere (Johnson 1991; Nelson & Davidson 1998; Rowley et al 1998). With the onset of calc-alkalic volcanism, prior to Iron Axis magmatism that produced laccoliths, a sequence of regional Oligocene and Miocene calc-alkaline andesite to rhyolite ash-flow tuffs of the Wah Wah Springs Formation (30 Ma), Isom Formation (27 Ma) and Quichapa Group (24 to 22.5 Ma) were spread over the area from sources to NW and west (mostly from the Indian Peak and Caliente caldera complexes; see Best et al 1989 and Rowley el al. 1995). This sequence of pre-Iron Axis volcanic rocks overlies fluvial and lacustrine sedimentary rocks of the Upper Palaeocene-Oligocene Claron Formation, that in turn, unconformably overlie Cretaceous and Jurassic sedimentary rocks deformed during the Sevier orogeny (Late Cretaceous to Early Tertiary) that produced east and SE verging thrusts.

253

During Iron Axis magmatic activity, ascending magma from an inferred batholith complex intruded along one or more NE-striking Sevierage thrust faults before being emplaced as bulbous laccoliths within Mesozoic and Tertiary sedimentary strata (Mackin 1960; Blank & Mackin 1967; Blank el al 1992; Rowley et al 1998; Rowley 1998). More than a dozen exposed intrusions have been mapped within the magmatic province and others are inferred from structures in the cover-rock and aeromagnetic signatures, or are known from drilling and seismic data. Thus, the intrusions are distributed along a NE-trending zone that follows the trend of the Sevier orogenic front (Fig. Ib). These intrusions were forcibly emplaced into sedimentary rocks at depths ranging mostly between 2.5 and 0.25km from the surface and deformed their roofs by upward folding and faulting (Corry 1988; Johnson & Pollard 1973). The intrusions in the eastern Bull Valley Mountains (Bull Valley, Hardscrabble Hollow and Big Mountain) and the Iron Springs mining district (Iron Mountain, Granite Mountain and Three Peaks) intruded into limestone strata of the Jurassic Carmel Formation and produced major replacement magnetite and hematite ore bodies. Only three plutons in the central part of the Iron Axis Iron Springs mining district were extensively mined and they constitute the largest iron ore production in the western United States (Mackin 1960; Blank & Mackin 1967; Barker 1995). Iron-rich solutions were derived from deuteric breakdown of ferromagnesian minerals in the outer selvage-joint phase of the intrusions (Mackin & Ingerson 1960; Rowley & Barker 1978; Barker 1995). Plutons of the central part of the Iron Axis are quartz monzonite to granodiorite porphyries characterized by phenocrysts of plagioclase (andesine-labradorite), biotite, hornblende and/ or pyroxene (diopsidic augite), and magnetite in a groundmass (1/3 to 1/2 total volume) of very fine-grained quartz and potassium feldspar. The only exceptions to this composition range are the Mineral Mountain intrusion at the SW end of the trend, which is a granite porphyry (Morris 1980; Adair 1986), and the Iron Peak intrusion at the NE end, which is a gabbrodiorite porphyry (Spurney 1984). The exposed plutons yield K-Ar dates and 40Ar/39Ar age spectra values of 20-22 Ma (Armstrong 1970; Hacker et al 1996; McKee et al 1997; Rowley et al, in press) indicating a two million year period of Iron Axis magmatic activity. Both structural relief and current topographic relief have been produced by the emplacement of the Iron Axis laccoliths. The nature of the

254

M. S. PETRONIS ET AL.

Fig. 2. (a) Simplified geological map.

igneous intrusions in the Iron Axis province has been historically controversial, in part, because the current level of erosion makes it difficult to distinguish the intrusions as stocks or laccoliths. Only the Pine Valley and Iron Peak intrusions have been eroded to expose their planar, subhorizontal floors. However, geophysical and drill hole data, together with detailed field mapping, confirmed that the intrusions have nearly planar floors and irregular but concordant

roofs (Van Kooten 1988; Hacker 1998; Hacker et al, 2002). Cross-sections show that most of the laccoliths have a typical plano-convex shape (Corry 1988), but are not circular in map view (Hacker 1998). During the emplacement of the laccoliths, several (e.g. Pine Valley, Bull ValleyBig Mountain, Pinto Peak, Stoddard Mountain and Iron Mountain) were partially unroofed by catastrophic gravity sliding and vented to the surface (except Iron Mountain) forming related

Fig. 2. (b) cross-section (A-A' and B-B') of the Stoddard Mountain laccolith showing its estimated extent, vertical thickness, and intrusive and extrusive features in cross-section delineates the sampling region within the laccolith.

256

M. S. PETRONIS ET AL.

ash-flow tuffs and minor lava flows that buried the gravity slide masses (Mackin 1960; Blank et al 1992; Hacker 1995, 1998; Hacker et al 2002) (Fig. 2). Field relations The Stoddard Mountain Laccolith, which underlies the Stoddard Mountain area in the northern Pine Valley Mountains, is one of the quartz monzonite porphyry laccoliths of the central part of the Iron Axis (Fig. Ib). The exposed part of the laccolith is roughly elliptical in shape, about 9 km long N-S and 6 km wide EW, and about 54km 2 in area. The unexposed part of the laccolith extends westward at least another 5km beneath the Richie Flat anticline (Fig. 2a) as delineated by drill cores (Cook 1957). It intruded clastic sedimentary rocks of the Cretaceous Iron Springs Formation, which, along with Tertiary volcanic rocks, dip steeply away from the exposed part of the intrusion and more gently over the Richie Flat anticline. At the NE margin of the laccolith the country rock is steeply overturned. Although erosion has partially removed some of the country rocks to expose the upper crest and flank of the intrusion (the laccolith is well exposed over a topographic relief of about 700m), the floor is not exposed. Based on geological mapping and cross-section construction, the intrusion has a remaining thickness of about 1.5km where exposed and 1.0km beneath Richie Flat (Fig. 2b). The laccolith consists of homogeneous medium-grained quartz monzonite porphyry forming two gradational phases. Marking the

outermost 20 to 60 m of the intrusion is a chilled margin zone consisting of fresh, resistant, lightgrey, pink and light yellow porphyry. This zone consists of about 45% phenocrysts (approximately 29% plagioclase, 2% clinopyroxene, 1% biotite, trace amphibole, 1% Fe-Ti oxides, and 14% aggregates of plagioclase, sanidine, clinopyroxene, chlorite, and calcite as pseudomorphous after amphibole and biotite phenocrysts) in a fine granular groundmass of quartz and alkali feldspar (Barker 1995) (Fig. 3). The phenocrysts reach lengths of 8 mm and show no visible preferred orientation. The interior zone, which makes up most of the intrusion, consists of the same minerals as the peripheral chill zone except that the groundmass is slightly coarser (medium grained) and the minerals are more altered (Fig. 3). This zone contains abundant irregular miarolitic cavities, consistent with its shallow emplacement depth. The intrusion post-dates deposition of the 22 Ma Harmony Hill tuff of the Quichipa Group and its age is approximated by a biotite K-Ar age of 21.5 ± 0.9 Ma (Mckee et al. 1997) and a 40Ar/39Ar age determination of 21.86 ±0.09 Ma (Rowley et al. 2003, in press). Field evidence from the surrounding laccoliths show that the Pinto Peak and Iron Mountain laccoliths were emplaced concurrently; prior to Stoddard Mountain emplacement (Hacker 1998). Both of these intrusions produced gravity slides, and the Pinto Peak intrusion vented a thick sequence of ash-flow tuff and lava flows (Fig. 2). Stoddard Mountain and The Dairy laccoliths are interpreted as coeval intrusions, because they both deformed volcanic units that vented from the Pinto Peak laccolith. The Stoddard Mountain laccolith also vented

Fig. 3. Thin section of representative texture from chilled zone rocks (a) plane light (b) polarized light. Inner zone rocks (not shown) have a similar texture but the groundmass and Fe-Mg silicates are more altered, mt, magnetite; plag, plagioclase feldspar; bio, biotite; hbd, hornblende.

MAGMATIC FLOW PATHS AND PALAEOMAGNETISM

ash-flow tuffs and lava flows following several gravity sliding events from its flank. At least two vent areas are interpreted along its flank, one on the east side where the intrusion bulges out from the main pluton (Fig. 2), and on the SW side where lava flows grade downward into the intrusion (Hacker 1998; Rowley et al., in review). Recent mapping of the Stoddard Mountain and The Dairy laccoliths (Hacker 1995, 1998; Hacker et al. 2002) suggests that the intrusions were emplaced as sills, which spread laterally for several kilometres before inflating to laccolithic shapes (Johnson & Pollard 1973). This is based on field relations of intrusive anticline fold axes NE (Richie Flat) and west of the Pinto Peak laccolith that have a characteristic arcuate trend wrapping around the Pinto Peak laccolith and its volcanic pile of over 500 m of ash-flow tuffs and lavas lying E-SE of the Pinto Peak laccolith (see Fig. 2a). The intrusive cored anticlines clearly deformed part of the Pinto Peak intrusive-volcanic complex (Fig. 2a), but did so only around the thinner edges of the pile. Hacker (1998) proposed that magma of both Stoddard Mountain and The Dairy laccoliths likely emanated from a common buried conduit north of Pinto Peak. Finding the area to the south blocked by the older Pinto Peak laccolith, the magma migrated to the east (for Stoddard Mountain) and west (for The Dairy). The sills, emplaced at about 1.2km depth within the Cretaceous Iron Springs Formation, are inferred to have spread laterally, east and west, for several kilometres around the northern edge of the thicker Pinto Peak intrusive-volcanic complex. As The Dairy sill migrated westward around the intrusive-volcanic complex, it turned southward and stepped up to a higher structural level within the Tertiary Claron Formation (Fig. 2a). As magma flowed east, beneath Richie Flat, it gently deformed the overlying country rock to produce the Richie Flat anticline. The Stoddard Mountain sill continued to migrate to the east past the intrusive-volcanic barrier, then turned southward before inflating fully. Thus, it appears that differences in lithostatic pressure caused by thickness variations of the Pinto Peak intrusivevolcanic overburden were influential in the lateral magma emplacement of both the Dairy and Stoddard Mountain laccoliths. The major focus of this research is to test the proposed lateral flow pattern hypothesis for development of the Stoddard Mountain laccolith. Because the quartz monzonite does not display any discernable mineral preferred orientations on the outcrops or in thin sections, we undertook a detailed AMS fabric study to provide flow-direction data for the laccolith.

257

Sampling and Analytical Methods Sampling methods We confined our sampling to the lower 250 m of the exposed laccolith, in part because of accessibility limitations, but also in an effort to sample a relatively narrow magma flow level or zone. Thirty-two palaeomagnetic sites were sampled with 12 sites in the chilled zone and 20 sites in the inner zone. Seventy-six AMS stations were sampled in a relatively tight grid network over the 54km exposure with sampling stations spaced at about 1km to 0.5km (Fig. 4). At either a palaeomagnetic site or an AMS station, from 4 to 10 independent samples (typically six) were drilled over an area of about 25m . We observed a relatively high dispersion in AMS data, at the station level, between samples in the Stoddard Mountain intrusion, which is common in granites (Rochette et al. 1992; Tarling & Hrouda 1993). Analysis of six or more samples per station (with at least 2 specimens per sample) yielded a statistically robust estimate of the orientations of each of the principal susceptibility axes and/or a planar fabric. In most cases, the dispersion between specimens from the same sample was low (k greater than 50°), however, at several stations, station-level dispersion was high (k less than 15°). Analytical methods Palaeomagnetism Oriented samples from all sites were collected using a portable, gasoline-powered drill with a non-magnetic, diamond-tipped drill bit. Each sample was oriented always using both solar and magnetic compasses. Each core sample was cut into 2.2 x 2.5cm right cylinder specimens, using a diamond tipped, non-magnetic saw blade. Up to four specimens per sample were obtained. Remanent magnetizations of all samples were measured using a static, three-axis 2G Enterprises magnetometer with an integrated alternating field (AF) demagnetizing unit. Specimens were progressively AF demagnetized, typically in 15 to 20 steps to a maximum field of 140mT. Samples of high coercivity were further treated with thermal demagnetization until less than 10% of NRM remained; typically, up to 615-635 °C depending on rock type and behaviour. Thermal demagnetization of replicate specimens, to compare with AF behaviour, was conducted using a Schonstedt TSD-1 thermal demagnetizer, with an ambient field of less than 20 nT in the heating chamber and 3nT in the

258

M. S. PETRONIS ET AL.

Fig. 4. Outline of the Stoddard Mountain laccolith. Numbered sample locations indicated by black dot correspond to locations numbered in Table 1 and Table 2. Dashed line: boundary between northern and southern zones.

cooling chamber. Principal component analysis (PCA; Kirshvink 1980) was used to determine the best-fit line through selected demagnetization data points for each sample. For most samples, a single best-fit line could be fit to the demagnetization data points. Best-fit magnetization vectors, typically, involved 10 to 15 data points. Sample

maximum angular deviation (MAD) values for linear data, typically, were less than 4°, but ranged from less than 1° to 5°. For less than 10% of the demagnetization results, it was necessary to anchor the magnetization vector to the origin. Data from individual specimens were considered outliers and rejected from the site

MAGMATIC FLOW PATHS AND PALAEOMAGNETISM

mean calculation if the angular distance between the specimen direction and the estimated site mean direction was greater than 15°. Rock magnetism To characterize the magnetic mineralogy of the Stoddard Mountain intrusion, standard rock magnetic experiments were conducted with the goal of identifying the magnetic phases carrying the remanence and the overall ability of these rocks to faithfully record an ambient field. The tests include AF demagnetization of anhysteretic remanent magnetization (ARM), d.c. acquisition of a saturation isothermal remanent magnetization (IRM), d.c. demagnetization of the saturation IRM to yield a backfield IRM (coercivity of remanence), AF demagnetization of saturation isothermal remanent magnetization (SIRM), bulk susceptibility in varying fields between 3 A/m and 450 A/m in 21 steps, monitoring bulk susceptibility changes during thermal demagnetization, continuous susceptibility vs. temperature experiments, thermal demagnetization of three-component orthogonal IRM (Lowrie 1990), and analysis of low temperature magnetic susceptibility. Anhysteretic remanent magnetization (ARM) was imparted in a d.c. field of 1 mT and a peak AF field of 95 mT, and then AF demagnetized in about 13 steps to 95 mT. An IRM acquisition experiment was then conducted where the samples are treated in a stepwise fashion to higher induction fields until they reach a saturation magnetization. A backfield-IRM experiment was then conducted by applying specimens to a 1.33 T isothermal induction field along the +z axis and treating the specimens to increasing induction fields along the -z axis until the +Zremanence was reduced to zero (i.e. the sign flipped). Finally, the samples were again saturated in a 1.33T induction field and then AF demagnetized in about 13 steps to 95 mT. For the thermal demagnetization of three-component IRM, we used fields of 3.0, 0.3 and 0.03 T, imparted using an impulse magnet. Analysis of the low-temperature magnetic susceptibility involved cooling samples to 77 K and continuously measuring susceptibility during heating to 298 K to obtain a temperature vs. susceptibility curve. Anisotropy of magnetic susceptibility The magnetic susceptibility (K) is defined by M = KH, where M is the induced magnetization of the material and H is the inducing magnetic field. The volumetric susceptibility (K) is dimensionless, because both M and H are expressed in A/m. The output of an AMS

259

measurement of one rock specimen is an ellipsoid of magnetic susceptibility defined by the length and orientation of its three principal axes, K! > K2 > K3, which are the three eigenvectors of the susceptibility tensor. Depending on magnetic mineralogy, granite rocks may be generally characterized as Mt (magnetite)-bearing or FeMg silicate dominated. In Fe-Mg silicate dominated granites, the silicates such as biotite or amphibole largely control AMS. In Mt-bearing granites, the contribution of Fe-Mg silicates is negligible with respect to the magnetite contribution because of the high intrinsic magnetic susceptibility of magnetite. The magnetic fabric, in that case, results primarily from the shape anisotropy of the grains of magnetite. This effect was postulated by Stacey (1960), and examined by microscopic observations on various rock types (Khan 1962; Uyeda el al. 1963) including granites (Ellwood and Whitney 1980; Gregoire et al. 1995). The AMS parameters are at times difficult to interpret in Mt-bearing granites, because of the unknown relationship between the fabric of magnetite (low volume %) and the fabrics of the volumetrically dominant phases (feldspars, amphibole, biotite). The usual observation, when magnetite crystallizes in association with K-feldspar, hornblende and biotite, is that the shape of the magnetite grains is controlled by the crystalline faces of those minerals (Hrouda et al. 1971). The magnetite grains use the Fe-Mg silicates as a 'template' during grain growth; therefore, the magnetic fabric is a good image of the rock fabric. This 'template' behaviour has been demonstrated on ferromagnetic granites with quantitative image analysis of thin sections (Cruden & Launeau 1994; Archanjo et al. 1995). Reflected light petrography confirms the presence of magnetite in association with Fe-bearing silicate phases in the Stoddard Mountain laccolith (Fig. 3). We measured the AMS of 959 specimens prepared from samples collected at 76 sites distributed across the intrusion (Fig. 4). AMS measurements were preformed on a Kappabridge KLY-4S, operating at low alternating field 3.7x 10~ 5 T (300A/m) at 875 Hz. These measurements give the orientation and magnitude of the K 1? K 2 , K3 axes of the ellipsoid; maximum, intermediate, minimum, respectively. The magnetic lineation, K 1? and the magnetic foliation, perpendicular to K 3 , characterize the fabric (Borradaile 1980; Hrouda 1982; Bouchez 1997, 2000). The relationships between tensor axes, normalized by the method of Jelinek (1977), were studied using the anisotropy degree, P', and the shape parameter, T. T varies between -1 (prolate ellipsoids) and +1 (oblate

260

M. S. PETRONIS ET AL.

ellipsoids) (Jelinek 1981). The orientation data (summarized in Table 2) are the foliation, which is the plane (K1? K2) perpendicular to the smallest axis (K3) of the ellipsoid, and the lineation, which is the long axis of the AMS ellipsoid (K!). The average, or bulk, susceptibility is given by K - (K^ + K2 + K 3 )/3. Palaeomagnetism General demagnetization behaviour Of the 32 sample sites, 30 sites yield acceptable results. The two excluded sites (SMS and SM20) had a high dispersion (a95 > 15°, k < 15) and did not yield interpretable results. These sites are not discussed further. We note that at many of the sites only 6 samples were collected, because the principal reason for carrying out this study was to obtain and evaluate AMS data to infer emplacement fabrics. With a more conventional number of samples (e.g. seven to ten) at each site, the confidence and precision estimates would probably improve. A structural correction is not available for any parts of the laccolith. Therefore, all results are reported in 'in situ' (geographic) coordinates. In general, 11 sites in chill zone rocks, distributed around the perimeter of the intrusion, yield south declination and shallow to moderate negative and positive inclination mean directions; and 19 sites in the inner zone rocks yield north to NW declination, moderate positive inclination mean directions. Two sites in the inner zone rocks yielded south declination, moderate steep inclination means; these sites are antipodal to the remaining inner zone sites and we have therefore inverted them through the origin. Overall, progressive alternating field (AF) response of these rocks is characterized by high quality results (Fig. 5). Duplicate specimens treated with thermal demagnetization yielded directional data similar to those resolved in AF demagnetization (Fig. 5). A clear difference exists in demagnetization behaviour and remanence data (see below), between chilled zone rocks and inner zone rocks. Representative demagnetization diagrams show the variation between the two rock types and the overall high quality of the demagnetization response (Fig. 5a, b). Chill zone Chill zone rocks show a relatively complex demagnetization behaviour characterized by superimposed magnetization components (Fig.

5b). Most specimens yielded a low coercivity N-NE declination, steep, positive inclination overprint that is readily removed by about 20 mT. The remaining component, defined by roughly 30% of the NRM intensity, decays along a near univectoral path to the origin and is well grouped at the site level with 10 to 40% of the NRM intensity remaining after AF treatment to 140mT. Median destructive fields range from about lOmT to 40mT. Selected treatment of high coercivity specimens with thermal demagnetization up 635 °C did not isolate additional magnetization components (Fig. 5b). The eleven accepted sites in the chill zone rocks yield S-SE declination, shallow negative or positive inclination magnetizations that are well grouped (eight sites with a95 < 15°) at the site level. Two sites (not included with the eleven chill zone sites), collected within a gradational, transition zone between the chill and inner zones, yield south declination, moderate negative inclination mean directions (Table 1) and, because the mean directions are essentially antipodal (less than one angular standard deviation from the overall inner group mean) to the results characteristic of the inner zone, they have been inverted and included to define a grand inner zone mean. The eleven sites provide an overall, well-defined, estimated group mean (£>=178.5°, 7 = 0.9°, a95 = 10.0°, k = 24.1) (Fig. 6). The result from the eleven sites lies well over 90° from a Mid-Miocene normal polarity expected field and over 45° from a reverse polarity expected field, based on the palaeomagnetic pole of Mankinen et al. (1987) (Fig. 6). Inner zone Demagnetization of inner zone rocks reveals an essentially single magnetization component that is well grouped at the site level with median destructive fields (MDF) from about 30 mT to about 80 mT (Fig. 5a). Some samples, however, also contain a superimposed random magnetization component that is readily removed by about 15mT or by 300 °C, which we interpret to be a low-coercivity viscous overprint (VRM) (Fig. 5a). After removing the VRM, the characteristic remanent magnetization (ChRM), which we interpret as the primary thermoremanent magnetization (TRM) and of normal polarity, decays along a roughly univectoral path to the origin with less than 10% of the NRM intensity remaining after treatment in 140 mT fields or by 620 °C (Fig. 5a).

Fig. 5. Representative modified demagnetization diagrams (Zijderveld 1967; Roy & Park 1974) for rocks collected from the Stoddard Mountain laccolith (in situ coordinates), (a) Inner zone. Solid (open) symbols represent the projection onto the horizontal (true vertical) plane. AF demagnetization steps are given in milliTesla and thermal demagnetization steps in degrees Centigrade. Typically, AF and thermal demagnetization results from two specimens of the same sample are shown for comparison. Diagrams are designated by a site number (SMI), specimen letter (-b), method of treatment (AF or TH). Magnetization intensity (A/m), is shown along one axis for each sample.

Fig. 5. Representative modified demagnetization diagrams (Zijderveld 1967; Roy & Park 1974) for rocks collected from the Stoddard Mountain laccolith (in situ coordinates), (b) Chilled zone. Solid (open) symbols represent the projection onto the horizontal (true vertical) plane. AF demagnetization steps are given in milliTesla and thermal demagnetization steps in degrees Centigrade. Typically, AF and thermal demagnetization results from two specimens of the same sample are shown for comparison. Diagrams are designated by a site number (SMI), specimen letter (-b), method of treatment (AF or TH). Magnetization intensity (A/m), is shown along one axis for each sample.

263

MAGMATIC FLOW PATHS AND PALAEOMAGNETISM

Table 1. Anisotropy of magnetic susceptibility data from the Stoddard Mountain laccolith, in geographic (in situ) coordinates. (For symbol description, see explanation attached to data table) Site

No

SM-1 SM-2 SM-3* SM-4 SM-5 SM-6 SM-7 SM-8 SM-9 SM-10 SM-11 SM-12 SM-13 SM-14 SM-15* SM-16 SM-17 SM-18 SM-19 SM-20 SM-21* SM-22 SM-23 SM-24 SM-25 SM-26 SM-27 SM-28 SM-29* SM-30 SM-31 SM-32 SM-33 SM-34* SM-35* SM-36 SM-37 SM-38 SM-39 SM-40* SM-41 SM-42 SM-43 SM-44 SM-45 SM-46 SM-47 SM-48 SM-49 SM-50 SM-51 SM-52 SM-53* SM-54 SM-55 SM-56 SM-57 SM-58 SM-59

6 7 7 6 6 6 6 7 6 6 6 6 6 6 6 4 6 7 6 6 6 5 6 6 7 7 6 5 5 6 6 5 4 6 5 5 5 6 7 8 6 6 7 8 5 5 8 7 6 7 8 6 6 5 6 8 5 6

N

Km

Kl

K3

L

22 23 20 19 18 18 14 14 12 13 13 12 12 11 12 12 12 14 12 12 12 13 12 12 12 14 12 19 11 12 8 12 12 12 12 12 12 12 12

18 19 10 13 14 12 14 11 8 7 12 9 8 9 10 6 9 11 9 10 12 8 10 8 8 10 11 8 7 10 8 8 10 12 10 8 10 11 10

1.73 2.77 1.31 1.70 26.19 14.01 20.76 21.32 14.81 15.39 23.70 1.95 2.04 1.59 6.96 6.59 1.89 5.31 20.53 17.86 1.21 10.13 8.20 31.41 23.07 18.81 24.68 18.83 27.25 24.35 24.40 8.28 6.23 30.27 19.44 17.65 23.84 8.18 10.89

171/16 181/2 73/38 175/1 144/25 123/14 303/8 148/16 311/2 160/7 5/42 130/8 168/1 225/16 50/26 20/1 63/4 49/38 195/49 100/11 223/51 29/3 43/7 193/27 214/4 3/34 290/19 181/28 259/13 114/42 268/25 65/7 51/25 11/6 336/4 329/15 119/19 100/8 224/21

262/3 287/84 184/25 84/34 280/57 353/69 44/52 38/50 47/73 292/80 153/43 320/82 77/52 75/71 145/10 114/71 155/32 154/19 91/11 324/76 90/29 297/35 296/67 330/56 124/9 171/56 96/70 65/39 358/34 325/48 136/56 230/83 302/36 202/84 76/71 166/74 13/39 268/82 327/30

1.009 1 .001 1 .009 1.010 1.005 1 .002 1.007 .007 1.006 1.003 1.008 .009 1.009 1 .003 1.012 .012 1.007 1 .013 1.021 .021 1.012 1 .015 1.027 .027 1.003 1 .012 1.015 .015 1.008 ] .014 1.022 1.022 1.005 ] .012 1.017 1.018 1.013 1 .006 1 .019 1.019 1.004 ] .013 1.016 1.017 1.002 ] .009 1.011 1.011 1.004 1.006 1.009 1.009 1.003 ] .003 1.007 1.007 1.005 ] .007 1.012 1.012 1.009 ] .003 1.012 1.012 ] .002 1.004 1.006 1.006 ] .008 L.002 1.010 1.011 ] .009 1 .003 1.012 1.012 1.008 1.010 1.016 1.016 1.001 1.001 1.002 1.002 1.012 1.005 1.017 1.018 1.007 1.006 1.013 1.013 1.017 1.004 1.021 1.022 1.008 1.008 1.016 .016 ] .009 1.007 1.016 .016 1.011 1.010 1.021 .021 1.014 1.013 1.027 .027 1.007 1.004 1.011 .001 1.006 1.011 1.017 .017 1.006 1.018 1.024 .025 1.006 1.007 1.013 .013 ] .008 1.008 1.016 .016 ] .003 1.016 1.019 .020 ] .008 1.004 1.013 .013 ] .004 1.006 1.010 .011 1.007 1.004 1.011 .011 1.007 1.004 1.011 .011 1.007 1.008 1.015 .015

13 12 13 11 12 12 12 12 12 12 8 12 12 12 12 12 12 12 12

10 12 7 8 12 10 9 12 10 12 7 11 9 12 9 11 12 9 12

2.71 3.57 12.83 23.13 15.06 21.35 23.09 2.37 4.89 18.24 2.04 13.52 12.58 4.27 7.91 9.09 6.67 23.84 4.12

70/9 99/20 29/7 353/10 274/64 131/21 88/62 2/5 40/32 70/13 73/59 52/22 153/8 128/48 191/60 156/59 128/45 211/13 115/31

256/81 1.003 227/59 1.001 122/26 1.009 89/31 1.009 11/4 1.008 253/54 1.020 205/14 1.010 266/51 1.003 247/55 1.006 286/75 1.012 235/30 1.008 211/67 1.006 1.003 6/81 302/42 1.008 313/17 L.010 289/22 1.006 356/34 1.007 317/52 1.009 303/58 L006

F

1.003 1.007 1.010 1.011 1.005 1.013 1.002 1.015 1.006 1.005 1.008 1.014 1.015 1.013 1.010 1.006 1.015 1.008 1.007

P

1.007 1.008 1.019 1.020 1.014 1.034 1.012 1.017 1.013 1.016 1.016 1.020 1.018 1.022 1.020 1.012 1.022 1.017 1.013

Pj

.007 .008 .019 .020 .014 .034 .013 .019 .013 .017 .016 .021 .019 .022 .020 .012 .023 .017 .013

T

Shape

Rk Type

-0.870 -0.310 -0.354 -0.469 0.283 0.089 0.611 0.254 0.435 -0.326 0.543 0.637 0.205 -0.044 0.131 -0.492 0.218 -0.575 -0.571 0.128 0.233 -0.388 -0.150 -0.639 -0.010 -0.074 -0.021 -0.023 -0.276 0.299 0.474 0.113 0.003 0.664 -0.309 0.152 -0.310 -0.200 0.130

P P P P 0 O 0 0 O

Inner Inner Inner Inner Inner

-0.033 0.736 0.038 0.119 -0.224 -0.223 -0.619 0.692 0.002 -0.433 -0.044 0.371 0.667 0.224 -0.025 -0.005 0.353 -0.047 0.103

P

P

0 O O

P 0 P 0 P P 0 0

P P P P

P P

P P 0 O O 0 O

P

O

P

P 0 0 O 0

P P P

O 0

P

P O O O

P P

0

P

0

Chilled Chilled Chilled Chilled Chilled Inner Inner Inner Inner Inner Inner Inner Chilled Inner Inner Inner Inner Inner Inner Chilled Chilled Chilled Chilled Chilled Chilled Inner Inner Inner Inner Inner Inner Inner Inner Inner Inner Inner Inner Chilled Inner Chilled Chilled Inner Inner Inner Inner Inner Inner Inner Inner Inner Inner Inner Inner

M. S. PETRONIS ET AL,

264

Table 1. (cont.) Anisotropy of magnetic susceptibility data from the Stoddard Mountain laccolith, in geographic (in situ) coordinates. (For symbol description, see explanation attached to data table) Site

No

SM-60 SM-61 SM-62 SM-63* SM-64* SM-65 SM-66 SM-67 SM-69 SM-70* SM-71 SM-72* SM-73 SM-74 SM-75 SM-76 SM-77

6 5 10 6 5 5 6 8 7 6 7 8 8

11 5 6 7

12 12 12 13 12 12 12 14 14 12 13 14 9 14 7 12 12

N

Km

Kl

K3

L

F

P

Pj

T

Shape

Rk Type

9 9 12 10 8 8 8 13 12 8 13 11 9 14 6 10 11

8.99 3.40 28.52 8.82 11.63 24.00 16.31 27.53 2.16 3.19 4.33 2.19 23.33 31.03 25.31 28.20 3.43

172/23 186/25 82/2 347/17 185/4 61/36 34/32 353/9 295/23 264/35 38/46 340/16 257/9 140/20 3/3 38/14 163/7

263/1 285/20 205/87 102/55 67/83 253/54 168/49 105/67 152/62 154/26 258/37 233/47 124/77 32/41 94/28 190/74 67/40

1.012 1.009 1.004 1.006 1.009 1.006 1.009 1.006 1.004 1.012 1.000 1.004 1.011 1.013 1.009 1.014 1.009

1.004 1.009 1.016 1.002 1.001 1.009 1.009 1.014 1.008 1.002 1.002 1.001 1.003 1.007 1.002 1.012 1.005

] .016 1 .018 1.020 1.008 1.010 L015 1.019 1.020 1.012 1.014 1.003 1.005 1.013 1.020 1.011 1.026 1.014

1.016 1.018 1.021 1.008 1.011 1.015 1.019 1.020 1.012 1.015 1.003 1.005 1.014 1.020 1.012 .026 .014

-0.537 -0.050 0.620 -0.515 -0.826 0.248 -0.026 0.355 0.304 -0.735 0.624 -0.500 -0.606 -0.313 -0.664 -0.103 -0.332

P P 0 P P 0

Inner Inner Inner Inner Inner Inner Chilled Chilled Inner Inner Inner Inner Inner Inner Inner Inner Inner

P

0

o

P o P P P P P P

No, number of samples collected at each site; N, number of accepted specimens at each site, typically, one to three specimens per sample; Km, magnitude of susceptibility (in 10~3SI); Kl, azimuth and plunge (in degrees) of magnetic lineation; K3, azimuth and plunge (in degrees) of normal to magnetic foliation plane; L, magnetic lineation ((Kl - K2)/Km); F, magnetic foliation ((K2 - K3)/Km), P, anisotropy degree (K1/K3); Pj, magnitude of anisotropy, corrected anisotropy (Jelinek 1981); T, shape

The 19 acceptable sampling sites established yield interpretable demagnetization data of NNW declination and moderate, positive inclination values that are well grouped (18 sites with a95 < 10°) at the site level (Table 1). The 19 results are used to calculate an overall group mean (D = 351.1°, 7-57.8°, a95 = 3.7°, k = 82.2) that is well defined and slightly discordant, in an anticlockwise sense, to the Mid-Miocene expected direction (358°, 58°), based on the palaeomagnetic pole of Mankinen et al. (1987) (Fig. 6).

(< 1 m) and then sampling away from the intrusive contact in exposed sandstone layers. All modified tests yield positive results. Sites closest to the intrusive contact (< ~lm) appear to be thermally annealed and yield remanence directions similar to the intrusive rocks, regardless of the orientation of the Cretaceous strata. Sites located greater than 1 m from the contact yield poorly defined remanence directions and exhibit demagnetization behaviour characteristic of low-coercivity, multidomain magnetite, which is typical of the Iron Springs Formation in areas well away from the intrusion.

Palaeomagnetic field tests

Chilled margin test At the eastern margin of the intrusion, a road cut allowed for a non-standard field test to be conducted. We informally refer to this test as a chilled margin test (Fig. 7). The field test is based on the observation that the inner zone and chill zone rocks yield significantly different remanence directions. As noted above, chill zone rocks are considered to form a thin carapace to the intrusion, between 20 and 60m thick, which is easily recognized in the field by a low degree of oxidation of Fe-bearing silicate minerals. The inner zone rocks, however, are only moderately oxidized. The nature of the test involved a traverse across the inner zonechill zone boundary that started at the intrusive-country rock contact and ended about

Contact test Comprehensive contact tests were attempted at three locations on the margin of the intrusion in order to assess the antiquity of the magnetization. Unfortunately, the laccolith was emplaced into the Cretaceous Iron Springs Formation, a very fine- to medium-grained sandstone interbedded with black shale, and, where exposed, it intruded into the black shale layer, which prevented us from collecting samples directly adjacent to the intrusion. Therefore, an ideal contact test could not be conducted and a modified contact test was attempted. This test entailed collecting oriented samples from the sandstone as close as possible to the intrusive contact

MAGMATIC FLOW PATHS AND PALAEOMAGNETISM

265

Fig. 6. Equal area projections of in situ site mean directions from the Stoddard Mountain laccolith, (lower hemisphere projection = solid circle; upper hemisphere projection = open circle; ellipse, 95% confidence ellipse; normal (reversed) polarity Late Oligocene-Early Miocene expected direction = solid (open) square)

75m into the intrusion; a distance that should have included the chill zone-inner zone boundary (Fig. 7). A complete palaeomagnetic site was established at either end of the traverse (SM45 west side; SM44 east side) and a total of 28 samples were collected along the traverse at a spacing of about two to ten metres and, in the region where the degree of oxidation of Febearing silicates increased, samples were collected every half metre. The test did not yield a 'positive' result. Based on field observation, it was anticipated that the chill zone-inner zone transition should be relatively abrupt and the south declination/very shallow inclination characteristic remanence direction of the chill zone rocks should progressively, over a short

distance, change to that characteristic (north declination/moderate positive inclination) of the inner zone rocks. The directional data from the 28 specimens collected along the traverse yield consistent south-directed declinations, but a general pattern of inclination steepening from east to west into the 'main' body of the laccolith (Fig. 7). Data from the two palaeomagnetic sites also reflect this general trend (Fig. 7). Site SM44, located at the extreme eastern boundary of the laccolith, yields a south declination, shallow inclination mean (D = 174.7°, 7 = -9.6°) and SM45, located about 75m west of SM44, yields a south declination, more moderate inclination mean (D = 177.8°, 7 - -30.9°) (Fig. 7). Based on these data, the chill zone

266

M. S. PETRONIS ET AL.

Fig. 7. Palaeomagnetic chilled margin test. Samples (indicated by black triangles in upper diagram) collected at increasing distances from the intrusive-country rock contact yield progressively steeper inclination values and basically form a loop that trends towards the expected reversed polarity field direction (lower diagram). Lower diagram partial lower hemisphere equal area projection.

boundary, defined by the alteration degree of the oxide mineral phases, does not correspond with the remanence-defined boundary. It appears that the remanence-defined inner zone-chill zone boundary is characterized by intermediate remanence directions. Rock magnetism and AMS results

Rock magnetic properties Standard rock magnetic tests indicate that the remanence and the AMS fabrics in the Stoddard Mountain laccolith are likely defined by a ferromagnetic mineral phase, probably multidomain (MD) to pseudo-single domain (PSD) magnetite, although some variations were

observed. Analysis of the low temperature magnetic susceptibility reveals that paramagnetic and ferromagnetic minerals both contribute to the susceptibility, and heating curves of most samples are the result of the superposition of the paramagnetic and ferromagnetic effect (Richter & van der Pluijm 1994), although some samples exhibit temperature independent behaviour, indicating the presence of single domain (SD) magnetite (Fig. 8). Susceptibility magnitudes for paramagnetic mineral phases never exceed 3 x 1(T4SI (Uyeda et al 1963; Hrouda 1982; Borradaile 1988; Archanjo et al 1995). In general, ferromagnetic granites (Mtbearing) have a specific susceptibility about three orders of magnitude higher than that of paramagnetic mineral species (Borradaile 1988). We attribute the high average bulk susceptibility

MAGMATIC FLOW PATHS AND PALAEOMAGNETISM

Fig. 8. Normalized reciprocal magnetic susceptibility (ko/k) as a function of temperature. Ideal ferromagnetic curve would be a straight line (no change in susceptibility with temperature), ideal paramagnetic curve is a line described by the CurieWeiss law (kpara = C/T - 0, where C= Curie constant, 9 = paramagnetic Curie temperature; T = temperature in kelvin).

(13.46 x 10"3 SI ±9.48 x 1(T3SI) to magnetite as the principal mineral phase contributing to the overall susceptibility of the Stoddard Mountain intrusion (Table 1). AF demagnetization of the NRM reveals that MD magnetite grains of intermediate coercivity are characteristic of the chill zone rocks, as suggested by median destructive fields (MDF) of the NRM, when it is essentially single component in nature, between 10 and 20mT (Fig. 9). The typical behaviour of the inner zone rocks implies a dominance of somewhat higher coercivity phases, with MDFs of about 30-80 mT (Fig. 9). Stepwise thermal demagnetization of duplicate specimens essentially give the same demagnetization results and yield laboratory unblocking temperatures

267

between 540 °C and 635 °C (Fig. 5). This behaviour is indicative of the presence of low Ti to nearly pure magnetite and maghemite, both of which carry a important fraction of the natural remanence and low-coercivity magnetizations are likely viscous in origin (O'Reilly 1984; Dunlop & Ozdemir 1997). IRM acquisition curves show a narrow spectrum of responses, ranging from MD-magnetite-dominated behaviour to that influenced by somewhat higher-coercivity PSD magnetite and, on the basis of other observations, some maghemite (Fig. 9). All samples are likely dominated by a cubic phase (magnetite-type curves) and these show steep acquisition curves and nearly reach saturation by 0.20T. Most specimens reach complete saturation by 1.25T, yet some did not reach a stable plateau by 1.25T (e.g. SM6, Fig. 9). Typically, IRM acquisition curves climb steeply below 0.1 T and display moderately to strong inflections around 0.15T. The curves for the inner zone rocks reach saturation by about 0.15T, and these rocks are likely dominated by a lower coercivity assemblage of grains than that typical of the chill zone rocks, which display curves that reach saturation at slightly higher fields than inner zone rocks. This result may indicate that multiple generations of ferromagnetic phases are present in the chilled zone rocks, which is consistent with the general multicomponent composition of the NRM (Fig. 5). Back-field IRM acquisition curves show a very narrow range of responses, with coercivity of remanence values between about 0.030T and 0.065T. Chill zone rocks yield values between 0.045 T and 0.065 T, and inner zone rocks yield values between 0.030 T and 0.065 T (Fig. 9).

Fig. 9. Representative normalized IRM acquisition and back-field IRM demagnetization curves.

268

M. S. PETRONIS ET AL.

Fig. 10. Modified Lowrie-Fuller test (Johnson et al. 1975) comparing AF demagnetization response of natural remanent magnetization (NRM), anhysteretic remanent magnetization (ARM) and saturation isothermal remanent magnetization (SIRM).

Modified Lowrie-Fuller test Following the modified Lowrie-Fuller test (Johnson et al. 1975) the AF decay of normalized NRM, ARM and SIRM was compared. The test is based on the experimental observation that normalized AF demagnetization curves of weak-field TRM (i.e. NRM) and strong-field TRM (i.e. SIRM) have different relationships for SD and MD grains of magnetite (Dunlop & Ozdemir 1997). Laboratory investigations typically use weak-field ARM as a proxy for TRM, and we adhere to that procedure. In large MD grains, SIRM requires larger destructive fields than ARM to reach the same normalized level (i.e. the MDF of SIRM is larger than MDF of ARM). AF demagnetization of strong- and weak-field IRM show a limited spectrum of responses. All curves suggest MD magnetitedominated behaviour, with a limited range in grain size likely responsible for carrying the remanence (Fig. 10). The chill zone rocks generally yield curves where ARM is more resistant than SIRM. In inner zone rocks, SIRM is generally more resistant to AF demagnetization than ARM.

Continuous susceptibility vs. temperature Continuous susceptibility vs. temperature measurements on a Kappabridge KLY-3S instrument from room temperature to about 700 °C

were made to further evaluate the magnetic mineralogy. The measurements were conducted in air, because maintaining an AR atmosphere with high-susceptibility igneous rocks during the experiment does not yield significantly different results. Continuous susceptibility vs. temperature measurements show a consistent decrease in room temperature susceptibility, of about 5 to about 20%, after heating to near 700 °C (Fig. lla). Inferred Curie temperatures range from about 560 °C to about 635 °C. In some cases, the presence of a second ferromagnetic phase (maghemite) is indicated by a decrease in susceptibility over a temperature range of 300 °C to 400 °C in the heating curve, with no indication of its presence in the cooling curve. Inferred Curie temperature ranges are similar for inner and chill zone rocks.

Bulk susceptibility with temperature Bulk susceptibility measured on a Sapphire Instruments SI-2 system between steps in the progressive thermal demagnetization of the NRM shows similar results for chilled and inner zone rocks, although some variations do occur (Fig. lib). Most inner zone rocks are characterized by a rapid drop (about 40 to 60%) between 350 °C and 450 °C, suggesting that maghemite is converted to hematite (Fig. lib) (Dunlop and Ozdemir 1997). Magnetite is inferred to remain stable to about 580 °C (i.e.

MAGMATIC FLOW PATHS AND PALAEOMAGNETISM

269

Fig. 11. Continuous susceptibility vs. temperature experiments and bulk susceptibility as a function of thermal demagnetization, (a) Continuous bulk susceptibility measurement during heating (solid lines) to about 700 °C and cooling (dotted lines) to 25 °C. (b) Normalized bulk susceptibility during stepwise progressive thermal demagnetization of the NRM.

<10% change in x to 580°C). At higher temperatures (>580°C), susceptibility decays uniformly by another 20%, suggesting that maghemite continues to break down and is an important carrier of the NRM of the rocks. Most chill zone rocks are characterized by a rapid drop (about 10%) between 350 °C and 450 °C, with magnetite remaining stable to about 580° C (i.e. <5% change in x to 580 °C). No change in susceptibility occurs at temperatures above 580 °C. It is likely that maghemite is converted to hematite in the oxygen-rich environment of the furnace and that magnetite carries the natural remanence up to 580 °C, with most of the NRM unblocked by this temperature. Thermal demagnetization of three component IRM The test of Lowrie (1990) revealed considerable heterogeneity in the overall magnetic mineralogy of the materials studied from the Stoddard Mountain intrusion (Fig. 12). In some cases, the IRM acquired at 3.0T is similar to (but never greater than) that acquired at 0.3T (e.g. specimen 13F, Inner Zone, north part of pluton). The 3.0T IRM persists to well above 600 °C, indicating that hematite is an important phase in these rocks, despite the fact that it does not carry appreciable remanence, based on thermal demagnetization. The 0.3T IRM component varies in intensity by over an order of

magnitude, and is always greater than that acquired at 0.03T. For some specimens (e.g. 12G, Inner Zone, north part of pluton) this IRM component is not completely unblocked by about 580 °C, suggesting that fine-grained maghemite may be an important phase. Thermal demagnetization of specimens from samples of this site, for example, show that the NRM is fully unblocked by about 620 °C. Overall, we see no systematic relationship between inner zone and chilled zone rocks and their response to thermal demagnetization of three component IRM. For example, those sites from the chill zone that yield the south-directed and shallow inclination remanence (i.e. specimens 7D, 9F, 1 IB, 19E, and 26B) exhibit a range of behaviours comparable to those from sites in inner zone rocks. Anisotropy of magnetic susceptibility In rocks where the main magnetic minerals are MD low-titanium magnetites, as is the case for the Stoddard Mountain laccolith, the shape of the grains controls the AMS. The susceptibility anisotropy results from an alignment of elongated particles or a planar/linear distribution of the particles. All AMS measurements were performed prior to palaeomagnetic analysis, because AF treatment can impart a domain anisotropy superimposed on the shape anisotropy (Park et al 1988; Potter & Stephenson 1990;

270

M. S. PETRONIS ET AL.

Fig. 11. Bulk susceptibility as a function of thermal demagnetization, and continuous susceptibility vs. temperature experiments.

Rochette et al. 1992), especially for low anisotropy rocks. In total, 959 specimens were analyzed from 76 stations with 763 specimens yielding interpretable results. Excluded specimens (196) were rejected due to a high within sample dispersion, meaning the individual specimen results from the same sample did not overlap at one standard angular deviation, or

an individual specimen result was greater than 2a from the overall mean of the station. In addition, 12 stations were rejected because of high between-specimen dispersion; the confidence interval about the three mean principal tensor axes were all greater than 15° (Fig. 13). The magnetic susceptibility (Km) magnitudes from 763 specimens vary from 31.41 to 1.21 x 10~3 SI,

MAGMATIC FLOW PATHS AND PALAEOMAGNETISM

271

Fig. 12. Representative thermal demagnetization curves of three-component orthogonal IRM (Lowrie 1990).

magnetic anisotropy (Pj) ranges from 1.034 to 1.001, and magnetic ellipsoids are both prolate and oblate (Fig. 14a; data at the site level are summarized in Table 2). The bulk susceptibility of all samples is high and indicates that the magnetic susceptibility is mainly produced by ferromagnetic minerals (Rochette 1987; Bouchez 1997), with some contribution by Fe-Mg bearing silicates. The susceptibility magnitude is lower for inner zone rocks (mean of 11.02 x 10~3 SI) than that of chill zone rocks (mean of 21.18 x 10~3SI). The anisotropy degree (Pj) is low, but also varied between inner and chill zone rock (mean inner =1.014; mean chill zone= 1.018), and the average magnitude of Pj generally increases with the increase of the bulk susceptibility (Table 2). The shape of the AMS ellipsoid, defined by the parameter of T (Jelinek 1978), is not correlated with Pj (Table 2; Fig. 14b). AMS fabrics, denned by magnetic lineations (K! axes) and magnetic foliation planes (K! — K2 plane), are spatially variable through

the general extent of the laccolith and are inferred to correlate with the magmatic fabrics in specific parts of the intrusion that developed during emplacement (Table 2; Fig. 15). In general, magnetic lineations throughout the inner zone have low to moderate plunge values with N-NE to S-SW trends, while the chilled zone localities are characterized by magnetic foliations or lineations that parallel the country rock contact (Fig. 15a). To simplify the description of the magnetic fabric distribution, the AMS data are divided into three zones, each of which reveals consistent lineations and/or planar fabrics: (1) chilled zone (19 stations), (2) northern zone (27 stations) and (3) southern zone (31 stations). Chilled zone In general, all chilled zone stations, regardless of their location on the margin of the intrusion, are characterized by a well-developed, steep to moderate dipping planar magnetic fabric that parallels the sub-circular contact with the

272

M. S. PETRONIS ET AL.

Fig. 13. Lower hemisphere equal area projections of AMS principal susceptibility axes (max Kj, intermediate K2, and minimum K3) for representative sites in the (a) chill zone, (b) northern zone, and (c) southern zone rocks of the Stoddard Mountain laccolith.

country rock (Fig. 15b), with magnetic lineations also trending roughly parallel to the country rock contact (Fig. 15b). One station was rejected due to a high between-sample dispersion. More specifically, magnetic lineations along the western margin of the laccolith trend roughly parallel to the contact with the country rock

(NW-SE) and plunge gently (about 21°). A moderate- to steep-dipping planar magnetic fabric parallels the country rock contact at most stations (11/13 stations). Two stations near Richie Flat, however, yield magnetic lineations that trend east-west and have a moderate south-dipping planar fabric that are

MAGMATIC FLOW PATHS AND PALAEOMAGNETISM

273

Fig. 14. Relationship between the degree of magnetic anisotropy, Pj, and (a) bulk susceptibility, Km (b) shape of the susceptibility ellipsoid, T.

perpendicular to the country rock contact (Fig. 15). The trend of the magnetic lineations near Richie Flat seem to reflect the flow pattern, while the remainder of the chilled zone stations seem to be influenced by the laccolith margin. Five stations on the NE and eastern margin of the intrusion yield a well-defined moderate to near vertical dipping planar fabric that parallels the country rock contact (Fig. 15b). Northern zone Of the 27 stations established in the northern zone, 20 stations yield interpretable results. Magnetic lineations in the northern zone trend generally E-NE and plunge gently to moderately (Fig. 15a), although two stations yield a SE trend and one station yields a NW trend with plunges for the three stations between 1° and 23° (Fig. 15a). Sub-horizontal planar fabrics, which are generally well defined, characterize stations along the traverse through the northern zone (Fig. 15b).

Southern zone Of the 31 stations established in the northern zone, 27 stations yield interpretable results. Magnetic lineations in the southern zone trend south to SE and plunge gently to moderately. Four stations in the eastern-part of the laccolith yield magnetic lineations that trend nearly east towards the location from which the laccolith vented (Fig. 4; Fig. 15a). One station yields a magnetic lineation direction that trends NE and plunges gently (14°) and one station trends NW and plunges gently (15°). Overall, sub-horizontal planar fabrics characterize stations in the southern zone (Fig. 15b). Discussion Our study of the AMS and palaeomagnetism of the shallow level Stoddard Mountain intrusion shows that the petrographically distinct zones of the quartz monzonite exhibit different charac-

M. S. PETRONIS ET AL.

274

Table 2. Palaeomagnetic data from the Stoddard Mountain laccolith Site

n/N

R

a95

K

D

I

VGP LAT

LONG

RK Type

SMI SM2 SM3 SM4 SMS SM6 SM7 SMS SM9 SM10 SM11 SM12 SM13 SM14 SM15 SM16 SM17 SM18 SM19 SM20 SM21 SM22 SM23 SM24 SM25 SM26 SM28 SM31 SM44 SM45 SM66 SM73

6\6 7\7 5\7 5\6 6\6 6\6 6\6 reject 5\6 6\6 6\6 6\6 4\6 5\6 6\6 6\6 5\6 7\7 4\6 reject 6\6 5\6 6\6 5\6 6\6 6\7 5\6 5\5 5\5 7\7 4\5 5\6

5.949 6.952 4.956 4.966 5.948 5.932 5.758

6.8 5.4 8.1 7.1 6.8 7.8 15.1

97.61 125.73 89.85 116.45 97.05 73.95 20.70

340.6 340.7 348.9 357.1 0.9 179.3 181.2

64.6 60.9 65.7 52.2 54.8 -68.9 21.1

73.1 74.5 76.8 84.7 87.6 -75.1 -41.5

-165.3 179.3 212.1 94.1 47.8 64.9 -114.9

4.892 5.918 5.882 5.904 3.990 4.966 5.973 5.851 4.971 6.924 3.909

12.7 8.7 10.4 9.4 5.4 7.1 4.9 11.7 6.5 6.9 16.3

37.15 60.88 42.54 51.97 287.87 116.36 118.20 33.56 139.79 78.69 32.85

180.6 182.7 193.2 336.6 344.5 350.3 349.2 349.6 10.7 147.1 181.5

13.0 12.8 4.8 46.9 48.4 51.1 55.4 51.0 49.5 -73.1 -12.3

-45.8 -45.9 -48.2 68.3 74.7 80.2 81.2 79.7 78.6 -60.0 -58.7

-114.2 -117.3 -133.4 137.8 124.2 115.3 140.1 125.3 10.3 32.4 -116.3

5.952 4.957 5.969 4.980 5.987 5.909 4.9437 4.9778 4.8867 6.886 3.9817 4.9591

6.6 8.0 5.3 5.5 3.5 9.1 9.1 5.7 13.1 8.4 7.2 7.8

104.26 93.06 159.46 197.95 376.10 55.06 71.0093 179.9127 35.3153 52.6411 163.6567 97.6953

349.3 358.1 354.8 349.1 355.0 172.4 187.5 158.3 174.7 177.8 174.1 358.8

55.7 55.2 58.8 55.9 57.8 -8.6 11.4 3.8 -9.6 -30.9 3.4 66.6

81.3 87.6 85.5 81.2 86.0 -56.1 -46.0 -45.6 -56.8 -68.9 -50.3 78.5

151.2 108.2 184.4 152.9 171.0 -99.7 -128.5 -85.9 -108.0 -111.9 -108.6 -121.7

Inner Inner Inner Inner Inner Inner Chilled Chilled Chilled Chilled Chilled Inner Inner Inner Inner Inner Inner Chilled Chilled Inner Inner Inner Inner Inner Inner Chilled Chilled Chilled Chilled Inner Chilled Inner

Sites with (*) were rejected in this study. n/N = ratio of samples accepted for statistical analysis to samples collected; R = resultant vector length; a95 = 95% confidence interval about the estimated mean direction, assuming a circular Fisher distribution; K = best estimate of (Fisher) precision parameter; D/I = in situ declination and inclination; VGP = in situ latitude and longitude of the virtual geomagnetic pole; RK Type, rock type, inner zone or chilled zone.

teristics in both their palaeomagnetic remanence behaviour and their anisotropy of magnetic susceptibility.

Palaeomagnetism Palaeomagnetic data from the 30 sites yield two distinct populations of group means. Most chill zone rocks yield a south declination and shallow positive and negative inclination mean CD =178.5°, 7-0.9°, a95 = 10.0°, A: = 24.1, N = 11). Most inner zone rocks yield a N-NW declination and moderate positive inclination mean (£ = 351.1°, 7 = 57.8°, a95 = 3.7°, k = 82.2, N = 19), which we interpret to be of normal polarity. Two of these sites yield a south-directed and moderate negative mean,

essentially antipodal from the normal polarity results. The mean direction from inner zone rocks is statistically distinguishable from a Mid-Tertiary normal polarity expected direction (D -358°, 7 = 58°; D = 178°, 7 = -58°) for west-central Nevada (based on the palaeomagnetic pole of Mankinen et al 1987). The inner zone rocks may reflect a very small magnitude (<10°) anticlockwise rotation of the Stoddard Mountain laccolith, which is consistent with similar age rocks in the area (Hudson et al. 1998). Most sites in chill zone rocks, however, yield an unusual magnetization direction (south-directed and very shallow inclination) that is more than 90° from an expected Late Miocene normal polarity field and over 45° from an expected reverse polarity field. This discordant magnetization cannot be readily explained by

Fig. 15. Summary of accepted AMS data from the Stoddard Laccolith, (a) magnetic lineation (solid arrows), (b) magnetic foliation (solid line with triangle). Rejected stations are indicated by black circle. See text for rejection criteria.

276

M. S. PETRONIS ET AL.

some magnitude of tilting, or a complex combination of tilting and rotation, in particular because the data are from sites distributed around the intrusion. The near uniform directions from sites around the laccolith margin argue against a physical process, such as differential tilting, as a viable explanation of the discordant magnetization. Assuming that the magnetization characteristic of the chill zone rocks is a faithful record of the field, a plausible explanation is that these chill zone rocks record some form of short-lived transitional field direction. The rocks were emplaced at roughly 1 km depth and the chilled zone would have crystallized very rapidly relative to secular variation and lock-in a remanence direction quickly. If reasonable, then this explanation may imply that the chill zone rocks were emplaced early as a sill, which was later intruded by the main body of the intrusion, in a pre-magnetization acquisition state, as the laccolith inflated to its full extent. In this case, the chill zone rocks have been fractured by the main body of the intrusion and have been tilted, prior to magnetization acquisition, about a horizontal axis by a magnitude equal to the adjacent steeply dipping country rock. Subsequently, the chill zone rocks, which presumably would have cooled

earlier than the inner zone rocks, acquired their unusual field direction. Additional investigation is needed to evaluate this possibility. Notably, inner zone rocks cooled over a sufficiently long period to time to capture both magnetic field polarities.

Anisotropy of magnetic susceptibility The high magnetic susceptibility (average 11 x 10 SI) indicates that a ferromagnetic phase dominates the susceptibility in the Stoddard Mountain intrusion. In comparison, for example, Pyrenean granites yield susceptibilities two orders of magnitude less than the Stoddard Mountain intrusion (Gleizes et al. 1993). Because magnetite is the main carrier of the AMS, the fabric is defined by both the crystalline anisotropy and shape preferred orientation of these particles dispersed in the rock matrix (Stacey 1960; Uyeda et al. 1963). The directional regularity of the magnetic fabric, both within individual stations and between stations over several kilometres of the laccolith, suggests that an efficient mechanism operated to control the AMS. In reflected light studies, magnetite is slightly inequant in shape and associated with

Fig. 16. Lower hemisphere equal area projections of maximum susceptibility axes (K}) from stations in the north zone (black squares) and south zone (black circles). Solid symbols indicate station mean Kj axes; square with X, mean K} direction of northern zone; open circle with dot, mean Kj direction of southern zone.

MAGMATIC FLOW PATHS AND PALAEOMAGNETISM

biotite and amphibole, either as inclusions or along grain boundaries. The magnetite grains, when in contact with the biotite or amphibole, tend to have their long axes parallel to the crystallographic axes of these minerals (Fig. 3). It seems reasonable to propose that if biotite and amphibole grains have a shape preferred orientation (SPO) in the intrusion, then the alignment of the long axes of magnetite grains will control the magnetic fabric. This kind of 'mimetic' behaviour, where the orientation of magnetite is controlled by the SPO of biotite and amphiboles, has been recognized as an important mechanism for producing consistent AMS fabrics over large areas in other ferromagnetic intrusions (Hrouda et al. 1971; Archanjo et al. 1992; Launeau & Bouchez 1992; Cruden & Launeau 1994). It is likely that the AMS fabric in the Stoddard Mountain laccolith is controlled by ferromagnetic mineral phases in a preferred crystallographic orientation dictated by the paramagnetic mineral phases in the rock. Inner zone The AMS fabrics, which we infer to correlate with magmatic fabrics, are interpreted to reflect two principal flow paths (N-NE and roughly south) (Fig. 16). AMS lineations in the northern zone of the laccolith trend N-NE and plunge gently (mean flow direction = 044°/34°, N = 20) and in the southern zone lineations trend south and plunge gently (mean flow direction = 1390/ 30°, N = 27). Chill zone The AMS fabrics reflect a principal flow path defined by a well-developed foliation parallel to margin of the intrusion, except in the NW region, where the foliation is perpendicular to the margin of the intrusion and lineations trend east-west. We hypothesize that the significantly different AMS fabric in the chill zone, relative to the inner zone reflects an early, pre-inflation flow pattern that was controlled by the margin of the intrusion. Emplacement mechanisms Studies of several Late Oligocene to Early Miocene Colorado Plateau laccoliths suggest that magma emplacement occurred from below by a central axial feeder pipe or dyke(s) from which they grew both laterally and vertically (Corry

277

1988; Jackson & Pollard 1988; Ross 1998). In contrast, geological relations in the area surrounding the Stoddard Mountain laccolith suggest that this intrusion may have been fed laterally from near the Pinto Peak area west of Stoddard Mountain. We hypothesize that the magma flowed eastward beneath Richie Flat and around a southern barrier of Tertiary volcanic rocks and then spread laterally south to north and inflated into a laccolith (Fig. 18). In this model, inferred magma flow paths expected for the development of a laterally fed laccolith whose source area is to the west are considerably different from those for an axial fed laccolith (Fig. 17). In the axial feeder system (Fig. 17a), flow paths in cross-section vary from vertical near the central source to sub-horizontal near the base. In map view, a symmetric, near radial flow path would be expected as the pluton spread laterally in all directions as a sill. In contrast, for the lateral feeder system (Fig. 17b), magma flow paths would tend to be only slightly or moderately inclined given the high length to thickness ratio of the laccolith. In map view, the magma flow path would be asymmetric, tending to radiate outwards from a point source in dominantly one direction (i.e. eastward in the case of Stoddard Mountain laccolith). In map view, the expected flow directions are markedly different for the western side of the laccolith. The magnetic anisotropy axes and thus magnetic fabrics determined here are interpreted to represent indicators of the principal axes of strain undergone by the Stoddard magma at the end of its emplacement. The fabric that is preserved likely represents the latest stage of magma emplacement prior to solidification. In map view, the orientations of AMS axes in the Stoddard intrusion are spatially variable across the short, east-west axis of the intrusion (Fig. 18). In the north part of the laccolith, AMS lineations trend north to NE (mean flow direction trend/plunge of about 044°/34°, TV = 20) and in the south part of the laccolith, lineations trend S-SE (trend/plunge of about 139°/30°, N = 27). The dominance of NE, east, and SE trending directions in the north, east, and south regions, respectively, is best explained by the lateral intrusion model (Fig. 17b). The plunge of K! is generally sub-horizontal to moderate. The main body of the laccolith lacks any vertically oriented magnetic lineations, which we infer to indicate the absence of a source area feeding the pluton from below. In the chill zone regions of the pluton margin, magnetic foliations are everywhere parallel to the contact except along the inferred source area in the NW part of the intrusion (Fig. 15).

Fig. 17. Idealized schematic models for laccolith formation, (a) Magma flow paths for a laccolith fed by an axial feeder dyke system; (b) Magma flow paths for a laccolith fed by a lateral feeder system. Arrows delineate expected magma flow paths in (1) cross-section and (2) map view.

MAGMATIC FLOW PATHS AND PALAEOMAGNETISM

279

Fig. 18. Interpretative diagram depicting our inferred magma flow model that led to the formation of the Stoddard Mountain laccolith, based on the available AMS data and geological field relations. Arrows indicate hypothesized magma flow direction.

Here, the magnetic foliation is approximately east-west, perpendicular to the contact and likely parallel to the magma flow direction. The observed variation in trend of magnetic fabrics and the overall shallow plunge of magnetic lineations is most consistent with the lateral feeder hypothesis for the emplacement and growth of the Stoddard Mountain laccolith (Figs 17b, 18). Conclusions Based on palaeomagnetic and AMS data reported here we reach the following conclusions bearing on the emplacement history of the Stoddard Mountain laccolith. The palaeomagnetic data from the inner zone rocks reveal that the main mass of the intrusion has not been affected by any appreciable deformation, either local tilting or vertical axis rotation, since emplacement. Based on field relations and thin section observations, we suggest that the magmatic flow fabrics are predominately controlled by a preferred orientation of paramagnetic mineral phases. Superimposed on the paramagnetic

fabric is one attributable to fine-grained, MD to PSD magnetite grains, which control the AMS directions, contribute substantially to the overall high magnetic susceptibility (greater than 10~3SI), and carry the remanent magnetization. The Stoddard Mountain laccolith displays well-organized magnetic fabrics that reflect magma flow during emplacement (Fig. 18). Our rock magnetic data are interpreted to indicate the dominance of low-titanium coarse-to intermediate grained magnetite as the highsusceptibility fraction within the intrusion. MD magnetite grains control the low-field magnetic fabric of the intrusion. The MD nature of magnetite, with its tendency to statistically align its magnetic moments parallel to its grain elongation suggests a correlation between the shape of the grains and the magmatic fabric. The similarity of the magnetic and magmatic fabrics is a consequence of the fact that magnetite is likely mimicking the orientation of the ferromagnesian phases in the rock. The magnetic anisotropy data correspond well with predicted magma flow directions associated with a laterally fed shallowly-emplaced laccolith.

280

M. S. PETRONIS ET AL.

The overall absence of NW and SW-trending, and steeply plunging magnetic lineations argues against emplacement via a central axial feeder system. Further AMS studies in the Iron Axis laccoliths may prove valuable for resolving conflicting data on magma emplacement models (i.e. Three Peaks area: Van Kooten 1988) and for better understanding the relationship between laccolith emplacement and ore genesis. We thank the US National Forest Service for providing lodging during our stays in the Pine Valley Mountains, Forest Service Volunteers W. Swenson and J. Seals, mules Stumpy and Durango, and horses Duke and Gus in assisting in sample collecting in the rugged areas of Stoddard Mountain. We also thank M. Lister and E. Roth for assistance with sample preparation. A special thank you goes to P. D. Rowley for supplying unpublished geological data on the Stoddard Mountain area. The manuscript was greatly improved by the constructive comments provided by the two reviewers: K. Benn and P. Launeau. This research was supported by grants from Sigma XI and the Geological Society of America (Petronis) and the Kent State University Research Council (Holm and Hacker).

References ADAIR, D. H. 1986. Structural setting of the Goldstrike district, Washington County, Utah, In: GRIFFIN, D. T. & PHILLIPS, W. R. (eds), Thrusting and extensional structures and mineralization in the Beaver Dam Mountains, southwestern Utah. Utah Geological Association Publication 15, 129135. ARCHANJO, C, OLIVIER, P. & BOUCHEZ, J. 1992. Granite plutons of the Serido Belt (northeastern Brazil) - belt-parallel magmatic flow as revealed by their magnetic-anisotropy. Bulletin De La Societe Geologique De France, 163(4), 509-520. ARCHANJO, C., LAUNEAU, P. & BOUCHEZ, J. 1995. Magnetic fabric vs magnetite and biotite shape fabrics of the magnetite-bearing granite pluton of Gameleiras (northeast Brazil). Physics of the Earth and Planetary Interiors, 89(1-2), 63-75. ARMSTRONG, R. L. 1970. Geochronology of Tertiary igneous rocks, eastern Basin and Range province, western Utah, eastern Nevada, and vicinity, USA. Geochimica et Cosmochimica Acta, 34, 203-232. BALSEY, J. R. & BUDDINGTON, A. F. 1960. Magnetic susceptibility anisotropy and fabric of some Adirondack granites and orthogniesses, American Journal of Science, 285A, 6-20. BARKER, D. S. 1995. Crystallization and alteration of quartz monzonite, Iron Springs Mining District, Utah: Relations to associated iron deposits, Economic Geology, 90, 2197-2217. BENN, K., ROCHETTE, P., BOUCHEZ, J. & HATTORI, K. 1993. Magnetic-Susceptibility, Magnetic Mineralogy And Magnetic Fabrics In: A Late Archean Granitoid Gneiss Belt. Precambrian Research, 63(1-2), 59-81.

BEST, M. G., CHRISTIANSEN, E. H. & BLANK, R. H. Jr. 1989. Oligocene caldera complex and calc-alkaline tuffs and lavas of the Indian Peak volcanic field, Nevada and Utah. Geological Society of America Bulletin, 101, 1076-1090. BLANK, R. H. & MACKIN, J. H. 1967. Geologic interpretation of an aeromagnetic survey of the Iron Springs district, Utah. US Geological Survey Professional Paper 516-B. BLANK, R. H., ROWLEY, D. & HACKER, D. B. 1992. Miocene monzonitic intrusions and associated megabreccias of the Iron Axis region, southwestern Utah, In: WILSON, J. R. (ed.) Field guide to geologic excursions in Utah and adjacent areas of Nevada, Idaho, and Wyoming, Rocky Mountain section, Geological Society of America Guidebook. Utah Geological Survey Miscellaneous Publication 92-3, 399-420. BORRADAILE, G. J. 1988. Magnetic susceptibility, petrofabrics and strain. Tectonophysics, 156, 1-20. BOUCHEZ, J. L. 1997. Granite is never isotropic: an introduction to AMS studies of granitic rocks. In: BOUCHEZ, J. L., HUTTON, D. H. W. & STEPHENS, W. E. (eds) Granite: From Segregation of Melt to Emplacement Fabrics. Kluwer, Dordrecht, 96-112. BOUCHEZ, J. L. 2000. Anistropie de susceptibilite magnetique et fabrique des granites. Comptes Rendues de I'Academie des Sciences, 330, 1-14. BOUCHEZ, J. L., GLEIZES, G., DJOUADI, M. T. & ROCHETTE, P. 1990. Microstructures and magnetic susceptibility applied to emplacement kinematics of granites: the example of the Fox pluton (French Pyrenees), Tectonophysics, 184, 157-171. BOUILLIN, J. P., BOUCHEZ, J. L., LESPINASSE, P. & PECHER, A. 1993. Granite emplacement in an extensional setting: an AMS study of the magmatic structures of Monte Capanne (Elba, Italy). Earth and Planetary Science Letters, 118, 263-279. CLEMENS, J. D., PETFORD, N. & MAWER, C. K. 1997. Ascent mechanisms of granitic magmas: causes and consequences. In: HOLNES, M. (ed.) Deformation-enhanced Fluid Transport in the Earth's Crust and Mantle. Chapman & Hall, London, 144-171. COOK, E. F. 1957. Geology of the Pine Valley Mountains. Utah. Utah Geological and Mineralogical Survey Bulletin 58. CORRY, C. E. 1988. Laccoliths: Mechanics of Emplacement and Growth. Geological Society of America, Special Paper, 220. CRUDEN, A. R. 1990. Flow and fabric development during the diapiric rise of magma. Journal of Geology, 98, 681-698. CRUDEN, A. R. & LAUNEAU, P. 1994. Structure, magnetic fabric and emplacment of the Archean Lebel Stock, SW Abitibi Greenstone belt, Journal of Structural Geology, 16, 677-692. DE SAINT BLANQUAT, M. & TIKOFF, B. 1997. Development of magmatic to solid-state fabrics during syntectonic emplacemnt of the Mono Creek granite, Sierra Nevada Batholith. In: BOUCHEZ, J. L. (ed.) Granite: From Segregation of Melt to Emplacement Fabrics. Kluwer, Dordrecht.

MAGMATIC FLOW PATHS AND PALAEOMAGNETISM DUNLOP, D. J. & OZDEMIR, O. 1997. Rock Magnetism: Fundamentals and Frontiers, Cambridge Studies in Magnetism, 3, Cambridge University Press, New York. ELLWOOD, B. B. & WHITNEY, J. A. 1980. Magnetic fabric of the Elberton Granite, Northeast Georgia. Journal of Geophysical Research, 85, 1481-1486. GILBERT, G. K., 1877. Geology of the Henry Mountains, Utah. US Geographical and Geological Survey of the Rocky Mountain Region. GLEIZES, G., NEDELEC, A., BOUCHEZ, J. L., AUTRAN, A. & ROCHETTE, P. 1993. Magnetic susceptibility of the Mont-Louis Andorra ilmenite-type granite (Pyrenees): a new toll for the petrographic characterization and regional mapping of zoned granite plutons. Journal of Geophysical Research, 98, 4317-4331. GREGOIRE, V., BLANQUAT, M., NEDELEC, A. & BOUCHEZ, J. 1995. Shape anisotropy versus magneticinteractions of magnetite grains-experiments and application to AMS in granitic-rocks. Geophysical Research Letters, 22(20), 2765-2768. GUILLET, P., BOUCHEZ, J. & WAGNER, J. 1983. Anisotropy of magnetic-susceptibility and magmatic structures in the Guerande Granite Massif (France). Tectonics, 2(5), 419-429. HACKER, D. B. 1995. Deformational structures related to the emplacement and growth of the Pine Valley laccolith, southern Iron Axis region, Washington County, Utah. Eos, Transactions, American Geophysical Union, 76, 625. HACKER, D. B. 1998. Catastrophic gravity sliding and volcanism associated with the growth of laccoliths - examples from Early Miocene hyp abyssal intrusions of the Iron Axis magmatic province, Pine Valley Mountains, southwest Utah. Kent, Ohio, Kent State University, Ph.D. dissertation. HACKER, D. B., ROWLEY, P. D., BLANK, H. R. & SNEE, L. W. 1996. Early Miocene catastrophic gravity sliding and volcanism associated with intrusions of the southern Iron Axis region, southwest Utah. Geological Society of America Abstracts with Programs, 29, 511. HACKER, D. B., HOLM, D. K., ROWLEY, P. D. & BLANK, H. R. 2002. Associated Miocene laccoliths, gravity slides, and volcanic rocks, Pine Valley Mountains and Iron Axis region, southwestern Utah, In: LUND, W. R. (ed), Field Guide to Geologic Excursions in Southwestern Utah and adjacent areas of Arizona and Nevada. US Geological Survey Open-File Report 02-172, 236-283. HELLER, F. 1973. Magnetic anisotropy of granite rocks of the Bergell massif (Switzerland), Earth and Planetary Science Letters, 20, 80-188. HORSMAN., E. & TIKOFF, B. 2002. Three-dimensional magmatic fabric of a sheeted laccolith, Henry Mountains, Utah. Geological Society of America Abstracts with Programs, 35, 271. HROUDA, F. 1982. Magnetic anisotropy of rocks and its applications in geology and geophysics. Geophysical Survey, 5, 38-82. HROUDA, F., JANAK, F., REJL, L. & WEISS, J., 1971. The mimic fabric of magnetite in some foliated granodiorites, as indicated by magnetic

281

anisotropy. Earth and Planetary Science Letters, 11, 381-384. HUDSON, M. R., ROSENBAUM, J. G., GROMME, C. S., SCOTT, R. B. & ROWLEY, P. D. 1998. Paleomagnetic evidence for counter clockwise rotation in a broad sinistral shear zone, Basin and Range province, southeastern Nevada and southwestern Utah. In: FAULDS, J. E. & STEWART, J. H. (eds) Accommodation Zones and Transfer Zones: The regional segmentation of the Basin and Range Province, Geological Society of America Special Paper 323, 149-180. HUTTON, D. H. W. 1988. Granite emplacement mechanisms and tectonic controls: inferences from deformation studies. Transactions of the Royal Society of Edinburgh: Earth Sciences, 79, 245-255. HUTTON, D. H. W. 1992. Granite sheeted complex: evidence for the dyking ascent mechanism. Transactions of the Royal Society of Edinburgh: Earth Sciences, 83, 377-382. JACKSON, M. D. & POLLARD, D. D. 1988. The laccolithstock controversy - New results from the southern Henry Mountains, Utah. Geological Society of America Bulletin, 100, 117-139. JELINEK, V. 1977. The statistical theory of measuring anisotropy of magnetic susceptibility of rocks and its applications. Brno Geofyzika, 1-88. JELINEK, V. 1978. Statistical processing of anisotropy of magnetic susceptibility measured on groups of specimens. Studies of Geophysics and Geodesy, 22, 50-62. JELINEK, V. 1981. Characterization of the magnetic fabric of rocks. Tectonophysics, 79, 63-70. JOHN, B. E. & BLUNDY, J. D. 1993. Thermal and mechanical aspects of granitoid emplacement in the southern Adamello massif, Italy. Geological Society of America Bulletin, 105, 1517-1541. JOHNSON, C. M. 1991. Large-scale crust formation and lithosphere modification beneath Middle to Late Cenozoic calderas and volcanic fields, western North America. Journal of Geophysical Research, 96(B8), 13485-13507. JOHNSON, H. P., LOWRIE, W. & KENT, D. V. 1975. Stability of anhysteretic remanent magnetization in fine and coarse magnetite and maghemite particles, geophysical journal of the royal astronomical society, 41, 1-10. KHAN, M. A. 1962. The anisotropy of magnetic susceptibility of some igneous and metamorphic rocks. Journal of Geophysical Research, 67, 28732885. KING, R. F. 1966. The magnetic fabric of some Irish granites. Geological Journal, 5, 43-66. KIRSHVINK, J. L. 1980. the least-squares line and plane and the analysis of palaeomagnetic data. Royal Astronomical Society Geophysical Journal, 62, 699-718. LAUNEAU, P. & BOUCHEZ, J. 1992. Mode and shape preferred orientation of granites by numerical image-analysis. Bulletin De La Societe Geologique De France, 163(6), 721-732. LOWRIE, W. 1990. Identification of ferromagnetic minerals in a rock by coercivity and unblocking

282

M. S. PETRONIS ET AL.

temperature properties. Geophysical Research ROCHETTE, P. 1987. Magnetic susceptibility of the rock Letters, 17, 159-162. matrix related to magnetic fabric studies, Journal MACKIN, J. H. 1960. Structural significance of Tertiary of Structural Geology, 9, 1015-1020. volcanic rocks in southwestern Utah. American ROCHETTE, P., JACKSON, M. & AUBOURG, C. 1992. Journal of Science, 258, 81-131. Rock magnetism and the interpretation of anisoMACKIN, J. H. 1968. Iron deposits of the Iron Springs tropy of magnetic susceptibility. Reviews of Geodistrict, southwestern Utah In: Ore deposits of physics, 30, 209-226. the United States 1933-1967. American Institute Ross, M. L. 1998. Geology of the Tertiary intrusive of Mining, Metallurgy and Petroleum Engineers, centers of the La Sal Mountains, Utah - influence 2, 992-1019. of pre-existing structural features on emplacement MACKIN, J. H. & INGERSON, E. 1960. An hypothesis for and morphology, In: FRIEDMAN, J. D. & HUFFMAN, the origin of ore-forming fluid, in Short papers in A. C. Jr. (eds) Laccolith complexes of southeastern the geological sciences. US Geological Survey Utah - Time of emplacement and tectonic setting Professional Paper 400-B, B1-B2. Workshop proceedings. US Geological Survey BulMANKINEN, E., LARSON, E., GROMME, C. S., PREVOT, letin 2lS&,6\-%3. M. & COE, R. S. 1987. The Steen mountain ROWLEY, P. D. 1998. Cenozoic transverse zones and (Oregon) geomagnetic polarity transition, 3. Its igneous belts in the Great Basin, western United regional significance. Journal Of Geophysical States: their tectonic and economic implications, Research, 92, 8057-8076. In: FAULDS, J. E. & STEWART, J. H. (eds) AccomMCKEE, E. H., BLANK, H. R. & ROWLEY, P. D. 1997. modation zones and transfer zones - The regional Potassium-Argon ages of Tertiary igneous rocks segmentation of the Basin and Range Province. in the eastern Bull Valley Mountains and Pine Geological Society of America Special Paper 323 Valley Mountains, southwestern Utah, In: MAL195-228. DONADO, F. & NEALEY, L. D. (eds) Geologic ROWLEY, P. D. & BARKER, D. S. 1978. Geology of the Studies in the Basin and Range-Colorado Plateau Iron Springs mining district, Utah. Geological transition in southeastern Nevada, southwestern Association of Utah Publication 1, 49-58. Utah, and northwestern Arizona, 1995. US Geolo- ROWLEY, P. D., NEALY, L. D., UNRUH, D. M., SNEE, L. gical Survey Bulletin 2153, 241-251. W., MEHNERT, H. H., ANDERSON, R. E. & MORRIS, S. K. 1980. Geology and ore deposits of GROMME, C. S. 1995. Stratigraphy of Miocene Mineral Mountain, Washington County, Utah. ash-flow tuffs in and near the Caliente caldera Brigham Young University Studies in Geology, 27, complex, southeastern Nevada and southwestern 85-102. Utah, In: SCOTT, R. B. & SWADLEY, W. C., (eds) MUTSCHLER, F. E., LARSON, E. E. & Ross, M. L. 1998. Geologic Studies in the Basin and RangePotential for alkaline igneous rock-related gold Colorado Plateau transition in southeastern deposits in the Colorado Plateau laccolithic Nevada, southwestern Utah, and northwestern centers. In: FRIEDMAN, J. D. & HUFFMAN, A. C. Arizona, 1992. US Geological Survey Bulletin JR. (eds) Laccolith complexes of southeastern 2056, 43-88. Utah - Time of emplacement and tectonic setting ROWLEY, P. D., CUNNINGHAM, C. G., STEVEN, T. A., - Workshop proceedings. US Geological Survey MEHNERT, H. H. & NAESER, C. W. 1998. Cenozoic Bulletin 2158, 233-252. igneous and tectonic setting of the Marysvale volNELSON, S. T. & DAVIDSON, J. P. 1998. The petrogenesis canic field, and its relation to other igneous centers of the Colorado Plateau laccoliths and their relain Utah and Nevada, In: FRIEDMAN, J. D. & HUFFtionship to regional magmatism. In: FRIEDMAN, MAN, A. C. Jr. (eds) Laccolith complexes of southJ. D. & HUFFMAN, A. C. Jr. (eds) Laccolith comeastern Utah - Time of emplacement and tectonic plexes of southeastern Utah - time of emplacesetting - Workshop proceedings. US Geological ment and tectonic setting - workshop Survey Bulletin 2158, 167-202. proceedings. US geological Survey Bulletin 2158, ROWLEY, P. D., WILLIAMS, V. S., ET AL., in press 2003, Geologic map of the Cedar City 30' x 60' quad85-100. O'REILLY, W 1984. Rock and Mineral Magnetization, rangle, Iron and Washington, Counties, Utah. Blackie & Son, Glasgow. Utah Geological Survey. PARK, K., TANCZYK, E. I. & DESBARATS, A. 1988. ROY J. L. & PARK, J. K. 1974. The magnetization Magma fabric and its significance in the 1400 process of certain red beds: Vector analysis of Ma Mealy diabase dykes of Labrador, Canada. chemical and thermal results, Canadian Journal Journal of Geophysical Research 93, 13689-13704. of Earth Sciences, 11, 437-471. PITCHER, W. S. 1993. The Nature and Origin of Granite. SPURNEY, J. C. 1984. Geology of the Iron Peak intrusion, Chapman & Hall, London. Iron County, Utah. Kent, OH, Kent State UniverPOTTER, D. & STEPHENSON, A. 1990. Field-impressed sity, M. S. thesis. magnetic-anisotropy In Rocks. Geophysical STACEY, F. D. 1960. Magnetic anisotropy of igneous Research Letters, 17(13), 2437-2440. rocks. Journal of Geophysical Research 65, 2429RICHTER, C. & VAN DER PLUUM, B. A. 1994. Separation 2442. of paramagnetic and ferromagnetic susceptibilities TARLING, D. H. & HROUDA, F. 1993. The Magnetic using low temperature magnetic susceptibilities Anisotropy of Rocks, Chapman & Hall, London and comparison with high field methods, Physics UYEDA, S., FULLER, M. D., BELSHE, J. C. & GIRDLER, of the Earth and Planetary Interiors, 82, 113-123. R. W. 1963. Anisotropy of magnetic susceptibility

MAGMATIC FLOW PATHS AND PALAEOMAGNETISM of rocks and minerals. Journal of Geophysical Research 68, 279-291. VAN KOOTEN, G. K. 1988. Structure and hydrocarbon potential beneath the Iron Springs laccolith, southwestern Utah. Geological Society of America Bulletin, 100, 1533-1540. VIGNERESSE, J. L. & CLEMENS, J. D. 2000. Granitic magma ascent and emplacement: neither diapirism nor neutral buoyancy. In: VENDEVILLE, B., MART,

283

Y. & VIGNERESSE, J. L. (eds) Salt, Shale and Igneous Diapirs in and around Europe. Geological Society, London, Special Publications, 174, 1-19. ZIJDERVELD, J. D. A. 1967. A. C. Demagnetization of rocks: Analysis of results, In COLLINSON, D. W., CREER, K. M. & RUNCORN, S. K. (eds), Methods In Palaeomagnetism. Elsevier, Amsterdam, 254286.

This page intentionally left blank

Magma flow inferred from preferred orientations of plagioclase of the Rio Ceara-Mirim dyke swarm (NE Brazil) and its AMS significance CARLOS J. ARCHANJO1 & PATRICK LAUNEAU2 1

Universidade de Sao Paulo jIGc, Rua do Lago 562, 05508-080 Sao Paulo, Brazil (e-mail: [email protected]) 2 Universite de Nantes, Laboratoire de Planetologie et Geodynamique, Nantes, France (e-mail: [email protected]) Abstract: Low-field magnetic and plagioclase fabrics were compared in Mesozoic mafic dykes of the Rio Ceara-Mirim swarm. Coarse titanomagnetites with pervasive ilmenite lamellae constitute the main carrier of the magnetic anisotropy. The hysteresis parameters of the mafic dykes fall in the pseudo-single domain field. The resulting AMS ellipsoid is usually oblate and has a very low anisotropy (<3%). Textures indicate that the oxi-exolution processes and size reduction of the ferrimagnetic domains occurred at subsolidus temperatures on cooling of the dykes. The magmatic fabric was determined by the shape preferred orientation of plagioclase laths. It rarely matches the magnetic fabric. Besides their contrasting shape ellipsoids, prolate and oblate respectively, their corresponding principal directions diverge from each other or exchange their positions depending on the symmetry of the ellipsoids. These discrepancies are attributed principally to small differences in the net shape of Ti-poor magnetite after exsolution of ilmenite and in the inherently oblique fabric of grains with different shapes. These results draw attention to the need to use independent methods to confirm the conclusions about flow fabrics of weakly anisotropic mafic dykes based only in AMS.

Structural studies in apparently isotropic mafic dykes have frequently used the anisotropy of low-field magnetic susceptibility (AMS) to infer flow directions and geodynamic settings of several dyke swarms (e.g. Ernst & Baragar; Raposo & Ernesto 1995; Tauxe et al. 1998; Abelson et al. 2001). AMS is the anisotropy of a physical property of a rock that depends on the preferred orientation of ferrimagnetic grains, usually titanomagnetite in mafic rocks (see Rochette et al. 1992 and Borradaile & Henry 1996 for recent reviews). In dykes, the maximum susceptibility direction (k\, where k\ > k2 > £3) has been equated to the mineral lineation, hence to the magma flow direction. When the minimum susceptibility direction (&3) is perpendicular to the dyke wall, AMS is considered to be of 'normal' type, the magnetic foliation being coplanar to the flow plane. Besides 'normal' AMS types, dykes may present 'intermediate' (&2 perpendicular to the dyke plane) and 'inverse' type (&3 perpendicular to the dyke plane) magnetic fabrics (Rochette et al. 1999). When interpreting the regional flow pattern in dykes the so-called 'abnormal' AMS fabrics are usually discarded, even though they may correspond to more than 50% of the studied samples. They are attributed to the influence of considerable amounts (>25%) of single domain ferrimagnetic particles, which have an intrinsic inverse magnetic anisotropy (Potter &

Stephenson 1988); variable proportion of interacting magnetic particles (textural anisotropy, Canon-Tapia 2001); or interactions during the crystallization of silicates and ferrimagnetic oxides in the mafic magma (Archanjo et al. 2002). After removing the intermediate and inverse fabrics from the data set, the remaining 'normal' fabrics are then used to infer flow dynamics of the magma within the swarm, to detect its feeder centres and to explore its tectonic significance in volcanic provinces (Glen et al. 1997; Callot et al. 2001). However, such an 'ad hoc' criterion of magnetic fabric selection has its drawbacks. Recent studies based on image analysis have found large angular departure between the magnetic lineation (k\) and the plagioclase linear fabric (Archanjo et al. 2002; Geoffrey et al. 2002). These findings led Geoffrey et al. (2002) to discard the use of k\ as a kinematic indicator on Tertiary dykes from Greenland. Here, we examine more deeply such a relation by using a three-dimensional approach that consists in integrating the petrofabric from three mutually perpendicular sections cut from cores used in the AMS study of the Rio CearaMirim swarm. The results confirm the strong differences between rock and magnetic fabrics revealed in previous studies, and highlight the need for an independent check of magnetic fabric results when interpreting the signal from weakly anisotropic mafic dykes.

From: MARTIN-HERNANDEZ, F., LUNEBURG, C. M., AUBOURG, C. & JACKSON, M. (eds) 2004. Magnetic Fabric: Methods and Applications. Geological Society, London, Special Publications, 238, 285-298. 0305-8719/04/ $15.00 © The Geological Society of London 2004.

286

C. J. ARCHANJO & P. LAUNEAU

Geological setting and regional AMS pattern The Rio-Ceara Mirim mafic dykes form a nearly 400 km long, E-trending swarm that intruded the Precambrian basement of northeastern Brazil between 140 Ma and 120 Ma (Mizusaki et al. 2002) (Fig. 1). In a pre-drift reconstruction, the swarm lies opposite to the Benoue Trough (Nigeria) where two Mesozoic igneous activities, at 147-106 Ma and 97-81 Ma, have been recognized (Maluski et al, 1995). In NE Brazil the dykes are tholeiitic in composition, being extracted from a lithophile enriched lithospheric source (Bellieni et al. 1992). They correspond to the initial magmatic pulses before the opening of the Equatorial Atlantic in the Late Cretaceous. The main swarm was divided into four subswarms (I to IV) according to their different magnetization directions and location (Fig. 1; Bellieni et al 1992; Ernesto et al 2003). Individual dykes occur as vertical bodies of 1 m up to 150m in width. Dykes from subswarms I and II have mean lengths of 4 km, while those from subswarms III and IV have lengths of around 3 km. Dykes from the western part tend to be thicker, between 60 and 150m in width,

while in the eastern part they have mean widths around 20m. A few dykes form an en echelon pattern, in particular some dyke sets from the central part, indicating that the emplacement occurred at a high crustal level (Oliveira 1992). Otherwise, they form linear, anastomosed to parallel sets running close to the southern border of the Mesozoic Potiguar Basin. An AMS study along the whole swarm revealed a large proportion of abnormal magnetic fabric types (Archanjo et al 2002). From 59 AMS sites, 29 dykes showed inverse to highly oblique to intermediate magnetic fabrics, while 30 dykes presented normal fabrics. The normal fabrics have been called type-A, while the inverse to highly oblique have been called type-B magnetic fabric. The bulk magnetic susceptibility for all the studied dykes clusters in the 40-60 x 10"3 SI range (average — 56.1 x 10~3 SI or mSI). These high susceptibility values are due to the high content in magnetic oxides, which varies from 2.6% to 6% by volume of the mafic dykes. Very weak mean anisotropies (P) are found in dykes with both type-A (P= 1.024, a0.014) and type-B (P = 1.019, crO.017) fabrics. Regionally magnetic lineation tends to be sub-horizontal, except for a zone of

Fig. 1. Geological map of the mafic rocks of northeastern Brazil, with location of the Rio Ceara-Mirim dyke swarm. Subswarms I-IV are labelled after Bellieni et al (1992). Along the swarm magnetic lineations in the western, central and eastern domains plunge gently (K\> mean lineation). A zone of steep lineations (dashed line) separates the central from the eastern domain. Schmidt diagrams of !%/!% area (central and western domain) and of 2%/l% area for the steep zone (SZ) and eastern domain.

MAGMA FLOW DIRECTION FROM PLAGIOCLASE ORIENTATION

287

Fig. 2. (a) Magnetic lineation of the 'normal' (type-A) fabric and the model for the acquisition of the flow line structure (b). The lineation trajectory would form a fan-like pattern converging to the magmatic feed zone (after Archanjo et al. 2000); (c) Magnetic lineation and foliation pole of inverse (type-B) fabric. K\ and K^ are, respectively, the mean lineation and foliation pole (azimuth, plunge). Schmidt diagrams: contours of 2%/l% area.

steep plunging lineations which separate the eastern from the central domain. This peculiar zone is informally called 'Steep Zone' (SZ). By isolating the normal AMS configuration (typeA) from the bulk fabric, a relatively simple fanlike lineation trajectory appears along the swarm (Fig. 2a). The lineations plunge gently (eastern and western domains) to moderately (central domain) to converge into the SZ. This latter region was interpreted the feeder zone of the dykes. The fan-like flow line pattern would be produced as the magma moves away from its source and spread laterally (Fig. 2b). Type-B magnetic fabrics occur in 17 dykes distributed mainly on the central and western domains. They show sub-horizontal lineations perpendicular, or at high angle, to the dyke wall and gently dipping foliations (Fig. 2c). They were interpreted as unrelated to flow, being attributed to competition of interstitial magnetite crystallizing on cooling of the dykes. Magnetic mineralogy and oxide texture The dykes exhibit a range of textures from subophitic to seriate to intergranular, with 0.42.5mm long plagioclase laths, partially enclosed within large, 0.3-2.Omm, equant clinopyroxene grains. Opaque grains usually occupy the spaces between plagioclase and clinopyroxene.

Cryptocrystalline material is frequently observed in variable amounts in thin section, and seems to be associated with the chemical alteration of the silicates. SEM examination of the fine-grained material indicates the presence of carbonate and minute magnetite particles, the latter also included within crystals of plagioclase and clinopyroxene (Archanjo et al 2002). Large titanomagnetite grains are quite abundant and may reach up to 2mm in diameter (Fig. 3). Ilmenite is comparatively scarce and individual crystals normally occur as subhedral to anhedral grains or as small elongated laths. Ilmenite intergrowths in titanomagnetite form trellis to sandwich type microstructures, with ilmenite occupying the {111} planes of the host titanomagnetite (Fig. 3). The width of ilmenite lamellae tends to be fairly uniform, but large trellis (10-20 jam) occur locally. The conspicuous exsolution lamellae within titanomagnetite grains is attributed to a high-temperature oxidation during the dyke's cooling, unmixing primary titanomagnetite crystals into composite grains with Ti-poor and Ti-rich domains. Rock magnetic studies indicate that Ti-poor magnetite is the main carrier of magnetization and susceptibility of the Rio Ceara-Mirim dykes (Bucker et al. 1986; Bellieni et al. 1992; Archanjo et al. 2000). Thermomagnetic curves revealed that the susceptibility of crushed 0.4cm3 basaltic samples falls abruptly in the

288

C. J. ARCHANJO & P. LAUNEAU

Fig. 3. Oxide microstmctures of dykes. Euhedral to subhedral titanomagnetite showing conspicuous, fine ilmenite trellis occupying octahedral sites. Coarser ilmenites occur: (a) as subhedral to anhedral grains or as elongated laths in contact with titanomagnetite and, (b) as subhedral grains with embayed contacts with the silicate matrix, locally preserving daktilitic textures.

range of 550-570 °C, decreasing to near zero above 580 °C. Evidence of low temperature oxidation and hydrothermal alteration leading to maghemite formation was also found in several dyke samples. Hysteresis data showed narrowwaisted loops with magnetization saturating at applied fields lower than 300 mT and coercivity (He) ranging from 5.5 to 21.6mT (mean He = 12.2, a4.1). The ratio of saturation remanence (Mrs) to saturation magnetization (Ms) shows a more or less linear proportionality with He (Fig. 4a) that fits in with hysteresis parameters of magnetite precipitated in glass ceramic samples with sizes between 2um to 0.5um (Worm & Markert 1987). The hysteresis ratios (Her: He, Mrs: Ms) are characteristic of a pseudo-single domain (PSD) behaviour for all samples (Fig. 4b), with a trend towards the multidomain (MD) field of titanomagnetites in the Day et al. (1977) plot, which may reflect a mixture of fine-grained single-domain (SD) magnetite and MD grains (Dunlop 2002). Alternatively, the ubiquitous exsolution in the magnetite grains, reducing the effective grain size of the ferrimagnetic domains, may be responsible for the PSD behaviour of the magnetic grains (e.g. Hodych 1996).

in this study. When imaged by a matrix of pixels (picture elements), a crystal section is a set of adjacent numbers coding a mineral phase. Various techniques are then available for identifying and classifying each set of numbers as a particular phase (Launeau et al. 1994). A major improvement of the technique was given by the development of computerized stages (Fueten & Goodchild 2001) allowing the segmentation of crystals with polarized light without light extinction. The final image is made of pixel sets having a specific code for each mineral phase and crystal tag number. Each of them can then be analysed independently. If x and y are the coordinates of each pixel of a crystal j in the image xy, its twodimensional gravity centre is given by the mean x and y coordinates called xc and yc. The analysis of the square distance of each pixel i to that gravity centre, can be used to build the following inertia tensor of the jih crystal:

where

Plagioclase fabric Determination of shape preferred orientation (spo) by image analysis Launeau & Cruden (1998) and Launeau (2004) give a detailed account of the methods employed

X; and yt are coordinates of the /th pixel, and A is the number of pixels. The exact size of all those data may be reached by multiplying each pixel

MAGMA FLOW DIRECTION FROM PLAGIOCLASE ORIENTATION

289

Fig. 4. (a) Ratio of saturation remanence (Mrs) to saturation magnetization (Ms) versus coercive force (He) for dykes of the Rio Ceara-Mirim swarm and grain size estimates (arrows) for some non-interacting magnetites crystallized in glass ceramic samples (Worm & Markert 1987). The solid line (1) represents the trend for mid-ocean ridge TM60 titanomagnetite (Wang & Van der Voo 2004). (b) Hysteresis ratios (Mrs/Ms and Her/He) and respective fields of single (SD), pseudosingle (PSD) and multidomain (MD) after Day et al. (1977).

edge by their corresponding size dx and dy. The greatest eigenvector of Mj gives the long axis direction 0y. The corresponding eigenvalue \{ is a function of the long length a of the inertia tensor ellipse and the second eigenvalue A2 is a function of the short length b of the same ellipse (Rink 1976).

with

and with two other mutually perpendicular sections xz and yz, giving coefficients bxx, bxz, bzz and byy, byz, bzz respectively, one can write (Launeau & Cruden 1998):

The two-dimensional shape preferred orientation of a population of TV crystals is calculated by averaging the inertia tensor as follows: with As for one crystal j with M7, the mean shape parameters , a and b are given by the corresponding eigenvector and eigenvalues of M. As proposed by Shimamoto & Ikeda (1976), three mutually orthogonal, two-dimensional quadratic shape tensors can be combined to calculate one three-dimensional quadratic shape tensor. When sizes are available, the quadratic shape tensor of a section xy can be written:

The angles a, 0 and 9 are the Euler angles of the azimuth, strike and dip of a three-dimensional ellipsoid with diameter lengths A, B, C with A > B > C. Axis [A] is the shape preferred orien-

290

C. J. ARCHANJO & P. LAUNEAU

tation lineation and the plan (A, B) is its foliation. When the consistency between sizes in perpendicular sections is not preserved, as when the modal fraction varies, or when the sections are not perpendicular to each other, the Robin (2002) method is used to calculate an ellipsoid from any set of sections (minimum 3). The consistency of the sections with the ellipsoid is measured with an 'incompatibility index' yF. When the sections are not sufficiently scattered, the method does not converge to an ellipsoid but to an open hyperboloid. In the last case the data are discarded. When the data fit to an ellipsoid its shape ratio (SR = A/Q, symmetry (T > 0, oblate; T < 0 prolate; T = 0, neutral with T = [2(\nB - In C)/(\nA - In C)] - 1) and orientation of its principal axes [A], [B] and [C] are then compared to the equivalent AMS parameters.

Sampling and analytical methods To check the relation between the magnetic lineation and the magmatic fabric of the dykes, the shape preferred orientation of plagioclase

was studied in thin sections. The samples with type-A AMS configuration come from the SZ (cm 14 and cm28) and Central (cm34 and cm45) domains. The two samples with type-B (cm30 and cm41) also come from the central domain. All samples used for image analysis but samples from dyke cm28 were collected more than a metre away from the margin of 30m to 50m thick dykes. The dyke cm28 has well-exposed contacts with its host rocks (Fig. 5a). It was used to test the application of imbricated magnetic fabrics against chilled margins to fix the kinematics of the magma flow (Knight & Walker 1988). Samples from its chilled margins (cm28M (n, northern border) and cm28N (s, southern border)) and from its centre (cm28K (c)) were analysed (Figs 5b, c). Thin sections were prepared from slabs cut parallel to the xy, xz and yz frame of the cylindrical specimens used for the AMS determinations (Fig. 6a-c). From each thin section, two consecutive high-resolution images were digitized through a rotating polarizer stage fixed to a microscope (Fueten 1997). The resulting images allow, by 23 combinations, the statistical deter-

Fig. 5. (a) Mafic dyke trending E-W (North = left) with fresh, unweathered contacts with its host gneissic rock. The cores to AMS and mineral preferred orientation studies (cm28) come from the margins and from the middle of the dyke. The fine plagioclase laths of the chilled margins display a better alignment on horizontal (b) than on vertical plane (c), in agreement with the shape ratio (sr) determined by the inertia tensor method. n = number of laths. Base of photos in (b) and (c) = 2 mm.

291

MAGMA FLOW DIRECTION FROM PLAGIOCLASE ORIENTATION

Fig. 6. Analytical procedures for image analysis study of thin sections from specimens used for AMS measurements, (a) Geographically oriented cylindrical core; (b) Typical specimens used for AMS determination and respective reference frame axes (.x, y and z); (c) Mutually orthogonal sections cut parallel to the reference frame; (d) Grains of plagioclase of digitized images (2.2 x 2.5cm) from mutually orthogonal thin sections and respective rose of directions (shaded) and sectional ellipse describing the plagioclase distribution (n, number of grains; sr, ellipse shape ratio).

confidence ellipses (Icr) around the mean principal directions is lower than 28° and when the incompatibility index is lower than 15%. Otherwise, the fabric is poorly defined. Finally, the inertia tensor ellipsoid describing the plagioclase fabric is rotated to the geographic coordinates and then compared to the AMS.

mination of the three-dimensional inertia tensor ellipsoid of each specimen. The plagioclase is easily separated from clinopyroxene and opaque minerals by its lower relief and distinctive twin planes. Once isolated, the shape preferred orientation is determined with the inertia tensor calculation and the rose diagrams of long axis directions are presented (Fig. 6d) as a check of crystal orientation distribution homogeneity in each section, xy, xz and yz. The sectional ellipses (2-D) are then combined to find the best-fit ellipsoid (3-D) using the program Ellipsoid.exe without scale factor (Robin 2002; Launeau 2004). The fabric is considered well-defined when the semi-angle of the

Results The AMS data of the specimens studied for image analysis are shown in Table 1 and the corresponding spo parameters are summarized in Table 2. The mean plagioclase grain size varies

Table 1. AMS parameters of the specimens used for determining the plagioclase fabric Specimen

km (1(T3 SI)

P

T

kl

k3

normal

cm!4C cm34D cm28M (n margin) cm28K (centre) cm28N (s margin) cm45E

71.0 43.1 24.2 33.4 24.6 45.7

1.06 1.03 1.05 1.02 1.05 1.02

0.15 0.67 0.32 0.60 0.65 0.35

178,79(6.4) 264, 5(1.7) 182,64(1.5) 058, 39 (3.9) 051,62(4.8) 278, 20 (5.2)

057, 5 (5.1) 173,8(1.7) 041,21 (2.0) 150,2(1.4) 148, 4 (4.4) 175, 33 (4.8)

inverse

cmBOC cm41A

47.9 45.6

1.03 1.01

-0.06 0.60

001, 10(2.8) 170, 9 (7.5)

107, 58 (4.9) 055, 69 (6.0)

km, magnetic susceptibility; P, anisotropy degree (= kl/k3); T, shape parameter; kl and k3, lineation and magnetic foliation pole, respectively. In parentheses the semi-angle (measured in degrees) of the l
292

C. J. ARCHANJO & P. LAUNEAU

Table 2. Plagioclase fabric parameters of the Rio Ceard-Mirim dykes Specimen

n

(p (mm)

SRplag

T

vF

A

C .

normal

cm!4C cm34D cm28M(n) cm28K(c) cm28N (s) cm45E

266 (35.2%) 372 (32.2%) 768 (26.7%) 518 (22.2%) 465(15.3%) 228 (37.5%)

0.50 0.48 0.03 0.12 0.03 0.47

1.26 1.17 1.42 1.16 1.78 1.21

-0.10 -0.23 0.19 -0.23 -0.47 -0.02

3.6 6.4 2.8 5.3 4.6 9.9

090, 50 (9.7) 266, 5 (17.2) 318, 22 (4.4) 054, 37 (13.4) 066, 12 (2.5) 062, 60 (25.2)

356, 3 (4.4) 010, 74 (23.5) 049, 3 (3.2) 152,20(6.7) 332, 19 (6.0) 252, 27 (25.5)

inverse

cm30C cm41A

566 (27.8%) 588 (34.5%)

0.28 0.27

1.13 1.12

-0.35 -0.49

3.9 5.6

166,26(8.5) 097, 10 (14.8)

315,66(22.7) 240, 78 (38,6)

n, number of grains by thin-section and its contend in the image; 0, mean grain size; SRplag, shape ratio of plagioclase fabric, T, shape parameter; vF (%), incompatibility index; A and C, azimuth and plunge of the lineation and the foliation pole, respectively. In parentheses the semi-angle (measured in degrees) of the la cone of data scattering around the the mean axes.

between 0.5 mm and 0.3 mm in the thickest dykes displaying type-A and type-B magnetic fabrics respectively. The grain size is smaller in the thinnest dyke (cm28); around 0.12mm in the centre and around 0.03mm in the chilled margins (Table 2). The calculated three-dimensional mineral fabric fits well to an ellipsoid, as indicated by the incompatibility index \ff being lower than 10% in all samples. The plagioclase shape ratio (SRplag) of dykes with type-A magnetic fabric is higher than that of dykes with type-B. In all specimens SR values are much higher than the corresponding magnetic anisotropy (P) (cf. Tables 1 and 2). These results are consistent with those obtained by Archanjo et al (2002) in other dykes from the Rio CearaMirim swarm using the same methods. They show SRplag values usually much higher than P and a tendency for an increase in the magnetic anisotropy, dominated by titanomagnetite, as the magnitude of the plagioclase fabric increases (Fig. 7). Moreover, the shape of the AMS and the plagioclase fabric ellipsoids are usually quite different from each other; oblate (T > 0) in the

Fig. 7. Magnetic anisotropy (P) versus plagioclase shape ratio (SRplag) of the Rio Ceara-Mirim dykes.

former and neutral to prolate (T < 0) in the latter. The oblate symmetry of AMS gives a better-defined magnetic foliation that usually dips steeply in the type-A magnetic fabric (Figs. 8 & 9) or dips gently in the type-B (Fig. 10). Conversely, the prolate ellipsoid symmetry indicates that the plagioclase fabric of the Ceara-Mirim dykes tends to be linear. The AMS of dykes from the Central Domain (cm34 and cm45) have steeply dipping magnetic foliations. On both dykes the mean lineation is sub-horizontal, but k\ axes tend to be scattered on the foliation plane in agreement with the oblate symmetry of the magnetic fabric (Fig. 8). The corresponding plagioclase lineation is either sub-horizontal (cm34) or plunges moderately to the NE along the foliation dip (cm45). In contrast to the AMS, the plagioclase foliation is sub-horizontal in the dyke cm34, and steep, north-trending in the dyke cm45. The dykes from the Central-Eastern (cm 14 and middle part of cm28) have kj axes plunging steeply, nearly down dip. The plagioclase lineation of the inner parts of each dyke plunges 50° and 37° to the east and NE, respectively, indicating that the magma moved instead at a high to moderate angle with the surface. On the other hand, the magnetic and plagioclase lineations are quite different from each other at the chilled margins of dyke cm28 (Fig. 9). The fine plagioclase laths form a very well-defined piano-linear fabric, where the foliation is imbricated close to the dyke margin and the lineation plunges gently. The plagioclase alignment close to the horizontal plane is remarkably evident in thin section (Figs. 5b, c), where the plagioclase shape ratio (SRplag) is stronger on the horizontal section than on the vertical section (Fig. 5d). In contrast to the AMS indication of highly inclined flows, the imbrication of plagioclase against the margins of the dyke suggests a lateral

294

C. J. ARCHANJO & P. LAUNEAU

Fig. 11. Shape preferred orientation of fine and coarse plagioclase laths and opaque grains of the dyke cm45 (Schmidt diagrams, lower hemisphere). ^/F, incompatibility index; SRplag and SRop are, respectively, the shape ratio of plagioclase and opaque distribution; T, shape parameter (see text for discussion). Symbols as Fig. 8.

To check for the influence of different shape and grain size of mineral populations, the spo orientation of opaque grains and fine and coarse plagioclase of dyke cm45 were analysed by the Inertia Tensor method (Fig. 11). Around 173 ± 17 opaque grains (mean size 0.31mm) per thin section were classified by image analysis. They have euhedral to subhedral shapes, welldefined contacts and a homogeneous content, around 5.5% on each thin section. The plagioclase grains were split into fine (0.21 < (/> (mm) < 0.47) and coarse (0.47 < (mm) < 1.44) grains, each comprising approximately 110 grains per thin section. The spo ellipsoid obtained for the opaque fraction agrees with that obtained by AMS both in shape and orientation. The spo of opaque grains is moderately oblate (T = 0.35) and the mean shape ratio (SRop) is 1.11. Orientation of magnetic and mineral foliation are undistinguishable (see error ellipses in Fig. 11), and the maximum and intermediate axes form a girdle in accordance with the oblate shape of the magnetic and mineral ellipsoids. The fine plagioclase grains show a very well-defined subfabric (^/F = 7.4%; semi-angle of error ellipses <18° for Icr), while the coarse plagioclase grains do not show any shape preferred orientation at this scale of measurement. Although the lineations of opaque and fine plagioclase are coaxial, the foliations are almost orthogonal to each other. Discussion

AMS significance In spite of their original 'normal' or 'abnormal' character, the AMS fabric of the Rio CearaMirim dykes has remarkable differences from

the plagioclase spo, both in orientation as well in shape of their ellipsoids. The ellipsoid symmetry is dominantly linear in the plagioclase fabric and planar in the magnetic fabric. The spo strength in the silicate fabric is systematically higher than the total AMS anisotropy (P). Moreover, the orientation of the plagioclase and magnetic fabric axes is usually oblique or even perpendicular to each other. The strong differences between the rock and magnetic fabrics should be explained in terms of their mineral carriers and fabric development processes. The plagioclase fabric, which typically is used to document the magmatic fabric of mafic igneous rocks (Nicolas 1992), is defined by the preferred orientation of its crystal shapes. If they were perfectly aligned with no interaction between adjacent grains, the fabric magnitude (SR) would be equivalent the mean axial ratio of the grain population. The AMS is controlled by the strong magnetic susceptibility of Ti-poor magnetite grains. A perfectly aligned assembly of magnetite grains may give contrasting AMS results depending on the size and shape of the grains. Relatively coarse, multidomain (MD) grains (<j) > 10 um) have maximum and minimum susceptibility axis parallel to the long and short grain dimensions respectively; the anisotropy degree (P) closely matches the mean shape ratio of the magnetite grains (Gregoire et al. 1998). For a single domain (SD) particle (0 < 0.3 jim) (Banerjee 1994), the maximum susceptibility is perpendicular to the long axis giving rise to an inverse fabric (k{ exchange its direction with £3). A mixture of SD and MD grains may form 'intermediate' fabrics with k2 exchanging its orientation with k\ or k3 (Rochette et al. 1992). Although the hysteresis data may be interpreted in terms of a mixture of SD and MD

MAGMA FLOW DIRECTION FROM PLAGIOCLASE ORIENTATION

Ti-poor magnetite grains, our data suggest that the differences between rock fabrics and type-B magnetic fabric are not due to the influence of large amounts (>50%) SD particles. The AMS of dyke cm30 may be simply explained by the alignment of coarse magnetite with long axes (sub)parallel to the linear fabric of plagioclase. A similar shape lineation was found in the opaque and plagioclase grains of a dyke with typical inverse fabric of the central domain (cm27, see Archanjo et al. 2002). In these dykes with type-B configuration the anisotropy is assumed to be dominated by MD grains. For dyke cm41, no fabric information was obtained on the opaque grains, apparently because its variable content in the images (1.5% to 4%) which shaped a poorly defined ellipsoid. Note that this dyke has a strongly oblate magnetic fabric (7 = 0.6, k{/k2 = 1.002) close to the mean plagioclase foliation, which may indicate that the AMS directions are also controlled by the silicate fabric. Alternatively, competing mineral fabrics may have been acquired during rotation at magmatic stage of grains with different aspect ratios (Ildefonse et al. 1992). In fact, spo measurements of opaque grains and two plagioclase subfabrics of dyke cm45 give three different ellipsoids in terms of both shape and orientation. The fabric of opaque grains as given by spo or indirectly from the AMS is oblate and weakly anisotropic (P < 3%), enabling the easy exchange of k\ and k2 axes by minute modifications in the shape and orientation of magnetic oxides. It is noteworthy that the cm41 and cm28n show ki — k2 exchanges, while cm34 show k2 — A:3 exchanges. We speculate that the subsolidus unmixing of ilmenite (paramagnetic at room temperature) and Ti-poor magnetite (ferrimagnetic) within the coarse grains may cause these exchanges as a result of small differences in the net shape of ferrimagnetic lamellae after exsolution. Furthermore the tendency of P to increase as SRplag strengthens (Fig. 7) suggests that the control of magnetic anisotropy is given by MD titanomagnetite. Alternatively, moderate amounts of SD particles (<50%) formed by secondary processes could dominate the magnetic fabric of dyke cm34. They would change an earlier intermediate fabric (coaxial with the plagioclase spo) to a strongly oblate type-A configuration. Whichever the processes that affect the 'primary' magnetic mineralogy of the dykes, i.e. high-temperature oxidation or/and low-temperature recrystallization of fine magnetite particles, they introduce important deviations in AMS with direct consequences on its interpretation.

295

Magma flow along the swarm Chemical and isotopic data of the Rio CearaMirim dykes indicate that the mafic magma was extracted from a lithospheric mantle when the Gondwana continent broke-up in the Early Cretaceous (Bellieni et al. 1992). Crustal extension gave rise to magma emplacement filling in fractures that cut at high angle the structural grain of the host wall rocks. The resulting dyke swarm spreads over 400km and is formed by several dykes of some kilometres in length, bearing an arrangement that varies from anastomosing to en echelon, this latter found particularly in the central domain. The plagioclase shape preferred orientation recorded on the central domain of the swarm is shown in Figure 12. In the SZ region they show neutral to linear fabrics, with foliations nearly parallel to the trend of the dyke and lineations that plunge steeply. This fabric pattern is consistent with a vertical magma flow in a 'feeder zone', although sub-horizontal to inclined flows were recorded in the chilled margins and in the centre of dyke cm28. Out of the SZ, fabric ellipsoids vary from moderately prolate to oblate. The foliation tends to dip gently and the lineation is parallel to very oblique to the dyke wall. A model for the magmatic infilling of fractures and fabric development on dykes of the central domains is represented in Figure 13a. A hidden master fracture of approximately E-W direction would have channelled the magma in depth and infilled the vertical secondary fractures above. The magmatic fabric in the secondary fractures would be similar to those obtained in experiments that simulated the slow rising of a plastic mass flowing into an open cavity (Fig. 13b; Nadai 1931). The cavity is filled upwards by squeezing the plastic mass laterally. The finite strain within the confined material predicts a vertical stretching direction near the bottom of the cavity, which progressively flattens towards its top. The strain pattern within the upper part of the flowing mass may be compared with experiments that simulate the ascent of a diapir that moves up by contrast of viscosity with its host (Cruden 1990). Oblate flattening strains occupy the lateral margin and top of the diapir while prolate strains forms within its inner parts. As the diapir rise, the peripheric flattening strain decreases in volume and the intervening transition zone to prolate strains approaches the margin of the diapir. These combined models are able to describe the plagioclase fabric variation found in the central domain. The infilling of the dykes would occur due to overpressure in the magma chamber

296

C. J. ARCHANJO & P. LAUNEAU

Fig. 12. Plagioclase fabric of the Rio Ceara-Mirim dyke swarm. The dashed line correspond the 'feeder zone' detected by the type-A magnetic fabric. The superscripts 'n' and *c' in the stereonets of the Steep Zone correspond, respectively, the chilled north margin and the middle of the dyke cm28.

Fig. 13. Model for emplacement of dykes along an E-trending fracture system. A master fracture at depth pumps the magma upward (a) and infills secondary fracture sets above. The flow lines of a plastic mass (magma) infilling a cavity (b) predicts vertical strains in the neck and horizontal strains in the roof of the ascending mass (after Nadai 1931). The silicate fabric pattern of the central domain of the Ceara-Mirim would record the exposures of different sections of the dyke set.

MAGMA FLOW DIRECTION FROM PLAGIOCLASE ORIENTATION enabling upward fracture propagation in the host rocks. The variable orientations and symmetries of the plagioclase fabric would correspond to the different levels of exposure/ erosion of the dyke sets. The transition from gentle dipping magmatic fabrics in the Central Domain to steeply dipping fabrics at the feeder zone would be controlled by the depth of the master fracture that would become shallower to the east (Fig. 13a).

Conclusions Our results indicate that the weakly anisotropic oblate AMS ellipsoids obtained on the Rio Ceara-Mirim mafic dykes can be rarely used to indicate the magma flow directions. Such magnetic fabric seems to reproduce the alignment of magnetite grains, which may be parallel to the dyke plane, hence 'normal', but may be highly oblique to the actual magmatic fabric (= silicate alignment). Moreover, in some cases, inverse (type-B) magnetic fabrics seem to be grounded on a formerly 'abnormal' magmatic fabric; thus they are not related to the presence of large amounts of SD ferrimagnetic particles in the dyke. We propose that the contrasting magnetic and mineral fabrics arise from the different timing and mechanisms of fabric acquisition. While the plagioclase fabric is frozen in the latest stages of the magmatic flow, the AMS may be strongly influenced by sub-solidus reactions that modify the shape and anisotropy of the 'primary' titanomagnetite grains. In addition, small to moderate amounts of SD grains formed by low-temperature oxidation processes may have modified the total anisotropy to form intermediate fabrics or to develop normal magnetic fabrics settled on an abnormal silicate fabric (cm34). These fabric inconsistencies were also documented in a chilled margin of a dyke. Thus, an imprecise picture of the flow structure would appear if AMS were the unique parameter used to reveal the fabric of the swarm. Data from complementary methods need to be taken into account for a fine interpretation in terms of emplacement and flow directions of the Rio Ceara-Mirim dykes. This research was supported by grants from Funda9ao de Amparo a Pesquisa do Estado de Sao Paulo (2001/ 14154-2) and Conselho Nacional de Pesquisa e Tecnologia (300889/96-8). We thank R. Trindade who made constructive comments on an earlier draft of the manuscript, and P. Rochette and an anonymous reviewer for helping to improve this work.

297

References ABELSON, M., BAER, G. & AGNON, A. 2001. Evidence from gabbro of the Troodos ophiolite for lateral magma transport along a slow-spreading midocean ridge. Nature, 409(4), 72-75. ARCHANJO, C. J., TRINDADE, R. I. F., MACEDO, J. W. P. & ARAUJO, M. G. S. 2000. Magnetic fabric of a basaltic dyke swarm associated with Mesozoic rifting in northeastern Brazil. Journal of South American Earth Sciences, 13, 179-189. ARCHANJO, C. J., ARAUJO, M. G. S. & LAUNEAU, P. 2002. Fabric of the Rio Ceara-Mirim mafic dyke swarm (northeastern Brazil) determined by anisotropy of magnetic susceptibility and image analysis. Journal of Geophysical Research, 107(B3), 10.1029/2001JB0000268. BANERJEE, S. K. 1994. Contributions of fine particle magnetism to reading the global paleoclimate record. Journal of Applied Physics, 75(10), 5925-5930. BELLIENI, G., MACEDO, M. H. F., PETRINI, R., PICCIRILLO, E. M., CAVAZZINI, G., COMIN-CHIARANONTI, P., ERNESTO, M., MACEDO, J. W. P., MARTINS, G., MELFI, A. J., PACCA, I. G. & DE MIN, A. 1992. Evidence of magmatic activity related to Middle Jurassic and Lower Cretaceous rifting from northeastern Brazil (Ceara-Mirim): K/Ar age, paleomagnetism, petrology and Sr-Nd isotope characteristics. Chemical Geology, 97, 932. BORRADAILE, G. J. & HENRY, B. 1997. Tectonic applications of magnetic susceptibility and its anisotropy. Earth Science Reviews, 42(1/2), 49-93. BUCKER, C., SCHULT, A., BLOCK, W. & GUERREIRO,

S. D. C. 1986. Rockmagnetism and paleomagnetism of an Early Cretaceous/Late Jurassic dike swarm in Rio Grande do Norte, Brazil. Journal of Geophysics, 60, 129-135. CALLOT, J. P., GEOFFROY, L., AUBOURG, C., Pozzi, J. P. & MEGE, D. 2001. Magma flow directions of shallow dykes from East Greenland volcanic margin inferred from magnetic fabric studies. Tectonophysics, 335, 313-329. CANON-TAFIA, E. 2001. Factors affecting the relative importance of shape and distribution anisotropy in rocks: theory and experiments. Tectonophysics, 340, 117-131. CRUDEN, A. R. 1990. Flow and fabric development during the diapiric rise of magma. Journal of Geology, 98, 681-698. DAY, R. M., FULLER, M. & SCHMIDT, V. A. 1977. Hysteresis properties of titanomagnetites: grain size and composition dependence. Physics of the Earth and Planetary Interiors, 13, 260-267. DUNLOP, D. J. 2002. Theory and application of the Day plot (Mrs/Ms versus Her/He) 2. Application to data for rocks, sediments and soils. Journal of Geophysical Research, 107(B3), 10.1029/ 2001JB0000487. ERNESTO, M., BELLIENI, G., PICCIRILLO, E. M., MARQUES, L. S., DE MIN, A., PACCA, I. G., MARTINS, G. & MACEDO, J. W. P., 2003. Paleomagnetic and geochemical constraints in the timing and duration of the CAMP activity in Northeastern

298

C. J. ARCHANJO & P. LAUNEAU

Brazil. , In: The Central Atlantic Magmatic Province: insights from fragments of Pangea. Geophysical Monograph 136, American Geophysical Union, 10.1029/136GM07, 129-149. ERNST, R. E. & BARAGAR, W. R. A. 1992. Evidence from magnetic fabric for the flow pattern of magma in the MacKenzie giant radiating dyke swarm. Nature, 356, 511-513. ERNST, R. E. & BARAGAR, W. R. A. 1992. Evidence from magnetic fabric for the flow pattern of magma in the MacKenzie giant radiating dyke swarm. Nature, 356, 511-513. FUETEN, F. 1997. A computer controlled rotating polarizer stage for the petrographic microscope. Computer & Geosciences, 23, 203-208. FUETEN, F. & GOLDCHILD, J. S. 2001. Quartz c-axes orientation determination using the rotating polarizer microscope. Journal of Structural Geology, 23, 895-902. GEOFFROY, L., CALLOT, J. P., AUBOURG, C. & MOREIRA, M. 2002. Magnetic and plagioclase linear fabric discrepancy in dykes: a new way to define the flow vector using magnetic foliation. Terra Nova, 14, 183-190. GLEN, J. M. G., RENNE, P. R., MILNER, S. C. & COE, R. S. 1997. Magma flow inferred from anisotropy of magnetic susceptibility in the coastal ParanaEtendeka igneous province: evidence for rifting before flood volcanism. Geology, 25(12), 11311134. GREGOIRE, V., DARROZES, J., GAILLOT, P. & NEDELEC, A. 1998. Magnetite grain shape fabric and distribution anisotropy vs rock magnetic fabric: a three-dimensional case study. Journal of Structural Geology, 20(7), 937-944. HODYCH, J. P. 1996. Inferring domain state from magnetic hysteresis in high coercivity dolerites bearing magnetite with ilmenite lamellae. Earth and Planetary Science Letters, 142, 523-533. ILDEFONSE, B., LAUNEAU, P., FERNANDEZ, A. & BouCHEZ, J. L. 1992. Effect of mechanical interactions on development of shape preferred orientations: a two-dimensional experimental approach. Journal of Structural Geology, 14, 73-83. KNIGHT, M. D. & WALKER, G. P. L. 1988. Magma flow directions in dikes of the Koolau Complex, Oahu, determined from magnetic fabric studies. Journal of Geophysical Research, 93(B5), 4301-4319. LAUNEAU, P. 2004. Mise en evidence des ecoulements magmatiques par analyse d'images 2-D des distributions 3-D d'orientations preferentielles des formes. Bull. Geol. Soc. Fr. 175, 331-350 LAUNEAU, P. & CRUDEN, A. R. 1998. Magmatic fabric acquisition mechanisms in a syenite: results of a combined anisotropy of magnetic susceptibility and image analysis study. Journal of Geophysical Research, 103(B3), 5067-5089. LAUNEAU, P., CRUDEN, A. R. & BOUCHEZ, J. L. 1994. Mineral recognition in digital images of rocks: a

new approach using multichannel classification. Canadian Mineralogist, 32, 919-933. MALUSKI, H., COULON, C., POPOFF, M. & BAUDIN, P. 1995. 40Ar/39Ar chronology, petrology and geodynamic setting of Mesozoic to Early Cenozoic magmatism from Benoue Trough, Nigeria. Journal of Geological Society, London, 152, 311-326. MlZUSAKI, A. M. P., THOMAZ-FlLHO, A., MlLANI, E. J.

& CESERO, P. 2002. Mesozoic and Cenozoic igneous activity and its tectonic control in northeastern Brazil. Journal of South American Earth Sciences, 15, 183-198. NADAI, A. 1931. Plasticity: A Mechanics of the Plastic State of Matter. McGraw-Hill Book Company, New York. NICOLAS, A. 1992. Kinematics in magmatic rocks with special reference to gabbros. Journal of Petrology, 33(4), 891-915. OLIVEIRA, D. C. 1992. O papel do enxame de diques Rio Ceard-Mirim na evolucao tectonica do Nordeste Oriental (Brasil): implicacoes na formacao do rifle Potiguar. Ms. C Thesis, Universidade federal de Ouro Preto, Ouro Preto. POTTER, D. K. & STEPHENSON, A. 1988. Single-domain particles in rocks and magnetic fabric analysis. Geophysical Research Letters, 15, 1097-1100. RAPOSO, M. I. B. & ERNESTO, M. 1995. Anisotropy of magnetic susceptibility in the Ponta Grossa dike swarm (Brasil) and its relationship with magma flow direction. Physics of the Earth and Planetary Interiors, 87, 183-196. RINK, M. 1976. A computerized quantitative image analysis procedure for investigating features and an adapted image precess. Journal of Microscopy, 107, 267-286. ROBIN, P. -Y. F., 2002. Determination of fabric and strain ellipsoids from measured sectional ellipses - theory. Journal of Structural Geology, 24, 531544. ROCHETTE, P., JACKSON, M. & C. A. 1992. Rock magnetism and the interpretation of anisotropy of magnetic susceptibility. Reviews of Geophysics, 30(3), 209-226. ROCHETTE, P., AUBOURG, C. & PERRIN, M. 1999. Is this magnetic fabric normal ? a review and case studies in volcanic formations. Tectonophysics, 307, 219234. SHIMAMOTO, T. & IKED A, Y. 1976. A simple algebraic method for strain estimation from deformed ellipsoidal objects. 1. basic theory. Tectonophysics, 36, 315-337. TAUXE, L., GEE, J. S. & STAUDIGEL, H. 1998. Flow directions in dikes from anisotropy of magnetic susceptibility data: the bootstrap way. Journal of Geophysical Research, 103(68), 17,775-17,790. WORM, H. U. & MARKET, H. 1987. Magnetic hysteresis properties of fine particle titanomagnetites precipitated in a silicate matrix. Phys. Earth Planet. Inter., 46, 8^92.

Anisotropy of magnetic susceptibility (AMS): magnetic petrofabrics of deformed rocks GRAHAM J. BORRADAILE1 & MIKE JACKSON2 1

Geology Department, Lakehead University, Thunder Bay ON Canada P7B 5El (e-mail: [email protected]) 2Institute for Rock Magnetism, Winchell School of Earth Sciences, University of Minnesota, 100 Union St SE, Minneapolis, MN 55455, USA (e-mail: [email protected]) Abstract: For 40 years magnetic anisotropy has provided successful geological interpretations of magnetic ellipsoid orientations; in contrast the interpretation of anisotropy magnitudes is far more convoluted. This is due to complexities at various levels within rocks, including different physical magnetic responses of different minerals, grain-scale magnetic anisotropy, the anisotropy of interacting ensembles, the mineralogical constitution of rocks and the processes and mechanisms that align minerals in nature. The chief factors determining the magnetic fabrics of tectonized rocks include: mineral-physics properties, crystal symmetry, mineral-abundances, tectonic symmetry and crystal orientation-distribution, strain or stress, kinematic history and certain tectono-metamorphic processes (e.g. diffusion, crystal plasticity, dynamic recrystallization, particulate flow, neomineralization). AMS ultimately provides an integrated record of some combination of these factors. Subfabrics due to distinct processes or events may be expressed in different mineral and/or grain-size fractions, and are superposed in the conventionally observed AMS. Their discrimination may be achieved by various laboratory techniques such as magnetization and torque measurements in weak and strong applied fields, anisotropy of ARM and IRM, gyroremanence, Rayleigh magnetization, chemical leaching. However, under limited circumstances, statistical approaches such as differential analysis, tensor standardization, symmetry of confidence regions for the principal axes may partly isolate different subfabric orientations.

This review attempts to present the main compo- additional anisotropy may be superposed in nents of magnetic petrofabricsas a springboardrocks with very large remanent magnetizations for graduate students and for workers from or in those that have previously been exposed other geological sub-disciplines. At the same to high fields. Properly measured, at low fields, time, since the field of magnetic petrofabrics is in rocks without strong permanent magnetizaitself interdisciplinary (mineral magnetism, rock tions, AMS is unaffected by contributions from mechanics, petrology), we aim to provide a sum- irreversible magnetization processes. Careful mary of principles, methods and interpretation calibration and holder corrections are important guidelines that will be useful to experienced for weakly susceptible and/or weakly anisotropic workers in this field. Greatest benefit derives materials; the typical noise environment for a from the use of anisotropy of induced magnetiza- routine AMS instrument is <0.1uSI (Sagnotti tion in a low field (on the order of 102A/m, et al. 2003). 10~4T or 1 oersted), from measurements along Quite different supplementary techniques at least six suitably oriented axes through a speci- measure the anisotropy of permanent (remanent) men. This is anisotropy of magnetic susceptibility magnetization. These isolate the ODs of sub(AMS); Table 1 presents a glossary of related fabrics of independent mineral grains and inclumagnetic and petrofabric terms. Susceptibility sions of magnetically ordered species such as anisotropy is occasionally measured in high magnetite, hematite or pyrrhotite. The techfields that saturate and therefore discount niques are time-consuming because artificial ordered phases (most usually magnetite). All permanent magnetizations must be applied and induced magnetizations are ephemeral responses then removed along numerous different axes to an applied field; the ratio of induced magneti- through the specimen. Anisotropy of anhysteretic zation to inducing field is the susceptibility, remanent magnetization (AARM, McCabe et al. whose dependence upon orientation is controlled 1985; Jackson et al. 1988, 1989a, b) is favoured almost exclusively by the combined orientation- since the d.c. bias field is small; measurable distribution (OD) of all the minerals in the remanence results from the combined action of specimen. A relatively small (<1%, P < 1.01) the weak bias and a large superposed AF. Less From: MARTIN-HERNANDEZ, F., LUNEBURG, C. M., AUBOURG, C. & JACKSON, M. (eds) 2004. Magnetic Fabric: Methods and Applications. Geological Society, London, Special Publications, 238, 299-360. 0305-8719/04/ $15.00 © The Geological Society of London 2004.

300

G. J. BORRADAILE & M. JACKSON

Table 1. Glossary of magnetic and tectonic fabric terminology Magnetism & anisotropy terminology Magnetic dipole moment [Am2]; a measure of the strength of a magnet. Volumetric magnetization intensity [A/m]; dipole moment per unit volume. Mass-specific magnetization intensity [Am2/kg]; dipole moment per unit mass. Magnetic field [A/m]. In the absence of free currents, the tangential component of H is continuous across a material boundary. B Magnetic induction [T] = [Wb/m2] = [J/(Am 2 )]. B = ^(H + M), where //0, the permeability of free space, is a fundamental physical constant (^0 = 4?r x 10~7 H/m). The normal component of B is continuous across a material boundary. K Magnetic susceptibility [dimensionless]; the ratio of induced magnetization to magnetizing field K — M/H, measured in a low direct field (e.g. SQUID magnetometer) or alternating field (e.g. induction coil). To preserve linear magnetization-field response (M = K,H), the field must not affect the remanence of any permanently magnetizable content, i.e. usually <10mT. Induction coil methods use low frequencies (<20kHz) to suppress conductivity responses. K is frequency dependent for conductive materials and for viscous superparamagnetic grains. AMSAnisotropy of magnetic susceptibility with principal values INT > KMIN> tt MAX > ft described as a second-rank tensor. Where the principal values have the same sign, this may be visualized as magnitude ellipsoid. AMS ellipsoid The AMS tensor possesses a representation surface, a quadric of radii r. Only where the principal susceptibilities have the same sign and where none are zero can we define a magnitude ellipsoid (the AMS ellipsoid} with radii 1 />/*"• K Bulk susceptibility; commonly given as arithmetic mean of the principal susceptibilities, (^MAX + ^INT + «MiN)/3, if all principal values have same sign. The geometric mean is preferred, especially if the anisotropy is large, as in AARM or finite strain; requires nonzero, same-sign principal susceptibilities. Standardized Principal values (KMAX^ Q^c) maY be standardized simply by dividing by K. For the finite strain tensor, principal axes must be standardized using their geometric mean since anisotropies may be large. P Anisotropy degree, P = KMAX/^MIN L Magnetic lineation L = ^MAX/^INT cf- Flinn's a F Magnetic foliation F = /^INT/^MIN cf. Flinn's b Pj or Pf Jelinek's (1981) 'corrected' anisotropy degree, reflecting eccentricity of the magnitude ellipsoid for a second-rank tensor where all principal values have the same sign. Ranges from 1.0 (sphere, isotropy) upward, original definition involved normalization with arithmetic mean of susceptibility but for higher anisotropies, e.g. AARM or strain the geometric mean may be preferable. T Jelinek's (1981) shape parameter for the magnitude ellipsoid, measuring the range from prolate (T = +1) through neutral (T = 0) to oblate (T = +1) ellipsoids, independent of eccentricity; identical to Lode's parameter (1926), v of structural geology. X Mass-specific susceptibility [m3 kg"1]; x — J/H = ft/density. d.c. Direct current used to apply a steady field, for IRM or as a bias field to generate an ARM. AF Alternating field; laboratory applied sine wave field damped so that its amplitude decays progressively and usually linearly with time. Remanence Magnetization that persists in the absence of an applied magnetic field. In contrast induced magnetization, measured during the study of AMS, disappears when the applied field is removed. A^ Intrinsic magnetic susceptibility [dimensionless]; not directly measured for highsusceptibility materials, where it is 'screened' by self-demagnetization. fi (Relative) magnetic permeability. In linear materials and/or in weak applied fields, ^ = 1 + «i, and absolute permeability /^abs = ^0 relates magnetic induction B and (internal) magnetic field H

p M J H

ARM AARM K ARM or ^

B = ^(H + M) = pv(H + KiHi) = /J 0 #(l + «i) = fJLQfJLH

= //abs//

Anhysteretic remanent magnetization; laboratory remanence due to small d.c. field biasing the moments scattered during the decay of a simultaneously decaying large alternating field. Anisotropy of anhysteretic remanent magnetization Anhysteretic susceptibility, M ARM /// dc [dimensionless]. Mean value is A = ^ARM = ( K ARM,MAX^ARM,INT^ARM,MIN) > where ^ARM,MAX > ^ARM,INT > «ARM,MiN- (Geometric mean is wiser, since anisotropies are large and ARM's are always positive values.)

AMS-PETROFABRIC OF DEFORMED ROCKS

301

Table 1. (cont.) Magnetism & anisotropy terminology pARM pAARM MD SD SP

PSD

Polydomain TM IRM AIRM Inverse AMS fabric Blended AMS fabric

Diamagnetism

Paramagnetism

'Ferro '-magnet

Partial anhysteretic remanent magnetization (that due to grains in a selected range of coercivities). Anisotropy of pARM. Multidomain (remanence-bearing) grain: for magnetite usually >5 domains and >2um diameter but shape-dependent Remanence-bearing grain with a unique direction of spontaneous magnetization (in one domain); for magnetite 20-100 nm but shape-dependent Superparamagnetic grains are SD but thermally unstable; for magnetite <20 nm but shape-dependent Pseudo-single domain; remanence-bearing grain that has some remanence properties like SD and therefore desirable in routine palaeomagnetism (for magnetite 2-5 domains, ~0.1-2um in diameter, depending on shape. Umbrella term for MD & PSD characteristics. Titanomagnetite series Fe3_xTixO4 e.g. TM60 = Fe2.4Ti0 6O4 Isothermal remanent magnetization; remanence acquired in a large d.c. field Anisotropy of isothermal remanent magnetization Due to (a) SD magnetite grain in which KMIN must be parallel to the long axis OR (b) the presence of some mineral (e.g. calcite, tourmaline) in which K MIN is parallel (or ^parallel) to the crystallographic axis associated with the long habit Rock AMS fabric that combines magnetic subfabrics without a simple one-to-one mapping or parallelism of their principal magnitudes, e.g. A^MIN f°r one sur> fabric is inclined to K MIN for another subfabric. The subfabrics may be due to some combination of minerals with inverse and normal AMS, OR to some combination of competing petrofabrics. Universal negative susceptibility response, usually ~ — 14uSI. Loosely refers to some minerals, which if pure, would only show a diamagnetic response, e.g. quartz, calcite. However, natural examples commonly have overwhelming positive responses from inclusions. Rocks such as quartzite and limestone are rarely diamagnetic. Diamagnetism is not temperature-dependent. Positive response to an applied field with 0 < K < 2000 uSI for most pure rock-forming minerals. (Unfortunately, matrix minerals are rarely pure and K may be much higher due to high-^ microscopic-submicroscopic inclusions or exsolutions (e.g. iron-oxide; see 'ferro'magnetic). Paramagnetic susceptibility decreases inversely with temperature. Umbrella term for magnetically ordered phase, more specifically and more accurately described as, e.g. ferrimagnetic (magnetite) or antiferromagnetic (hematite). Sometimes stoichiometrically controlled, e.g. pyrrhotite (Fe^^S) is ferrimagnetic for x > 0.125, otherwise antiferromagnetic. Ferromagnetism mostly decreases non-linearly with increasing temperature and disappears above Curie temperature (ferrimagnets) or Neel temperature (antiferromagnets). Structural & petrofabric terminology

Strain

X > Y >Z Flinn 's a, b Flinn's k Ramsay's K

Change of shape; homogenously, keeping parallel lines parallel and straight lines straight; or heterogeneously. Homogeneous finite strain, a second-rank tensor, may always be represented by a magnitude ellipsoid since principal magnitudes are always positive. Strain history (accumulation of incremental strains) may not be deduced from a single state of homogeneous finite strain. Principal stretches or finite strains, axes of the finite strain ellipsoid (Ramsay 1967). [Note that Flinn's original work (1962) used the convention, Z > Y > X, which was never widely adopted]. Stretch = (new length)/(old length). a = X/ 7, b = Y/Z describes the magnitude ellipsoid of the finite strain tensor and the orientations-distribution tensor; (x,y) coordinates of the Flinn and Woodcock plots. Comparable older terms in AMS are L and F. k = (a — \)/(b — l ) ; a measure strain ellipsoid shape from oblate (k = 0) through plane strain (k = 1) to prolate k = oo. K = \n(a)/ \n(b): Iog10 scaling of the Flinn plot disperses low anisotropy data points for greater clarity; points separated by similar distances on the log-plot have similar strain-differences.

302

G. J. BORRADAILE & M. JACKSON

Table 1. (cont.) Structural & petrofabric terminology L-S scheme

OD

Orthorhombic

Material line/plane Non-material line

Deformation Deform. Mechanism Co-axial

Vorticity PDO PCO

Flinn's (1965b) qualitative scheme; associates shape or mineral fabrics for a homogenous sample of specimens with the appropriate finite-strain or orientation-distribution ellipsoid. L = prolate, L — S neutral; S = oblate, general cases, e.g. L > S or L < S. Importantly recognizes the smooth spectrum of possibilities vis-a-vis historically incorrect notion of discrete symmetry classes. Orientation Distribution: frequency distribution of axes or directions in three-dimensions, illustrated on a stereogram or by an orientation tensor that is visualized by its magnitude ellipsoid. Mirror symmetry across three mutually orthogonal planes: no axial-symmetric rotation the planes' intersections; e.g. a magnitude ellipsoid representing for finite strain or AMS. End-member prolate and oblate ellipsoids possess higher, rotational symmetry about one axis. AMS ellipsoids never have less than orthorhombic symmetry whereas most rockforming minerals have lower crystal symmetry. Physically characterized line, usually in a continuum, that may be tracked through a strain history, e.g. bedding. A finite strain axis (X, 7, or Z) or some other distinct linear property (e.g. line of no finite strain) defined by the state of strain at any point in a strain history. While their orientation may be determined they are rarely tied to any material line (one exception would be incremental strain axes during simple shear strain history). Changes in location of parts of a material with respect to an external reference frame. May comprise any combination of: (i) translation (ii) rigid body rotation (iii) strain (iv) volume change. Material processes by which internal relative displacements are achieved; including crystal-plasticity, diffusion, pressure solution, particulate flow, etc. A strain-history or fabric-evolution in which the orthogonal principal axes remain parallel to the same material lines as events progress; including pure shear or compaction (X = Y > 1 > Z) or general coaxial flattening X > Y > 1 > Z); cf. noncoaxial. Older terms irrotational/ rotational may be confused with the internal rotation of material lines during coaxial strain history or even with external rigid-body rotation. The rate and sense of finite strain axis rotation (in some complex three-dimensional manner), usually with respect to the rate of accumulation of strain increments. Usually very difficult to relate to the orientation of material lines. Preferred dimensional orientation, for example of strained objects such as pebbles, grains or any other clasts, fossils, lapilli (L-S scheme). Preferred crystallographic orientation, for example an alignment of minerals by growth, rotation or solid-state plastic deformation into a linear-planar fabric (L-S scheme).

We generally follow the usage of Collinson (1983) and Hunt et al. (1995); symbols differ to varying degrees in standard references (Stacey & Banerjee 1974; O'Reilly 1984; Dunlop & Ozdemir 1997; Tauxe 1998).

commonly, anisotropy of isothermal remanent magnetization (AIRM, Daly & Zinsser 1973, Jelinek 1993) may be used but the larger direct fields required may risk a non-linear relationship between each magnetization and its applied field, so that the second-rank tensor structure is invalid. Without jeopardizing the integrity of our goal, this review focuses on tectonically deformed rocks. We justify the restriction because the geophysical and petrofabric principles used by students of igneous and sedimentary petrofabrics are a subset of the procedures treated here and reviewed elsewhere (Jackson & Tauxe 1991; Tarling & Hrouda 1993). The newcomer is welcomed to this field knowing that AMS provides a measure of the net OD of all minerals in

a ~10.5cm3 core-specimen of any rock in ~3 minutes. Traditional petrofabric analysis relied on optical universal-stage techniques to determine the OD of small samples of grains (typically <200) from specially prepared microscope mounts, preferably several of different orientation per specimen (Sander 1930), requiring several hours per specimen. Simpler X-ray approaches (Kisch 2002) and more elaborate X-ray goniometry and automation of the optical universal-stage have streamlined dataacquisition (Price 1980; Oertel 1983) but the conceptual basis of conventional petrofabric data is essentially unchanged; some representation of preferred lattice orientation is obtained. In contrast, genetic interpretations of ODs have leapt ahead through the assimilation of

AMS-PETROFABRIC OF DEFORMED ROCKS

knowledge from materials science and metallurgy (Nicolas & Poirier 1976; Poirier 1985; Blenkinsop 2000) whereas AMS fabric interpretations have not yet been integrated into our understanding of fabric evolution, to the same extent. For historical convenience, progress in magnetic petrofabrics can be measured against a few key landmarks (a)

Graham's recognition of AMS as a petrofabric proxy (1954). (b) Fuller's (underappreciated) note that AMS is due to the frequency distribution of minerals (1963). Furthermore, Fuller's unpublished dissertation (Cambridge, 1961) recognized, among other things, the importance of spatial distributions of ferromagnets ('supergrains'). (c) Jelinek's (1978) fundamental contribution to the correct statistical characterization of a sample of tensors. His parameters for the description of any anisotropy are superior to those used in structural geology (Jelinek 1981). Hrouda (1982) summarized this first campaign of analytical applications. Our review is weighted towards the post-1982 campaign, but individual landmarks are noted, regardless of vintage. Almost 100 years ago it was appreciated that the orientation of crystals was expressed and detectable through magnetic properties (Voight & Kinoshita 1907), principally their magnetic susceptibility in low fields (<10mT). However, precise rules relating the orientation of principal susceptibility axes to the mostly non-orthogonal crystallographic axes were only fully appreciated after Nye's (1957) description of susceptibility anisotropy of induced magnetization as a second-rank tensor. Susceptibility axes' orientations are controlled by crystal symmetry and only retain parallelism with crystal axes for the high-symmetry groups (Fig. 1). In fact, 50% of all crystalline substances are monoclinic, and feldspars (the most abundant crystal mineral group) are largely triclinic. Composition controls angles between crystal and principal susceptibility axes in pure minerals and Figure 1 shows possible symmetry-compatible relationships for the principal crystal symmetry groups. The only constraint is that the anisotropy's symmetry elements must be included in those of the crystal system (Neumann's principle). In nature, AMScrystal symmetry relationships are reduced further and non-systematically due to impurities, especially where ferromagnetic inclusions increase K (Table 2) beyond the host's theoretical maximum (Fig. Ih). Despite the obvious

303

practical complications of precisely interpreting AMS axes in terms of mineral orientations, early investigators confirmed that induced magnetization was easier in certain significant directions, e.g. parallel to bedding or schistosity (Ising 1942; Graham 1954). However, the first routine applications of AMS were to safeguard palaeomagnetism, assuring that natural remanence vectors accurately record palaeofield orientations (e.g. Hargraves 1959; Fuller 1963; Uyeda et al. 1963; Kodama 1991). Second, although bulk susceptibility (K) was always required to interpret magnetic surveys at any scale, its anisotropy (AMS) may cause further anomaly shifts (Hrouda 1982; Rochette 19946), especially in pyrrhotite-rich rocks (e.g. Rochette 19876; Clark et al 1988, 1992; Clark & Tonkin 1994). Third, and independent of any theoretical basis, the simple patterns of AMS axes were obviously related to petrofabric and, initially, attributed largely to magnetite since early studies were of some igneous protoliths in which K increased rapidly with magnetite-content (Balsley & Buddington 1960) (Fig. 2a). However, the necessity of plotting the large range of K logarithmically produces a deceptive sense of linear correlation and obscures the misleading contribution of high-ft outliers. Graham (1954) established the concept that the AMS axes could be associated with subtle or cryptic grain alignments, adding a new dimension to petrofabrics (Sander 1930; Turner & Weiss 1963). AMS could 'average' the OD and intensity of alignment of all the grains in a ~llcm core-specimen, more comprehensively, quickly and reproducibly than any other petrofabric technique before or since. The most general caveats to AMS as a petrofabric proxy are sampling issues (Fig. 2b-d). We now know that the differences in mineral-ft are so great (Table 2) that a few grains of, say magnetite, in a low-ft matrix (e.g. quartzfeldspar) may dominate the AMS. However, petrofabric analyses require >100 grains to define reproducible orientations ('stable OD'); a measurable AMS does not guarantee a stable OD (Fig. 2b) (Borradaile & Stupavsky 1995). Special sampling issues arise where there are low abundances of minerals with extremely high K or in low-ft rocks, e.g. where ft<100|iSI,although this depends on the strength of the OD. Even equi-dimensional and thus individually isotropic magnetite grains may yield a strong net AMS, due to their high susceptibility, usually ft > 2 SI and interaction effects, if arranged in a suitable location fabric (Fig. 2d). These special clusters of grains were termed 'supergrains' by

304

G. J. BORRADAILE & M. JACKSON

Fig. 1. For pure minerals, AMS principal axes are controlled by crystal structure and fixed in orientation according to the crystal's symmetry group (Nye 1957). (Magnetite is an exception and certain other minerals show counterintuitive relationships, see Fig. 3). The maximum symmetry of the AMS arrangement may equal that of the crystal, (a) For orthorhombic crystals each crystal axis is parallel to a principal susceptibility axis, (b) For monoclinic crystals only one crystal axis defines one of the three principal susceptibilities; >50% of minerals are monoclinic, including many of interest in tectonic-metamorphic studies, (c) Triclinic crystals show no correspondence of crystal and AMS axes; these include most feldspars, the most abundant rock forming mineral group, (d)-(g) Possible arrangements in pure minerals; these are dependent on stoichiometry but, in nature, they are disturbed by inclusions, (h) For example, crystallographic control of AMS in some natural phlogopite crystals is disturbed by the AMS of magnetite inclusions; the subfabric of the latter was isolated by AARM (Borradaile 1994). (N.B. all stereograms are lower hemisphere, equal area in this paper.) Fuller (1961). Problems also arise in low-susceptibility rocks in which diamagnetic responses almost cancel positive susceptibilities. AMS may now be measured reproducibly in induction coil instruments with a routine precision of

±0.1 uSI; however calibration is rather important (Sagnotti et al 2003). Rocks with \K\ -> 0, due to paramagnetic-diamagnetic mixtures, impede interpretation rather than measurement. An elementary consideration of their combined,

AMS-PETROFABRIC OF DEFORMED ROCKS

305

Table 2. Selected susceptibilities and anisotropy of dominantly paramagnetic ( + ve) or minerals that are expected to be perfectly diamagnetic (—ve) when pure. (These are not to be used as standard reference values, since most matrix minerals contain ferromagnetic inclusions. All paramagnetic minerals also possess a diamagnetic response.)

Biotite Biotite Biotite Phlogopite Phlogopite Chlorite Chlorite Muscovite Muscovite Clay minerals Epidote Epidote Hornblende Pargasite Orthopyroxene Orthpyroxene Serpentine Siderite Dolomite Tourmaline** Calcite** Quartz** Plagioclase Feldspar Magnetite* Hematite Maghemite* Pyrrhotite (Fe7Ss) Goethite**

n

AV M : uSI

30 10 12 12

1042 1502 1070 273 1180 552 490 140 110 300-450 639 968 903 59 500-5000 10485 10485 3980 40 1690 -13.6 -13.4 -14.0

41 8 19 9 23 5 9 2 5 4

Pj

T

Source

106 240

1.72

0.15

135 447

3.36 1.14 3.77

0.69 0.37 0.07

347

1.27

0.56

BW Z M BW Ba BW M BW M HMB BP L F L RJA L L RJA B RJA, B RJA H B H H H H H

K F : |iSI

Tj

Pj

1.38

0.62

1.34 1.54 1.31 1.19 1.30 1.27 1.49

0.88 0.78 0.95 0.28 0.63 0.56 0.68

1.33

0.11

1.14

0.53

*

* *

*

834 65

71.30

7

8485 75700 7

1.56

1.01 1.11 1.01 <1.01?

-0.9 7 1.0 1.0 1.0 7 2.0-5.7 x 106 <40000 2.0-2.5 x 106 <3.2 x 106 1100-12000

KM = for rock-forming (matrix) minerals; paramagnetic (+ diamagnetic) susceptibility for BW data, otherwise low field susceptibility from all sources, including any 'ferro'-magnetic contamination. Kf = 'ferro'-magnetic susceptibility component, due to inclusions in the silicates and carbonates listed, where determined by alternating gradient force magnetometer. = may increase with grain size and internal stress, AMS inverse for SD due to shape, not crystallographic control. ** = AMS is crystallographically inverse n = different sources of mineral, where known. Italic = inverse fabric; |& MAX |// or closest to long crystal habit ( / / c in quartz , calcite & tourmaline), see Fig. 1. B = Borradaile (1987); Ba = Borradaile (1994); BW - Borradaile & Werner (1994); F = Friederich (1995); H = Hrouda 1986; L-Lagroix & Borradaile (20000); RJA = Rochette et al. (1992); Z = Zapletal (1990). M = Martin-Hernandez & Hirt (2003) provide the most complete AMS data, in particular high-field paramagnetic tensors for biotite, chlorite and muscovite examples. HMB = Hunt et al. (1995) provide a recent comprehensive summary, mainly for bulk susceptibility.

opposing magnetic contributions shows that AMS orientations and magnitudes will be highly sensitive to idiosyncratic specimens (Fig. 3a, b). This commonly arises in a calcite matrix with some mica grains arranged with their respective crystal axes as shown in Figure 3(c,d), see Hamilton et al., this volume. Despite the long habit of euhedral quartz and calcite crystals //c (Figs. 1, 3), anhedral matrix crystals align with c at a high angle to S, under many deformation mechanisms. Although quartz and calcite are trigonal minerals, the precision of

AMS measurements has not yet confirmed their AMS is described by a uniaxial ellipsoid but here we shall accept that possibility. The use of 'maximum' and 'minimum' with negative principal susceptibilities is confusing (see Hrouda, this volume for further discussion), so we adopt the crystallography convention of describing the (apparently) unique susceptibility axis as the extraordinary principal susceptibility ( k e / / c ) ; the possibly equal susceptibilities in the basal plane (J_ to c) are the ordinary susceptibilities (ku) (Fig. 3c, e). For pure calcite and quartz, ke

306

G. J. BORRADAILE & M. JACKSON

Fig. 2. (a) K increases with magnetite-concentration and in some igneous and plutonic rocks it has been implied magnetite controls « (Balsley & Buddington 1960). However, the range of K permits outliers to give considerable spurious weight and logarithmic scaling conceals a non-linear relationship. Such observations fostered the early belief that AMS was mostly due to magnetite, (b) In low-ft rocks the number of paramagnetic/magnetite mineral grains may be too few to define a meaningful petrofabric. (c) Grain-size and the strength of the OD determine whether or not AMS is faithfully sampled by a standard-sized specimen, (d) A weak location fabric (Turner & Weiss 1963) of isotropic magnetite grains ('supergrains') may produce an AMS due to magnetic interactions (Fuller 1961).

is the most negative principal susceptibility (Fig. 3c) whereas in paramagnetic ferroan calcite it is the greatest positive susceptibility (Rochette et al. 1992). Available approximate measurements suggest the following values for pure quartz and pure calcite; some reports determine (ke — ku) from torque measurements, others determine principal values directly: Calcite A:,, = -14.4 fcw =-12.9 (e.g. Nye 1957; Owens & Rutter 1978). Quartz £, = -13.7 fcw =-12.5 (e.g. Hrouda 1986 & this volume). For interpretation and modelling work, measurements of synthetic specimens are more useful. For example, a simple arithmetic mean of natural quartz specimens, some of which are

inevitably contaminated, may yield the misleading conclusion that quartz is magnetically isotropic (e.g. Hrouda 1986, Table 1; but see Hrouda, this volume). Having understood that AMS-petrofabric relationships involve crystallography, magnetic responses and rock-composition, we must now address the role of the crystal-aligning process. Finite strain is the end-result of a deformation history that incorporates many overlapping or sequential fabric-forming events. Discrete subfabrics may form at different stages, by different mechanisms, involving different mineral and/or grain-size fractions. Since Hrouda's (1982) landmark review, attention has been focused on isolating mineral contributions to AMS, and partitioning AMS between distinct subfabrics

Fig. 3. (a), (b) In weakly susceptible limestone (K < + 100uSI), the intrinsic inverse diamagnetic (K < 0) anisotropy of calcite may be neutralized by very low concentrations of paramagnetic minerals (K > 0). Their blended AMS may have very unstable axial orientations (see Hamilton et al., this volume). (c,d) competition of crystallographic control on AMS due to the common orientations of calcite (in matrix) and clay/mica, in typical sympathetic orientations, i.e. calcite c axes normally align normal to phyllosilicate basal planes. Note that in calcite (and in quartz) the euhedral habit shown is not expressed in rock matrices, (e) Similar effects might also be due to matrix quartz in quartz-rich rocks, e.g. tonalite and quartzite. (f) Although isometric, magnetite has a crystalline magnetic anisotropy but, due to its high intrinsic susceptibility, grain-shape is more important. AMS thus varies with grain-shape from nearly rounded clasts through to needle-like inclusions, but also with grain-size, (g) In multidomain magnetite, maximum induced magnetization defines KMAX along the maximum dimension, parallel to maximum magnetization (^MAX)(h) However, single domain grains are spontaneously saturated along their long dimension so that during AMS measurement, KMAX appears along the short dimension, giving rise to an inverse grain-shape. [In contrast, the inverse AMS of calcite, quartz and goethite is an intrinsic crystallographic property, shared by few other minerals.]

308

G. J. BORRADAILE & M. JACKSON

(Henry 1983, 1989; McCabe et al 1985; Rochette 19870; Jackson et al. 1988; Rochette & Pillion 1988; Borradaile et al 1990, 19990; Hrouda & Jelinek 1990; Jackson 1991; Johns & Jackson 1991; Ellwood et al. 1993; Housen et al. 19930; Borradaile & Werner 1994; Hrouda et al. 1997; Lagroix & Borradaile 20000; Borradaile & Gauthier 2001; Martin-Hernandez & Hirt 2001, 2003; Hrouda 20026; Kelso et al 2002; Ferre et al 2004). Inevitably, high-susceptibility inclusions make it difficult to determine K or AMS for natural rock-forming minerals. This explains discrepancies between reports for the 'same' mineral. The widely varying stoichiometry of many rock-forming 'pure' silicates requires ranges of values to describe their AMS or K anyway. Thus, values quoted for rock-forming minerals are merely examples, not estimates of intrinsic crystal properties (e.g. Table 2). Physical basis of magnetic fabric characterization Before discussing in more detail the characterization methods and interpretation of magnetic fabrics, it is useful to begin with a concentrated review of salient physical principles. For much more detailed coverage, the reader is referred to standard texts (Nagata 1961; Stacey & Banerjee 1974; O'Reilly 1984; Fuller 1987; Tarling & Hrouda 1993; Dunlop & Ozdemir 1997; Tauxe 1998), from which the following summary was distilled.

about -5 x 10~9 and -8 x 10~ 9 m 3 /kg); bismuth is an extreme case, with KJ ~ —165 uSI. Paramagnetism Paramagnetism is related to the permanent magnetic moments of certain atoms with unpaired electrons, which become partially aligned in an applied field. Atomic moments (p) have electron-spin and electron-orbital components that are quantized in multiples of /?B (the Bohr magneton, equal to 9.274 x 10~24 Am 2 ): where h is Planck's constant. The transitionmetal ions Fe2+, Fe3+ and Mn2+ are the major sources of paramagnetism in minerals; only rarely are magnetic ions of Ni, Cr, Co or rare earths sufficiently abundant to contribute significantly. In a crystal lattice the electron orbitals are affected by an electrostatic crystal field, and the electron-orbital components are suppressed ('quenched'); the net moments come primarily from electron spins. For Fe2+, p is about 5ApE and for Fe3+ and Mn 2+ , p is close to 6/?B. An applied field H tends to align the atomic moments; this is opposed by randomizing thermal energy. In classical theory, the balance is given by the Langevin function, based on Boltzmann thermodynamics:

where n is the number of magnetic ions per unit volume,

Classes of magnetic behaviour and the approximation is valid for a < 1. Thus

Diamagnetism Diamagnetism is a linear function of applied field (M = K.H), with K small, negative, and essentially independent of temperature and applied field strength. It is a property of all matter, related to the precession of electron orbits in an applied field. The classical Langevin expression relates diamagnetic susceptibility to the mean square orbital radius {r2}, the atomic number Z (i.e. number of electrons per neutral atom), and the number n of atoms per unit volume:

The Curie constant C (= n/j,Qp2/(3kE)) is a measure of the concentration and 'strength' of magnetic ions in a material. At room temperature (~300K), when these magnetic ions are present in non-negligible quantities (~0.5% or more by mass) their positive susceptibilities outweigh the diamagnetic susceptibilities of the bulk material. Syono's (1960) formula is equivalent to

The additional physical constants in the expression are the permeability of free space ^0> and the electron charge e and mass me. Because n is inversely related to both (r2) and Z, the product is nearly constant: for most materials Kd is between about -10 and -20uSI (xd between

where />Fe2+ w 5.3/?B is the moment per Fe2+ ion, .xFe2+ is the number of Fe + ions per kilogram of sample, and so on for the other ions (/?Fe3+ ~/?Mn 2 + ~ 5.9/?B). Similarly, Collinson (1983) gives susceptibility contributions in

AMS-PETROFABRIC OF DEFORMED ROCKS

terms of the concentrations of the oxides of Fe2+ and of Fe3+ in a specimen at room temperature:

309

Thus

and

where /(FeO) is the mass fraction of ferrous oxide and /(Fe2O3) is that of ferric oxide. Measured K values for selected minerals are given in Table 2. Because thermal fluctuations disrupt the alignment of atomic moments in an applied field, paramagnetic susceptibility diminishes with increasing temperature (Fig. 4). Equation (5) applies to 'pure' paramagnets, where the atomic moments are completely independent of one another. The behaviour is modified in most paramagnetic minerals by weak exchange coupling; this is known as Curie-Weiss paramagnetism. The coupling is equivalent to interaction through a molecular field, whose intensity is proportional to the magnetization of the material Hm = XM.

Fig. 4. Susceptibility has a characteristic temperaturedependence for antiferromagnets and for disordered materials (including ferro-, antiferro- and ferrimagnets above their ordering temperatures, as well as for paramagnets). K is temperatureindependent for diamagnets and for metallic ('weak') paramagnets; it varies inversely with (T — 0) for Curie-Weiss paramagnets and for ferro-, antiferroand ferrimagnets above their ordering temperatures.

9 is called the paramagnetic Curie temperature or the Weiss temperature; it is a measure of the nature and strength of exchange coupling (Fig. 4). A different form of paramagnetism is called Pauli paramagnetism or weak paramagnetism. It is due to conduction electrons in metallic materials, and among natural materials it is important only in some iron sulphide minerals (e.g. pyrite (Miyahara & Teranashi 1968; Tebble & Craik 1969)). The susceptibility of these materials is weak and independent of temperature. Ferromagnetism (sensu lato) In ferromagnetic materials the atomic magnetic moments are systematically ordered. In a true ferromagnet (e.g. iron) adjacent moments are exchange-coupled in parallel orientations; there is consequently a strong net spontaneous magnetization. On a larger scale, magnetic domains may form, with different magnetization orientations, but within each domain the atomic moments are all parallel. True ferromagnets are exceedingly rare in rocks and sediments Although it is also magnetically ordered, a pure antiferromagnet has no net spontaneous magnetization, because atomic moments are arranged in antiparallel sublattices (or other more exotic arrangements, e.g. helical antiferromagnetism in some rare earths). Impure antiferromagnets have a parasitic spontaneous magnetization, either due to imperfect antiparallelism of sublattices ('canted antiferromagnetism', e.g. hematite) or due to imperfect sublattice cancellation resulting from lattice irregularities ('defect moment', e.g. goethite). A special class of antiferromagnetism is ferrimagnetism, in which a moderately strong spontaneous magnetization results from a structurally controlled imbalance of sublattices (e.g. magnetite, maghemite, pyrrhotite, greigite). For magnetically ordered materials K(T) is generally not an intrinsic property, but depends on grain size, structure, etc. Above their ordering temperatures (Curie temperature Tc for ferromagnets; Neel temperature Jn for antiferromagnets), these materials become magnetically disordered (paramagnetic), and K(T) follows the Curie-Weiss law, with 9 = Tc > 0 for

310

G. J. BORRADAILE & M. JACKSON

Fig. 5. Magnetocrystalline anisotropy arises from the nature of the effective interactions between magnetic ions as a function of orientation. For this actinolite sample, there is a weak antiferromagnetic coupling in the a axis direction (0a = -3.5 K, xa (290K) = 9.7 x 10~8 m /kg), and weak ferromagnetic couplings along the b (0b = +3.6K, Xb (290 K) = 10.8 x 10~ 8 m 3 /kg) and c axes (0C = +3.3K, xc (290K)= 10.7 x 10~ 8 m 3 /kg) (Friederich 1995).

ferromagnets and 0 < 0 < Tn for antiferromagnets (Fig. 4). Ferrimagnets have a more complex behaviour above their ordering temperatures, with I/K increasing nonlinearly.

Grain-scale anisotropy With the common exception of ferrimagnetic saturation magnetization, all magnetic properties are anisotropic for individual grains of almost any mineral, due to crystallography, stress and/or grain shape. Magneto crystalline anisotropy Magnetic ions in a crystal lattice interact with each other and with neighbouring nonmagnetic atoms, through exchange coupling and electrostatic interactions. The nature and strength of these interactions varies systematically with orientation in the lattice. A detailed treatment is beyond the scope of this paper; the reader is referred to standard texts (Chikazumi & Charap 1964; Stacey & Banerjee 1974; O'Reilly 1984; Fuller 1987; Dunlop & Ozdemir 1997). For paramagnets, the Weiss temperature Op is a measure of the nature and strength of the interactions between magnetic ions. Values greater than or less than zero respectively indicate positive ('ferromagnetic') or negative ('antiferromagnetic') interactions. The Weiss temperature is directly related to the molecular field coefficient A (eq. 8) and should therefore generally be expected to be anisotropic. The Curie constant C is a product of non-directional scalar quantities (eq. 5) and might thus be

expected to be isotropic, but experimentally it is found to vary with orientation in various minerals (e.g. Rochette et al, 1992); this anisotropy is due to preferred electron-orbital orientations for Fe2+ in the crystal lattice. In general, therefore, the room-temperature AMS of individual paramagnetic crystals is closely related to directional differences in both C and 0p, and in the latter case paramagnetic anisotropy increases at low temperatures (Fig. 5; see also Rochette et al, 1992; Friederich 1995; Pares & van der Pluijm 2002). In magnetically ordered minerals, magnetocrystalline anisotropy coexists at the grain scale with magnetoelastic and shape-related magnetostatic anisotropies. Magnetocrystalline anisotropy is generally most important in antiferromagnetic minerals (e.g. hematite and goethite) and in ferrimagnets with relatively low crystal symmetry (e.g. monoclinic pyrrhotite), and it is less significant (though not negligible) in the strongly magnetic cubic minerals (e.g. magnetite and maghemite). Here we will briefly consider two illustrative cases: magnetite and hematite. The cubic lattice body diagonals (111) are the minimum-energy directions of magnetization for magnetite; in the absence of other anisotropies and of applied fields, the magnetization orients itself in one of these directions. Energy varies with orientation in a cubic crystal with volume V according to

AMS-PETROFABRIC OF DEFORMED ROCKS

where KI and K2 are the anisotropy constants ([J/m3]) and the as are direction cosines with respect to the cubic axes. An applied field perpendicular to the easy axis rotates the sublattice and net magnetizations into an equilibrium orientation with higher magnetocrystalline energy but lower magnetostatic field energy; the induced magnetization corresponds to

which for magnetite corresponds to about 20 SI. This is one component of the intrinsic susceptibility; the other major component is domainwall translation, which dominates intrinsic x/f for multidomain grains. Measured external susceptibilities for magnetite are significantly less than the intrinsic susceptibilities, due to self-demagnetization, which we discuss in the following section. Hematite has a corundum crystallographic structure, with hexagonal basal-plane symmetry. At temperatures below the Morin transition (r M ~260K) it is a 'pure' antiferromagnet, with the Fe3+ atomic moments aligned parallel to the easy c axis, and alternating polarity in successive basal-plane layers. Like the magnetite case, an induced magnetization may be generated normal to the easy axis by application of a perpendicular field, but unlike the ferrimagnetic case the sublattice magnetizations do not rotate together to turn a net spontaneous magnetization toward alignment with the field. Instead, the applied field has to disturb the perfect antiparallelism of the layered sublattices, in opposition to the antiferromagnetic exchange coupling, and rotate each sublattice separately toward the field orientation; consequently x± is very weak (~240uSI; Stacey & Banerjee 1974). The spin-parallel c-axis susceptibility x// *s even weaker: because there is no angular gradient of energy, an applied field has no 'leverage' to rotate the atomic moments and x// = ® (at least at OK, where the atomic moments are perfectly aligned; at higher temperatures thermal energy randomly misaligns the atomic moments, and x// increases linearly with temperature as in Fig. 4). In the 'pure' antiferromagnetic state, Xj_ > X//> and grain-scale anisotropy increases at lower temperatures. Similarly, goethite's antiferromagnetic sublattice moments are aligned with the c axis, which is therefore the minimum susceptibility orientation (Hedley 1971; Ozdemir & Dunlop 1996). Because goethite's crystal habit is elongate parallel to c; its AMS is 'inverse' but alignment mechanisms are unknown.

311

On warming through TM, hematite's atomic moments abruptly rotate into the basal plane ('spin flop'), but retain the sublattice structure with ferromagnetic coupling within layers and antiferromagnetic coupling between layers. The antiparallelism is imperfect, and the 'spin canting' results in a net basal-plane spontaneous magnetization (or 'parasitic ferromagnetism'), nearly orthogonal to the sublattice magnetizations. There are thus three orientations to consider. Maximum susceptibility is in the direction parallel to the atomic spins (i.e. in the easy basal plane and normal to the spontaneous magnetization); as in the ferrimagnetic case an applied field can rotate the sublattice magnetizations and the net spontaneous magnetization coherently against the moderate basal-plane anisotropy (which is a combination of hexagonal magnetocrystalline and uniaxial magnetoelastic components). This is the dominant contribution to the observed susceptibilities of hematite (500-40 000 uSI; Hunt et al. 1995). Susceptibility in the direction of the spontaneous magnetization is much lower: dM/dH has no parasitic ferromagnetic component and is solely due to opposite rotations of sublattice magnetizations (increasing the canting angle) against the exchange coupling. This is comparable to x± in the pure antiferromagnetic state below TM (~240jiSI). Finally, along the c axis, both mechanisms operate (albeit weakly): rotation of the sublattice moments and the weak parasitic ferromagnetism out of the basal plane is opposed so strongly by exchange forces and by the hard c-axis anisotropy that the spontaneous magnetization is considered to be restricted to the basal plane in all experimentally applied fields (see e.g. Rochette & Pillion 1988). Monoclinic pyrrhotite is similar (Clark 1984; Rochette & Pillion 1988). For most minerals, there is some correspondence between crystal habit, crystallography and AMS (Fig. 1); in general, crystallography determines both AMS and grain shape. Chain and sheet silicates are important rockforming minerals that mostly form elongate and platy grains, respectively, and are mostly paramagnetic. Generally, they have ^MAX subparallel to their long dimension and AVMIN a^ a high angle to a c-basal plane. This approximation suffices for petrofabric interpretations with most chlorites, clays and micas but the shapesymmetry-AMS angular relationship must be considered individually for each pure mineral. Moreover, there are three significant complicating factors. First, 'inverse' anisotropy is found in some paramagnetic minerals. Tourmaline, for

312

G. J. BORRADAILE & M. JACKSON

example, is c-elongate but the c-axis is characterized by antiferromagnetic interactions and KMW, and cordierite may also have inverse AMS (Rochette et al. 1992). Unfortunately, only a few studies relate AMS to crystallography and composition for common rock-forming, stoichiometrically variable matrix minerals such as amphibole and phyllosilicates (Friederich 1995; Martin-Henandez & Hirt 2003). Second, and more generally, there are precise angular relationships between AMS axes and crystal axes for pure minerals. AMS and crystal axes are parallel, though not necessarily with any particular one-to-one mapping for the high-symmetry tetragonal, hexagonal, trigonal and orthorhombic systems. However, many rock-forming minerals are monoclinic, so that only one AMS axis is parallel to the Z?-crystal axis. For triclinic minerals (e.g. plagioclase) all AMS axes are inclined to crystal axes (Fig. 1). The AMS-crystal angular relationships for pure minerals in the monoclinic and triclinic systems depend on chemical composition. Finally, ferrimagnetic inclusions are common in nature and their role in particle-level anisotropy may be of paramount importance (Borradaile & Werner 1994; Johns & Jackson 1991, Borradaile 1994; Cottrell & Tarduno 1999; Lagroix & Borradaile 2000a; Renne et al. 2002). In some cases, the orientations and/or spatial distribution of these inclusions (and thus the grain-scale AMS) is controlled by the crystallography of the host mineral, but that is not always the case.

Self-demagnetization and shape anisotropy Magnetic properties and behaviour are strongly controlled by the boundary conditions at the surfaces of grains composed of highly magnetic materials (e.g. magnetite, maghemite, pyrrhotite). The effect is usually explained in terms of magnetic 'poles' that form at the discontinuity terminating the magnetization, in the same way that electrical charges form on the terminating surfaces of a polarized material. These 'poles' generate a demagnetizing field within the material, proportional and antiparallel to the magnetization: The proportionality constant N is called the demagnetizing factor. The net internal field in a grain exposed to an external field H0 is then Measured susceptibility KO (=M/H0) is thus related to intrinsic susceptibility «, (=M/Hi) through the demagnetizing factor:

For weak intrinsic susceptibilities (/^ 1), the observed susceptibility reaches the 'self-demagnetization limit' of I/TV (Fig. 6). TV is a tensor with a mean value in SI units of 1/3 (for any 3 orthogonal axes N{ + TV2 + TV3 = 1). Thus high intrinsic-susceptibility minerals like magnetite have mean external susceptibilities of around

Fig. 6. Self-demagnetization becomes significant for intrinsic susceptibilities greater than about 0.1 SI, resulting in diminished externally-observed susceptibilities. Observed susceptibility reaches a limiting value of I/TV for high intrinsic susceptibilities.

AMS-PETROFABRIC OF DEFORMED ROCKS

3 SI (increasing with shape anisotropy, as described below). M and Hd are in general nonuniform in arbitrarily shaped grains, but they are uniform in ellipsoidal grains. Stoner (1945) calculated TV for uniaxial ellipsoidal grains (i.e. grains having rotational symmetry about a unique axis). For such grains the shape can be described by the dimension ratio 0 = dp/de = polar axis/equatorial axis. The demagnetizing factor along the unique polar axis is:

313

easy axes, even in the absence of an applied field, it has essentially zero low-field susceptibility parallel to its long dimension. Weak perpendicular fields, however, are able to rotate the grain moment slightly away from the easy axis, and thus there is a finite susceptibility normal to the easy axis (Fig. 3h). This 'single-domain effect' (Potter & Stephenson 1988) is one of the most important mechanisms producing 'inverse AMS' in natural materials. Anisotropy of remanence, in contrast is 'normal' (i.e. the shape of the magnetic ellipsoid resembles that of the grain) for SD as well as for MD grains (Fig. 3g; Stephenson et al 1986).

Anisotropic interactions in grain ensembles

Because of the rotational symmetry, Np 4- 2Ne = 1, so Ne = (1 - Np)/2 follows immediately when Np is calculated by the above. Limiting cases are (Np,Ne) = (0,0.5) for needle-like grains, (1/3,1/3) for spheres and (1,0) for flat disks (Fig. 7a). For high intrinsic susceptibilities, the observed directional and mean susceptibilities change systematically with grain shape; the self-demagnetization limit in SI units is (ftp,« e ) — (^i>2) for needles, (3,3) for spheres, and (1,/q) for disks (Fig. 7b). The grain-scale AMS ratios due to self-demagnetization thus depend on both intrinsic susceptibility and grain shape. Conveniently, for dimension ratios near unity (0.1 < j3 < 10) and sufficiently high Ac is the AMS ratios are roughly equal to the dimension ratios (Fig. 7c). Remanence is almost always more anisotropic than susceptibility. For example, TRM theory for MD grains (Stacey & Banerjee 1974) predicts

where the approximations are valid for intrinsic susceptibilities of K^-IO2 SI or greater (e.g. magnetite) (Stephenson et al 1986; Cogne 1987). SD grains in weak applied fields are an interesting and important special case. Because an SD grain is magnetized to saturation along its

Magnetostatic interactions among neighbouring ferrimagnetic grains may be anisotropic if the grains are not uniformly distributed throughout a dia-/paramagnetic matrix; such spatial distributions are referred to as 'location fabrics' (Turner & Weiss 1963) and the resulting AMS has been termed 'distribution anisotropy' (Margraves et al. 1991). A grain with magnetic moment p produces a dipole field h that locally modifies an externally applied field H. In the polar coordinate system of the dipole, the field at (r, 6) due to p is:

Note that since p = MV = TtMd3/6, interaction fields scale as M(d/r)3, where dis grain diameter. In the equatorial plane of the dipole (r, 90°), h is antiparallel to />; along the dipole axis (r, 0°) h and p are parallel. Thus for an interacting collinear ensemble (a chain of grains), when a field is applied parallel to the chain axis, the induced magnetizations of the grains are mutually reinforcing, i.e. the interaction fields are parallel to the applied field. The effective susceptibility along the chain axis is therefore larger than that of the same ensemble when dispersed (without interactions). Conversely, when a field is applied perpendicular to the chain axis, the induced moments produce fields that oppose the applied field and diminish the effective susceptibility. Thus, to a certain extent, a cluster of individual grains behaves like a single 'supergrain' having the same shape as the cluster (Fuller 1961). Stephenson (1994) used simple dipole-dipole calculations to model interactions in linear and planar arrays of isotropic spherical particles with K{ > 1, and found that interactions become significant when d/r > 0.5 (i.e. when

314

G. J. BORRADAILE & M. JACKSON

Fig. 7. Shape anisotropy results from differences in self-demagnetization in different orientations, (a, b) for uniaxial prolate or oblate spheroids anisotropy is governed by the demagnetizing factors along the polar (Np) and equatorial (Ne) axes, which depend on the dimension ratio (3 = dp/de. (c) For low to moderate dimension ratios (0.1 to 10), the susceptibility ratio is approximately equal to the dimension ratio or its reciprocal (i.e., L = Kp/Ke ~ dp/de and F = KC/KP ~ de/dp).

the centre-to-centre separation between particles is less than about twice the particle diameter and the edge-to-edge gap between particles is less than about one diameter). Maximum anisotropy, when adjacent particles touch, was P = L = 2.5 for linear 'supergrains' and P = F = 2.0 for

planar ensembles (Stephenson 1994). These are in fact underestimates, since the dipole-dipole approximation breaks down at close spacings, and since only nearest-neighbour interactions were considered. Similar geometric groupings were considered by Canon-Tapia (1996), who

AMS-PETROFABRIC OF DEFORMED ROCKS

extended the model to anisotropic grains. General confirmation of the model results was found experimentally by Gregoire et al. (1995). The increase in susceptibility along the chain is larger than the perpendicular decrease, so the mean susceptibility (i.e. the average of the values measured in any three mutually orthogonal directions) increases as a result of the interaction, in much the same way that mean susceptibility grows with increasing shape anisotropy. In Stephenson's (1994) model, mean susceptibility increases by up to 20% and 10% respectively for linear and planar arrays of grains (approximately simulating L and S fabrics). Interestingly, the generally constructive effects of interactions on susceptibility (increasing K) contrast strongly with the negative average effects of interactions on remanence. For weakfield remanences (e.g. ARM and TRM) especially, acquisition efficiency (intensity per unit mass of remanence carriers) is depressed significantly by magnetostatic interactions (Dunlop & West 1969; Sugiura 1979; Yamazaki & loka 1997; Fukuma 2002). The strong-field IRM acquisition process for SD ensembles is also on average opposed by magnetostatic interactions (Henkel 1964; Cisowski 1981; Muxworthy et al. 2003). No controlled experimental work to date has dealt with evaluating the effects of interactions on AARM or AIRM; the first theoretical treatment is in Muxworthy and Williams (this volume).

Anisotropic (second-rank tensor) bulk magnetic properties With the common exception of ferrimagnetic saturation magnetization, all magnetic properties are anisotropic on the bulk rock scale, due to preferred orientation of anisotropic grains and/or to anisotropic interactions among strongly magnetic grains. The simplest case is when magnetostatic interactions can be neglected and grain-scale anisotropy is uniaxial (i.e. it has rotational symmetry around a unique easy or hard axis) and can be described by a secondrank tensor. Then the anisotropy of the rock is also described by a tensor, depending only on the properties and orientation distribution of the constituent grains (Owens 1974):

where [tt]grain(#5 0) is the tensor, in the rock coordinate system, for an individual particle whose easy axis is oriented in the direction (0,0). For

315

the more general case of triaxial grain-scale anisotropy, the unique-axis OD /(0, 0) is replaced by an orientation tensor. Weak-field susceptibility (a.c.fd.c.) Although astatic and fluxgate spinner magnetometers were widely used historically, the magnetization induced in weak fields (typically <103 A/m) is now most commonly measured by a.c. induction instruments, with frequencies typically between 102 and 104Hz. For conductive materials (e.g. graphite, sulphides), the a.c. signal contains a complex electromagnetic eddy-current response in addition to the magnetic susceptibility response, especially at higher frequencies (e.g. Vincenz 1965; Worm et al. 1993). Anisotropy of the conductive response is related to shape and/or crystallographic orientation of the conductive particles (Borradaile et al. 1992; Ellwood et al. 1993). The d.c. susceptibility can be measured by SQUID magnetometers with precision and speed comparable to those of a.c. measurements (Schmidt et al. 1988). AMS can be determined from sets of directional susceptibility measurements using these instruments, or by weak-field torque measurements using a torsion magnetometer. Low-field (<10 3 A/m) induced magnetization is linear (M = K,H) for strong, soft ferrimagnets with susceptibilities near the self-demagnetization limit I/TV ~ 3 SI (e.g. magnetite), as well as for dia-, para- and pure antiferromagnets. Nonlinear behaviour of multidomain titanomagnetites, pyrrhotite and hematite begins around 101-102A/m, i.e. within the operating range of many modern instruments. In fields up to a few hundred A/m the behaviour of these materials can be reasonably described by Rayleigh's quadratic law:

where a is known as the (second) Rayleigh coefficient. In a field oscillating between peak intensities of ±H0, Rayleigh behaviour is characterized by parabolic minor hysteresis loops of the form:

where the negative sign applies to the ascending branch of the minor loop (H increasing from —H0 to +HQ). A.c. magnetization in the Rayleigh regime follows precisely the same function; a.c. susceptibility measurements in this regime are therefore characterized by both inphase and quadrature components, and by an applied-field dependence of the apparent a.c.

316

G. J. BORRADAILE & M. JACKSON

susceptibility,

Anisotropic Rayleigh magnetization is discussed in a subsequent section. It is important because of the nonlinear behaviour of the common minerals hematite, titanomagnetite and pyrrhotite in the a.c. fields typically used for AMS measurements (e.g. Pokorny et at., this volume). Weak-field torque (a.c./d.c.) The torque exerted on a specimen by a weak applied field is directly related to the anisotropy of the specimen:

That is, the torque is due to the non-parallelism of the induced magnetization and the applied fiejd. The_ third equality in this expression (M — [K]H), again, holds for most geological materials for H < 103 A/m, with the exceptions of titanomagnetites, hematite and pyrrhotite. Note that when magnetization is a linear function of applied field, torque is proportional to H2. High-field magnetization/susceptibility (d.c.) In the fields typically used in rock magnetic labs (IJLQH < 7 T; H < 5.5 MA/m), paramagnetic minerals remain far from saturation at room temperature (although they begin to approach saturation in these fields at lower temperatures). Therefore, practically speaking, room-temperature paramagnetic susceptibility is independent of applied field strength. The same is not true of ferrimagnetic susceptibility, which changes dramatically as a saturating field is applied or removed. For this reason, various types of susceptibility are defined (Fig. 8), and have different values according to situation with respect to the ferrimagnetic hysteresis loop (Chikazumi & Charap 1964). The total susceptibility (Ktot ~ M /H) is simply the ratio of magnetization to field; the differentialsusceptibility (Kdif — dM/d//) is the derivative of the initial magnetization curve; and the reversible susceptibility (ftrev = (dM/d//) rev ) is the slope of a minor loop (Fig. 8) generated by interrupting the initial magnetization process at some field HI, decreasing the field to H2 < H\, and then increasing again to larger positive fields; ttrev is effectively what would be measured by a weakfield a.c. instrument with a strong superposed d.c. field. In the weak-field regime below the onset of Rayleigh behaviour and other

Fig. 8. The dashed curve M shows the in-field magnetization of a previously-demagnetized ferromagnetic material as the applied field H is increased. Susceptibility may be defined in different ways. The differential susceptibility ttdif is the derivative of the M curve; total susceptibility is the ratio M/H; reversible susceptibility KKV is the slope of a minor loop. In weak fields these all equal the initial susceptibility, and when M is saturated, Kdi{ and Krev are zero.

irreversible processes, all three of these are equal to the initial susceptibility KQ. In fields strong enough to saturate the ferrimagnetic moment, «dif = ftrev — 0. In natural materials containing a mixture of dia-, para- and ferrimagnetic mineral grains, the differential and reversible susceptibilities in sufficiently strong fields are due only to the dia- and paramagnetic minerals. Imperfect antiferromagnets like hematite contribute both a parasitic ferromagnetic susceptibility in low fields and an approximately constant differential antiferromagnetic susceptibility in strong applied fields. The applied field intensity required for ferrimagnetic saturation depends on mineralogy, grain size (domain state), and grain-scale anisotropy. For SD magnetite with shape anisotropy, the critical field #K// required to reverse the moment of a single particle aligned with the field is The same field also suffices to rotate the moment of a perpendicular grain into alignment with the field, along the short dimension of the grain. For all intermediate orientations, the saturating field //s < ffja//' Thus an ensemble of SD magnetites, whether aligned or randomly oriented, will always reach saturation in fields

even with extreme planar shape anisotropy

AMS-PETROFABRIC OF DEFORMED ROCKS

((7Vp,7V e ) = (1,0)). Maximum acicular shape anisotropy ((Np,Ne) = (0,0.5)) gives saturating fields half as large (240kA/m, 300 mT). When shape anisotropy is weak and other anisotropies (magnetocrystalline or stress) dominate, saturating fields for magnetite are lower. MD grains of any magnetic species dominated by shape anisotropy also saturate in fields of the order NM$ (i.e. on the order of 105 A/m for magnetite). Hematite is a somewhat more complicated case, due to the combination of antiferromagnetism and parasitic ferromagnetism. The strong c-axis magnetocrystalline anisotropy keeps magnetization confined to the basal plane, except for reversible rotations which produce a field-independent susceptibility along the c axis. Anisotropy within the basal plane is much weaker, with a hexagonal magnetocrystalline component and a uniaxial magnetoelastic component. In a nonrandomly oriented ensemble of hematite grains, rotation of the parasitic ferromagnetism of each grain into the basal-plane orientation closest to the applied field produces an anisotropic saturation magnetization, usually in fields below 1 MA/m, along with an approximately isotropic high-field slope («dif = ^rev > 0).

High-field torque In sufficiently strong fields, the cubic ferrimagnetic minerals in a specimen are saturated, i.e. they are magnetized precisely parallel to the applied field, regardless of its orientation, and there is no directional variation in their magnetization intensity. There is still a torque, however, related to angular energy gradients associated with the anisotropy of the ferrimagnetic assemblage, i.e. r = -dE/d0. (Energy is minimized for each grain when magnetization is parallel to an easy axis.) High-field torque measurements on single crystals enable determination of their magnetocrystalline anisotropy. For an ensemble of ferrimagnetic grains with shape anisotropy, high-field torque is related to the average or effective demagnetization tensor of the ensemble. For saturating applied fields (H > Hs) the torque due to ferrimagnetic anisotropy is independent of field (dr/dH = 0). When an anisotropic paramagnetic ensemble is also present, its contribution to the high-field torque varies (as it does in weak fields) in proportion to H2, and measurements in different high fields thereby allows separation of ferrimagnetic and paramagnetic deviatoric tensors (e.g. Martin-Hernandez & Hirt 2001). Hysteresis and isothermal remanence Hysteresis loops measured in high fields (sufficient to saturate ferrimagnets) yield various

317

important pieces of information about a rock specimen: Saturation magnetization Ms is usually calculated as the intercept of a linear fit to the high-field part of a hysteresis loop. It is a direct measure of the specimen's ferrimagnetic content. When a single ferrimagnetic mineral (e.g. magnetite) is dominant (as shown, e.g. by thermomagnetic analysis), the volume fraction in the specimen can be calculated as /mgt,voi = ^s,specimen/^s,mgt and similarly the mass fraction is /mgt,mass = ^.specimen M,mgtFor cubic ferrimagnets Ms is isotropic, i.e. independent of measurement orientation. High-field slope /%f (reversible or differential) in fields greater than //s is due to dia-, paraand antiferromagnetic susceptibility, and is thus a tensor. Saturation remanence Mrs (or SIRM) is the magnetization that remains after the saturating field is removed. It is anisotropic (but not tensorial), and is determined by ferrimagnetic content, grain size (domain state), and grainscale anisotropy. When Mrs is measured as part of a hysteresis loop, usually only the parallel component of magnetization is measured. SIRM measurements made with spinner or SQUID magnetometers most commonly measure all 3 components of the vector. Coercivity (or coercive force) Hc is the reverse field required to change the sign of the magnetization after saturation in a positive field. It is an anisotropic property controlled by mineralogy, grain size (domain state) and grain-scale anisotropy. For MD material with dominant shape anisotropy,

and although neither Mrs nor Hc is properly considered a tensor, their ratio is closely related to the ferromagnetic susceptibility tensor. Various approaches have been used in measuring anisotropy of (S)IRM. Measurements can be made parallel and perpendicular to visible structural/fabric elements, and ratios computed, without reference to tensor calculations (Fuller 1963; Jackson & Borradaile 1991). For relatively weak d.c. fields, IRM may in some cases be accurately described by tensor mathematics, and principal axis directions and magnitudes may be computed from sets of directional measurements (Stephenson et al. 1986). And if, by assumption or by direct demonstration, IRM is taken to be a quadratic function of applied field, the Rayleigh coefficients may be treated by tensor methods (Daly & Zinsser 1973; Markert & Lehmann 1996).

318

G. J. BORRADAILE & M. JACKSON

Anhysteretic and thermal remanence These fall (together with chemical and depositional remanences, among others) under the classification of weak-field remanences, which can be acquired by even very hard magnetic ensembles in fields comparable to the surface geomagnetic field (~40A/m or 50 uT). In these magnetization processes, the work of overcoming energy barriers is accomplished not by the weak d.c. field, but by thermal and/or alternating-field energy, and the weak d.c. field simply imposes a slight bias on a process that would otherwise result in demagnetization. Like weak-field induced magnetization, ARM and TRM are commonly linear anisotropic (tensor) functions of applied d.c. field. TRM theory for both SD and MD grains predicts Mtr oc HQ/N^ in weak applied fields (see, e.g. Dunlop & Ozdemir (1997)). SD ARM theory (Egli & Lowrie 2002) also predicts linear weakfield behaviour; no theory yet exists for ARM in MD ensembles. For SD carriers, both ARM and TRM are observed experimentally to be very sensitive to magnetostatic interactions, which oppose the magnetization process, effectively stretching the acquisition curve out to much higher fields before saturation begins. For MD carriers, selfdemagnetization has the same effect. In general, nonlinearity is evident below ~10 3 A/m only for noninteracting SD ensembles, for which ARM and TRM may approach saturation in fields as weak as 102 A/m. Most of the time, however, ARM and TRM are linear functions of applied d.c. field, and can therefore justifiably be treated with tensor methods.

The matrix is symmetric (KJJ — Kjt). For remanent magnetizations, all three components of the magnetization vector can often be measured, and the elements of the susceptibility tensor can then be calculated directly; for example, after applying a field along the x axis (H = (// x ,0,0)) and measuring the resulting remanence (Mx,My,Mz\ direct calculation gives KXX = MX/HX, Kxy — My/Hx, and KXZ = MZ/HX (e.g. Stephenson et al 1986). For induced magnetizations, typically only the component parallel to the applied field is measured:

Torsion magnetometers measure the torque related to the transverse induced magnetization:

Tensorial representation and measurement/ computation of anisotropy A linear anisotropic magnetization (induced or remanent)^ is related to the magnetizing field by M = [K]H , where the K matrix is generally referred to as the susceptibility. In this section we deal only with linear tensorial magnetization; nonlinear properties are treated later. Fig. 9. In two dimensions, the tensorial induced magnetization vector M (square endpoints) describes an ellipse as the magnetizing field goes through the range of possible orientations. In general M is not exactly parallel to the magnetizing field, but is composed of parallel (M//) and transverse (Afj_) components. Only along the principal axes is ML = 0 and is M exactly aligned with the applied field. A/// (open circle endpoints) does not describe an ellipse; it varies as cos(2# — (/>).

AMS-PETROFABRIC OF DEFORMED ROCKS

Measurements are typically made in coplanar sets for experimental convenience; this also simplifies the mathematics. For example, for a set of measurements in the xy plane, with Hx = Hcos0 and Hy = Hsin9, the parallel magnetization component is:

Note that unlike the full magnetization vector, the parallel component does not describe an ellipse in two dimensions, but a double-lobed curve (Fig. 9). Similarly, for fields applied in different orientations in the xy plane, the perpendicular torque is:

The in-plane parallel magnetization data can be expressed as a Fourier series:

Comparison of (29) and (31) shows that the Fourier coefficients and tensor elements are related by:

Fourier analysis of each coplanar data set thereby yields an estimate of the two related

319

diagonal tensor elements and the related offdiagonal element, and the complete susceptibility tensor may thus be obtained by collecting measurements in three orthogonal planes. Torque data, however, do not provide sufficient information to calculate the complete susceptibility tensor; similar analysis yields the deviatoric susceptibility tensor, and a separate determination of mean susceptibility is required. In practice, although data are most often collected in orthogonal planar sets, Fourier analysis is routinely used only for torque data, and susceptibility data are treated by the least-squares methods of Hext (1963) and Girdler (1961) (see also Borradaile & Stupavsky (1995) and Tauxe (1998)). For any arbitrary set of n applied field orientations, a 'design matrix' A (n x 6) relates the corresponding array of parallel directional susceptibilities K// (n x I) to the 6 independent tensor elements:

In the inverse problem, where we have n measured values of directional susceptibility, the best-fit values for the tensor elements are given by (A T A)~ 1 A T K//. Eigenvalues and eigenvectors are then calculated by iterative diagonalization methods (e.g. Press 1992). At least six measurements must be made, over a suitable range of orientations(, to resolve the six tensorelements. Schemes in common use include n = l (Borradaile & Stupavsky 1995), 9 (Girdler 1961), 12 (Hext 1963), 15 (Jelinek 1976), or more directional measurements. Any such scheme must use a broad distribution of measurement orientations (e.g. Owens 2000&), and should strike an optimum balance between minimizing random errors (by increasing n) and minimizing drift-related errors (by decreasing n). The primary advantage of the least-squares approach, as opposed to calculation by Fourier analysis, lies in the calculation of uncertainties in the principal axis lengths and orientations (Hext 1963; Jelinek 1976, 1981; Tauxe 1998; Owens 2000a,6).

320

G. J. BORRADAILE & M. JACKSON

These linear tensor methods are strictly valid only when magnetization is a linear function of applied field. This is generally true for weakfield a.c. susceptibility, for which most commercial instruments use applied fields <500A/m. However nonlinear a.c. behaviour begins to occur in fields of K^-IO2 A/m for coarse-grained pyrrhotite (Worm et al. 1993), titanomagnetite (Jackson et al. 1998) and hematite (Kletetschka & Wasilewski 2002), so linearity may not hold for these materials in some instruments (Hrouda 20020). Similarly, weak-field remanences (ARM, TRM, DRM) are commonly linear functions of applied field for H < 100 A/ m. Strong-field differential susceptibility (in applied fields sufficient to saturate the ferromagnetic susceptibility of a specimen) is due to dia-, para- and antiferromagnetic components, and is therefore also linear and tensorial (e.g. Rochette et al. 1983; Rochette & Pillion 1988; Hrouda & Jelinek 1990; Kelso et al. 2002).

Nonlinear magnetizations and nontensorial anisotropy Rayleigh regime For material that follows the Rayleigh law, the IRM acquisition curve is initially parabolic (Mr - HQ). Daly & Zinsser (1973) formulated anisotropic IRM acquisition (in relatively weak d.c. fields H
effort required for measurements of Rayleigh anisotropy may therefore be justified by better accuracy, due to the 'amplified' anisotropy. In-field magnetization in the Rayleigh regime involves both the linear initial susceptibility and the quadratic Rayleigh coefficients. To our knowledge, the only study to date in which the linear and quadratic anisotropies were explicitly and separately determined is that of Markert & Lehmann (1996). They measured parallel and transverse components of in-field magnetization directly, using a specially constructed two-axis vibrating-coil magnetometer (which operates on the same principle as a vibrating-sample magnetometer (VSM), but which generates a timevarying flux through the pickup coils by vibrating the coils rather than the sample). They measured both initial magnetization curves and Rayleigh loops and calculated linear and quadratic coefficients for the parallel and transverse magnetizations, thereby directly determining the corresponding tensor elements. The same approach could be adapted to single-axis instruments (e.g. standard VSMs) by least-squares tensor fitting of the parallel-component linear and quadratic coefficients (thereby also allowing goodness-of-fit evaluation). An interesting result of the study was that the linear and quadratic tensors had a systematic quantitative relationship of the form atj = CK^. In other words, the initial susceptibility and Rayleigh ellipsoids were coaxial, but had different axial ratios, the latter being more anisotropic (PRayieigh ex: ^AMS)Both of these approaches allow empirical validation of both the Rayleigh law and the tensorial nature of the Rayleigh coefficients, by measuring in different applied fields. Most a.c. susceptibility bridges use a fixed amplitude, and AMS calculations using directional susceptibilities obtained from such instruments implicitly assume linearity. Because titanomagnetites, pyrrhotite and hematite are now known to behave nonlinearly in the fields used in many commercial instruments (>10A/m), Hrouda (20020) evaluated the effects of applying linear tensor theory to nonlinear anisotropic magnetizations, experimentally and through numerical models. He concluded that when magnetization is significantly nonlinear, it invariably results in large fitting errors in the least-squares tensor calculations (i.e. susceptibility does not vary with orientation according to equation (29)). Although the principal axis orientations that he obtained were essentially correct, corresponding closely to those of the initial susceptibility, the axial ratios were not. This conclusion is reinforced by Pokorny et al. (this volume). A practical corollary of these studies is that when fitting errors

AMS-PETROFABRIC OF DEFORMED ROCKS

are small, one may be assured that the behaviour is linear, even when measurements are made with only one a.c. field amplitude. Gyroremanence Consider an individual uniaxial SD grain in a strong alternating field H > 77k, with its easy axis inclined at an angle (say, 45°) to the field. When the AF is at its peak, the grain moment is parallel to the applied field, and as the field cycles down to zero the moment rotates to the grain's easy axis. As the field then builds up in the opposite direction, the moment continues to rotate in the same sense, eventually again aligning with the field. With continuing AF cycles, the magnetic moment continues rotating in a consistent sense, about an axis normal to both the AF and the grain easy axis. The forced continuous rotation of the magnetic moment results in a gyroscopic precession, equivalent to an effective d.c. field parallel to the rotation axis (Stephenson 1981<2,6; Edwards 1982a; Stephenson & Potter 1987). In a randomly oriented ensemble of such grains the effects are mutually cancelling, but in an aligned ensemble a gyroremanence (GRM) may be acquired, perpendicular to both the AF and the grain alignment axis. Gyroremanence is closely related to rotational remanence (RRM) (Stephenson 1976; Edwards 19826, 1986). Because GRM occurs even in the absence of an applied d.c. bias field, it can be a serious source of error in palaeomagnetic studies using AF demagnetization (Bankers & Zijderveld 1981). When a bias field is applied (parallel to the AF), the resulting remanence consists of a parallel ARM component and transverse superposed ARM and GRM components; for this reason ARM anisotropy studies should only use the parallel remanence component, as originally proposed by McCabe et al (1985) (see also Trindade et al. 2001). For an ensemble of uniaxial particles, oriented with an ellipsoidal OD having principal magnitudes nx > ny > nz, Stephenson (198la) calculated GRM magnitudes as:

(36)

where 7 is a constant for each peak AF intensity, and clearly Q + C2 + C3 = 0. GRM can be used to characterize the magnetic fabric of geological specimens, by repeated AF treatment in different orientations and by iterative forward-modelling calculations, using the relationships above, to find the principal axis orientations (Stephenson

321

198 la). Like torque measurements, GRM measurements are sensitive to the differences between principal values, rather than to the values directly. Greigite has an unusually strong penchant for acquiring gyroremanence (Snowball 1991 a, b; Stephenson & Snowball 2001). This is useful both for identifying its presence and characterizing its preferred orientation. Strong-field remanence Although manifestly nontensorial, strong-field IRM usually has an easily detectable and geologically meaningful anisotropy. The primary disadvantage of the inapplicability of tensor methods is that maximum and minimum magnitudes cannot be accurately predicted from sets of directional measurements in a coordinate system that is arbitrarily oriented with respect to the 'principal' axes. Often this is not a problem, as values can be readily measured parallel and perpendicular to obvious elements such as bedding or schistosity. An advantage of high-field IRM is that it reflects the anisotropy of high-coercivity remanence-carriers such as hematite, and AIRM is therefore useful for evaluating possible deviations of NRM from palaeofield orientations (e.g. Tan et al. 2002). Variation in AIRM as a function of magnetizing field also provides insight into the fabric of different coercivity fractions of the remanence-carrying assemblage (Jackson & Borradaile 1991; Tan et al. 2002). Tensor parameters and representation Several parameters have been used to represent the anisotropy of second-rank tensors, in continuum mechanics and in strain analysis (Nadai 1963), structural geology (Flinn 1962), and still more in magnetic fabrics (Tarling & Hrouda 1993; Canon-Tapia 1994; Tauxe 1998); we discuss a few that have stood the test of time. The first item of concern is the shape of the magnitude ellipsoid expressed as its degree of prolateness (rod-shape) or oblateness (discshape). Second, we need a measure of the eccentricity of the ellipsoid; how far removed it is from the sphere, which represents the isotropic state. Various plots are available and all may represent ellipsoids describing sample ODs, as well as the ellipsoidal shapes that represent individual specimen anisotropy (Fig. 10), (Woodcock 1977; Pares et al. 1999). All plots compromise some aspect of presentation; for the purposes of AMS and strain-analysis we offer a new plot, which overcomes some problems of earlier graphs.

322

G. J. BORRADAILE & M. JACKSON

Fig. 10. (a, b) Anisotropy plots for magnitude-ellipsoids of tensors in structural geology and strain analysis include the Flinn plot (1965) and its logarithmic equivalent (Ramsay 1967). (c) Jelinek's parameters (1978) distinguish shape and eccentricity for an ellipsoid on separate Cartesian axes, but all near-isotropic states are spread misleadingly along the Pj — 1 vertical axis, (d) Our new polar plot overcomes this and the negative field facilitates comparison of dia/paramagnetic susceptibilities, (e) This plot may improve understanding of the petrofabric significance of inverse-AMS minerals like calcite and quartz, for which the KQ//C axes (most negative susceptibility) align perpendicular to S (e.g. cleavage or schistosity).

Shape is described in a continuous spectrum from prolate ellipsoid, through the neutral ellipsoid ('plane strain' of the structural geologist, X/Y = Y/Z) to the perfect disc shape of the oblate ellipsoid: (prolate) (neutral) (oblate)

(37)

In structural terminology, constricted ellipsoid shapes lie between prolate and neutral, and flattened ellipsoids lie between neutral and oblate. However, such genetic terms are meaningless in the context of magnetic anisotropy; geometrical descriptors like rod- and disk-shaped are preferable.

Whereas crystals fall into discrete symmetry classes, their physical anisotropies do not (Nye 1957). Nor do petrofabrics, as Flinn (19656) first recognized; a single event homogeneous fabric may only have 'orthorhombic' symmetry in the general case and rotational symmetry about one axis in the two end-member cases, prolate and oblate. Flinn plotted a = max/int and b = int/min giving a fabric/anisotropy plot with origin at (b,d) = (1,1). Shape was represented by the slope from the origin to the datapoint by an unfortunately asymmetric parameter k — (a- I)/ (b - 1); disc-ellipsoids have 1 > k > 0 whereas rod-shaped ellipsoids range over o o > A : > l . Flinn's plot unfortunately clusters weak anisotropies near the origin (1,1). Ramsay's (1967) logarithmic graph with

AMS-PETROFABRIC OF DEFORMED ROCKS

K = \n(a)/\n(b) overcame that, with the advantage that strains combine as a product (Fig. lOb). Ramsay's K is no more advantageous than Flinn's but it relates more simply to a modern magnetic fabric parameter (below). Ramsay's convention X > Y > Z for the principal stretches (e.g. X = ^NEW/^OLD) is now universally adopted whereas Flinn's pioneering studies (1962, 1965a, b) used the optical mineralogy convention (Z > Y >X). Early strain analysis described the shape of finite and incremental strain ellipsoids with Lode's parameter, simplified here by substituting recognizable Flinn ratios (b,a).

323

preserve that equivalence. In rock magnetism this quantity is mysteriously termed the anisotropy degree although it is readily quantifiable. Nagata's (1961) simple anisotropy degree is: (42) A 'corrected' (sic) anisotropy degree P' or Pj (Jelinek 1981) is now widely accepted, not least because it includes a reference to ftINT. Pj normalizes the principal susceptibilities to their mean. Jelinek used the arithmetic mean. However the geometric mean may be wiser, especially for large anisotropies, e.g. finite strain or AARM):

(38) (Lode 1926, Nadai 1963). A few other shape descriptors still find occasional use (Khan 1962): L (lineation)= ^MAX/^INT analogous to Flinn's a; not to be confused with 'L' in his qualitative field-fabric L-S scheme used by structural geologists (Fig. 12a). F (foliation) = ^INT/^MIN analogous to Flinn's b These are quantifiable ratios; confusingly, in structural geology 'lineation' and 'foliation' are fabric elements, or their orientations. Jelinek (1981) introduced a shape parameter 71, defined identically to Lode's parameter (is) of continuum mechanics but for our purposes easily remembered in terms of F and L:

(39)

(43)

It is similarly designed to the logarithmic shear strain parameter, £S, of continuum mechanics

(44)

(Nadai 1963), which usefully considers the three possible principal axial ratios, although its principal values are inconsistently normalized. Pj has numerous advantages for us (and for structural geology). Moreover, it may be retrieved simply from Nagata's older P-values with

Like Lode's parameter, and in contrast to Flinn's (45) k, it ranges symmetrically over the spectrum of ellipsoid shapes: T = +1 for oblate, T = — 1 for prolate and +1 > T > -1 for the general Jelinek's (P7,7") plot assigns eccentricity and ellipsoid (Fig. lOc). Jelinek's T and Ramsay's shape uniquely and conveniently to different shape parameter K may be simply related since Cartesian axes. However, for near isotropic eccentricities (Pj —» 1), slight shape differences Ramsay's plot uses logarithmic space: (T) appear just as significant as very large shape-differences at higher Pj (Fig. lOc). In geo(40) logical strain analysis a superior but under-used annular plot shows shape (is) along 60e arcs where K = ln(a)/ \n(b) or K = ln(L)/ ln(F) The distance of the point from the origin versus eccentricity (es) as radii (Hossack 1967; (b = l]a = 1) represents some aspect of total Ramsay & Huber 1983). We introduce a 7r/4strain/anisotropy, e.g. using Ramsay's logarith- segment polar-plot, on which Pj is radius and T mic plot: is arc-length; this representation has further advantages for strain and AMS. For example, shapes (T) are not dispersed for low Pj, and near-isotropic states plot unambiguously, close to (41) the origin. For AMS, the radial structure is Such concepts are application-dependent; e.g. in easily reflected to plot diamagnetic anisotropies. mechanics different finite strain states may have The negative field also facilitates the re-plotting been achieved by equal work and the logarithmic of AMS for counterintuitive inverse petrofabrics shear strain parameter (es, eq. (44) below) may due to intrinsic inverse AMS in minerals such as

324

G. J. BORRADAILE & M. JACKSON

Fig. 11. (a) Orientation distributions (ODs) may be described by second-rank tensors and magnitude-ellipsoids (Woodcock 1977), just like AMS or strain ellipsoids. The OD stereograms are for linear elements, e.g. for long axes of acicular grains or for AMS axes. At low anisotropy degree (Pj ~ 1.0) similar, weak ODs scatter too much along the T axis, (b) Our new polar plot clusters low anisotropy ODs together, near the origin, (c) Jelinek plot shows two distributions of specimen AMS-ellipsoid shapes; they spread along the T axis for low Pj [Data of Werner & Borradaile 1996]. (d) The same data in our new polar projection.

calcite, SD magnetite and tourmaline (Rochette 1988; Rochette et al 1992, 1999). Our polar plot is just as beneficial with ODs and samples, as with individual specimen ellipsoids. Weak ODs cluster unfavourably near the origin of the Flinn plot (Woodcock 1977) or disperse unnaturally along the T axis of the Jelinek plot (Fig. 11 a); near isotropic ODs cluster close to the origin on the polar plot (Fig. lib). This is illustrated with a sample-suite of AMS tensors, for Archaean greywackes (data of Werner &

Borradaile 1996) whose misleading dispersion in Cartesian coordinates (Fig. lie) becomes more meaningful in polar projection (Fig. lid).

Rock composition, bulk susceptibility (K) and AMS ellipsoid-shape It cannot be stressed too strongly that AMS ellipsoid shapes rarely correlate with finite strain magnitudes (Borradaile 1991) or with a causal

AMS-PETROFABRIC OF DEFORMED ROCKS quantity. In contrast, the principal AMS axial orientations usually show some causal relationship with a geological process or event. Some elementary observations serve to illustrate this point; consider a suite of Archaean upper greenschist facies greywackes with mean low field

325

susceptibility (K) mostly controlled by biotite and pyrrhotite, with a somewhat bimodal frequency distribution, Figure 12a (Borradaile et al. 1988, 1993; Borradaile & Sarvas 1990; Borradaile & Spark 1991; Werner & Borradaile 1996). Simple comparison of the eccentricity

Fig. 12. AMS magnitude ellipsoid shape is dictated primarily by rock composition, expressed partly in terms of K,. (a) Bimodal frequency distribution of K for Archaean greenschists, in which biotite and pyrrhotite mainly control K (Borradaile et al. 1988, 1993; Borradaile & Sarvas 1990; Borradaile & Spark, 1991; Werner & Borradaile 1996). (b) For the same specimens, Pj (eccentricity of AMS ellipsoid) is controlled by K (i.e. rock composition) for both high and low-ft subgroups. Both correlations are significant at the 95% level, due to favourable sample-sizes. The high-ft sub-sample represents pyrrhotite-rich specimens, (c) The symmetry or shape (7) of AMS ellipsoids as well as Pj, may be more influenced by rock-type and mineralogy than by metamorphism, strain or other secondary processes (some of data from Jackson & Borradaile 1991).

326

G. J. BORRADAILE & M. JACKSON

(Pj) of the AMS ellipsoid with K shows a general dependence of Pj on rock composition, since K proxies for rock-composition (Fig. 12b). All specimens considered globally give a deceptively strong correlation with rock-composition (~ft), K appearing to account for 72% of the variance in PJ (R2 = 0.72). Analysing the specimens in two, more homogeneous subgroups (K < 300 jiSI; K > 300 jiSI) that are dominated by pyrrhotite and biotite respectively reveal slightly weaker associations between rock composition (proxied by K) and Pj, however both correlate significantly at the 95% confidence level, in view of the sub-sample's sizes (Borradaile 2003). We cannot reject the hypothesis that rock composition controls anisotropy degree (Pj). Similar observations have been recorded elsewhere (Tarling & Hrouda 1993; Rochette 1988; Rochette et al 1992). The next logical question concerns the shape or symmetry of the AMS ellipsoid; is T controlled by factors other than strain or other secondary alignment processes? Compare the data for non-deformed Oligocene sandstone, specimens of which have been studied for other purposes (Borradaile & Mothersill 1991) with specimens of Cambrian Welsh Slate (Fig. 12c). The latter has an extremely penetrative, continuous fine cleavage and strains reported in excess of 60% shortening perpendicular to cleavage, whether it is purple or reduced to show greenspots or completely made green (Wood et al 1976). AMS and K merge contributions from matrix chlorite (the green pigment) and hematite (purple-red pigment). Magnetite is important in green spots and in bedding-controlled green reduction patches (Jackson & Borradaile 1991). Despite its extreme metamorphic fabric and strain, the magnetite-bearing slate has almost complete overlap in Pj — T space with a nondeformed typical sedimentary rock (Fig. 12c). Only the non-reduced, completely purple Welsh slate is distinct in Pj — T space from the sedimentary rock. The logical conclusion is that metamorphism and strain may be far less significant than mineralogy in determining AMS ellipsoid shape (T) and eccentricity (Pj). It will be clear subsequently that AMS axial orientations are far more useful and more readily interpreted in terms of secondary processes. Interpreting AMS in terms of mineral ODs

Determining AMS and K/or specific minerals The roles of individual minerals may be approached along several different routes; some

are restricted to estimates of «, while others permit an estimate of AMS or another magnetic anisotropy. Some permit the exclusion or isolation of contributions from ordered phases. Some are mechanically destructive and most preclude subsequent palaeomagnetic work on the same specimens. Exposures to high fields (Potter & Stephenson 1990; Stephenson & Potter 1996) and to low or high temperatures may change K and AMS (Dunlop & Ozdemir 1997). Although mineral-specimens may be measured just as standard rock-specimens, it is sometimes forgotten that no natural mineral is pure. The lower its intrinsic K, the more difficult it is to determine its AMS. Inclusions and exsolutions of magnetite and other magnetically ordered minerals are most troublesome; a few SD grains, undetectable by any microscopy may swamp the measurement of AMS in weakly paramagnetic or diamagnetic minerals (e.g. Fig. 2b, Table 2). Calculation: Ideally, and for a pure mineral phase, K may be calculated. Syono's (1960) formula (eq. 6) approximates this for a general paramagnet. This indicates upper limits near ~2000jiSI for stoichiometric, hypothetically pure, paramagnetic Fe-Mg silicates (e.g. chlorite, amphibole, biotite, pyroxene, epidote). Single crystal measurements: Individual crystals, where suitably shaped, may be mounted in universal-orientation holders for AMS and for AARM measurement. Shape-effects are a minor consideration for most /^-values. Hysteresis parameters including Khf may be measured in selected orientations, for microscopic specimens (<100mg), using an alternating-gradient magnetometer (Princeton Measurements MicroMag) (Borradaile & Werner 1994; Lagroix & Borradaile 2000a). Torque measurements sensitively detect crystalline anisotropy (Banerjee & Stacey 1967; Martin-Hernandez & Hirt 2001, 2003). Mineral separations: crushed or picked mineral aggregates may be density separated or magnetically separated to yield quite pure concentrations of individual minerals. Individual grains may be assembled and aligned mechanically or magnetically (Borradaile et al. 1985, 1987, 1990; Johns & Jackson 1991; Johns et al. 1992). Leaching: Oxides and carbonates may be preferentially removed; especially where the solid specimens have been suitably perforated with holes or saw cuts; the same technique is used for 'chemical demagnetization' of palaeomagnetic specimens (Henry 1979). This may preferentially remove certain subfabrics according to grain-size or composition (Borradaile et al. 1990; Jackson & Borradaile 1991).

AMS-PETROFABRIC OF DEFORMED ROCKS

High-field techniques: High-field magnetization measurements show field-dependent anisotropy for ordered phases, with a maximum in fields near the coercivity, declining to zero as the phases saturate (Rochette & Pillion 1988). AARM and AIRM exclusively characterize the ferromagnetic mineral fabric (Fuller & Kobayashi 1964; Daly 1967; Daly & Zinsser 1973). High-field torque measurements in different fields allow separate characterization of the deviatoric anisotropy of the linear (dia-, paraand antiferromagnetic) and nonlinear (ferriand parasitic ferromagnetic) phases (Jelinek 1985; Martin-Hernandez & Hirt 2001). Low-temperature and high temperature techniques: Remanence-bearing phases may be identified non-destructively by low-temperature thermomagnetic measurements, which show discrete changes in magnetization associated with structural phase transitions, isotropic points, and spin-flop transitions, e.g. magnetite (~120K Verwey transition, ^130K isotropic point), hematite (~260K Morin transition), monoclinic pyrrhotite (Fe7S8; ~34K) (O'Reilly 1984; Thompson & Oldfield 1986; Rochette et al. 1990; Richter & van der Pluijm 1994; Dunlop & Ozdemir 1997; Moskowitz et al. 1998). Ti-rich titanomagnetites and titanohematites, including the ilmenite and ulvospinel end-members are magnetically disordered at room temperature. Curie or Neel temperatures (200-700 °C) characterize most common remanence-bearing minerals, although heating may alter the mineralogy. Some minerals, especially SD magnetite, may show enhanced susceptibility just prior to complete thermal demagnetization ('Hopkinson peak'). At high and low temperatures, K is relatively easily measured but AMS determination requires sophisticated laboratory facilities (Rochette & Pillion 1988). Temperature equilibration during measurements along different specimen axes is difficult, even in a relatively stable liquid nitrogen environment (77 K) but some data are available (Ihmle et al. 1989; Luneburg et al. 1999, Pares & van der Pluijm 2002). Cryogenic AMS measurements may offer the most promise to partition anisotropic contributions between the paramagnetic, diamagnetic, and Terro'-magnetic responses although the mineral sources might still be ambiguous.

Anomalous orientations of AMS in certain minerals Although MD titanomagnetite-magnetite, or pyrrhotite is usually present in traces (~0.2wt%), depending on the other minerals,

327

their high K may outweigh their concentration and they may dominate the rock's K and the rock's AMS (Borradaile 1988). First, it is necessary to identify special orientation effects here, like Fuller's (1961) 'supergrain' interaction effect (Fig. 2d). Recently, Hargraves et al. (1991) termed this 'distribution anisotropy\ which may confuse students of petrofabrics for which the term location fabric less ambiguously, and with historical precedent, describes the fabric contribution from any lithological component discretely located in space, regardless of its internal isotropy or anisotropy (Turner & Weiss 1963). Magnetite causes a more common and serious special fabric effect; in SD-form it shows an inverse fabric with its long axis //«MIN (O'Reilly 1984; Stephenson et al. 1986; Potter & Stephenson 1988). Few rocks show a net inverse AMS (i.e. /^MAX -L S) since unusually high proportions of SD magnetite are required. However, lesser concentrations cause blended fabrics more commonly, in which an SD subfabric interferes with a 'normal' matrix (Rochette et al. 1992; Borradaile & Gauthier 2001, 20036). Matrix-forming calcite shows intrinsic inverse fabrics where there is insufficient competition from paramagnets and ferromagnets, as in many limestones. Calcite is diamagnetic, with the most negative susceptibility, Ke//c (Rochette 1988; Ihmle et al 1989). K is not precisely reported (-13 to -14|iSI, Nye 1957) but precise values for anisotropy (K€ - Ke) ~ 1.172 ± 0.028 jiSI were obtained by torque magnetometry (Owens & Rutter 1978). Common metamorphic-deformation mechanisms align c axes for calcite (and also for quartz) steeply with respect to the XY-plane or to the plane of maximum shear (later, Fig. 22). In nature, calcite (and quartz) grains are flat-shaped in the basal plane (_L to c). Thus, an inverse fabric results with the most negative susceptibility Ke//c and _L to S-fabric. The elongate habits of calcite and of quartz (Fig. 3c, e) are not expressed in the matrix. No other major rock-forming minerals are yet reported with intrinsic inverse AMS; although in some leucocratic granites there is sufficient accessory tourmaline to yield inverse or blended fabrics (Rochette et al. 1994).

The Orientation Distribution (OD) of specimen AMS principal axes Two principal axes (usually ftMAX and ftMIN) suffice to define the orientation of a few similarly oriented specimens since the orthogonality of the third principal axis follows naturally from

328

G. J. BORRADAILE & M. JACKSON

the second-rank tensor structure of AMS. However, there are distinct visual and interpretative advantages to viewing all three axes for each specimen on the stereoplot. The orientation of K INT with respect to other structures may be a clue to blended subfabrics (Fig. 17a later); in general, permutations of all subfabric axes must be considered in combination (Borradaile &

Sarvas 19906; Rochette et al 1992; Ferre 2002). For similarly oriented specimen-tensors, density contours are interpretable and the sample's OD is revealed by the cluster-girdle patterns associated with the L-S fabric scheme (Flinn 19656; Woodcock 1977) (Fig. lla) though their distribution in Pj - T space is possibly more meaningful in polar coordinates (cf. Fig. lib, d; 12c).

Fig. 13. (a) Flinn's (1965) qualitative L-S fabric scheme describes ODs of preferred crystallographic or preferred dimensional orientations. It is now also used to describe ODs of magnitude-ellipsoids of tensors, for example finite strain or AMS. (b) ODs of axes of specimen AMS-tensors; above perfect L-fabric, below perfect S-fabric. (c) For a sample of AMS specimen-tensors, Jelinek's confidence regions about the mean axes have a shape and symmetry that characterizes the sample's OD in the L-S scheme. Here the OD is for an S > L fabric since the confidence regions about mean K MAX and mean /%IIN are elongate in the plane perpendicular to KMIN(d) Werner's (1996) comparison of confidence regions for the sample's mean axes, determined by different calculation methods for the same sample of A ARM measurements.

AMS-PETROFABRIC OF DEFORMED ROCKS

When using fabric parameters and fabric plots, it is important to recall that (PjT) values for the sample-OD may be unrelated to any or even the average specimen. For example, an L-tectonite fabric (point-concentration of ftMAx)> with rOD = —1.0 may be comprised of individual specimen-ellipsoids of any shape, even J' = +1.0, provided their K^T and A^MIN axes are dispersed in a self-cancelling arrangement in the great-circle girdle _L to KMAX- (We discussed previously that in general PJOD < ^SPECIMEN)The L-S character of an OD is also represented in the shape of the confidence regions for the axes of Jelinek's mean-tensor (Fig. 13). ODs for L- or S-tectonites produce characteristic clusters and great-circle girdles of /^MAX an<^ ^MIN» respectively (Fig. 13b). Using Jelinek statistics, the confidence region for each mean-tensor axis (maximum, intermediate and minimum) is determined compatibly, retaining the essential meanaxes' orthogonality. The confidence regions thus retain a symmetry which characterizes the OD's prolateness versus oblateness in Flinn's L-S scheme (Fig. 13c). Whereas confidence regions possess axial symmetry for the endmember L (r0p = -l) and S (TOD = +1) cases, a hypothetical general orthorhombic case has elliptical confidence regions which show the OD has S > L in, e.g. TOD ~ +0.5 (Fig. 13c). The mean tensor and its confidence regions faithfully characterize a sample of tensors. However, the use of traditional density contours on stereograms is still useful for reconnoitring ODs, e.g. identifying sample-homogeneity. Density contours assume no distribution-model so that each group of axes (ACINT, etc.) is treated as though they were independent of the other principal axes (Fig. 14a). That example shows that individual specimens have poorly concentrated principal axes, neither forming elliptical clusters, nor elliptical girdles; this strongly suggests that the sample is heterogeneous and that it should be analysed in more homogeneous sub-samples. Any raw sample, however homogeneous, will generally show non-orthogonality of peak concentrations for KMAX f°r ^INT and for ^MIN? even mean vectors will generally be nonorthogonal (cf. Fig. 14b). In this sample, the AMS peak densities' non-orthogonality and their inclination to the AMS mean-tensor are due to two competing subfabrics in the AMS; one of which is due to a magnetite subfabric here isolated by A ARM. The differently oriented subfabrics shown by the AMS/AARM disagreement cause the peak-densities of AMS to be so unreliable (Fig. 14b). Confidence regions for a mean AMS axis (e.g. ^MIN) have also been determined using Fisher or

329

Bingham statistics, borrowed from palaeomagnetism where they are applied to vectors or directions. They assume circular or elliptical clusters on the sphere for directions that act as independent variables. In reality, the distribution of AVMIN axes is constrained by the distribution of ^MAX and ftiNT and their confidence regions may only be reliably determined by Jelinek statistics. Werner (1997) compared confidence regions for a homogeneous sample of AARM specimentensors and showed that Jelinek-statistics were superior and possessed the expected orthorhombic symmetry (Fig. 13d). One approach evaluated by Werner used the boot-strap method to determine confidence regions for sample-mean. This may be used for confidence regions of directions, axes or indeed any variable (Fisher et al. 1987). Bootstrapping randomly re-samples the original sample of n measurements, usually at least 200 times. Each of the 200 re-samples contains n orientations that duplicate some of the original measurements and omit others. The mean orientation of each re-sample is then plotted; the 200 pseudo-mean orientations produce a beguilingly simple elliptical concentration whose density distribution may be contoured to define the confidence region for a certain mean axis. This technique finds proponents and critics in AMS and palaeomagnetic applications (e.g. Werner 1997; Tauxe 1998; Borradaile 2003). It is subject again to the criticism that all the ftMAX orientations (and then all AVINT, «MN) must be treated independently, as if each «MAX orientation was unconstrained by the other principal axes of the specimen tensor. It is more worrisome from a general statistical viewpoint that all bootstrapping requires that the sample be strictly representative of the population. If that is the case, one might just as well use a traditional, less deceptive approach, to calculate the confidence regions, e.g. by assuming a bivariate-Normal distribution on the sphere (Henry & Le Goff 1995) or a Fisher-Bingham distribution (Fisher et al. 1987). Sub-orthorhombic symmetry of bootstrapped confidence regions need not have the same interpretative significance as those derived by Jelinek statistics (below). Jelinek (1978) provides a rigorous alternative procedure to determine the mean axial orientations and their confidence regions for a sample of tensors. Small samples, or unusually large confidence regions may limit their validity but sample-size rules may not be simply formulated as in classical frequency-distribution statistics. Orientation-statistics are harassed by the closure-problem, and the disadvantages of small sample-sizes may be offset by strong preferred

330

G. J. BORRADAILE & M. JACKSON

Fig. 14. (a) The peak-density for a certain set of axes, say ftMAX, for a sample of AMS specimen-tensors are usually non-orthogonal. Although inspection of density contours is useful to verify sample homogeneity, they do not characterize ODs for tensors faithfully [sample of Archaean greenschists]. (b) In contrast, Jelinek's (1978) mean tensor retains the required orthogonality of the sample's mean principal axes [same data], (c) Confidence regions for mean-tensor axes need not show orthorhombic symmetry [q.v. Fig. 13c], The confidence regions may be distorted due to (d) multiple subfabrics or (e) specimens with anomalous-^. When the sample is

AMS-PETROFABRIC OF DEFORMED ROCKS orientation. Apart from providing the only valid determination of mean axes for a homogenous sample of specimen-tensors, Jelinek statistics permit the determination of confidence regions about the mean axes that recognizes the mutual orthogonality of principal axes for each specimen. For this reason, the symmetry of Jelinek's confidence-cones for the mean-axial orientations defines the sample-OD in the L-S scheme and is ideally orthorhombic, for an homogeneous sample due to a single-event coaxial fabricforming process (Fig. 13). However, for heterogeneous samples the confidence regions may be sub-orthorhombic, two or more of them inclined to the symmetry planes (Borradaile 2001, 2003) (Fig. 14c). This is due to competing contributions from differently oriented subfabrics, or the distracting effects of some specimen-outliers (Fig. 14d, e). Jelinek statistics permit these competing contributions to be partially suppressed where the subfabrics, or the outliers, are significantly different in K from the remainder of the rock. Distortion of the sample OD by the specimens of anomalous K may be reduced and orthorhombicity may be restored by equal-weighting of all specimens. Standardization divides K;MAX> ^INT and «MIN f°r eacn specimen by its K ; thus, all specimens are reduced to unit-susceptibility. In this way the orientation of a minor subfabric, or of a few outlying specimens, is much less significant in the OD and the mean-tensor's orientation and the shapes of its confidence regions may be changed. For example, a few high-tt specimens are responsible for the AMS bedding fabric in an arkose; standardization reveals a cryptic cleavage-fabric due to the lower susceptibility matrix (Fig. 14f). Even without standardization, data-processing reconnaissance may also partially isolate subfabrics if ODs are compared for sub-samples of different K. The AMS OD for high K specimens may be as diagnostic as measuring AARM in the same rocks. The OD of \OW-K specimens better defines the silicate matrix petrofabric than the OD for high K specimens (Borradaile & Gauthier 2001), (Fig. 14g). Standardization may suppress an anomalous subfabric, re-establishing the orthorhombic symmetry expected for a single-generation L-S

331

fabric. However, it may also suppress a tectonic component, e.g. revealing the bedding fabric in a slaty matrix (Fig. 15a, b). In rocks with multiple subfabrics, standardization may enhance tectonic fabric symmetry (Fig. 15c) or reveal a different-symmetry subfabric, for example a cryptic L-S fabric where the raw AMS OD suggested an L-fabric (Fig. 15d). Orientation distribution (OD) of AMS specimen tensors and magnetic fabric The L-S fabric concept combines the contributions of shape and orientation of three-dimensional objects or tensor magnitude-ellipsoids to define an OD. A single-event, homogenous L-S fabric must show orthorhombic symmetry for its OD (Fig. 13b, c). For simplicity, consider a sample of specimens, each with the same anisotropy. With a perfect, saturation alignment, the sample has the same anisotropy (T,Pj) as individual specimens. Any imperfection in the OD requires P/SAMPLE < ^/SPECIMEN; it may sometimes also subdue ellipsoid-shape (I ^SAMPLEI < I ^SPECIMEN I ) -

The arguments applied to a sample of many specimens apply equally to the AMS of one specimen and the AMS of its constituent aligned minerals. It is also a conceptual objection to the correlation of finite-strain magnitudes and AMS magnitudes for a specimen (Borradaile 1988, 1991). Even a well-aligned tectonite specimen usually has a modest P;, much less than that of the dominant anisotropic mineral, due to self-cancelling effects of unfavourably oriented examples of that mineral, and also due to other minerals. In nature, the relationships of mineral-AMS to specimen-AMS are obfuscated in several ways: (i) Mineral AMS axes may only be coaxial with crystal axes for high symmetry; orthorhombic (Fig. Id), trigonal (Fig. 3c, e) as well as cubic and hexagonal classes. Most minerals are monoclinic; they may have only one AMS axis parallel to a crystal axis. For triclinic minerals (e.g. plagioclase), all AMS axes are inclined to crystal axes.

standardized by dividing each specimen's principal susceptibilities by «, the contribution of the anomalous subfabric or outliers to the OD is suppressed, (f) Standardizing specimen-tensors [dividing their principal susceptibilities by their K] may neutralize the distracting contribution of a subfabric or outliers of anomalous K. Thus a cryptic S-fabric is revealed when the AMS of weakly cleaved arkose is standardized; the masking contribution of a high-K bedding subfabric is suppressed, (g) Without resort to tensor-standardization or other techniques like AARM, inspecting sub-samples of AMS tenors according to their ^-ranges may also reveal different subfabrics [ophiolite dykes, Cyprus; Borradaile & Gauthier 2001].

332

G. J. BORRADAILE & M. JACKSON

Fig. 15. (a) AMS of Borrowdale volcanic slate defines a good S >> L fabric, (b) Mean-tensors of raw data emphasizes field observations of S ^> L fabric but standardization reduces the symmetry and strength of the Scomponent; this may be revealing a matrix-bedding fabric, (c) Standardization enhances fabric symmetry and slightly changes interpreted mean-orientations, (d) Gneisses show an overall S-fabric for non-standardized AMS, mainly due to a few high-«: specimens. Standardization shows that the silicate matrix fabric possesses an L-fabric. However, the L-directions of both fabrics are similar.

(ii) For low-symmetry minerals with poorly corresponding susceptibility axes and crystallographic axes, an OD of crystals may fail to define consistent AMS orientations. (iii) A one-to-one mapping of magnitudes with petrofabrically convenient crystal axes is not mandatory (Fig. 3c, e). This is due to intrinsic-cry>stallographic inverse AMS, apparently shown also by goethite,

tourmaline and possibly by cordierite (Rochette et al. 1992). (iv) For SD magnetite, /CMIN // the long dimension due to the inverse AMSshape-effect (Fig. 3h). Other high susceptibility minerals that may show an inverse AMS include maghemite (Borradaile & Puumala 1989), suitable titanomagnetite compositions and greigite (Aubourg & Robion 2002).

AMS-PETROFABRIC OF DEFORMED ROCKS

(v) Physical alignment mechanisms (crystalplastic mechanisms etc.) are related to crystallography in complex ways; they do not automatically align the long habit of the crystal. For example, quartz and calcite c axes align steeply to S, sometimes in cones. Moreover Ke//c so that inverse fabrics ensue. (vi) Weakly aligned crystal axes and partially cancelling contributions from a diamagnetic matrix and low abundance minerals with K > 0 compound the above problems, especially for Pj —» 1.0 (Fig. 3). Any specimen with K < 100 uSI may be suspect. Thus, AMS orientations may at best only approximate the minerals' OD. To follow from this monomineralic reductio ad absurdum, consider polymineralic rocks, with incompletely developed or even multiple sub-fabrics, with additional AMS contributions attributable to shape-controlled magnetite accessories or inclusions, and with potentially self-cancelling diamagnetic and paramagnetic components in low-susceptibility rocks. These confounding complications lead to the undeniable conclusions: (a) AMS fabrics that characterize the overall OD in a tectonically meaningful manner are probably attributable to a dominant subfabric of a single mineral phase of large K; with orthorhombic symmetry of its sample-mean tensor (e.g. Fig. 13c). These need not be due to magnetite; mafic silicates such as chlorite, biotite, serpentine or amphibole may dominate K and usually possess stronger ODs than magnetite. (b) A sub-orthorhombic OD is probably due to a subfabric or a few extreme-^ specimens in the sample (Fig. 14d, e). Subfabrics may be partially isolated by comparing standardized to non-standardized mean tensors (Figs. 14f, 15c, d), by comparing subsamples of different K (Fig. 14g). More definitive answers may be obtained by separate rock magnetic experiments that determine AARM or pAARM, since these completely isolate the OD due to remanence-bearing minerals (Fig. 14b). (c) Mathematical associations between ODs and AMS are applicable only to idealized scenarios (e.g. Hrouda & Schulmann 1990; Henry 1992; Hrouda 1993; Benn 1994). They usefully constrain thought-experiments and sampling strategies but few examples are directly applicable in nature (Housen et al 19930). (d) AMS studies would have been discouraged if these complications had been realized at

333

the time of Graham (1954). Fortunately, AMS axes commonly proxy for petrofabric orientations, in tectonic, sedimentary or igneous rocks, which is attributed mainly to (a) and (b) above (Hrouda 1982; Jackson & Tauxe 1991; Tarling & Hrouda 1993; Borradaile & Henry 1997).

Mineral abundances: AMS and K for an 'aligned' rock Consider AMS for a monomineralic rock; if its mineral-alignment is less than perfect ^/SPECIMEN < ^/MINERAL

an

d

^SPECIMEN

W1

U

differ from TM INERAL . Rock specimen's AMS is a subdued version of the constituent mineral's AMS. Now, consider the addition of a second mineral, similarly aligned, for example dispersed accessory MD magnetite. /^SPECIMEN *s approximated from the volume proportions of the two minerals. With 1% magnetite, ^SPECIMEN ~ 0'99«BiOTiTE + 0.01ttMAGNETiTE. Conservatively, assume the maximum bulk susceptibility for biotite (2000 fiSI) and a value of 2 500 000 uSI for magnetite: ^SPECIMEN is dominated by the magnetite-trace, ^SPECIMEN « 1980 + 25000 = 26980uSI. [The diamagnetic contributions of minerals has been overlooked here since it usually only affects AMS orientations and magnitudes in weakly aligned and low-ft rocks (Rochette 19870, 19940)]. Having recognized the affect of the dilution principle on K, we must now consider its effects on AMS (i.e. Pj and T). Combinations of coaxial AMS ellipsoids in this model show that AMSSPECIMEN migrates from biotite towards that of magnetite with increasing magnetite abundance, measured in ppm (Fig. 16a). A similar model is shown for a calcite-magnetite aggregate which is a realistic model of magnetite concentrations (1-10 ppm) in some limestone (e.g. Hamilton et al., this volume); here one notes how the AMS moves from the diamagnetic field to the field of positive-/^ which involves a change of AMS symmetry (T changes sign) and requires an inversion of principal axial orientations. Fuller (1963) first appreciated the essence of polymineralic problems, although their farreaching consequences for petrofabric and AMS interpretation could not be foreseen at that time. For example, on their own, such arguments throw doubt on any generally valid causal relationship between AMS and the magnitudes of alignment processes (Borradaile 1988, 1991; Borradaile & Henry 1997). Indeed, in nature, AMS and finite strain usually correspond

334

G. J. BORRADAILE & M. JACKSON

Fig. 16 The AMS of a specimen depends on the proportions of its constituent minerals, other factors being equal. These thought-experiments assume a saturation coaxial alignment of accessory magnetite and a monomineralic silicate matrix, (a) Increasing concentration of magnetite (1 ppm, lOppm, etc.) displaces specimen-AMS from that of the matrix towards that of the magnetite (Borradaile 1988). Rock composition may influence AMS more than mineral-alignment or strain. (Specific mineral values obtained from mineral separations; Borradaile et al. 1986.) (b-d) Mineral proportions may influence the OD of KUAX in a hypothetical mixing-model of phlogopite with varying magnetite-abundance. The orientation of specimen-KMAx depends on the magnetite-abundance.

poorly (Borradaile & Mothersill 1984; Borradaile 1991) although there are a few well documented examples to the contrary, where lithologies are uniform, strain gradients are small and metamorphic effects are minimal (Siddans et al. 1984; Hirt et al. 1988, 1993). More optimistically and actually more usefully, AMS orientations and finite strain axes do correspond well in nature (Hrouda 1982) and also in the laboratory (Borradaile & Alford 1987, 1988; Jackson et al. 1993). Finally, where the different minerals have different alignments, mineral abundances also influence the orientations of the gross AMS axes for the specimen. For example, increasing

concentrations of magnetite move ^MAX[SPECiMEN] progressively from the orientation of ^MAx[PHLOG] toward the orientation of «MAX[MAGNETITE] ( Fi g- 16c-e). With 1000 ppm

magnetite, the differently oriented phlogopite matrix is expected to show negligible influence on the sample's AMS orientations (Fig. 16e).

Blended sub-fabrics and the rock's gross AMS orientations The fortunate recognition of 'anomalous' principal orientations associated with discrete subfabrics in pressure-solution cleavage led to the

AMS-PETROFABRIC OF DEFORMED ROCKS

recognition of blended AMS subfabrics (Fig. 17a). In the case of the pressure-solution anomaly ^MAX is parallel to the intersection of cleavage with bedding, the /3-lineation, rather than X, due to the summation of orthogonal subfabrics from the bedding plane and from the styloliteplane (Borradaile & Tarling 1981). Penetrative schistosity may also produce KMAX//& which is disconcerting since it is not as predictable from petrographic or field observations (Borradaile & Sarvas 1990), Figure 17b. The discrete location subfabrics may include crenulation-cleavage surfaces versus their microlithons, many types of metamorphic differentiation, mylonitic lamination and S-C fabrics and pervasive subfabrics are also common but may require microscopic identification (e.g. slaty cleavage cutting massive beds). Provided the specimen samples the mineral-ODs in each subfabric representatively, the specimen-AMS will blend the subfabrics' differently oriented AMS. Of course the manner in which the subfabrics blend depends on their angular relationships as well as their anisotropy (Pj, T) and K; for convenient simplicity we consider subfabrics with nearly orthogonal AMS ellipsoids. Depending on the subfabrics' AMS-ellipsoid-shapes (Pj, 7), their relative contributions in terms of volume and K, numerous permutations exist, of which but one example is shown in Figure lie. Rochette et al. (1992, 1999) evaluated possible specimen AMS axes from varied volumetric proportions of coaxial normal and inverse subfabrics (Fig. 17d). The orientations of the specimen's principal AMS axes switch dramatically in orientation as the volume-fraction of the inverse subfabric increases, and the shape of the specimen's AMS ellipsoid changes dramatically and non-progressively, as shown on the polar Pj — T plot (Fig. lid). Ferre (2002) considers more complex situations. A matrix algebra approach to this problem is elegant (Hrouda 1992) but its sensitivity to commutativity may be overlooked in nature and it may be best reserved for cases where the sequence of truly discrete subfabrics is known. Although considerations of coaxial/orthogonal subfabrics provide useful interpretative principles, the AMS subfabrics will generally be inclined to one another in nature. Occasionally, one or more subfabrics may contribute counterintuitively to the gross AMS producing K;MAX at a high angle to S. These deserve consideration and three distinct categories reported: (a)

Minerals with intrinsic crystalline-inverse AMS, e.g. tourmaline and goethite have ftMIN//c, which is their long axis. Their

335

ODs thus have c//X lineation characterized by ^MIN> not ^MAX(b) Diamagnetic minerals like quartz and calcite. Although KQ//c and KQ is the most negative ('maximum', sic) principal susceptibility, common crystal-plastic deformation mechanisms (e.g. c-basal glide) align KC _L toS. (c) SD magnetite shows an inverse-effectAMS, its measured «MIN // grain long axes. Occasionally, ambiguous subfabrics occur in counter-intuitive orientations, e.g. /-CMAX steeply inclined to bedding in weakly deformed sediment (Borradaile et al. 1999a; Aubourg & Robion 2002) or J_ to XY in experiments (Borradaile & Puumala 1989). The present state of knowledge does not exclude unknown alignment processes or unverified in verse-AMS mineral-responses (e.g. greigite, maghemite). In a perfect inverse-situation, an S-petrofabric will have a perpendicular L-fabric AMS. Conversely an L-petrofabric mineral alignment will possess an orthogonal S-symmetry AMS. In other words, rPETROFABRic and TAMS are of different sign and the OD and AMS ellipsoids have perpendicular major axes. The importance and potential of AMS for the common matrixforming diamagnetic minerals, calcite and quartz was appreciated early (Owens & Bamford 1976; Owens & Rutter 1978), though not their inverse-fabric complication. Inverse AMS has been recognized for some limestone (inter alia Rochette 1988; Ihmle et al. 1989; de Wall et al. 2000; Hamilton et al., this volume) and is predicted for quartzite (Hrouda 1986) and is reported rarely for quartz-rich rocks like tonalite-gneiss [author, unpublished data]. These rocks' AMS may be diamagnetic or, in the presence of impurities, weakly paramagnetic. It is very important to recognize that AMS axial orientations may be difficult to interpret due to self-cancelling sources of positive and negative susceptibility, especially for specimens in the range - 1 5 < f t < 5 0 u S I (Fig. 3a,b) and the presence of any ferromagnetic contamination may hinder interpretation (Borradaile & Stupavsky 1995). The inverse-effect fabrics of SD magnetite are due to shape anisometry, the measured ftMAX and ttMIN correspond to short and long grain axes' orientations respectively. SD grains may occur as lattice-controlled inclusions/exsolutions in silicates, and therefore at best indirectly related to tectonic fabric. In ophiolite, copious SD production by ocean-floor metamorphism provides examples of inverse-subfabric contributions (Rochette et al. 1992; Borradaile

336

G. J. BORRADAILE & M. JACKSON

Fig. 17. (a) Combined AMS subfabrics due to bedding (So) and cleavage (Sj spaced and stylolitic due to pressure solution). The combination yields KUAX//& (ft = intersection lineation of bedding and cleavage) not // X (Rheinisches Schiefergebirge; Borradaile & Tarling 1978). (b) A penetrative schistosity (S) subfabric pervades and blends with a bedding (So) subfabric also to give ^MAX///3-Hneation & fold-hinge; the two subfabrics are penetrative (Archaean schists; Borradaile et al. 1988). (c) One of many possible permutations of coaxial subfabrics; however, subfabrics may also combine non-coaxially in nature, (d) The coaxial blending of a normal and an inverse AMS; the inverse fabric may be due to a counter-oriented subfabric or due to an

AMS-PETROFABRIC OF DEFORMED ROCKS

& Gauthier 2001, 20036). However, perfect inverse AMS responses due to the SD effect are rare since the AMS of ~106 SD grains may be cancelled by the AMS of a single lOum MD magnetite grain. More commonly, an inverseSD subfabric partly cancels a normal subfabric to produce a blended fabric. Interpreting AMSROCK is still more challenging where it comprises non-coaxial subfabric permutations. Certain subfabrics may be isolated, enhanced or suppressed by experimental or statistical procedures. These may partly resolve ambiguous interpretations, as follows. Experimentally, AARM (or even AIRM), isolates the subfabric OD of all 'ferro'magnetic grains, or part of the ferromagnetic population, if it is selected using a coercivity range in the pAARM technique. Fortunately remanenceanisotropy for all magnetite, including SD, shows a normal-fabric with ^MAX//l° n g axis and AmN //short axis (Jackson & Tauxe 1991), compatible with the associated AMS (e.g. Fig. 3g, h). Hematite and pyrhhotite also have compatible, 'normal' AARM and AMS. Goethite is reported to have an inverse-crystalline AMS (Ozdemir & Dunlop 1996) but it is rarely abundant and its alignment mechanisms are unknown. Mineral-abundances: the rock's K and principal susceptibility magnitudes The rock's magnitudes ftMAX > ^INT > ^MIN are influenced by mineral-proportions and their ODs (e.g. eq. 18; Owens 1974). For initial simplicity, assume two minerals are coaxially aligned. Henry (1983) and Henry & Daly (1983) showed simple relationships ensue for susceptibility-magnitude contributions with slightly varying mineral concentrations. The model predicts that for a suite of specimens principal magnitudes (KMAX> etc.) are linearly dependent on K, and this is verified in nature (Fig. 18a) if the premises are reasonably satisfied. 'False positives' may also be observed, in which linear correlations are observed but the model assumptions do not hold, and different interpretations are required (Johns & Jackson 1992; Rochette et al 1992). A similar concept underlies comparisons of K and A for ARM in granulometry studies,

337

Figure 18b (King et al 1982; Jackson et al 1988). Extended to anisotropy (AMS versus AARM), different relationships occur with different concentrations of the constituent subfabrics (Fig. 18c) (Borradaile & Lagroix 2001). Matrix K and accessory-subfabric A magnitudes may be isolated and one may infer relative contributions of matrix versus ferromagnetics to AMS from the graphs. Non-linearity may imply subequal contributions from more than two phases (e.g. Fig. 18c). Henry's (1989) triangular tensor plot, adapted from one used in strain analysis (Hsu 1966), investigates the combined effects of Ac-variation and orientation. Closure restraints of the triangular plot and the usual restrictive premises, e.g. high-symmetry crystals, limit this approach (Fig. 18d). Triangular graph axes represent each of the three standardized principal magnitudes, e.g. 0 < (ft M Ax/KO < 1» with advantages and disadvantages. Ranges of permissible standardized ^MAX, ^INT and /^MIN values for the tectonite lie along lines in the triangular diagram. Isolating separate subfabrics experimentally Physical discrimination of mineralogical sources is less ambiguous than statistical procedures, especially where one subfabric is 'ferro'magnetic, with low or moderate coercivity. Suitable laboratory approaches include: (a) Anisotropy of complex electromagnetic susceptibility, measured in high-frequency a.c. fields may characterize the subfabric of high-conductivity minerals, e.g. sulphides and graphite (Vincenz 1965; Clark et al. 1988; Borradaile et al 1992; Ellwood et al 1993; Worm et al 1993). Eddy currents induced by the time-varying applied field generate a positive quadrature and negative in-phase a.c. response (the latter of which can be mistaken for diamagnetism in a single-frequency, single-temperature measurement). (b) Measuring K at various low and high temperatures may identify the predictably inverse-T paramagnetic response, or the temperature-independence of diamagnetics (Rochette & Pillion 1988; Richter & van

inverse-AMS mineral (Rochette et al. 1992). In geographic coordinates, the orientation-switching of principal axes is very clear as the volume fraction of the inverse component increases. The AMS ellipsoid undergoes complex, non-progressive changes in shape as shown on the Pj — T polar plot.

338

G. J. BORRADAILE & M. JACKSON

Fig. 18. (a) Simple AMS-magnitude relations may exist for two dominant minerals (Henry 1983) although such relationships do not guarantee a simple concentration-control (Johns & Jackson 1991). (b) ARM intensities versus K indicate the relative importance of remanence-bearing versus paramagnetic minerals (King et al. 1982). (c) AMS magnitudes versus AARM magnitudes in the same specimens; homogeneous coaxial subfabrics with a non-linear relationship. Mafic silicates, MD magnetite, ilmenite and chromite all contribute significantly to AMS whereas only magnetite contributes to AARM (data of Borradaile & Lagroix 2001). (d) The effect of mineral abundances on AMS magnitudes and axial orientations has been tentatively investigated for simple systems with saturation alignment and high-symmetry crystals (Henry 1989).

der Pluijm 1994; Pares & van der Pluijm 2002), Figure 4. For 'ferro'-magnets, temperature-dependence is sensitive to crystal structure and composition (e.g. Dekkers

(c)

1989a, b; Dekkers et al 1989; Moskowitz et al 1998). The practice of laboratory heating to enhance and characterize a subfabric

AMS-PETROFABRIC OF DEFORMED ROCKS

(Perarnau & Tarling 1985) is not generally endorsed and is discussed conveniently under metamorphism, below, although that is a natural process. (d) Isolating or suppressing magnetic responses from 'ferro'magnetic subfabrics. Magnetization measured in the presence of high fields (>300mT) ignores the response of minerals that are saturated at that level, such as magnetite. Consequently, the 'matrix' AMS is isolated (Rochette & Pillion 1988; Kelso et al. 2002; Ferre et al 2004). High-field torque measurements in different fields permit separate characterization of the saturated ferromagnetic anisotropy (field-independent above saturation) and the anisotropy of the linear high-field susceptibility of dia-, para- and antiferromagnets (Jelinek 1985; Martin-Hernandez & Hirt 2001). Alternatively, and more easily, the anisotropy of the 'ferro'magnetic subfabric is measured independently, if its coercivity is low enough to permit successive remagnetizations and measurements along different specimen-directions. The latter techniques include AIRM (Daly & Zinsser 1973; Stephenson et al 1986), AARM (McCabe et al 1985; Jackson & Tauxe 1991; Trindade et al 2001) and GRM (Stephenson 198la; Stephenson & Potter 1987). Hematite, goethite and some pyrrhotite are usually too coercive to respond to the latter two (AF-based) techniques with the fields available. (e) Where the tensor for one subfabric is isolated, it is rarely possible to isolate the remaining OD by subtraction from the whole-rock OD. For example, Hrouda et al (2000) show that in general the nonferromagnetic contribution to AMS is not isolated when an AARM-defined subfabric is subtracted from the 'parent' AMS fabric. The most reliable techniques are (d); particularly AIRM and AARM (see also Potter, this volume). AIRM requires the measurement of IRMs applied in different directions, usually with a pulse-magnetizer. Advantages are high signal/noise ratio and speed (rapid magnetization process; multi-component measurements allow tensor determination from three orthogonal magnetization steps). The procedure is more time-consuming if fields exceed the nonlinearity threshold since multiple measurements are needed to determine each directional Rayleigh coefficient. Experimentation with pilot specimens at various applied fields may show it is possible to determine sensible AIRMs in the

339

linear regime for <50mT. However, Stephenson et al (1986) recommended <5mT. Despite its weaknesses, the AIRM method may produce some sensible comparisons with AMS fabrics and field microstructures (Borradaile & Dehls 1993). AARMs offer more reliable fabric interpretations, which offset the demands in time and technology (Jackson 1991). AARM was first used by McCabe et al (1985) to isolate a magnetite-OD from the whole-rock AMS, thus identifying a feeble tectonic magnetite subfabric overprint on bedding. AARM also permits the isolation of subfabrics within the magnetite subfabric; this is partial AARM (pAARM). pAARM is possible since most magnetite has coercivities below the peak field in routine AF demagnetizers; each pAARM may then be determined within a selected coercivity range within the decaying AF from its peak value, in an AF demagnetizer. Thus, one may measure pAARM for magnetite grain-subfabrics with different coercivity (Jackson et al 1988, 1989a, b\ Jackson 1991; Trindade et al 2001; Aubourg & Robion 2002; Aubourg et al 2000; Nakamura & Borradaile 20010, b) (see Fig. 24 later). The pAARMs define subfabric ODs characterized mostly by different grain-sizes or different degrees of internal stress. This technique may be useful with other ferromagnetic minerals if they have coercivities within the range of routine laboratory AF demagnetization. Finally, AARM is especially useful in assessing palaeomagnetic recording fidelity (Kodama & Sun 1992; Kodama 1997; Gattacceca & Rochette 2002), since it is often tensorial even in AFs much larger than the d.c.-field nonlinearity threshold. Interpreting AMS in terms of tectonic processes The community has recognized broad interpretative categories for single-generation AMS fabrics, including: (1) Finite strain (2) Strain history (kinematics) (3) Deformation mechanisms (grain-scale lattice realignment) (4) Metamorphism (5) Stress (incremental strain) Finite strain Field studies and rock-mechanical laboratory experiments have revealed some useful

340

G. J. BORRADAILE & M. JACKSON

relationships between AMS and the orientation of the finite strain ellipsoid (axes X > Y > Z) both in coaxial strain histories and those involving a notable shear-component. In the simplest, coaxial case, there may be a one-to-one mapping of X, 7, Z orientations with those of ftMAX, ftINT and «MIN» (Fig- 15a, b; see Fig. 24 later). Such simple angular relationships may be expected where the tectonic fabric successfully overprints or may be clearly distinguished from an earlier fabric (inter alia Hrouda 1982; Rochette & Vialon 1984; Siddans et al 1984; Benn & Allard 1988; Hirt et al 1988, 1993) or where, as in experiments, the initial fabric may be nearisotropic by design (Borradaile & Alford 1987, 1988; Borradaile & Puumala 1989). However, a coaxial strain history requires finite strain axes remaining constant with respect to the material as strain accumulates. In detail, this cannot be true and the non-material X, F, Z axes may spin in response to changing strain increments caused by inconsistent stress axial orientations: all strain histories are non-coaxial to some extent (Flinn 1962; Ramsay 1967; Ramberg 1975). Fortunately, the degree of noncoaxiality is beneath the detection-level in many cases; obvious exceptions are shear zones and mylonite zones. Most examples assume an accumulation of strain in which successive strain ellipsoids become progressively more anisotropic. Unfortunately, for extreme vorticities, later incremental strain ellipsoids may de-strain the previous ellipsoid so that the strain history is non-progressive, leading even to pulsating strain histories (Ramberg 1975; Ramsay & Huber 1983). The latter are probably restricted to high shear-strain zones, e.g. mylonites. Increasing field evidence, experimental evidence and theoretical considerations cast doubt on any meaningful cause-effectcorrelation of strain magnitudes on AMS (P7, or 7). Occasional apparent graphical correlations may give the illusion of a process-caused relationship but the real, lurking variable (Borradaile 2003), may be mineral-abundance and orientation-distribution (Borradaile 1988; Johns & Jackson 1991; Johns et al. 1992, Rochette et al. 1992).

Strain history (kinematics) Few rocks possess strain markers that reveal XYZ orientations and shear sense. However, more often AMS axes are compatible with tectonic structures that have well-known relationships to finite-strain orientations (Aubourg et al. 1997, 1999, 2000; Frizon de Lamotte et al. 2002; Saint-Bezar et al. 2002). Thus, magnetic

petrofabrics usually identify principal strain axes (XYZ) and sometimes independently confirm shear-sense from multiple subfabrics generated during a non-coaxial strain history. The sequence of three-dimensional fabric ellipsoids will necessarily be limited to the small number of identifiable magnetic subfabrics, e.g. paramagnetic matrix, MD-magnetite and PSD-magnetite, possibly supplemented by a field-fabric, e.g. quartz-feldspar schistosity. Since these subfabrics develop sequentially, their relative angular relationships reveal the shear-sense (e.g. Fig. 19a). Studies along the dextrally transpressed Archaean terrane boundaries in northern Ontario reveal successive subfabric axes compatible with dextral shear accompanied by upwards extension to the E-NE (Fig. 19b) (Borradaile & Dehls 1993; Borradaile et al. 1993; Werner & Borradaile 1996). Higher metamorphic grade on the northern sides of the boundaries and some traditional field evidence for dextral shear support these conclusions. The emplacement of plutonic bodies (or igneous ones, for that matter) may also be investigated using asynchronous subfabrics. Origins suggested for Archaean gneiss domes include emplacement by diapiric inflation or secondary doming. Clearly, the latter would disturb AMS axes equally at each site. In contrast, syn-metamorphic inflation would produce different angular relationships between KMAX - /^INT and XY (schistosity) at different sites according to the amount of radial inflation. For the Ash Bay dome of northern Ontario, the AMS foliation-dome, due largely to a late magnetite subfabric, is more subdued than the gneissic-foliation-dome, favouring an inflation origin (Fig. 19c; Borradaile & Gauthier 20030).

Deformation mechanisms (grain-scale processes) Since AMS is mostly due to crystallographic alignment (shape alignment in the special case of a high-ft mineral like magnetite) it is important to consider deformation mechanisms, which realign crystal lattices. Structural geologists have long been aware that passive continuum mechanical models inadequately model the actual process by which crystalline aggregates develop preferred crystallographic orientations (PCO), (e.g. Nicolas 1987; Nicolas & Poirier 1976). For example, modern petrofabrics and material science recognizes that alignment processes are due to environmental conditions such as mean pressure, deviatoric stress, incrementalstrain (^stress) history, vorticity, temperature

AMS-PETROFABRIC OF DEFORMED ROCKS

341

Fig. 19. During regional metamorphism and tectonism, subfabrics are normally initiated in the sequence quartz-feldspar, phyllosilicates/mafic silicate and finally magnetically-ordered phases (magnetite, pyrrhotite). (a) Successive differently oriented mineral fabrics may be recognized from the relative angles of quartz-feldspar schistosity, AMS (~chlorite/mica OD) and late 'ferro'magnetic phases, (b) Regional studies of field fabrics, AMS and AARM in northern Ontario reveal dextral, north-side-up movements of Archaean terraneboundaries (Werner & Borradaile 1996). (c) Relative orientation of gneissic foliation and AMS foliation favour gneiss-dome inflation in an Archaean metamorphic pluton (Borradaile & Gauthier 2003).

and fluid pressure (Nicolas & Poirier 1976; Means et al 1981; Poirier 1985; Nicolas 1987; Blenkinsop 2000). They recognize also that many intrinsic material properties influence alignment: crystal symmetry and process-activation energies and dynamic quantities such as dislocation density, grain-boundary shape and mobility, grain-size and disaggregation. Single-crystal processes Rigid body rotation Dating from the earliest days of structural geology (Wettstein 1886), this notion considers an object aligning its long axis by freely spinning along a locus towards X. The trace of its projection, for example on the XY plane, is given by the change from initial angle with X (9} to final angle (Of) via X/Y = tan(0)/tan(^). Its locus in threedimensions is fixed by combining the line's projections on the principal planes (Fig. 20a-c; Flinn 1962). Rotation is not necessarily by the shortest route, depending on the 7-extension;

normals to planar elements spin along inverse paths toward Z. March's (1932) method of strain analysis assumes this alignment model; thus, in any given direction the density p of normals to 'mica flakes' is predicted by the stretch; pz = Z~3. The most unrealistic assumption is that the initial OD is uniform; its unwise application in AMS and field strain studies may be attributed to the over-interpretation of some excellent but specialized rock-mechanics laboratory experiments (Nicolas & Poirier 1976; Tullis 1976). Although this is a common supporting process at low strain, in nature, spinning objects impinge and are impeded by the matrix. The association of KMAX and ^MIN orientations with X and Z respectively in nature is not sufficient evidence to invoke 'Marchian rotation' or any other purely geometrical continuum mechanical model. A state of finite strain may be achieved by an infinite number of strainhistories and rheology (and rheological history) is poorly constrained. For example, simple experiments with analogue materials (Fig. 20d, h) or

342

G. J. BORRADAILE & M. JACKSON

Fig. 20. (a—c) March (1932) envisaged the re-distribution of orientations of linear elements in space during a coaxial strain history. This used strict continuum mechanics assumptions (uniform initial OD, nonimpingement of elements, non-material properties of elements, etc.). The model provides a limiting case to constrain models and thought-experiments but is a poor approximation to any Natural alignment process, (d-e) AMS experiments with model materials and rock deformation experiments show superficial similarities with the movement paths predict the by the March model but are very sensitive to initial weak anisotropies; shape changes (e.g. 7}) may be non-progressive (Borradaile & Puumala 1989). (g-h) Noncoaxial strain history experiments show complex movement paths for AMS axes (Borradaile & Alford 1987, 1988). The fact that certain principal axes approach compatible strain axes is insufficient to justify March model rotations as the sole mechanism.

AMS-PETROFABRIC OF DEFORMED ROCKS

rock mechanical experiments with rock-analogues (Fig. 20e-g) reveal complex movement paths of AMS axes and non-progressive changes in AMS ellipsoid shape (Borradaile & Alford 1987, 1988; Borradaile & Puumala 1989). Those experiments showed a strong sensitivity to the slightest initial anisotropy that even involved the rapid switching of orientations for AMS axes. The following processes usually accommodate more strain effectively and are common in nature. Preferred nucleation New minerals may grow at the expense of others and their orientations may be controlled in several ways. First, overgrowths will be favoured on suitably aligned crystals with long habits (e.g. chlorite, mica, and actinolite), enhancing any initial alignment (Oertel 1983; Fig. 21a). Alternatively, neomineralization may create new lattices with their compliances appropriately controlled by the prevailing stress system; this could pro-

343

duce perfect saturation alignments of nuclei in high-grade rocks (Fig. 21b). Unfortunately, stress-nucleation metamorphic alignments may only record an ephemeral state of stress during nucleation (Fig. 21c). Subsequent stress or even finite strain increments may fail to leave a record in the fabric. Differentiation processes Many metamorphic differentiation processes produce heterogeneities and location fabrics. Pressure solution commonly produces textures readily sampled within an AMS specimen. A low strain-rate process, it is most effective at low temperature (<300°C), requiring a preferably mobile fluid-phase. Solutes such as quartz and calcite diffuse through the fluid to some local or distant reservoir from sites of high impingement stress. In some cases 60% of a rock has been transferred by pressure solution. The insoluble residue of phyllosilicates, aligned along stylolitic cleavage, commonly contributes

Fig. 21. Metamorphic effects, sensu s trie to, have largely been neglected and under-appreciated in the analysis of AMS fabrics, (a) PCO may develop due to the suppression of overgrowths on unfavourable oriented nuclei (Oertel 1983). (b) Stress-controlled nucleation crystallization may develop a saturation alignment that is aligned incompatibly with subsequent stress/strain history, (c) Non-coaxial strain histories may produce differently oriented subfabrics. During progressive metamorphism, minerals such as chlorite and biotite may form at several times in differently oriented subfabrics. (d) A rare study of progressive metamorphism, AMS and AARM revealed and addressed many natural complexities that should be considered more often (Housen & van der Pluijm 1991, 1993). Clearly, chlorite OD, 'cleavage', AMS and AARM subfabrics develop at different times and in different orientations.

344

G. J. BORRADAILE & M. JACKSON

Fig. 22. A deformation mechanism map (rear wall of diagram) represents dominant crystal-deformation and grain-alignment mechanisms in terms of environmental variables (e.g., temperature and differential stress), usually for a fixed grain-size (e.g., Nicolas 1987). The mechanisms may operate in subdued form outside their designated field. In nature, most of these mechanisms may not proceed without particulate flow of the grains (Borradaile 1981), a phenomenon that may be enhanced by fluid pressure. In principle, particulate flow may be entirely dependent on the intracrystalline processes or independent of them.

a significant AMS subfabric. A bedding-cleavage blended fabric may result with KMAX//f3 not X, since the cleavage-stylolites are rarely penetrative, (Fig. 17a). Like other deformationmechanisms pressure solution conditions may be identified on a deformation mechanism map (Nicolas 1987), extended to include also the intergranular processes, collectively particulate flow (Borradaile & Tarling 1981, 1984), (Fig. 22). Diffusion Particularly at high homologous temperature, say > 0.5rM (rM is the melting temperature in K), cation-diffusion pervades the lattices (NabarroHerring Creep) whereas diffusion follows grain boundaries at lower temperatures (Coble Creep). Change of crystal shape may be significant and it is not limited by grain interactions

as in the March model. Effectively, all grain boundaries are mobile with the available thermal energy. This permits alignments of minerals of even moderately weak anisotropy, giving rise to well-defined AMS fabrics, in rocks with seemingly poor tectonic fabrics such as deep continental granulites or mantle harzburgite (Benn & Allard 1988; Borradaile et al 1999b; Borradaile & Lagroix 2001). Diffusive processes are particularly sensitive to grain-size and to grain-shape which affect the diffusion-route. Crystal plasticity Early petrofabric work recognized strongly aligned crystal optic axes for matrix minerals such as quartz, calcite, olivine and pyroxene despite their weakly anisometric shapes (Sander 1930). Until metallurgical concepts of dislocation

AMS-PETROFABRIC OF DEFORMED ROCKS

motion were available in the 1950s, alignment mechanisms were poorly understood. However, in geology, causal correlations with structural/ petrofabric elements (schistosity, mineral lineation, L-S fabric) were acknowledged (Turner & Weiss 1963). For example, many mechanisms align quartz or calcite c axes steeply to Splanes, otherwise expressed by mica-alignment. Crystal-dislocations rearrange and multiply during any strain history, assisting the validation of Von Mises' criterion, that five independent dislocation slip systems suffice to strain (~'align') any crystal outline arbitrarily. Fewer systems suffice in practice since one system accommodates most slip, accompanied by intercrystalline motion (particulate flow). Characteristic, simple c axis distributions (small/great circle dispersions, point-clusters, etc.) are computer-modelled readily for the ideal polycrystal but in nature grain-interactions produce less neat petrofabric stereograms that may relate weakly to strain (and AMS). Laboratory experiments confirm theoretical predictions for the most part, though not as neatly and still with the simplification that ltriaxial' rig experiments usually impose macroscopic coaxial axial-symmetric strain (X = Y > Z), limited to <40% Z-shortening. Strain rates are normally high (~ 10~5 to 10-6s~1); >106 times faster than regional orogenic deformation, so that petrofabrics are replicated only for extreme processes (hot diffusion versus cool cataclastic flow). Stress-relaxation test permit processes to be inferred at much lower strain rates but the strains and textures may be undetectable. Crystal-plastic ODs are mostly temperature-dependent so that metamorphic grade may be useful in the interpretation of AMS. Experimental petrofabric ODs are usually produced at higher strain-rates and temperatures than their natural counterparts, and experimental strain histories are usually coaxial (with X = Y ^ Z). Whereas the stereogram usefully represents AMS and petrofabric ODs in a general natural tectonic or geographical reference frame, a crystal-axis coordinate system clarifies the relationships between ODs and strain for high-symmetry minerals with the experimental strain-symmetry (Owens & Rutter 1978). For high symmetry minerals the crystal axis coordinate system also corresponds to the AMS axes; important experimental evidence is available for trigonal quartz and calcite; orthorhombic olivine and orthopyroxene (e.g. Nicolas & Poirier 1976). Using crystal axes as a reference frame, the inverse pole figure illustrates the OD of the Z axes, recalling its limitation to (X = Y)

345

and high-symmetry minerals in a polycrystal with an isotropic initial OD (Fig. 23) (Nicolas 1987). The inverse pole-figure shows that in a plastically aligned polycrystal, petrofabric patterns are already quite complex and not always simply unimodal. For example, the distribution of some critical AMS axis (e.g. Ke//c axis, for calcite/quartz) may be dispersed along a small circle and difficult to interpret in terms of strain or kinematics. Equating neat theoretical or experimental PCOs too closely with AMS is perhaps premature, especially since for single-crystals, lowsymmetry usually weakens the relation between AMS and crystal-controlled deformation mechanisms (Fig. 1). Higher symmetry may permit more slip systems and the greater potential for crystal-plasticity. It may even be difficult to validate known strain histories from laboratory-produced petrofabrics for quartz, calcite, olivine and orthopyroxene (Nicolas & Poirier 1976). A further layer of symmetry constraints obfuscates the interpretation of strain-orientations from AMS. In nature vis-a-vis most experiments, further complications ensue since the OD is influenced by metamorphic grade (~temp.), X ^ Y and the strain history may be non-coaxial.

Aggregate processes Dynamic recrystallization Deformation-rheology under metamorphic conditions is often governed by competition of work-hardening and recovery from that increased strain-energy state. Depending on temperature and strain-rate, recovery may reduce grain-size to different degrees, in the first instance forming subgrains, rather than truly independent grains. Elongate ribbon grains, subgrain and grain rotation may accompany these processes but the tendency is toward grain-size reduction. These processes are well known in crustal quartz and feldspar and mantle pyroxenes. However, AMS in such tectonites may be an indirect consequence of the alignment of other, higher-^ minerals. Deformation mechanisms, metamorphism and AMS Structuralists conveniently collect thoughts on deformation mechanisms on a deformation mechanism map (Fig. 22). Differential stress, temperature, strain-rate and perhaps fluid pressure are important physical controls on deformation mechanisms. The first two controls are commonly chosen for axes of the graph, and

346

G. J. BORRADAILE & M. JACKSON

Fig. 23. (a) Geographic or specimen coordinate systems are the usual and logical reference frame to study AMS applications, (b, c) However, the inverse pole figure is superior to interpret crystalline-deformation mechanisms in triaxial experimental deformation experiments (Tullis 1973; Owens & Rutter 1978). Crystal axes provide the reference frame in which we plot the Z axes (shortening-directions) responsible for crystal alignment (Nicolas 1987). (d, e) The procedure only has advantage for macroscopically coaxial strain, with X — Y, as in most experimental deformation. Its use is mostly limited to high symmetry minerals such as trigonal quartz & calcite; orthorhombic olivine and orthopyroxene (Nicolas & Poirier 1976). These are fortunately of great interest in tectonics and petrofabrics (Fig. 1). (f) Possible expected crystallographic orientations with respect to Z in Iherzolite (Nicolas & Poirier 1976). (g) In non-coaxial strain histories, the stereographic projection must still be used. Here AMS and calcite c axes are compared for specimens with S-C shear zone textures (Borradaile & Me Arthur 1990).

AMS-PETROFABRIC OF DEFORMED ROCKS

strain-rate contours are plotted on this plane (rear wall of Fig. 22). However, diffusive processes, including pressure solution are grain-size dependent so that is sometimes chosen as an axis. The maps may be standardized to compare isomechanical materials, e.g. shear stress may be normalized to the material's shear modulus and temperature may be expressed as a fraction of the melting temperature (7"M in K), or even reexpressed as a composite P-T crustal-depthfunction (Ranalli 1987). Simplified maps, valid for one mineral and one grain-size, like Fig. 22, show ranges of conditions in which certain processes are most competitive. The mechanisms shown are not restricted to their labelled field. As a thought-guiding and problem-solving tool, the concept of a deformation mechanism map assists interpretations of magnetic fabrics. Simple continuum-mechanical 'strain-response' models bear little resemblance to the processes that align crystal lattices. AMS orientations are mostly referred to some structural or petrofabric feature, which in turn, may be defined more-or-less directly by finite strain axes (X, Y, Z axes from an L-S fabric) or by a kinematic pattern (non-coaxiality from some special structure, e.g. S-C fabric, shear zone). However, in terms of material processes, multiple deformation mechanisms may operate. For example, dislocation creep may align lattices. Although the AMS might be a simple function of the PCO for that reason (Fig. 23) other mechanisms may operate simultaneously; e.g. in low temperature environments pressure solution may also occur. Moreover, the deformation mechanisms on the rear-plane of the map are not independently capable of reshaping grains in a natural aggregate. Heterogeneous grain-strain demands intergranular motion, discussed next. Paniculate flow In all rocks, grain-strain requires intergranular motion. Finite strain of a polycrystalline, polymineralic assemblage is impossible without intergranular motion. Variously described as grain-boundary sliding or neighbour switching in different sciences, it is broadly paniculate flow. This may be entirely dependent on graindeformation as implied in the metallurgical and high-grade metamorphic contexts; it is a geometrical necessity to accommodate the incompatibility of new contiguous grain shapes (e.g. Flinn 19650; Nicolas & Poirier 1976; Poirier 1985). However, for low temperature and highstrain rate environments particulate flow may be independent of the crystal-alignment processes and this may affect AMS interpretation (Borradaile 1981; Borradaile & Tarling 1984).

347

As an extreme case, soft-sediment deformation or fault-rock cataclasis involve extensive particulate flow that are respectively without and independent of grain-deformation. Fluid pressure may enhance independent particulate flow, particularly at low metamorphic grade, and suppress the onset of grain-strain that would otherwise be evident. Commonly, grain-strain may be less than that of the aggregate and in a general noncoaxial stain history, the strain-analysis technique may determine which of several possible sets of X, Y, Z axes are reported and which of them may be related to AMS. Of course, mineral fabrics always control AMS, prima facie. Thus, interpreting AMS orientations requires some consideration of deformations mechanisms (Fig. 22) and the choice of strain analysis technique, some of which exclude the contribution of intergranular motion (Ramsay 1967; Ramsay & Huber 1983). The choice of strain analysis technique depends on the requirements; many methods measure grain-strain (Rf/3> of grains), some measure intergranular strain (centre-tocentre method) and some may measure both (strain of large objects) (Borradaile 1981).

AMS and metamorphism AMS orientation or ellipsoid-shape may vary with metamorphic grade or facies (Robion et al 1997; Rochette 1987b; Nakamura & Borradaile, this volume). However, AMS is rarely interpreted in terms of metamorphic processes, due to their complexities. A few studies draw attention to the role of recrystallization/ neomineralization mostly in classic areas, or with classic lithologies (Rochette 19870; Bina & Henry 1990; Henry 1990; Housen & van der Pluijm 1991; Jackson & Borradaile 1991; Housen et al. 19936, 1995). The principal problem is the extent to which the pre-existing tectonic fabric is masked by metamorphic recrystallization. In high-grade metamorphic environments the new fabric may nucleate in a nearperfectly aligned, saturation OD that obscures evidence of earlier subfabrics (Fig. 2la, b). On the other hand, at low grade, metamorphic recrystallization may be localized and develop subfabrics affecting only certain minerals, or certain location sub-fabrics (e.g. pressuresolution seams, strain-shadows, S-C foliations) (Aubourg et al. 1995; Robion et al. 1995). These studies combine all the sophistication recommended to reduce interpretation-ambiguity; mineralogical/crystallographic considerations, subfabric ODs, multiple mineralogical controls, and subfabric isolation by AARM or

348

G. J. BORRADAILE & M. JACKSON

AIRM (Jackson & Borradaile 1991; Nakamura & Borradaile 200la, b). A shale-slate transition study is particularly instructive, demonstrating partly overlapping histories for differently oriented chlorite, 'cleavage', AMS and AARM subfabrics (Housen & van der Pluijm 1990,

1991; Housen et al 1993b) (Fig. 2Id) and the inappropriateness of a solitary, continuummechanical alignment mechanism (e.g. Fig. 20a-c). In the chlorite-slates of Borrowdale, numerous studies show AMS is attributed largely to the alignment of chlorite and unrelated to

Fig. 24. Comparisons of AMS, pAARM and finite strain illustrate differently shaped magnitude ellipsoids for the different tensors in the same specimens, (a) Borrowdale Volcanic Slate (Nakamura & Borradaile, 2001). The finite strain ellipsoid defined by cleavage and mineral lineation is differently oriented than the AMS or highcoercivity pAARM fraction (15-30mT). Confidence cones for mean tensor of the low-coercivity pAARM fraction (magnetite, 3-1 OmT) indicate a very poorly defined OD. (b) AMS and soft and hard pAARM ellipsoids have different shapes and may be unsystematically associated (note tie-line for one specimen traverses distribution of others), (c) AMS, soft and hard pAARM and strain ellipsoids are very differently shaped; clearly strain magnitudes!shapes are quite unrelated to magnetic fabrics (see tieline for same specimen as in (b). (d) Cambrian Welsh Slate (Nakamura & Borradaile 2002); AMS axes closely relate to finite strain and petrofabric orientations (slaty cleavage and mineral lineation) but the reduced slates (magnetite-enhanced) show less perfect correlation, (e) Three pAARMs in the magnetite-enhanced, reduced 'spotted' slate suggest three different subfabrics. The softest pAARM (MD magnetite, 0-3 mT) corresponds most closely with the visible petrofabric.

AMS-PETROFABRIC OF DEFORMED ROCKS

finite strain. Henry (1990) identified the shortcomings of the strain-response models, sensu stricto, and recognized the importance of metamorphism (e.g. Fig. 2la, b). Orientation relationships of finite strain and AMS axes are still largely preserved in such cases (Borradaile & Mothersill 1984; Borradaile 1991). In the Cambrian Penrhynn Slates of North Wales, variably reduced purple slate show interesting metamorphic complications when finite strain, AMS, AIRM and pAARM are examined (Jackson & Borradaile 1991; Nakamura & Borradaile 200 la) and similar analyses of complex situations (Aubourg & Robion 2002; Aubourg et al 2000). AMS axes correspond well with finite strain but pAARM, broadly corresponding to SD, PSD and MD magnetite reveal successively more poorly defined subfabrics (Fig. 24) due to late metamorphism and the formation of magnetite in reduction spots, in this region. Strain was determined using Rf/$ strain analyses of grain

349

shapes. The late metamorphic reduction spots may be invalid strain markers in some cases (cf. Wood et al. 1976); in north Wales they appear to be post-metamorphic indicators of diffusionanisotropy. Elsewhere, reduction spots may be primary objects and thus valid strain markers (Siddans et al. 1984). Heat-treated magnetic susceptibility was first routinely used to warn of undesired changes in mineralogy in palaeomagnetic laboratories, between steps of thermal demagnetization. Modern studies reveal remarkably complex processes during laboratory heating, notwithstanding atmospheric controls (Hirt & Gehring 1991; Hrouda et al. 1997), usually increasing K. It is of concern that heating has been recommended to deliberately increase K, with the goal that AMS magnitudes/orientations of a certain subfabric will be enhanced. Although this was a reasonable suggestion when measurements were formerly less sensitive (e.g. Perarnau & Tarling

Fig. 25. In general, laboratory heating changes AMS orientations and also the shapes of AMS ellipsoids, both at the specimen-level and for the mean-tensor of homogeneous samples of specimens. Such heating does not necessarily enhance a natural fabric or subfabric, an artefact-subfabric may be produced. Only by careful tensor analysis of incrementally heated specimens may one be certain that a primary subfabric has been isolated (Henry et al. 2003).

350

G. J. BORRADAILE & M. JACKSON

1985), modern instruments render this unnecessary (e.g. Sapphire Instruments SI2B; AGICO Kappabridge). Laboratory heating may occasionally achieve the desired goal (Jelenska & Ka^dzialko-Hofmokl 1990); more usually differently oriented artefact-subfabrics are developed or enhanced (Fig. 25). However, a tensorial decomposition of AMS fabrics produced during incremental laboratory heating may characterize a primary fabric that has been desirably enhanced (Henry et al 2003; Souque et al. 2002).

Interpreting AMS axes as stress trajectories: neotectonics Graham (1954) easily convinced us that AMS might reflect the orientations of fabrics produced by tectonic or magmatic flow, in other words due to large finite strain. More recently, it has been a little more surprising to learn that AMS may detect incipient fabrics in very young, nonmetamorphosed strata without traditional evidence of penetrative strain, e.g. cleavage or mineral lineation ('L-S' fabrics). Lacking formal definition, 'neotectonics' appears to encompass deformation studies of unconsolidated or other young rocks, in most examples Quaternary (<2Ma) and usually no older than Pliocene (<5Ma). Generally neotectonic environments are characterized by an absence of penetrative deformation structures or substantial metamorphism. A similar concept to neotectonics, in rocks of any age, is the association of a cryptic or magnetic fabric with low-strain micro structures which may even represent ephemeral stress trajectories rather than finite strain axes. For example, stress-effects on magnetite produce convincing causal relationships between fractures and AARM (Jackson et al. 19896), or on AMS (Borradaile & Kehlenbeck 1994) or clear correlations between AMS and the fractal dimension in fault-zone microfractures (Nakamura & Nagahama 2001). The most aggressive opportunities for mineral alignment or 'metamorphism' in neotectonic environments appear to include diagenetic chemical reactions (especially involving clay minerals, iron oxides and sulphides), groundwater movement, anisotropic compaction, clayaggregate dewatering and calcite-twinning. Magnetite precipitation and other highsusceptibility alteration (Brothers et al. 1996; Elmore & McCabe 1991) are reported under sub-metamorphic conditions elsewhere (Banerjee et al. 1997). Important due to its palaeomagnetic complications, 'chemical' (i.e. late diagenetic

or low-grade metamorphic) remagnetization usually involves ferromagnetic re-mineralization and is well documented from diagenesis, burial environments, hydrocarbon reservoirs and orogenic foreland basins (Elmore & Crawford 1990; Elmore et al. 1987, 1993, 2001). Calcite, the most stress-sensitive petrofabric indicator, may provide the only independent petrofabric evidence of crystalline-deformation in neotectonic environments. Its c axes align nearly // maximum compressive stress (
AMS-PETROFABRIC OF DEFORMED ROCKS

domain-wall rearrangements. The sensitivity of weakly consolidated sedimentary rocks also permits AMS fabrics to develop from weakly aligned clay aggregates. Examples are known from AMS (AARM) fabrics associated with the opening and closure of a Miocene-Pliocene rift valley (Borradaile & Hamilton 2004), Neogene flysch basin-compression (Hrouda et al. 2002), Pleistocene rifted sediments in a transtensional regime (Facenna et al. 1994) and modern accretionary prisms (Housen & Kanamatsu 2003). Current plate motion vectors or the accepted geometrical relationships between faults and palaeostress trajectories permit us to interpret the AMS axes as palaeostress trajectories, suitably restrained by problems that might be due to magnetic mineralogy and composite fabrics.

351

A large sample from the depositional cover to the Troodos domed ophiolite revealed AMS axes compatible with gravitational flow of the chalks from depositional highs into local basins; weakly aligned clays control the net AMS fabric (Lagroix & Borradaile 2000Z?) (Fig. 26). In the same study, the more sensitive AARM detected a later neotectonic overprint. The original study of incremental-strain level magnetite fabrics identified a weak tectonic compressional subfabric and a primary depositional fabric in shale from pAARM and AMS fabrics, respectively (Jackson et al. 1989&; Jackson 1991). The same team related microfracture patterns and calcite-twin microfabrics that have known relationships to stress trajectories in gently folded carbonates, to AMS and AARM,

Fig. 26. Calcite petrofabrics are sensitive to weak strains and thus correspond more closely to stress axes than to finite strain axes. In limestone, weak neotectonic deformation is expressed simultaneously in calcite and by transmission in the clay and magnetite accessory minerals. AMS and AARM identify these respectively, revealing S or SSE verging motion associated with northward subduction on the Cyprean Arc, south of Cyprus (Lagroix & Borradaile 2000; see also Hamilton et al., this volume).

352

G. J. BORRADAILE & M. JACKSON

validating their axes as palaeostress trajectories (Jackson et al. 1989&). Relationships for calcitetwin microfabrics, AMS and stress similar to those inferred in the field (Fig. 23g) have been verified by experimental deformation of oolitic limestone with controlled strain-rate and stress in the laboratory (Borradaile & McArthur 1990 and unpublished experiments). Finally, the presence of AMS and AARM secondary fabrics in weakly consolidated Quaternary and Pliocene strata indicates that magnetic fabrics were imposed since 2 Ma. Stratigraphic considerations tentatively suggest a detectable AMS subfabric may evolve in <200Ka. Grain alignment could be more rapid in poorly consolidated sedimentary rocks, due to fluid expulsion, anisotropic volume-reduction, clay compaction, stress-sensitive calcite twinning and neomineralization (especially iron oxides and sulphides). Where rocks are sufficiently consolidated, AMS might be attributed to stress-invoked domain-wall rearrangements in polydomain magnetite where time-constraints are of less concern. Conclusions AMS provides a measure of the net OD of all minerals in a 10.5cm3 core-specimen of any rock in ~3 minutes, more comprehensively, quickly and reproducibly than any other petrofabric technique. Interpretation of AMS measurements, on the other hand, involves numerous complications, beginning with grainscale anisotropy: non-parallelism of crystal and principal susceptibility axes in monoclinic and triclinic minerals; ferromagnetic inclusions, which weaken the host-crystallographic control on AMS; and inverse anisotropy in SD ferromagnetic grains and certain paramagnets. Rock-specimen AMS is a convolution of particle anisotropies, ^-values, orientation distributions, and magnetostatic interactions. Discriminating methods of measurement and analysis allow us to separate and characterize different minerals, their grain-size distributions and their orientation distributions with varying success, to varying degrees. High-field magnetization and torque methods permit rigorous isolation of linear (dia-, para- and antiferromagnetic) and saturated (ferromagnetic) components; remanence measurements (AARM, AIRM, GRM) characterize different remanence-carrying subsets of the ferromagnetic assemblage. Variabletemperature AMS measurements allow quantitative separation of a paramagnetic AMS component when other components are

T-independent (diamagnets, Pauli paramagnets; approximately true for certain ferromagnets). Analysis of AMS variations with mean susceptibility, and differences between standardized and nonstandardized tensor means, provide indications of coexisting mineral subfabrics, as does non-orthorhombic symmetry of angular confidence regions for mean tensors. In some cases, deconstruction of magnetic fabrics allows recognition and characterization of superposed subfabrics, which result from complex emplacement/metamorphic/deformational histories. We are indebted to P. Rochette, K. Kodama and F. Martin Hernandez for their helpful comments and criticisms. Pierre's advice on magnetocrystalline anisotropy was particularly valuable. The Institute for Rock Magnetism is supported by a grant from the Instruments and Facilities Program, Earth Science Division, the National Science Foundation. This is IRM contribution number 04-02. Graham Borradaile's work and laboratory was supported by the Natural Sciences and Engineering Research Council of Canada (NSERC, Ottawa).

References AUBOURG, C. & ROBION, P. 2002. Composite ferromagnetic fabrics (magnetite, greigite) measured by AMS and partial AARM in weakly strained sandstones from western Makran, Geophysical Journal International, 151, 729-737. AUBOURG, C., ROCHETTE, P. & BERGMULLER, F. 1995. Composite magnetic fabric in weakly deformed black shales. Physics of the Earth and Planetary Interiors, 87, 267-278. AUBOURG C., FRIZON DE LAMOTTE, D., POISSON, A. & MERCIER, E. 1997. Magnetic fabric and oblique ramp-related fold: a case study from the western Taurus (Turkey). Journal of Structural Geology, 19, 1111-1120. AUBOURG, C., ROCHETTE, P., STEPHAN, J. F., POPOFF, M. & CHABERT-PELLINE, C. 1999. The magnetic fabric of weakly deformed late Jurassic shales from the southern Subalpine chains (French Alps): Evidence for SW-directed tectonic transport direction. Tectonophysics, 308, 473-486. AUBOURG, C., HEBERT, R., JOLIVET, L. & CARTAYRADE, G. 2000. The magnetic fabric in a detachment shear zone: The Example of Tinos Island (Greece). Tectonophysics, 321, 219-236. BALSLEY, J. R. & BUDDINGTON, A. F. 1960. Magnetic susceptibility anisotropy and fabric of some Adirondack granites and orthogneisses. American Journal of Science, 258, 6-20. BANERJEE, S. K. & STAGEY, F. D. 1967. The high-field torque-meter method of measuring magnetic anisotropy in rocks. In: COLLINSON, D. W., CREER, K. M. & RUNCORN, S. K. (eds) Methods in Palaeomagnetism. Elsevier, Amsterdam, New York, 470-476.

AMS-PETROFABRIC OF DEFORMED ROCKS BANERJEE, S., ELMORE, R. D. & ENGEL, M. H. 1997. Chemical remagnetization and burial diagenesis: testing the hypothesis in the Pennsylvanian Belden Formation, Colorado. Journal of Geophysical Research, 102(611), 24825-24842. BENN, K. 1994. Overprinting of magnetic fabrics in granites by small strains: numerical modelling. Tectonophysics, 233, 153-162. BENN, K. & ALLARD, B. 1988. Preferred mineral orientations related to magnetic flow in ophiolite layered gabbros. Journal of Petrology, 30, 925946. BINA, M. M. & HENRY, B. 1990. Magnetic properties, opaque mineralogy, and magnetic anisotropies of serpentinized peridotites from ODP Hole 670A near the mid-Atlantic ridge. Physics of the Earth and Planetary Interiors, 65, 88-103. BLENKINSOP, T. 2000. Deformation Micro structures and Mechanisms in Minerals and Rocks. Kluwer, Dordrecht. BORRADAILE, G. J. 1981. Partlculate flow of rock and the formation of cleavage. Tectonophysics, 72, 305-321. BORRADAILE, G. J. 1987. Anisotropy of magnetic susceptibility: rock composition versus strain. Tectonophysics, 138, 327-329. BORRADAILE, G. J. 1988. Magnetic susceptibility, petrofabrics, and strain: a review. Tectonophysics, 156, 1-20. BORRADAILE, G. J. 1991. Correlation of strain with anisotropy of magnetic susceptibility (AMS). Pure and Applied Geophysics, 135, 15-29. BORRADAILE, G. J. 1994. Paleomagnetism carried by crystal inclusions: the effect of preferred crystallographic orientation. Earth and Planetary Science Letters, 126, 171-182. BORRADAILE, G. J. 2001. Magnetic fabrics and petrofabrics; their orientation distributions and anisotropies. Journal of Structural Geology, 23, 1581-1596. BORRADAILE, G. J. 2003. Statistics of Earth Science Data. Springer Verlag. BORRADAILE, G. J. & ALFORD, C. 1987. Relationship between magnetic susceptibility and strain in laboratory experiments. Tectonophysics, 133, 121-135. BORRADAILE, G. J. & ALFORD, C. 1988. Experimental shear zones and magnetic fabrics. Journal of Structural Geology, 10, 895-904. BORRADAILE, G. J. & DEHLS, J. F. 1993. Regional kinematics inferred from magnetic subfabrics in Archean rocks of northern Ontario, Canada. Journal of Structural Geology, 15, 887-894. BORRADAILE, G. J. & GAUTHIER, D. 2001. AMSdetection of inverse fabrics without AARM, in ophiolite dikes. Geophysical Research Letters, 28(18), 3517-3520. BORRADAILE, G. J. & GAUTHIER, D. 2003a. Emplacement of an Archean gneiss dome, northern Ontario, Canada: inflation inferred from magnetic fabrics. Tectonics, 22.. BORRADAILE, G. J. & GAUTHIER, D. 2003b. Interpreting anomalous magnetic fabrics in ophiolite dikes. Journal of Structural Geology, 25, 171-182.

353

BORRADAILE, G. J. & HAMILTON, T. 2004. Magnetic fabrics may proxy as neotectonic stress trajectories, Polis Rift, Cyprus. Tectonics, 23. BORRADAILE, G. J. & HENRY, B. 1997. Tectonic applications of magnetic susceptibility and its anisotropy. Earth-Science Reviews, 42, 49-93. BORRADAILE, G. J. & KEHLENBECK, M. M. 1996. Possible cryptic tectono-magnetic fabrics in 'post-tectonic' granitoid plutons of the Canadian Shield. Earth and Planetary Science Letters, 137, 119-127. BORRADAILE, G. J. & LAGROIX, F. 2001. Magnetic fabrics reveal upper mantle flow fabrics in the Troodos ophiolite complex, Cyprus. Journal of Structural Geology, 23, 1299-1317. BORRADAILE, G. J. & MCARTHUR, J. 1990. Experimental calcite fabrics in a synthetic weaker aggregate by coaxial and non-coaxial deformation. Journal of Structural Geology, 12, 351-363. BORRADAILE, G. J. & MOTHERSILL, J. S. 1984. Coaxial deformed and magnetic fabrics without simply correlated magnitudes of principal values. Physics of the Earth and Planetary Interiors, 56, 254-265. BORRADAILE, G. J. & MOTHERSILL, J.S. 1991. Experimental strain of isothermal remanent magnetisation in ductile sandstone. Physics of the Earth & Planetary Interiors, 65, 308-318. BORRADAILE, G. J. & PUUMALA, M. A. 1989. Synthetic magnetic fabrics in plasticine. Tectonophysics, 164, 73-78. BORRADAILE, G. J. & SARVAS, P. 1990. Magnetic susceptibility fabrics in slates: structural, mineralogical and lithological influences. Tectonophysics, 172,215-222. BORRADAILE, G. J. & SPARK, R.N. 1991. Deformation of the Archean Quetico-Shebandowan subprovince boundary in the Canadian Shield near Kashabowie, northern Ontario. Canadian Journal of Earth Sciences, 28, 116-125. BORRADAILE, G. J. & STUPAVSKY, M. 1995. Anisotropy of magnetic susceptibility: Measurement schemes. Geophysical Research Letters, 22, 1957-1960. BORRADAILE, G. J. & TARLING, D. H. 1981. The influence of deformation mechanisms on magnetic fabrics in weakly deformed rocks. Tectonophysics, 77, 151-168. BORRADAILE, G. J. & TARLING, D. H. 1984. Strain partitioning and magnetic fabrics in particulate flow. Canadian Journal of Earth Sciences, 21, 694-697. BORRADAILE, G. J. & WERNER, T. 1994. Magnetic anisotropy of some phyllosilicates. Tectonophysics, 235, 223-248. BORRADAILE, G. J., MOTHERSILL, J. S. & TARLING, D. H. 1985. Sources of magnetic susceptibility in a slate. Earth and Planetary Science Letters 76, 336-340. BORRADAILE, G. J., KEELER, W., ALFORD, C. & SARVAS, P. 1987. Anisotropy of magnetic susceptibility of some metamorphic minerals. Physics of the Earth and Planetary Interiors, 48, 161-166. BORRADAILE, G. J., SARVAS, P., DUTKA, R., STEWART, R. & STUBLEY, M. 1988. Transpression in slates along the margin ofan Archaean gneiss belt, northern Ontario-magnetic fabrics and

354

G. J. BORRADAILE & M. JACKSON

petrofabrics. Canadian Journal of Earth Sciences, 25, 1069-1077. BORRADAILE, G. J., TELLA, S. & MCARTHUR, J. 1989. Magnetic fabric as a kinematic indicator of faults: a test case. Annals de Tectonicae, 3, 3-11. BORRADAILE, G. J., MACKENZIE, A. & JENSEN, E. 1990. Silicate versus trace mineral susceptibility in metamorphic rocks. Journal of Geophysical Research, B, Solid Earth and Planets, 95, 84478451. BORRADAILE, G. J., PUUMALA, M. & STUPAVSKY, M. 1992. Anisotropy of complex magnetic susceptibility as an indicator of strain and petrofabric in rocks bearing sulphides. Tectonophysics, 202, 309-318. BORRADAILE, G. J., STEWART, R. A. & WERNER, T. 1993. Archean uplift of a subprovince boundary in the Canadian Shield, revealed by magnetic fabrics. Tectonophysics, 227, 1-15. BORRADAILE, G. J., FRALICK, P. W. & LAGROIX, F. 1999a. Acquisition of anhysteretic remanence and tensor subtraction from AMS isolates true palaeocurrent grain alignments. In: TARLING, D. H. & TURNER, P. (eds) Palaeomagnetism and Diagenesis in Sediments. Geological Society Special Publications, London, 151, 139-145. BORRADAILE, G. J., WERNER, T. & LAGROIX, F. 1999b. Magnetic fabrics and anisotropy-controlled thrusting in the Kapuskasing structural zone, Canada. Tectonophysics, 302, 241-256. BROTHERS, L. A., ENGEL, M. H. & ELMORE, R. D. 1996, The late diagenetic conversion of pyrite to magnetite by organically complexed ferric iron. Chemical Geology, 130, 1-15. CANON-TAFIA, E. 1994. Anisotropy of magnetic susceptibility parameters; guidelines for their rational selection. Pure and Applied Geophysics, 142, 365-382. CANON-TAFIA, E. 1996. Single-grain versus distribution anisotropy: a simple three-dimensional model. Physics of the Earth and Planetary Interiors, 94, 149-158. CASEY, M., RUTTER, E., SCHMID, S., SIDDANS, A. W. B. & WHALLEY, J. 1978. Texture development in experimentally deformed calcite rocks. In: GOTTSTEIN, G. & LUCKE, K. (eds) International Conference on the textures of materials 5, Aachen, 199-210. CHIKAZUMI, S. & CHARAP, S. H. 1964. Physics of Magnetism. John Wiley & Sons, New York. ClFELLI, F., ROSETTI, F., MATTEI, M., HlRT, A. M.,

RENATO, F. & TORTORICI, L. 2004. An AMS, structural and paleomagnetic study of Quaternary deformation in eastern Sicily. Journal of Structural Geology, 26, 29-46. CISOWSKI, S. 1981. Interacting vs. non-interacting single-domain behavior in natural and synthetic samples. Physics of the Earth and Planetary Interiors, 26, 77-83. CLARK, D. A. 1984. Hysteresis properties of sized dispersed monoclinic pyrrhotite grains. Geophysical Research Letters, 11, 173-176. CLARK, D. A. & TONKIN, C. 1994. Magnetic anomalies due to pyrrhotite: examples from the Cobar area,

N.S.W., Australia. Journal of Applied Geophysics, 32, 11-32. CLARK, D. A., EMERSON, D. W. & KERR, T. L. 1988. The use of electrical conductivity and magnetic susceptibility tensors in rock fabric studies. Exploration Geophysics, 19, 244—248. CLARK, D. A., FRENCH, D. H., LACKIE, M. A. & SCHMIDT, P. W. 1992. Magnetic petrology: application of integrated rock magnetic and petrological techniques to geological interpretation of magnetic surveys. Exploration Geophysics, 23, 65-68. COGNE, J.-P. 1987. TRM deviations in anisotropic assemblages of multidomain magnetites. Geophysical Journal of the Royal Astronomical Society, 90, 1013-1023. COLLINSON, D. W. 1983. Methods in Rock Magnetism and Palaeomagnetism: Techniques and Instrumentation. Chapman and Hall, New York. COTTRELL, R. D. & TARDUNO, J. A. 1999. Geomagnetic paleointensity derived from single plagioclase crystals. Earth and Planetary Science Letters, 169(1-2), 1-5. DALY, L. 1967. Possibilite d'existence dans les roches, de plusieurs anistropies magnetiques superposees et leur separation. Comptes rendus hebdomadaires des seances de I'Academic des Sciences (Paris), SerieB, 264, 1377-1380. DALY, L. & ZINSSER, H. 1973. Etude comparative des anisotropies de susceptibilite et d'aimantation remanente isotherme: consequences pour 1'analyse structurale et le paleomagnetisme. Annales de Geophysique, 29, 189-200. DANKERS, P. H. M. & ZIJDERVELD, J. D. A. 1981. Alternating field demagnetization of rocks, and the problem of gyromagnetic remanence. Earth and Planetary Science Letters, 53, 89-92. DE WALL, H., BESTMANN, M. & ULLEMEYER, K. 2000. Anisotropy of diamagnetic susceptibility in Thassos marble; a comparison between measured and modeled data. Journal of Structural Geology, 22, 1761-1771. DEKKERS, M. J. 1989a. Magnetic properties of natural goethite—II. TRM behaviour during thermal and alternating field demagnetization and lowtemperature treatment. Geophysical Journal, 97, 341-355. DEKKERS, M. J. 1989b. Magnetic properties of natural pyrrhotite. II. High- and low-temperature behavior of Jrs and TRM as a function of grain size. Physics of the Earth and Planetary Interiors, 57, 266-283. DEKKERS, M. J., MATTEI, J.-L., PILLION, G. & ROCHETTE, P. 1989. Grain-size dependence of the magnetic behavior of pyrrhotite during its lowtemperature transition at 34K. Geophysical Research Letters, 16, 855-858. DUNLOP, D. J. & OZDEMIR, 6. 1997. Rock Magnetism: Fundamentals and Frontiers. Cambridge University Press, Cambridge. DUNLOP, D. J. & WEST, G. 1969. An experimental evaluation of single-domain theories. Reviews of Geophysics, 1, 709-757. EDWARDS, J. 1982a. Gyroremanent magnetization produced by specimen rotation between successive

AMS-PETROFABRIC OF DEFORMED ROCKS alternating field treatments. Geophysical Journal of the Royal Astronomical Society, 71, 199-214. EDWARDS, J. 1982b. Studies of partial rotational remanent magnetization and rotational remanent magnetization at slow speeds of rotation. Geophysical Journal of the Royal Astronomical Society, 68, 609-624. EDWARDS, J. 1986. Directions of anhysteretic and rotational remanent magnetizations in rotating samples. Geophysical Journal of the Royal Astronomical Society, 85, 209-215. EGLI, R. & LOWRIE, W. 2002. Anhysteretic remanent magnetization of fine magnetic particles. Journal of Geophysical Research-Solid Earth, 108(B2), art. no. 2209. ELLWOOD, B. B. TERRELL, G. E. & COOK, W. J. 1993. Frequency dependence and the electromagnetic susceptibility tensor in magnetic fabric studies. Physics of the Earth and Planetary Interiors, 80, 65-73. ELMORE, R. D. & CRAWFORD, L. 1990. Remanence in authigenic magnetite: testing the hydrocarbonmagnetite hypothesis. Journal of Geophysical Research B: Solid Earth, 95, 4539-4550. ELMORE, R. D. & McCABE, C. 1991. The occurrence and origin of remagnetization in the sedimentary rocks of North America. Reviews of Geophysics 29, 377-383. ELMORE, R. D., ENGEL, M. H., CRAWFORD, L., NICK, K., IMBUS, S. & SOFER, Z. 1987. Evidence for a relationship between hydrocarbons and authigenic magnetite. Nature, 325, 428-430. ELMORE, R. D., LONDON, D., BAGLEY, D., FRUIT, D. & GAO, G. 1993. Remagnetization by basinal fluids: testing the hypothesis in the Viola Limestone, southern Oklahoma. Journal of Geophysical Research B: Solid Earth, 98, 6237-6254. ELMORE, R. D., KELLEY, J., EVANS, M. & LEWCHUK, M. T. 2001. Remagnetization and orogenic fluids: testing the hypothesis in the central Appalachians. Geophysical Journal International, 144, 568-76. FACENNA, C., FUNICIELLO, R., BRUNI, A., MATTEI, M. & SAGNOTTI, L. 1994. Evolution of a transferrelated basin: the Ardea basin (Latiou, central Italy). Basin Research, 6, 35-46. FERRE, E. C. 2002. Theoretical models of intermediate and inverse AMS fabrics. Geophysical Research Letters, 29, 31-1-4. FERRE, E. C., MARTIN-HERNANDEZ, F., TEYSSIER, C. & JACKSON, M. 2004. in press. Paramagnetic and ferromagnetic AMS in migmatites: Measurements in high and low fields and kinematic implications. Geophysical Journal Interntional. FISHER, N. L, LEWIS, T. & EMBLETON, B. J. J. 1987. Statistical analysis of spherical data. Cambridge University Press. FLINN, D. 1962. On folding during three-dimensional progressive deformation. Geological Society of London Quaterly Journal, 118, 385-433. FLINN, D. 1965a. Deformation in metamorphism. In: PITCHER, W. S. & FLINN, G. W. (eds) Controls of Metamorphism. Oliver and Boyd, Edinburgh, 46-72.

355

FLINN, D. 1965b. On the symmetry principle and the deformation ellipsoid. Geology Magazine, 102, 36-45. FRIEDERICH, D. 1995. Gefugeuntersuchungen an Amphiboliten der Bhomishen Masse unter besonderer Berucksichtigung der Anisotropie der magnetischen Suszetibiltat. Geotektonische Forschungen, 82, 1-118. FRIZON DE LAMOTTE, D., SOUQUE, C., GRELAUD, S. & ROBION, P., 2002. Early record of tectonic magnetic fabric during inversion of a sedimentary basin: examples from the Corbieres transfer zone. Bulletin de la Societe Geologique de France, 173: 461-469. FUKUMA, K. 2002. Effect of interaction on ARM. In: SAGNOTTI, L. & ROBERTS, A. P. (eds) Fundamental Rock Magnetism and Environmental Applications. Istituto Nazionale di Geofisica e Vulcanologia, Erice, Sicily, 45^6. FULLER, M. 1961. Magnetic Anisotropy of Rocks. Ph.D. thesis, Cambridge University. FULLER, M. 1963. Magnetic anisotropy and paleomagnetism. Journal of Geophysical Research, 68, 293-309. FULLER, M. 1987. Experimental methods in rock magnetism and paleomagnetism. In: SAMMIS, C. G. & HENYEY, T. L. (eds) Methods of Experimental Physics, 24A. Academic Press, Orlando, 303-471. FULLER, M. & KOBAYASHI, K. 1964. Identification of magnetic phases carrying natural remanent magnetization in certain rocks. Journal of Geophysical Research, 69, 2111-2120. GATTACCECA, J. & ROCHETTE, P. 2002. Pseudopaleosecular variation due to remanence anisotropy in a pyroclastic flow succcession. Geophysical Research Letters, 29, 4. GIRDLER, R. W. 1961. The measurement and computation of anisotropy of magnetic susceptibility in rocks. Geophysical Journal of the Royal Astronomical Society, 5, 34-44. GRAHAM, J. W. 1954. Magnetic susceptibility anisotropy, an unexploited petrofabric element. Geological Society of America Bulletin, 65, 1257-1258. GREGOIRE, V., DE SAINT BLANQUAT, M., NEDELEC, A. & BOUCHEZ, J.-L. 1995. Shape anisotropy versus magnetic interactions of magnetite grains: Experiments and application to AMS in granitic rocks. Geophysical Research Letters, 22, 2765-2768. HARGRAVES, R. B. 1959. Magnetic anisotropy and remanent magnetism in hemo-ilmenite from ore deposits at Allard Lake, Quebec. Journal of Geophysical Research, 64, 1565-1578. HARGRAVES, R. B., JOHNSON, D. & CHAN, C. Y. 1991. Distribution anisotropy: the cause of AMS in igneous rocks? Geophysical Research Letters, 18, 2193-2196. HEDLEY, I. G. 1971. The weak ferromagnetism of goethite (a-FeOOH). Zeitschrift fur Geophysik, 37, 409-420. HENKEL, O. 1964. Remanenzverhalten und Wechselwirkungen in hartmagnetischen Teilchenkollektiven. Physica Status Solidi, 7, 919-924. HENRY, B. 1983. Interpretation quantitative de 1'anisotropie de susceptibilite magnetique. Tectonophysics, 91, 165-177.

356

G. J. BORRADAILE & M. JACKSON

HENRY, B. 1989. Magnetic fabric and orientation tensor of minerals in rocks. Tectonophysics, 165, 21-28. HENRY, B. 1990. Magnetic fabric implications for the relationships between deformation mode and grain growth in slates from the Borrowdale Volcanic Group in the English Lake District. Tectonophysics, 178, 225-230. HENRY, B. 1992. Modelling the relationship between magnetic fabric and strain in polymineralic rocks. Physics of the Earth and Planetary Interiors, 70, 214^218. HENRY, B. & DALY, L. 1983. From qualitative to quantitative magnetic anisotropy analysis: the prospect of finite strain calibration. Tectonophysics, 98, 327-336. HENRY, B. & LE GOFF, M. 1995. Application de 1'extension bivariate de la statistique de Fisher aux donnees d'anisotropie de susceptibilite magnetique: integration des incertitudes de mesure sur 1'orientation des directions principales. Comptes Rendus de I'Academic des Sciences, Serie 2, Mecanique, Physique, Chimie, Sciences de I'Univers, Sciences de la Terre (1037-1042). HENRY, B., JORDANOVA, D., JORDANOVA, N., SOUQUE, C. & ROBION, P. 2003. Anisotropy of magnetic susceptibility of heated rocks. Tectonophysics, 366, 241-258. HENRY, B., JORDANOVA, D., JORDANOVA, N., SOUQUE, C. & ROBION, P., 2003. Anisotropy of magnetic susceptibility of heated rock. Tectonophysics, 366, 241-258. HENRY, S. G. 1979. Chemical demagnetization methods, procedures, and applications through vector analysis. Canadian Journal of Earth Sciences, 16, 1832-1841. HEXT, G. R. 1963. The estimation of second-order tensors, with related tests and designs. Biometrika 50, 353-373. HIRT, A. M. & GEHRING, A. U. 1991. Thermal alteration of the magnetic mineralogy in ferruginous rocks. Journal of Geophysical Research B: Solid Earth, 96, 9947-9953. HIRT, A., LOWRIE, W., CLENDENEN, W. & KLIGFIELD, R. 1988. The correlation of magnetic anisotropy with strain in the Chelmsford formation of the Sudbury Basin, Ontario. Tectonophysics, 145, 177-189. HIRT, A. M., LOWRIE, W., CLENDENEN, W. S. & KLIGFIELD, R. 1993. Correlation of strain and the anisotropy of magnetic susceptibility in the Onaping Formation; evidence for a near-circular origin of the Sudbury Basin. Tectonophysics, 225, 231-254. HOSSACK, J. R. 1967. Pebble deformation and thrusting in the Bygdin area (southern Norway). Tectonophysics, 5: 315-339. HOUSEN, B. A. & KANAMATSU, T. 2003. Magnetic fabrics from the Costa Rica margin; sediment deformation during the initial dewatering and underplating process. Earth and Planetary Science Letters 206, 215-228. HOUSEN, B. A. & VAN DER PLUIJM, B. A. 1990. Chlorite control of correlations between strain and aniso-

tropy of magnetic susceptibility. Physics of the Earth and Planetary Interiors, 61(3-4), 315-323. HOUSEN, B. A. & VAN DER PLUIJM, B. A. 1991. Slaty cleavage development and magnetic anisotropy fabrics. Journal of Geophysical Research, 96, 9937-9946. HOUSEN, B. A., RICHTER, C. & VAN DER PLUIJM, B. A. 1993a. Composite magnetic anisotropy fabrics: experiments, numerical models, and implications for the quantification of rock fabrics. Tectonophysics, 220, 1-12. HOUSEN, B. A., VAN DER PLUIJM, B. A. & VAN DER Voo, R. 1993b. Magnetite dissolution and neocrystallization during cleavage formation: paleomagnetic study of the Martinsburg formation, Lehigh Gap, Pennsylvania. Journal of Geophysical Research B: Solid Earth, 98, 13,799-13,813. HOUSEN, B. A., VAN DER PLUIJM, B. A. & ESSENE, E. J. 1995. Plastic behavior of magnetite and high strains obtained from magnetic fabrics in the Parry Sound shear zone, Ontario Grenville Province. Journal of Structural Geology, 17, 265—278. HROUDA, F. 1982. Magnetic anisotropy of rocks and its application in geology and geophysics. Geophysical Surveys, 5, 37-82. HROUDA, F. 1986. The effect of quartz on the magnetic anisotropy of quartzite. Studia Geophysica et Geodaetica, 30, 39-45. HROUDA, F. 1992. Separation of a component of tectonic deformation from a complex magnetic fabric. Journal of Structural Geology, 14, 65-71. HROUDA, F. 1993. Theoretical models of magnetic anisotropy to strain relationship revisited. Physics of the Earth and Planetary Interiors, 77, 237-249. HROUDA, F. 2002a. Low-field variation of magnetic susceptibility and its effect on the anisotropy of magnetic susceptibility of rocks. Geophysical Journal International, 150, 715-23. HROUDA, F. 2002b. The use of the anisotropy of magnetic remanence in the resolution of the anisotropy of magnetic susceptibility into its ferromagnetic and paramagnetic components. Tectonophysics, 347,269-281. HROUDA, F. & JELINEK, V. 1990. Resolution of ferrimagnetic and paramagnetic anisotropies in rocks, using combined low-field and high-field measurements. Geophysical Journal International, 103, 75-84. HROUDA, F. & SCHULMANN, K. 1990. Conversion of the magnetic susceptibility tensor into the orientation tensor in some rocks. Physics of the Earth and Planetary Interiors, 63, 71-77. HROUDA, F., JEL!NEK, V. & ZAPLETAL, K. 1997. Refined technique for susceptibility resolution into ferromagnetic and paramagnetic components based on susceptibility temperature-variation measurement. Geophysical Journal International, 129, 715-719. HROUDA, F., HENRY, B. & BORRADAILE, G. 2000. Limitations of tensor subtraction in isolating diamagnetic fabrics by magnetic anisotropy. Tectonophysics, 322, 303-310. HROUDA, F., PUTIS, M. & MADARAS, J. 2002. The Alpine overprints of the magnetic fabrics in the

AMS-PETROFABRIC OF DEFORMED ROCKS basement and cover rocks of the Yeporic Unit (Western Carpathians, Slovakia). Tectonophysics, 359, 271-288. Hsu, T. C. 1966. The charactersitics of coaxial and noncoaxial strain paths. Journal of Strain Analysis, 1, 216-222. HUNT, C. P., MOSKOWITZ, B. M. & BANERJEE, S. K. 1995. Magnetic properties of rocks and minerals. In: AHRENS, T. J. (ed.) A Handbook of Physical Constants, vol. 3. American Geophysical Union, Washington, DC, 189-204. IHMLE, P. F., HIRT, A. M., LOWRIE, W. & DIETRICH, D. 1989. Inverse fabric in deformed limestones of the Morcles Nappe, Switzerland. Geophysical Research Letters, 16, 1383-1386. ISING, G. 1942. On the magnetic properties of varved clay. Arkiv for Matematik, Astronomi och Fysik, 29A, 1-37. JACKSON, M. J. 1991. Anisotropy of magnetic remanence: a brief review of mineralogical sources, physical origins, and geological applications, and comparison with susceptibility anisotropy. Pure and Applied Geophysics, 136, 1-28. JACKSON, M. J. & BORRADAILE, G. 1991. On the origin of the magnetic fabric in purple Cambrian slates of north Wales. Tectonophysics, 194, 49-58. JACKSON, M. J. & TAUXE, L. 1991. Anisotropy of magnetic susceptibility and remanence: developments in the characterization of tectonic, sedimentary, and igneous fabric. Reviews of Geophysics, suppl. (IUGG Report-Contributions in Geomagnetism and Paleomagnetism) , 29, 371-376. JACKSON, M. J., GRUBER, W., MARVIN, J. A. & BANERJEE, S. K. 1988. Partial anhysteretic remanence and its anisotropy: applications and grainsize dependence. Geophysical Research Letters, 15, 440-443. JACKSON, M. J., CRADDOCK, J. P., BALLARD, M. M. & VAN DER Voo, R. 1989a. Anhysteretic magnetic anisotropy, and calcite twin strains in Devonian carbonates from New York State. Tectonophysics, 161, 43-53. JACKSON, M., SPROWL, D. & ELL WOOD, B. 1989b. Anisotropies of partial anhysteretic remanence and susceptibility in compacted black shales: grainsizeand composition-dependent magnetic fabric. Geophysical Research Letters, 16, 1063-1066. JACKSON, M., BORRADAILE, G. J., HUDLESTON, P. & BANERJEE, S. 1993. Experimental deformation of synthetic magnetite-bearing calcite sandstones; effects on remanence, bulk magnetic properties, and magnetic anisotropy. Journal of Geophysical Research, B, Solid Earth and Planets, 98, 383-401. JACKSON, M., MOSKOWITZ, B., ROSENBAUM, J. & KISSEL, C. 1998. Field-dependence of AC susceptibility in titanomagnetites. Earth and Planetary Science Letters, 157, 129-39. JELENSKA, M. & K^dzialko-Hofmokl, M. 1990. Dependence of anisotropy of magnetic susceptibility of rocks on temperature. Physics of Earth & Planetary Science International, 62, 19-31. JELINEK, V. 1976. The statistical theory of measuring anisotropy of magnetic susceptibility of rocks and its application. Geophyzika, Brno.

357

JELINEK, V. 1978. Statistical processing of anisotropy of magnetic susceptibility measures on groups of specimens. Studia geophisica et geodetica, 22, 50-62. JELINEK, V. 1981. Characterization of the magnetic fabrics of rocks. Tectonophysics, 79, T63-T67. JELINEK, V. 1985. The physical principles of measuring magnetic anisotropy with the torque magnetometer. Travaux Geophysiques, 33, 177-198. JELINEK, V. 1993. Theory and measurement of the anisotropy of isothermal remanent magnetization of rocks. Travaux geophysiques, 37, 124-134. JOHNS, M. & JACKSON, M. 1991. Compositional control of anisotropy of remanent and induced magnetization in synthetic samples. Geophysical Research Letters, 18, 1293-1296. JOHNS, M. K., JACKSON, M. J. & HUDLESTON, P. J. 1992. Compositional control of magnetic anisotropy in the Thomson formation, east-central Minnesota. Tectonophysics, 210, 45-58. KELSO, P. R., TIKOFF, B., JACKSON, M. & SUN, W. 2002. A new method for the separation of paramagnetic and ferromagnetic susceptibility anisotropy using low field and high field methods. Geophysical Journal International, 151, 345-59. KHAN, M. A. 1962. Anisotropy of magnetic susceptibility of some igneous and metamorphic rocks. Journal of Geophysical Research, 67, 2873-2885. KING, J., BANERJEE, S. K., MARVIN, J. & OZDEMIR, 6. 1982. A comparison of different magnetic methods for determining the relative grain size of magnetite in natural materials: some results from lake sediments. Earth and Planetary Science Letters, 59, 404-419. KISCH, H. J. 2002. XRD reflection ratios in slate slabs as detectors of incipient colcite fabric in slates. Schweiz. Mineral. Petrogr. Mitt., 82, 169-185. KLETETSCHKA, G. & WASILEWSKI, P. J. 2002. Grain size limit for SD hematite. Physics of the Earth and Planetary Interiors, 129, 173-9. KODAMA, K. P. 1991. Overview of geomagnetism and paleomagnetism, 1987-1990. Reviews of Geophysics, 29, 343-348. KODAMA, K. P. 1997. A successful rock magnetic technique for correcting paleomagnetic inclination shallowing: case study of the Nacimiento Formation, New Mexico. Journal of Geophysical Research 102(B3), 5193-5205. KODAMA, K. P. & SUN, W.-W. 1992. Magnetic anisotropy as a correction for compaction-caused paleomagnetic inclination shallowing. Geophysical Journal International, 111, 465-469. LAGROIX, F. & BORRADAILE, G. 2000a. Magnetic fabric interpretation complicated by inclusions in mafic silicates. Tectonophysics, 325(3-4), 207-225. LAGROIX, F. & BORRADAILE, G. J. 2000b. Tectonics of the circum-Troodos sedimentary cover of Cyprus, from rock magnetic and structural observations. Journal of Structural Geology, 22(4), 453469. LODE, W. 1926. Versuche iiber den Einfluss der mittleren Hauptspannung auf das Fliessen des Metalle Eisen, Kupfer, und Nickel. Zieitschrift fur Physikunde, 36, 913-939.

358

G. J. BORRADAILE & M. JACKSON

LUNEBURG, C. M., LAMPERT, S. A., LEBIT, H. D., HIRT, A. M., CASEY, M. & LOWRIE, W. 1999. Magnetic anisotropy, rock fabrics and finite strain in deformed sediments of SW Sardinia (Italy). Tectonophysics, 307, 51-74. MARCH, A. 1932. Mathematische Theorie der Regelung nach der Korngestalt bei amner Deformation. Zeitschrift fur Kristallographie, 81, 285-297. MARKERT, H. & LEHMANN, A. 1996. Three-dimensional Rayleigh hysteresis of oriented core samples from the German Continental Deep Drilling Program: susceptibility tensor, Rayleigh tensor, threedimensional Rayleigh law. Geophysical Journal International, 127, 201-214. MARTIN-HERNANDEZ, F. & HIRT, A. M. 2001. Separation of ferrimagnetic and paramagnetic anisotropies using a high-field torsion magnetometer. Tectonphysics, 337, 209-221. MARTIN-HERNANDEZ, F. & HIRT, A. M. 2003. The anisotropy of magnetic susceptibility in biotite, muscovite and chlorite single crystals. Tectonophysics, 367, 13-28. McCABE, C., JACKSON, M. J. & ELLWOOD, B. B. 1985. Magnetic anisotropy in the Trenton limestone: results of a new technique, anisotropy of anhysteretic susceptibility. Geophysical Research Letters, 12, 333-336. MEANS, W. D., HOBBS, B. E., LISTER, G. S. & WILLIAMS, P. F. 1981. Vorticity and non-coaxiality in progressive deformations. Journal of Structural Geology, 2, 371-378. MIYAHARA, S. & TERANASHI, T. 1968. Magnetic properties of FeS2 and CoS2. Journal of Applied Physics, 39, 986-897. MOSKOWITZ, B. M., JACKSON, M. & KISSEL, C. 1998. Low-temperature magnetic behavior of titanomagnetites. Earth and Planetary Science Letters, 157, 141-9. MUXWORTHY, A., WILLIAMS, W. & VIRDEE, D. 2003. Effect of magnetostatic interactions on the hysteresis parameters of single-domain and pseudo-single-domain grains, art. no. 2517. Journal of Geophysical Research-Solid Earth, 108(611), 10.1029/2003JB002588. NADAI, A. 1963. Theory of Flow and Fracture of Solids. McGraw-Hill, New York. NAGATA, T. 1961. Rock Magnetism. Maruzen, Tokyo. NAKAMURA, N. & BORRADAILE, G. 200la. Do reduction spots predate finite strain? A magnetic diagnosis of Cambrian slates in North Wales. Tectonophysics, 340, 133-139. NAKAMURA, N. & BORRADAILE, G. 200 Ib. Strain, anisotropy of anhysteretic remanence, and anisotropy of magnetic susceptibility in a slaty tuff. Physics of the Earth & Planetary Interiors, 125, 85-93. NAKAMURA, N. & NAGAHAMA, H. 2001. Changes in magnetic and fractal properties of fractured granites near the Nojima Fault, Japan. The Island Arc, 10, 486-494. NICOLAS, A. 1987. Principles of Rock Deformation. Kluwer, Dordrecht. NICOLAS, A. & POIRIER, J. P. 1976. Crystalline Plasticity and Solid State Flow in Metamorphic Rocks. J. Wiley & Sons, New York.

NYE, J. F. 1957. Physical Properties of Crystals. Oxford University Press, London. OERTEL, G. 1983. The relationship of strain and preferred orientation of phyllosilicate grains in rocks: a review. Tectonophysics, 100, 413-447. O'REILLY, W. 1984. Rock and Mineral Magnetism. Blackie, Glasgow. OWENS, W. H. 1974. Mathematical model studies on factors affecting the magnetic anisotropy of deformed rocks. Tectonophysics, 24, 115-131. OWENS, W. H. 2000a. Error estimates in the measurement of anisotropic magnetic susceptibility. Geophysical Journal International, 142, 516-26. OWENS, W. H. 2000b. Statistical applications to second-rank tensors in magnetic fabric analysis. Geophysical Journal International, 142, 527-38. OWENS, W. H. & BAMFORD, D. 1976. Magnetic, seismic, and other anisotropic properties of rock fabrics. Philosophical Transactions of the Royal Society of London, Series A, 283, 55-68. OWENS, W. H. & RUTTER, E. H. 1978. The development of magnetic susceptibility anisotropy through crystallographic preferred orientation in a calcite rock. Physics of the Earth and Planetary Interiors, 16, 215-222. OZDEMIR, 6. & DUNLOP, D. J. 1996. Thermoremanence and Neel temperature of goethite. Geophysical Research Letters 23, 921-924. PARES, J. M. & VAN DER PLUIJM, B. A. 2002. Phyllosilicate fabric characterization by Low-Temperature Anisotropy of Magnetic Susceptibility (LTAMS). Geophysical Research Letters. PARES, J. M., VAN DER PLUIJM, B. & DINARES-TURELL, J. 1999. Evolution of magnetic fabrics during incipient deformation of mudrocks (Pyrenees, northern Spain). Tectonophysics, 307, 1-14. PERARNAU, A. & TARLING, D. H. 1985. Thermal enhancement of magnetic fabric in Cretaceous sandstones. Journal of the Geological Society London, 142, 1029-1034. POIRIER, J. P. 1985. Creep of Crystals. Cambridge University Press. POTTER, D. K. & STEPHENSON, A. 1988. Single-domain particles in rocks and magnetic fabric analysis. Geophysical Research Letters, 15, 1097-1100. POTTER, D. K. & STEPHENSON, A. 1990. Field-impressed magnetic anisotropy in rocks. Geophysical Research Letters, 17, 2437-2440. PRESS, W. H. 1992. Numerical recipes in C: the art of scientific computing. Cambridge University Press, Cambridge. PRICE, G. P. 1980. The analysis of quartz c-axis fabrics by the photometric method. Journal of Geology, 88, 181-195. RAMBERG, H. 1975. Particle paths, displacement and progressive strain applicable to rocks. Tectonophysics, 28, 1-37. RAMSAY, J. G. 1967. Folding and Fracturing of Rocks. McGraw-Hill, New York. RAMSAY, J. G. & HUBER, M. I. 1983. The Techniques of Modern Structural Geology. Volume I: Strain analysis. Academic Press. RANALLI, G. 1987. Rheology of the Earth. Allen & Unwin, London.

AMS-PETROFABRIC OF DEFORMED ROCKS RENNE, P. R., SCOTT, G. R., GLEN, J. M. G. & FEINBERG, J. M. 2002. Oriented inclusions of magnetite in clinopyroxene; source of stable remanent magnetization in gabbros of the Messum Complex, Namibia. Geochemistry, Geophysics, Geosystems - G3, 3, 11. RICHTER, C. & VAN DER PLUUM, B. A. 1994. Separation of paramagnetic and ferrimagnetic susceptibilities using low temperature magnetic susceptibilities and comparison with high field methods. Physics of the Earth and Planetary Interiors, 82, 111-121. ROBION, P. & BORRADAILE, G. J., 2001. Stress remagnetization in pyrrhotite-calcite synthetic aggregates. Geophysical Journal International, 144, 96-104. ROBION, P., FRIZON DE LAMOTTE D., KISSEL C. & AUBOURG C., 1995. Tectonic versus mineralogical contribution to the magnetic fabrics of epimetamorphic slaty rocks: an example from the Ardennes massif (France-Belgium). Journal of Structural Geology, 17, 1111-1124 ROBION, P., KISSEL, C., FRIZON DE LAMOTTE, D., LORAND, J.P. & GUEZOU, J. C. 1997. Magnetic mineralogy and metamorphic zonation in the Ardenne massif (France-Belgium). Tectonophysics, 271: 231-248. ROCHETTE, P. 1987a. Magnetic susceptibility of the rock matrix related to magnetic fabric studies. Journal of Structural Geology, 9, 1015-1020. ROCHETTE, P. 1987b. Metamorphic control of the mineralogy of black shales in the Swiss Alps: toward the use of 'magnetic isograds'. Earth and Planetary Science Letters, 84, 446-456. ROCHETTE, P. 1988. Inverse magnetic fabric in carbonate-bearing rocks. Earth and Planetary Science Letters, 90, 229-237. ROCHETTE, P. 1994a. Comment on 'Magnetic fabrics, crystallographic preferred orientation and strain of progressively deformed metamorphosed pelites in the Helvetic zone of the Central Alps (Quartzenschiefer Formation) by C. Richter, L. Ratschbacher and W. Frisch. Journal of Geophysical Research, 99, 21825-21827. ROCHETTE, P. 1994b. Comments on 'Anisotropic magnetic susceptibility in the continental lower crust and its implication for the shape of magnetic anomalies' by G. Florio et al Geophysical Research Letters, 21, 2773-2774. ROCHETTE, P. & PILLION, G. 1988. Identification of multicomponent anisotropies in rocks using various field and temperature values in a cryogenic magnetometer. Physics of the Earth and Planetary Interiors, 51, 379-386. ROCHETTE, P. & VIALON, P. 1984. Development of planar and linear fabrics in Dauphinois shales and slates (French Alps) studied by magnetic anisotropy and its mineralogical control. Journal of Structural Geology, 6, 33-38. ROCHETTE, P., PILLION, G., MOLLARD, P. & VERGNE, R. 1983. Utilisation d'un magnetometre a effet Josephson pour 1'analyse de 1'anisotropie magnetique des roches. Comptes rendus de I'Academie des Sciences (Paris), Serie II, 296, 557-559. ROCHETTE, P., PILLION, G., MATTEI, J.-L. & DEKKERS, M. J. 1990. Magnetic transition at 30-34 Kelvin

359

in pyrrhotite: insight into a widespread occurrence of this mineral in rocks. Earth and Planetary Science Letters, 98, 319-328. ROCHETTE, P., JACKSON, M. J. & AUBOURG, C. 1992. Rock magnetism and the interpretation of anisotropy of magnetic susceptibility. Reviews of Geophysics, 30, 209-226. ROCHETTE, P., SCAILLET, B., GUILLOT, S., LE FORT, P. & PECKER, A. 1994. Magnetic properties of the High Himalayan leucogranites: Structural implications. Earth and Planetary Science Letters, 126, 217-234. ROCHETTE, P., AUBOURG, C. & PERRIN, M. 1999. Is this magnetic fabric normal? A review and case study in volcanic formations. Tectonophysics, 307, 219-234. SAGNOTTI, L. & SPERANZA, S. 1993. Magnetic fabrics analysis of the Plio-Pleistocene clayey units of the Sant' Arcangelo basin, southern Italy. Physics of the Earth and Planetary Interiors, 77, 165-176. SAGNOTTI, L., FACCENNA, C. & FUNICIELLO, R. 1994. Magnetic fabric and structural setting of PlioPleistocene clayey units in an extensional regime; the Tyrrhenian margin of central Italy. Journal of Structural Geology, 16, 1243-1257. SAGNOTTI, L., ROCHETTE, P., JACKSON, M., VADEBOIN, P., DlNARES-TURELL, J. & WlNKLER, A. 2003.

Inter-laboratory calibration of low-field magnetic and anhysteretic susceptibility measurements. Physics of the Earth & Planetary Interiors, 138, 25-38. SAINT BEZAR, B., HEBERT, R., AUBOURG, C., ROBION, P., SWENNEN, R. & FRIZON DE LAMOTTE, D., 2002. Magnetic fabric and petrographic investigation of hematite-bearing sandstones through ramp-related folds : examples from the South High Atlas front (Morocco). Journal of Structural Geology, 24, 1507-1520. SANDER, B. 1930. Gefugekunde der Gesteine. SpringerVerlag, Vienna. SCHILL, E., APPEL, E., GODIN, L., CROUZET, C., GAUTAM, P. & REGMI, K. R. 2003. Record of deformation by secondary magnetic remanences and magnetic anisotropy in the Nar/Phu Valley (central Himalaya). Tectonophysics, 377, 197-209. SCHMIDT, V. A., ELLWOOD, B. B., NAGATA, T., NOLTIMIER, H. C., HROUDA, F. & WAGNER, J.-J. 1988. The measurement of anisotropy of magnetic susceptibility using a cryogenic (SQUID) magnetometer and a comparison with results obtained from a torsion-fiber magnetometer. Physics of the Earth and Planetary Interiors, 51, 365-378. SIDDANS, A. W. B., HENRY, B., KLIGFIELD, R., LOWRIE, W., HIRT, A. & PERCEVAULT, M. N. 1984. Finite strain patterns and their significance in Permian rocks of the Alpes Maritmes (France). Journal of Structural Geology, 6, 339-368. SNOWBALL, I. F. 1997a. The detection of single-domain greigite (Fe3S4) using rotational remanent magnetization (RRM) and the effective gyro field (Bg): mineral magnetic and palaeomagnetic applications. Geophysical Journal International, 130, 704-16. SNOWBALL, I. F. 1997b. Gyroremanent magnetization and the magnetic properties of greigite-bearing

360

G. J. BORRADAILE & M. JACKSON

clays in southern Sweden. Geophysical Journal International, 129, 624-636. SOUQUE, C., ROBION, P. & FRIZON DE LAMOTTE, D., 2002. Cryptic magnetic fabric of tectonic origin revealed by heating of sedimentary samples from the Corbieres (France). Physics and Chemistry of the Earth, 27, 135-144. STAGEY, F. D. & BANERJEE, S. K. 1974. The Physical Principles of Rock Magnetism. Elsevier, Amsterdam. STEPHENSON, A. 198la. Gyromagnetic remanence and anisotropy in single-domain particles, rocks, and magnetic recording tape. Philosophical Magazine B, 44, 635-664. STEPHENSON, A. 1981b. Gyroremanent magnetization in a weakly anisotropic rock sample. Physics of the Earth and Planetary Interiors, 25, 163-166. STEPHENSON, A. 1994. Distribution anisotropy: two simple models for magnetic lineation and foliation. Physics of the Earth and Planetary Interiors, 82, 49-53. STEPHENSON, A. & POTTER, D. K. 1987. Gyroremanent magnetizations in dilute anisotropic dispersions of gamma ferric oxide particles from magnetic recording tape. IEEE Transactions on Magnetics MAG-23, 3820-3830. STEPHENSON, A. & POTTER, D. K. 1996. Towards a quantitative description of field-impressed anisotropy of susceptibility. Geophysical Journal International, 126, 505-512. STEPHENSON, A. & SNOWBALL, I. F. 2001. A large gyromagnetic effect in greigite. Geophysical Journal International, 145, 570-575. STEPHENSON, A., SADIKUN, S. & POTTER, D. K. 1986. A theoretical and experimental comparison of the anisotropies of magnetic susceptibility and remanence in rocks and minerals. Geophysical Journal of the Royal Astronomical Society, 84, 185-200. STEPHENSON, R. W. 1976. A study of rotational remanent magnetization. Geophysical Journal of the Royal Astronomical Society, 47, 363-373. STONER, E. C. 1945. The demagnetizing factors for ellipsoids. Philosophical Magazine, 36, 803-821. SUGIURA, N. 1979. ARM, TRM, and magnetic interactions: concentration dependence. Earth and Planetary Science Letters, 42, 451-455. SYONO, Y. 1960. Magnetic susceptibility of some rock forming silicate minerals such as amphiboles, biotites, cordierites and garnets. Journal of Geomagnetism and Geoelectricity, 11, 85-93. TAN, X., KODAMA, K. P. & FANG, D. 2002. Laboratory depositional and compaction-caused inclination errors carried by haematite and their implications in identifying inclination error of natural remanence in red beds. Geophysical Journal International, 151, 475-86. TARLING, D. H. & HROUDA, F. 1993. The Magnetic Anisotropy of Rocks. Chapman & Hall, London. TAUXE, L. 1998. Paleomagnetic Principles and Practice. Kluwer, Dordrecht.

TEBBLE, R. S. & CRAIK, D. J. 1969. Magnetic Materials. Wiley-Interscience, London. THOMPSON, R. & OLDFIELD, F. 1986. Environmental Magnetism. Allen and Unwin, London. TRINDADE, R. I. F., BOUCHEZ, J. L., BOLLE, O., NEDELEC, A., PESCHLER, A. & POITRASSON, F. 2001. Secondary fabrics revealed by remanence anisotropy: methodological study and examples from plutonic rocks. Geophysical Journal International, 147(2), 310-318. TULLIS, T. E. 1976. Experiments on the origin of slaty cleavage and schistosity. Geological Society of America Bulletin, 87, 745-753. TURNER, F. J. & WEISS, L. E. 1963. Structural Analysis of Metamorphic Tectonites. McGraw-Hill, New York. UYEDA, S., FULLER, M. D., BELSHE, J. C. & GIRDLER, R. W. 1963. Anisotropy of magnetic susceptibility of rocks and minerals. Journal of Geophysical Research, 68, 279-291. VINCENZ, S. A. 1965. Frequency dependence of magnetic susceptibility of rocks in weak alternating fields. Journal of Geophysical Research, 70, 1371— 1377. VOIGHT, W. & KINOSHITA, S. 1907. Bestimmung absoluter Werte von Magnetiserungszahlen, inbesondere fur Kristalle. Annals of Physics, 24, 492-514. WERNER, T. 1997. Experimental designs for determination of the anisotropy of remanence-test of the efficiency of least-squares and bootstrap methods applied to metamorphic rocks from southern Poland. Physics and Chemistry of the Earth, 22, 131-136. WERNER, T. & BORRADAILE, G. J. 1996. Paleoremanence dispersal across a transpressed Archean terrain; deflection of anisotropy or by late compression? Journal of Geophysical Research, B, Solid Earth and Planets, 101, 5531-5545. WETTSTEIN, A. 1886. Uber die Fishfauna des Tertiaren Glarner Schiefers. Schweiz. Paldont. Ges. Abhandl, 13, 1-101. WOOD, D. S., OERTEL, G., SINGH, J. & BENNETT, M. F. 1976. Strain and anisotropy in rocks. Philosophical Transactions of the Royal Society of London, Series A, 283, 27-42. WOODCOCK, N. H. 1977. Specification of fabric shapes using an Eigenvalue method. Geological Society of America Bulletin, 88, 1231-1236. WORM, H.-U., CLARK, D. & DEKKERS, M. J. 1993. Magnetic susceptibility of pyrrhotite: grain size, field and frequency dependence. Geophysical Journal International, 114, 127-137. YAMAZAKI, T. & IOKA, N. 1997. Cautionary note on magnetic grain-size estimation using the ratio of ARM to magnetic susceptibility. Geophysical Research Letters, 24, 751-754. ZAPLETAL, K. 1990. Low-field susceptibility anisotropy of some biotite crystals. Physics of the Earth and Planetary Interiors, 63, 85-97.

Phyllosilicate preferred orientation as a control of magnetic fabric: evidence from neutron texture goniometry and low and high-field magnetic anisotropy (SE Rhenohercynian Zone of Bohemian Massif) MARTIN CHADIMA,1 2 ANKE HANSEN,3 ANN M. HIRT,4 FRANTISEK HROUDA5 6 & HEINRICH SIEMES7 1

Palaeomagnetic Laboratory, Institute of Geology, Academy of Sciences of the Czech Republic, Rozvojovd 135, CZ-16502 Praha 6, Czech Republic (e-mail: [email protected]) 2

Institute of Geological Sciences, Masarykova Univerzita, Kotldrskd 2, CZ-61137 Brno, Czech Republic

3

GKSS Forschungszentrum Geesthacht GmbH, Max-Planck-Strafe, Geb. 03, D-21502 Geesthacht, Germany 4

InstitutfurGeophysik, ETH Zurich, CH-8093 Zurich, Switzerland 5

AGICO Inc., Jecnd 29a, CZ-62100 Brno, Czech Republic

^Institute of Petrology and Structural Geology, Charles University, 1

Albertov 6, Prague, Czech Republic Institut fur Mineralogie und Lagerstattenlehre, RWTH Aachen, Wullnerstr. 2, D-52056 Aachen, Germany Abstract: The low- and high-field magnetic anisotropy (AMS, HFA) of the Rhenohercynian mudstones and greywackes is compared to the theoretical anisotropy calculated from neutron texture goniometry measurements. The magnetic anisotropy is predominantly carried by the paramagnetic phyllosilicates in the form of chlorite/mica stacks and the ferromagnetic contribution is insignificant. The respective principal directions of the theoretical anisotropy and the AMS and HFA are sub-parallel; magnetic foliation reflects the orientation of the maximal concentration of phyllosilicate basal planes, magnetic lineation is subparallel to the intersection axis of those planes. For the purpose of quantitative comparison, the infrequently used standard deviatoric susceptibility as a measure of the HFA degree is employed. A very good linear correlation of the degree of theoretical anisotropy and the measured AMS and HFA is found. The prolate and oblate shapes of the respective fabric ellipsoids are reasonably well correlated. Neutron texture goniometry justifies the use of the conventional magnetic anisotropy technique for the assessment of the mineral fabric of studied rocks. When compared with other works relating the magnetic anisotropy to the mineral preferred orientation (examined by e.g. U-stage or X-ray texture goniometry) neutron texture goniometry seems to be a preferable and very precise method fabric analysis.

Anisotropy of magnetic susceptibility (AMS) measured at low field has been widely used as a powerful and fast indicator of mineral fabric for the various rock types (for reviews see Hrouda 1982; Tarling & Hrouda 1993; Borradaile & Henry 1997). Whereas the qualitative relationship between the AMS and rock fabric is largely understood, studies that investigate the quantitative relationship are rather scarce (e.g. Wood el al. 1976; Owens & Rutter 1978; Rathore 1979; Hrouda el al 1985; Hrouda & Schulmann 1990; Richter et al. 1993; Siegesmund et al. 1995; Lampert 1996; Hrouda et al. 1997b; Liineburg et al. 1999; Siemes el al. 2000; de Wall et al. 2000; Hrouda & Ullemeyer 2001;

Martin-Hernandez 2002). A better understanding of what is contributing to the magnetic fabric is necessary in order to use the AMS as a reliable proxy for the strain estimate. In weakly magnetic rocks, i.e. conventionally those with the bulk susceptibility less than 5 x 10~4 SI, the AMS is predominantly controlled by the preferred orientation of paramagnetic minerals, most frequently phyllosilicates (Borradaile el al 1986; Rochette 1987; Hrouda &Jelinek 1990). Phyllosilicates are common constituents of many sedimentary and low-grade metamorphic rocks, e.g. shales or slates. Using AMS, one may assess the strain in these rocks even if the conventionally used strain markers

From: MARTIN-HERNANDEZ, F., LUNEBURG, C. M., AUBOURG, C. & JACKSON, M. (eds) 2004. Magnetic Fabric: Methods and Applications. Geological Society, London, Special Publications, 238, 361-380. 0305-8719/04/ $15.00 © The Geological Society of London 2004.

362

M. CHADIMA ET AL.

are scarce or absent. In order to establish the relationship between the phyllosilicate preferred orientation and the AMS, an independent, nonmagnetic method of mineral fabric analysis is necessary. Different non-magnetic methods of fabric analysis have been used to study their quantitative relationship to the AMS for various phyllosilicates in various rock types, e.g. Universal stage (U-stage) for biotite in gneiss (Hrouda & Schulmann 1990), X-ray texture goniometry for phyllosilicates in slates (Rathore 1979; Liineburg et al. 1999), and biotite in granulite (Hrouda & Ullemeyer 2001). Only recently, the neutron texture goniometry has been combined with the AMS, e.g. on biotite in granulite (Ullemeyer & Weber 1999), hematite in banded iron ore (Siemes et al. 2000), and calcite in marble (de Wall et al. 2000). Although the U-stage has a long history, it remains a rather inexpensive and widely accessible method of fabric analysis. Disadvantages are the laborious and time-consuming nature of the work and the lower limit of measured grains being about 20 urn (Passchier & Trouw 1998). The volume of minerals with particular orientation is difficult to determine since normally one measurement per grain is taken. This fact may induce inaccuracies when the U-stage measurements are compared to the AMS as each mineral grain contributes to the bulk AMS according to its volume. The X-ray texture goniometry is most easily operated on monomineralic rocks, although bimineralic rocks can be studied in some cases (Braun 1994). Samples should be relatively finegrained (<200um). For the complete pole figure, the transmission and reflection measurement must be combined after appropriate intensity corrections (Casey 1981). For the measurements at very small diffraction angles, the additional defocusing correction should be applied (Ullemeyer & Weber 1994). Both the U-stage and the X-ray goniometer analyses are performed on thin or polished sections while the AMS is measured on the 'standard' palaeomagnetic cylinder or cubic samples. The neutron diffraction is very suitable for most geological materials (Brokmeier 1994, 1997; Schafer 2002). The penetration depth of neutrons is 100 to 10000 times greater than that of the conventionally used X-rays. Although the technique involves long measurement times and instruments are not widely available, the entire sample volume is measured. Since measurements can be done on exactly the same cylindrical (or cubic) sample, the texture can be directly compared to the AMS fabric. Due to the high beam cross section, the fabric of

coarse-grained rocks, i.e. grains larger than 1 mm, can be measured with reasonable accuracy. In this study the neutron texture goniometry is used to analyse the phyllosilicate preferred orientation of the selected mudstone and greywacke samples from the SE Rhenohercynian Zone of the Bohemian Massif. The theoretical anisotropy calculated from the neutron pole figures is compared to the AMS and high-field anisotropy (HFA). The AMS is a combination of the paramagnetic and ferromagnetic contributions to the anisotropy. As only the paramagnetic anisotropy should be correlated with the preferred orientation of phyllosilicate phases, the HFA is used for the separation of the paramagnetic and ferromagnetic contribution to the magnetic anisotropy (Hrouda & Jelinek 1990; MartinHernandez & Hirt 2001). In addition, the contribution of ferromagnetic minerals to the bulk susceptibility is investigated using measurements of the temperature variations of magnetic susceptibility and the microfabrics are observed using scanning electron microscopy. The combination of these different approaches enables the establishment of a more accurate qualitative and quantitative correlation between the phyllosilicate fabric and magnetic anisotropy and yields valuable information about the meaning of the magnetic fabric. Geological setting and sample description The study area is located in the SE Rhenohercynian Zone of the Bohemian Massif in the Czech Republic (Fig. 1). The Czech section of the Rhenohercynian Zone trends NNE-SSW, and consists mainly of the Lower Carboniferous flysch sediments. To a lesser extent, the Upper Devonian shallow water limestones, deep-water mudstones and basalts are incorporated into the thrust and fold structure (for detailed stratigraphy see Dvorak 1973; Hartley & Otava 2001). The Rhenohercynian sedimentary sequences were deformed during the Variscan orogeny. Ductile deformation is responsible for the evolution of the NNE-SSW trending folds with axial cleavage. The cleavage evolved preferentially in the incompetent mudstones and is less evident in the competent greywackes. According to the intensity of cleavage development, ductile deformation gradually increases from the SE to the NW (Dvorak 1973). The rocks were metamorphosed under the late diagenetic conditions (Francu et al. 1999). According to the illite crystallinity and vitrinite reflectance data, the

PHYLLOSILICATE CONTROL OF MAGNETIC FABRIC

363

Fig. 1. Simplified geological map of the SE part of the Rhenohercynian Zone in the Czech Republic. The position of the studied area within the Rhenohercynian Zone of the Bohemian Massif is marked in upper right. Studied sites are marked with open or closed circles representing greywacke or mudstone samples, respectively.

degree of metamorphism increases slightly from the SE to the NW (Dvorak & Wolf 1979; Francu et al 1999; Hrouda & Jezek 1999). The field area had been extensively sampled for the purpose of the AMS measurement. Samples were collected using a gasoline-powered portable drill. The samples were cylindrical in shape, 2.5cm in diameter and 2.2cm in height. Eighteen samples from thirteen different sites were selected for this comparative study (Fig. 1, Table 1). Samples were chosen to represent the lithological variety, i.e. greywackes and mudstone, and the different degrees of deformation and metamorphism in the study area.

Methods and data presentation

Temperature dependence of magnetic susceptibility The temperature variations of bulk magnetic susceptibility were determined on powder samples using a CS-3 Furnace Apparatus coupled with the KLY-3S Kappabridge (Parma & Zapletal 1991; Parma et al. 1993) at Agico, Inc., Brno, Czech Republic. All the measurements were performed in the air atmosphere. Temperature dependence of magnetic susceptibility allows the resolution of the room temperature susceptibility

Table 1. Sample names, lithology, measured mineral, AMS and HFP parameters, neutron diffraction parameters AMS

Lithology

Mineral

k

k2 k\ 4 *3 (1(T4) (io- ) (IO- 4 )

DV21-1-1 DV24-7-1 DV26-5-1 DV28-1-2 DV28-6-2 DV28-8-2

Gw

chl

Ms

chl

Gw Gw

chl chl

Gw

chl

2.26 3.75 2.10 2.17 2.74

Ms

chl chl

2.18 3.60 2.06 2.13 2.71 3.00 0.83 2.23 2.41 1.88

DV29-1-1 DV45-3-2 DV51-1-2 DV53-1-1 DV54-1-1 DV57-2-2 DV62-2-1 DV63-2-2 DV63-2-2 DV63-7-1 DV65-3 DV65-4 DV65-5

Gw Gw Gw

chl chl

Gw

chl

Gw Ms Gw

chl

Ms

chl

Ms Gw

3.06 0.85 2.28 2.44 1.89 2.04 3.60 2.24

Ms

bio chl chl

1.98 3.47 2.18 4.29 4.29 3.00 4.17

4.58 4.58 3.12 4.41

Ms Ms

chl chl

3.99 4.07

4.23 4.31

chl chl

M.r.d.

HFP

Specimen

(i
intensities, orientation tensor principal axes and theoretical anisotropy

^AMS

^AMS

(io- 5 )

2.20 3.58 2.06 2.13

2.09 3.47 2.03 2.10

0.69 1.12 0.27 0.27

1.079 1.079 1.032 1.031

2.70 2.99 0.83 2.21 2.43 1.88

2.69 2.94 0.80 2.19 2.37 1.86 1.92 3.37 2.13 3.92 3.92 2.89 3.91

0.22 0.50 0.21 0.38 0.31 0.14 0.50 0.94 0.46 2.73 2.73 0.93 2.04

1.018 1.041 1.064 1.041 1.030 1.018 1.062 1.068 1.053 1.167 1.167 1.078 1.127

3.72 3.79

2.07 2.13

1.136 1.137

1.96 3.46 2.18 4.36 4.36 2.98 4.18 4.02 4.11

Gw, greywacke samples; Ms, mudstone samples; chl, chlorite; bio, biotite.

U

k}-k

0.325 0.78 -0.175 1.63 -0.204 0.37 -0.180 -0.697 -0.231 -0.122

0.37 0.71 0.10

-0.463 0.41 0.645 0.044 0.15 -0.378 -0.245 -0.126 0.342 0.342 -0.215 0.107 0.158 0.213

k2-k

ks-k (10~ 5 )

1.87 -2.26 -0.29 -0.29 -0.77 -0.19 -0.58

6 (io~ 5 ) (io- )

Min

e\

^2


*T1

kJ2

k-ri

PT k'-r (IO- 5 )

-0.225 -0.121

2.55 1.46 1.55

0.16 0.08 0.37

0.466 0.374 0.382

0.346 0.348 0.350

0.188 0.277

3.75 2.09 2.17

3.59 2.05 2.12

3.46 2.04

1.16 0.24 0.29

-0.121 -0.175 -0.311 -0.226

1.36 2.12 1.71 1.67 1.61

0.49 0.37 0.21

0.372 0.414

0.280 0.240 0.257

0.31 0.56

0.399 0.395 0.382

0.348 0.346 0.344 0.348 0.333

0.258 0.285

2.18 2.39 1.97 5.36 2.70 2.71 3.81 3.97

0.12 0.24 0.27 0.00 0.05 0.09 0.00 0.00

0.449 0.461 0.419 0.614 0.443 0.459 0.531 0.556

0.340 0.345 0.345 0.256 0.340 0.357 0.335 0.327

0.212 0.195 0.236 0.130 0.217 0.184 0.134

^HFP

-0.97 -1.40 -0.34

0.73 1.25 0.29

-0.34

0.29 0.55 0.07 0.32

-0.63 -0.08 -0.35 -0.19

0.14

0.364

1.30

-2.46

-1.06

0.98

-0.311

3.14 3.14

6.22 6.22

-3.76 -3.76

2.85 2.85

0.270 0.270

2.10

-2.87

2.26

(io- 4 ) (io-4) (io-4)

Ur

0.321

0.41

2.66

Theoretical anisotropy

Max

^HFP

(io- 5 )

Orient ation tensor

0.114

0.268

0.118

2.75 3.08 0.85 2.27 2.45

2.70 2.99 0.83 2.22

2.05 3.61 2.24 4.54 4.43 3.12 4.40 4.23

2.10

0.30 0.61 0.14

2.41

2.68 2.93 0.81 2.19 2.38

1.97 3.46 2.17 4.38 4.28 2.98 4.16 4.00

1.91 3.35 2.13 3.95 4.16 2.89 3.93 3.74

0.54 1.07 0.47 2.50

0.36 0.27

1.12 0.97 1.92 2.02

1.082 1.028 1.033 1.026 1.051 1.041 1.040 1.028 1.070 1.078 1.053 1.149 1.066 1.081 1.119 1.133

-0.137 -0.464 -0.439 -0.478 -0.218 -0.225 -0.314 0.010 -0.080 -0.128 -0.191 0.479 -0.088 -0.258 -0.013 0.046

PHYLLOSILICATE CONTROL OF MAGNETIC FABRIC

365

into its paramagnetic and ferromagnetic components (Hrouda 1994; Hrouda et al. 1997a).

1; negative or positive values indicate prolate or oblate fabrics, respectively.

Scanning electron microscopy

High-field magnetic anisotropy

The microfabric observations were done on the polished sections using a Scanning Electron Microscope (SEM) CamScan at the Institute of Geological Sciences, Masaryk University, Brno, Czech Republic. All the observations were made in the backscattered electron image mode where the differences in the backscattered electron intensity reflect differences in the atomic number.

The high-field magnetic anisotropy (HFA) was measured using a high-field torque magnetometer in the Palaeomagnetic laboratory, Institute of Geophysics, ETH, Zurich, Switzerland (Bergmuller et al. 1994). Each sample was measured during a full rotation about each of three mutually orthogonal axes at angular increments of 20°. Four measurement fields, 1200, 1400, 1600 and 1800 mT were applied. These fields are strong enough to saturate most ferrimagnetic minerals and allow the separation of the paramagnetic and ferrimagnetic contributions to the magnetic anisotropy (Owens & Bamford 1976; Hrouda & Jelinek 1990; MartinHernandez & Hirt 2001); the torque increases linearly with the square of the applied field for paramagnetic minerals and is independent of the applied field above the saturation field of the ferrimagnetic minerals (Banerjee & Stacey 1967). The torque signal was analysed using the Fast Fourier transformation. The 2B term, which dominates the torque signal, was used to construct the paramagnetic and ferromagnetic HFA tensors.

Low-field magnetic anisotropy The low-field anisotropy of magnetic susceptibility (AMS) was measured with a KLY-3S Spinner Kappabridge (Jelinek & Pokorny 1997) at Agico, Inc., Brno, Czech Republic. The data processing was carried out by means of the SUSAR program, which calculates the eigenvectors (Ki,K2,K3) and eigenvalues (k\ >k2>k3) of the anisotropy tensors. Results are presented in terms of the mean susceptibility k, the degree of anisotropy P (Nagata 1961), the standard deviatoric susceptibility k' (Jelinek 1984) and the difference shape factor U (Jelinek 1981) defined as:

Neutron texture goniometry The texture goniometry measurements were carried out using a neutron texture diffractometer TEX-2 at the GKSS Research Centre in Geesthacht, Germany. TEX-2 is a conventional four-circle diffractometer optimized for the texture analysis of different types of materials (Brokmeier 1997). The wavelength of neutron radiation varies slightly with time and during the where k\ > k2 > k3 are the principal susceptibil- experiment the wavelength was 1.356 A. Slits of ities and (hi -k)> (k2 - k) > (k3 - k) are the width 22 mm were used to adjust the cross section deviatoric principal susceptibilities. The employ- of the neutron beam. Prior to texture analysis, two ment of infrequently used standard deviatoric sus- diffraction scans were run in two perpendicular ceptibility k' and shape difference factor U arises planes for each sample. The diffractographs from the intended comparison of the low-field obtained were used to search for the precise 20 AMS with the high-field anisotropy and with the- angle corresponding to the most pronounced oretical anisotropy. As the torque magnetometer diffraction intensities of the minerals studied. An used for the high-field anisotropy measures only appropriate 20 angle was then set as the diffracthe deviatoric component of anisotropy, only tion angle for the texture goniometer. The diffracthe k' and U parameters based on the susceptibil- tion angle 20 = 8.2° was set for the measurements ity differences can be calculated. Standard devia- of the background signal. Using an Eulerian toric susceptibility quantifies the intensity of cradle, the intensity of diffraction was measured anisotropy. The shape difference factor U is ana- in 679 distinct directions of the half-space. The logous to, even though not identical with, the T intensity of the background signal was measured shape factor (Jelinek 1981), ranging from -1 to in 16 directions.

366

M. CHADIMA ET AL.

The diffraction data were processed using the STARPLOT program (Siemes unpublished) modified for neutron texture goniometry. After subtraction of the background signal, the intensities of the diffracted signal were normalized to express densities as multiples of random distribution (m.r.d.) and plotted as pole figures in the lower hemisphere, equal area projection. The plane of projection is the bedding plane, except for sample DV65 where no bedding was observed and data were projected to the cleavage plane. (Note: The lower case, 29, and the upper case, 29, are used to denote the neutron diffraction angle and the 29 term of the torque signal, respectively.)

Calculation of theoretical anisotropy In order to compare the preferred orientation of phyllosilicates to the measured AMS, the orientation tensor of phyllosilicate oaxes was calculated using the AMSNEW program (Siemes unpublished). The orientation tensor, being defined as a 3 x 3 matrix of the normalized sums of cross-products of the directional cosines of measured linear element (Scheidegger 1965; Woodcock 1977), is a symmetric second-rank tensor whose eigenvectors (E{ ,E2,E3) and eigenvalues (e\ > €2 > £3) correspond to the axes of an ellipsoid. The AMS is also represented by a symmetric second-rank tensor, the susceptibility tensor. If the AMS is carried by a mineral with uniaxial anisotropy, the principal directions of the orientation tensor calculated from the distribution of its revolution axis are parallel to the AMS principal directions (Hrouda & Schulmann 1990). This is true for phyllosilicates possessing the oblate uniaxial anisotropy with its minimum susceptibility parallel to the oaxis (Zapletal 1990; Martin-Hernandez & Hirt 2003). In the case of the planar fabric element, the principal direction E\ of the orientation tensor corresponds to the minimum axis (K3) of the AMS. Similarly the principal direction £"3, representing the pole to the best-fit girdle to the oaxis distribution, corresponds to the maximum axis (Ki) of the AMS. The principal values of the theoretical AMS (£T1, &T2, &TS) are related to the principal values of the orientation tensor, as follows (Hrouda & Schulmann 1990):

where fcT1 > A;T2 > /cT3 are the theoretical principal susceptibilities, e\ > e2 > e3 are the principal values of the orientation tensor calculated from the phyllosilicate oaxis distribution, and PG = ki/k$ represents the degree of grain AMS. Results As already mentioned, the data for this comparative study were acquired in four different laboratories. Due to the relative unavailability of the instruments (especially the neutron texture goniometer and torque magnetometer) and a limited time for measurements, not all the outlined analyses could be performed for each sample. Consequently, 18 samples were measured for the AMS, 11 samples were measured for the HFA and 15 samples were measured for neutron diffraction (Table 1).

Temperature variations of magnetic susceptibility During laboratory heating the bulk susceptibility decreases with increasing temperature following a hyperbolic course (Fig. 2a). For pure paramagnetic minerals, the magnetic susceptibility is inversely proportional to temperature according to the Curie-Weiss law. The hyperbolic shape of the curve suggests that the dominant carriers of magnetic susceptibility are paramagnetic minerals. For some samples, however, a significant susceptibility increase during heating can be observed above 400 °C (Fig. 2b). This increase can be attributed to the artificial oxidation of original paramagnetic minerals induced by laboratory heating. Although susceptibility changes are induced in higher temperatures, the initial part of the thermomagnetic curves can by fitted with a paramagnetic hyperbola for all the studied samples. The mathematical resolution of the thermomagnetic curves based on hyperbola fitting (Hrouda 1994; Hrouda et al 1997a) shows that the ferromagnetic contribution to wholerock susceptibility is insignificant.

SEM observations For the purpose of electron microscopy, thin sections of mudstones from the sites with observed pencil structure (DV45 and DV57) were chosen. SEM images showed that platy minerals, i.e. chlorite and micas, dominated in the rock composition (Fig. 3). Other minerals identified were quartz and plagioclase. The size of the chlorite

PHYLLOSILICATE CONTROL OF MAGNETIC FABRIC

367

Fig. 2. Examples of the thermomagnetic heating curves of powder samples. The heating was rate 10°C/min, measurements were performed in the air atmosphere, (a) Mudstone sample with a paramagnetic hyperbola fitted to the thermomagnetic curve, (b) Greywacke sample.

and mica grains was of the order of tenths of a micrometre. In a number of cases the chlorite and mica crystals were interlayered and formed chlorite/mica stacks (Milodowski & Zalasiewicz 1991; Li et al. 1994). In the chlorite/mica stacks both minerals were deformed together. Frequent folding and kinking of the stack was observed. The orientation of the phyllosilicates ranged from bedding-parallel to cleavage-parallel. The new growth of phyllosilicates caused by partial dissolution and recrystallization was not observed at this scale of magnification.

Low-field anisotropy of magnetic susceptibility The mean magnetic susceptibility ranges from 8 0 x l ( T 6 S I to 430xl(T 6 SI (Table 1), with mudstones having higher values than greywackes, usually >300 x 10~6 SI. A significant correlation between bulk susceptibility and degree of anisotropy can be observed for mudstones samples whereas the degree of anisotropy of greywackes seems to be independent of bulk susceptibility (Fig. 4a).

368

M. CHADIMA ET AL.

Fig. 3. Backscattered images of the selected mudstone samples, (a) DV45, (b) DV57. The thin sections are cut approximately perpendicular to both bedding and cleavage planes. Micas and chlorite (Chi) are labelled. Orientation of bedding (B), and cleavage (C) is in lower left, scale bar are below each image. The chlorite and mica crystals are interlayered and form the chlorite/mica stacks. Frequent folding and kinking of the stack is observed. The orientation of the phyllosilicates ranges from bedding-parallel to cleavage-parallel.

The AMS degree ranges from P= 1.02 to P= 1.08 for the greywackes and P= 1.04 to P = 1.17 for the mudstones (Fig. 4b). The anisotropy ellipsoids are both oblate and prolate in shape. Except for the greywacke samples DV21-1-1, DV51-1-2 and DV53-1-1, the less anisotropic samples possess the prolate anisotropy shape whereas the more anisotropic samples have oblate magnetic fabric. The shapes of the anisotropy ellipsoid depend on the lithology. The greywackes have not undergone as much vertical compaction as mudstones and can respond to the tectonic shortening faster. The observed prolate fabric results from the superposition of the sedimentary and tectonic fabric. For more compacted mudstones the tectonic shortening was not sufficient to overprint the sedimentary fabric so the oblate bedding-parallel fabric can be still observed in some cases (e.g. DV63-2-2).

High-field anisotropy The HFA was measured on 11 samples (Table 1). All the samples show a very similar linear dependence of the torque on the square of the applied field (Fig. 5). The 29 term of all three perpendicular planes intersects at zero implying that the paramagnetic mineral is the exclusive carrier of magnetic anisotropy. The resolution of the HFA into paramagnetic (henceforth denoted as HFP) and ferromagnetic components using the technique of Martin-Hernandez & Hirt (2001) shows that the ferromagnetic fraction, if present, appears as insignificant. HFA tensors thus predominantly represent the anisotropy carried by paramagnetic minerals.

The standard deviatoric susceptibility ranges from kf = 7.41 x 10~7 to kf = 2.85 x 10"5. The anisotropy ellipsoids are both prolate and oblate in shape. Except for the greywacke samples DV21-1-1, and DV53-1-1, the less anisotropic samples possess the prolate anisotropy shape whereas the more anisotropic mudstone samples (DV63-2-2, DV65-5) have oblate magnetic fabric.

Neutron texture goniometry Neutron texture goniometry analysis was done after measurement of the AMS and HFA on 15 samples. It is evident from the neutron diffractograph (Fig. 6) that the maximum diffraction intensity of the supposed phyllosilicates (i.e. chlorite or micas) occurs at the diffraction angle 20 = 5.2°. This diffraction angle corresponds to the chlorite (001) basal plane, implying chlorite as the dominant phyllosilicate mineral. Thus the diffraction angle for the texture goniometry measurements was set as 20 = 5.2° in order to investigate chlorite basal plane preferred orientation. Nonetheless, as evident from SEM microscopy, micas contribute significantly to the composition of the rock samples studied. In contrast to the chlorite (001) diffraction peak, the peaks corresponding to biotite or muscovite are less pronounced. In order to compare both chlorite and mica textures the pilot sample DV63-2-2 with significant 29 = 7.6° diffraction corresponding to biotite (001) was measured. The observed phyllosilicate fabric, evaluated by the means of the multiple of random distribution (m.r.d.) values, is weak. The phyllosilicate

PHYLLOSILICATE CONTROL OF MAGNETIC FABRIC

369

Fig. 4. (a) The anisotropy degree (P) vs. the mean susceptibility (k) plot for selected samples. Open symbols represent greywackes, close symbols mudstones. (b) The shape parameter (U) vs. anisotropy degree (P) plot. Open symbols represent greywackes, close symbols mudstones.

fabric is weakly defined for most of the greywacke samples where the m.r.d. is less than two. Mudstones often possess higher m.r.d. (max = 5.36), reflecting the tighter grouping of c-axes. On the pilot sample (DV63-2-2) where both chlorite and biotite fabrics were measured, the chlorite fabric is noticeably stronger than the biotite fabric. Biotite oaxis distribution is more girdle-like implying rotation of biotite

grains between bedding and cleavage planes whereas chlorite basal planes are more confined to the bedding plane. In general, the c-axes of the phyllosilicate grains tend to show a girdle distribution between the principal structural foliation (i.e. bedding or cleavage, Fig. 7) sometimes with two indistinct maxima implying kinking of the phyllosilicate basal plane. The girdles are perpendicular to

370

M. CHADIMA ET AL. the structural lineation (i.e. strike of bedding, bedding-cleavage intersection, pencil structure or fold axis). The pole figures become more cluster-like with increasing value of the m.r.d. (e.g. DV65-3, DV63-2-2chlorite). The clusters are elliptical in shape with shorter axis parallel to the structural lineation.

Relationship between mineral and magnetic fabric

Fig. 5. Amplitude of the 2O term as a function of square of the applied field for the three perpendicular measurement planes, where black symbols represent the coefficients of the cosine term and grey symbols those of the sine term.

The principal directions of the orientation tensor, AMS tensor and HFP tensor are closely related and in good agreement with the observed bedding or cleavage planes and structural lineation. For all samples the minimum axis (£"3) of the orientation tensor corresponds very well to the maximum axis (K\) of both the AMS and HFP tensors. The intermediate and minimum axes (K2,K3) of the AMS and HFP tensors

Fig. 6. Example of a neutron diffractogram for two perpendicular planes (x = 0° and x = 90°). Various 20 angles corresponding to the peaks of different phyllosilicates are marked.

Fig. 7. Pole figures of the chlorite (001) or biotite (001), denoted as Chl(OOl) or Bi(OOl), respectively. The diagrams are equal-area, lower hemisphere projections plotted with respect to the bedding plane being horizontal, except for the sample DV65-3, which is plotted relative to the cleavage plane. Plots are contoured in multiples of random distribution (m.r.d.), contour interval is 0.4m.r.d. Maximum and minimum intensity values are in the lower right of each plot. The principal directions of the orientation tensor, HFP tensor and AMS tensor are represented with open, grey, and black symbols, respectively. Squares, triangles, and circles represent the maximum, intermediate and minimum directions, respectively. The great circles represent the cleavage plane, crosses represent either strike of bedding, bedding-cleavage intersection or fold axis, (a) Greywacke samples, (b) Mudstone samples.

PHYLLOSILICATE CONTROL OF MAGNETIC FABRIC

371

372

M. CHADIMA ET AL.

coincide reasonably well with the intermediate and maximum axes respectively (E2,E\) of the orientation tensor. The coincidence of above referred axes is less evident for the samples with more girdle-like distributions of oaxes, where the determination of E\ and E2 does not lead to clear results (Woodcock 1977). The anisotropy degree P, standard deviatoric susceptibility k!', and shape parameter U were calculated from the principal values of the AMS, HFP and theoretical anisotropy (Table 1). The theoretical anisotropy degree (PTch) calculated from chlorite preferred orientation correlates very well with the measured degree of anisotropy (PAMS)- Linear regression gives the correlation coefficient R = 0.978. Since the theoretical anisotropy depends on the grain anisotropy of the considered magnetic mineral, various grain anisotropy values of chlorite were used for PTch calculations. The grain anisotropy values for chlorite presented in the literature range from nearly 1.0 to almost 2.5 (Borradaile & Werner 1994; Martin-Hernandez & Hirt 2003). The use of variable PG values changes only the slope of the PTch vs. PAMS relation whereas the correlation coefficient (R) remains the same. The approximately one-to-one relation between PTch and PAMS *s reached if PG = 1.35 is used (Fig. 8a). Linear regression then gives the following equation:

A similar relation is observed between standard deviatoric susceptibility calculated from the theoretical anisotropy of chlorite (kfTch) an(i the AMS (&AMS) (Fig- &b). Linear regression then gives the following relationship:

The standard deviatoric susceptibility of HFP (&HFP), representing paramagnetic anisotropy, relates reasonably well to the /CAMS (Fig. 8c). Linear regression then gives the following relationship:

Standard deviatoric susceptibility of theoretical anisotropy calculated from chlorite fabric (&TCh) correlates well with the kfHFP (Fig. 8d). The correlation yields the following equation and correlation coefficient:

The shape of theoretical anisotropy derived from the chlorite fabric (UTch) is independent of the grain anisotropy (PG), exclusively depending on the distribution of crystallographic oaxes. In the f7Tch vs. C/AMS correlation, the shape parameter shows more significant variations but the prolate and oblate ellipsoids are still well defined (Fig. 9a). The shapes show higher variations for the less anisotropic samples (DV26-5-1, DV28-6-2) where the K2 and K-$ are poorly defined. For the greywacke sample DV51-1-2, the shape of measured AMS is distinctly oblate while the theoretical shape is neutral. The correlation of HFP and AMS shape (^HFP vs- ^AMS) represents paramagnetic vs. whole-rock anisotropy (Fig. 9b). The prolate and oblate fabrics are well defined, the slight discrepancies (see DV26-5-1, DV53-1-1) can be explained by very low anisotropy where the shape is not well defined. The correlation of theoretical anisotropy (^Tch) vs- paramagnetic anisotropy (t/HFP) shape is very strong and prolate and oblate shapes are well defined (Fig. 9c). The small discrepancies can be again attributed to the low anisotropy degree. Discussion A very good qualitative and quantitative correlation between experimental and theoretical anisotropy is found. As shown by the temperature variation curves, the principal carriers of magnetic susceptibility in the rocks studied are paramagnetic minerals. The contribution of the ferromagnetic fraction to the bulk susceptibility is not significant. Also the anisotropy, as seen from the HFA, is predominantly carried by the paramagnetic fraction. It has been further shown, using neutron diffraction scans and SEM images, that the most likely paramagnetic minerals are chlorites and micas. Both these minerals seem to be interlayered, forming chlorite/mica stacks showing coaxially oriented subfabrics. In addition, it can be assumed that both minerals would respond to progressive deformation in the same way and their subfabrics could be, under some simplification, regarded as one mineral fabric. As shown from neutron diffraction scans, the most suitable elements for neutron texture goniometry are the chlorite basal planes. Because of parallel orientation of chlorite and micas, the theoretical anisotropy calculated from the distribution of caxes of chlorite (hereinafter referred to as 'chlorite anisotropy' for the sake of brevity)

Fig. 8. (a) Correlation between theoretical degree of anisotropy calculated from chlorite oaxis distribution (P^h) and degree of AMS (P). (b) Correlation between the theoretical standard deviatoric susceptibility calculated from chlorite oaxis distribution (&Tch) and the standard deviatoric susceptibility of AMS (&AMS)-

Fig. 8. (c) Correlation between the standard deviatoric susceptibility of HFP (&HFP) an<^ the standard deviatoric susceptibility of the AMS (/CAMS)- (d) Correlation between the theoretical standard deviatoric susceptibility (&Tch) and the standard deviatoric susceptibility of HFP (k'HFP). Linear regression lines and correlation coefficients are added.

PHYLLOSILICATE CONTROL OF MAGNETIC FABRIC

375

Fig. 9. (a) Correlation between the theoretical shape parameter (C/jch) calculated from chlorite c-axis distribution and the shape parameter of AMS (C/AMS)- (b) Correlation between the shape of HFP (£7HFP) and the shape parameter of the AMS (£/AMS)- (c) Correlation between the theoretical shape parameter (C/Tch) and the shape of HFP (UHFP).

should be more or less equal to the paramagnetic as well as to the whole-rock anisotropy.

Qualitative correlation In most cases, the principal directions of the AMS, HFP and the orientation tensor derived from chlorite oaxes are sub-parallel (Fig. 7). No systematic deviation of paramagnetic fabric from whole-rock magnetic fabric can be observed. The poles figures of the least anisotropic and prolate samples (i.e. DV28-1-2, DV28-6-2, DV29-1-1, DV45-3-2) suggest a

poorly defined foliation. The respective directions of El9 K3AMS, K3HpP and E2, KIAMS, K2uFp do not coincide and lie on a girdle defined by oaxis distribution. On the other hand, the lineation is well defined with £3, #1AMS and possibly A^Hpp being sub-parallel. The prolate AMS as well as girdle distribution of oaxes of chlorite suggest that the lineation is an intersection axis of chlorite basal planes. For more anisotropic but still prolate samples (i.e. DV24-7-1, DV28-8-2, DV54-1-1, DV57-2-2, DV62-2-1, DV63-7-1), all three sets of principal directions are well grouped. The magnetic lineations (both AMS and HFP) are sub-parallel

376

M. CHADIMA ET AL,

to the £3 direction. The magnetic foliation ranges from bedding-parallel to cleavage-parallel. The most oblate samples (DV63-2-2, DV65-3) show cluster distribution of c-axes, oval in shape with short oval axis parallel to the magnetic lineations. All respective principal directions coincide within the limit of a few degrees.

Quantitative correlation The standard deviatoric susceptibility enables comparisons of the degrees of chlorite anisotropy, paramagnetic and whole-rock anisotropy to be made. Each of the three possible correlations has a very high correlation coefficient, R > 0.95, implying nearly the same degree of anisotropy for all the methods employed. It is evident that the chlorite anisotropy reasonably represents the whole-rock anisotropy. This fact further supports the assumption that chlorite is the dominant control of the magnetic fabric. The correlations of the shapes of all investigated anisotropies show more significant scatters compared to the degree of anisotropy. Despite this dispersion, the prolate and oblate shapes still remain to be well defined. It seems that the most noticeable differences from one-to-one relationship can be observed for the samples with the lowest anisotropy where the shape is poorly defined. It can be noticed that the variability in shapes of HFP fit to slightly narrower range (approx. —0.3 to 0.4) than those of AMS and chlorite anisotropy. Similarly to the previous, both extreme prolate or oblate values of AMS are reached for the samples with the lowest anisotropy degree. As the range of chlorite anisotropy shapes is similarly wide as that of AMS it seems that the HFA is less sensitive to the shape variability than AMS.

Model for two phyllo silicates As shown on the pilot sample DV63-2-2, the biotite texture is less pronounced than the chlorite one (2.5m.r.d. intensity for biotite and 5.36 for chlorite). This difference was also found on slates from Sardinia (Liineburg et al. 1999). Theoretical AMS calculated from biotite preferred orientation is less anisotropic and more prolate than chlorite one (Table 1). Knowing the orientation tensors for chlorite and biotite we can calculate the theoretical anisotropy of a rock composed of both minerals. The rock susceptibility model of Henry (1983) and Henry & Daly (1983) was adapted to our

situation:

or

where kr is the rock susceptibility tensor, k d , kch, kbi and kf are susceptibility tensors and cd, cch, cbi and Cf the amounts of the diamagnetic, chlorite, biotite and ferromagnetic fractions respectively. k n r> k nd , knch, knbi, knf are the normed susceptibility tensors of whole-rock, diamagnetic, chlorite, biotite and ferromagnetic fractions, while the A;r, fcd, £ch, fcbi, kf are the mean susceptibilities of the respective fractions. Since we do not know the exact values of mean susceptibility of chlorite and biotite (data presented in literature show large variability) we divide the previous equation by the mean susceptibility of rock and introduce p — ck/kr as the percentage of susceptibility of the respective fractions in the mean rock susceptibility. The equation can be expressed as: We assume the percentage of diamagnetic susceptibility to be very low so it can be, for the sake of simplicity, neglected. As already shown, the ferromagnetic fraction is also negligible. Thus the ferromagnetic term can be omitted as well. The resulting equation is:

Because the orientation tensors of chlorite and biotite are nearly coaxial, the diagonal form of the theoretical susceptibility tensors of the respective minerals is sufficient for the calculation of the whole-rock diagonal susceptibility tensor. Since grain anisotropy data for chlorite and biotite presented in the literature (Borradaile et al. 1987; Zapletal 1990; Borradaile & Werner 1994; Martin-Hernandez & Hirt 2003) vary significantly, the resulting equation has a wide range of solutions. Our model assumes various values of grain anisotropy of chlorite and biotite (Poch and PGbi). The degree of model rock anisotropy is linearly increased with increasing chlorite percentage on bulk rock susceptibility (Fig. 10). The anisotropy degree of model rock increases with increasing grain anisotropy of both minerals while the slope of linear dependence changes only slightly. Our pilot sample (DV63-2-2) has both degree of anisotropy and shape close to the model rock with 100% chlorite percentage. Although the presence of micas (either biotite or muscovite)

PHYLLOSILICATE CONTROL OF MAGNETIC FABRIC

377

Fig. 10. The degree of model of rock anisotropy (PT) composed of biotite and chlorite (with variable grain anisotropy) calculated from appropriate textures measured on sample DV63-2-2. Percentage of chlorite susceptibility in whole-rock susceptibility varies from 0 to 100%.

was proved in all the samples of our collection, it seems they have only a minor influence on whole-rock AMS, leaving chlorite as the dominant carrier of rock anisotropy. If micas possess substantially more prolate fabric than chlorite they will significantly influence the whole-rock anisotropy only if their grain anisotropy substantially differs from grain anisotropy of chlorite or if they dominate the whole-rock susceptibility.

Relationship of the AMS to the various methods of fabric analysis The intensity values of the observed phyllosilicate fabrics are very weak compared to slates investigated by, for example, Luneburg et al. (1999) and Martin-Hernandez (2002), where Xray texture goniometry was used for the fabric analysis. As already noticed by Ullemeyer & Weber (1999), quantitative comparison of the pole figure intensities obtained with different methods of fabric analysis is hard to achieve. The quantitative correlations between the AMS and phyllosilicate fabric presented in this work are more significant than the correlations published earlier (e.g. Hrouda & Schulmann 1990; Hrouda & Ullemeyer 2001). These studies were usually done on the higher grade metamorphic rocks with mica-controlled AMS. A better correlation was reached when the lower

grade rocks with chlorite-controlled AMS were studied (Hrouda et al. 1997b). There may be several reasons for the lower quantitative correlation observed in previous works. The preferential selection of the larger biotite grains during U-stage measurements may have played a role in the case of Hrouda & Schulmann (1990). The significant contribution of ferromagnetic fabric may have obscured the correlation in the case of Hrouda & Ullemeyer (2001). These obstacles were overcome by the use of neutron texture goniometry combined with high-field anisotropy. A very good correlation between the theoretical anisotropy and the AMS proved that the separation of the phyllosilicate fabric from the polymineralic rocks was very successful by the means of neutron texture goniometry. Using neutron texture goniometry, even the coarse-grained greywacke samples could be successfully analysed. For practical reasons it was found more convenient to sample the compact greywackes than highly fractured mudstones. For strain analysis of the larger area of flyschlike rocks it is definitely easier to collect extensive sets of greywacke samples than the same number of conventionally used mudstones. Conclusions The presented combination of infrequently used neutron texture goniometry and low- and

378

M. CHADIMA ET AL.

high-field magnetic anisotropy shows a very significant qualitative and quantitative correlation between mineral and magnetic fabric of studied rock. As evidenced by the temperature variation of magnetic susceptibility, and the high-field anisotropy, the magnetic susceptibility is predominantly carried by the paramagnetic minerals, i.e. chlorite and to a lesser extent by micas. As visualized by SEM, chlorite and mica grains usually form kinked chlorite/mica stacks. After separation of the paramagnetic and ferromagnetic contributions to the magnetic anisotropy using the HFA, ferromagnetic contribution to the magnetic anisotropy was proved to be insignificant. The paramagnetic minerals dominate the magnetic anisotropy of the rocks studied. The high-field paramagnetic tensor provides a more accurate gauge for the correlation with the theoretical anisotropy calculated from the chlorite or mica preferred orientation. In most cases, the principal directions of the AMS, HFP and the theoretical anisotropy are sub-parallel. With increasing degree of magnetic anisotropy the chlorite oaxis distribution ranges from the girdle-like to cluster-like. Magnetic foliation is normal to the direction of maximal intensity of chlorite oaxes, whereas the magnetic lineation is sub-parallel to the intersection axis of chlorite basal planes. No systematic deviation of the paramagnetic fabric from whole-rock magnetic fabric can be observed. The degree of magnetic anisotropy (& max /& min ) correlates well with the degree of theoretical anisotropy calculated from chlorite pole figure data. The correlation reaches approximately one-toone relation when the chlorite grain anisotropy PG = 1.35 is used for calculation of the theoretical anisotropy. The quantification of the HFA degree was facilitated by the employment of the standard deviatoric susceptibility as defined by Jelinek (1984). This infrequently used parameter proved to be a very effective measure whenever one needs to correlate the intensity of the highfield anisotropy with other types of anisotropy. The shape of the measured and theoretical anisotropy shows more significant variations but the prolate and oblate fabrics are still well defined. The shapes show higher variations for the less anisotropic samples. For comparative studies of the phyllosilicate and magnetic fabric of low grade rocks, neutron texture goniometry seems to be a preferable and more efficient method than other methods of fabric analysis. Neutron texture goniometry justifies the use of the conventional AMS technique for assessment of the rock fabric of the studied rocks.

F. Martin-Hernandez is thanked for providing us with her MatLab routines for processing the high-field anisotropy data. The SEM analyses were kindly done and evaluated by P. Sulovsky. The neutron texture goniometry measurements were financed by GKSS Forschungszentrum Geesthacht GmbH, Germany. GKSS also provided the travel grant for M.C. in order to visit their neutron facility. M.C.'s stay in Zurich was sponsored by the Czech Ministry of Education, Prague, and warmly supported by J. & M. McWilliams, Lauerz, Switzerland. H. de Wall and B. Henry are thanked for their criticism and comments, which significantly improved the presented paper.

References BANERJEE, S. K. & STAGEY, F. D. 1967. The high-field torque-meter method of measuring magnetic anisotropy of rocks. In: COLLINSON, D. W., CREER, K. M. & RUNCORN, S. K. (eds) Methods in Palaeomagnetism. Elsevier, Amsterdam, 470476. BERGMULLER, F., BARLOCHER, C., GEYER, B., GRIEDER, M., HELLER, F. & ZWEIFEL, F. 1994. A torque magnetometer for measurements of the high-field anisotropy of rocks and crystals. Measurement Science and Technology, 5, 1466-1470. BORRADAILE, G. J. & HENRY, B. 1997. Tectonic applications of magnetic susceptibility and its anisotropy. Earth-Science Reviews, 42, 49-93. BORRADAILE, G. J. & WERNER, T. 1994. Magnetic anisotropy of some phyllosilicates. Tectonophysics, 235, 223-248. BORRADAILE, G., KEELER, W., ALFORD, C. & SARVAS, P. 1987. Anisotropy of magnetic susceptibility of some metamorphic minerals. Physics of the Earth and Planetary Interiors, 48, 161-166. BORRADAILE, G., MOTHERSILL, J., TARLING, D. & ALFORD, C. 1986. Sources of magnetic susceptibility in a slate. Earth and Planetary Science Letters, 76, 336-340. BRAUN, G. 1994. A statistical geometric method for quantitative texture analysis. In: BURGE, H. J., SlEGESMUND, S., SKROTZKI, W.

& WEBER,

K.

(eds) Textures of Geological Materials. DGM Informationsgesellschaft, Oberursel, 29-60. BROKMEIER, H.-G. 1994. Application of neutron diffraction to measure preferred orientation of geological materials. In: BURGE, H. J., SIEGESMUND, S., SKROTZKI, W. & WEBER, K. (eds) Textures of Geological Materials. DGM Informationsgesellschaft, Oberursel, 327-344. BROKMEIER, H.-G. 1997. TEX-2 the texture diffractometer at the GKSS-research centre. Physica B, 234-236, 1144^1145. CASEY, M. 1981. Numerical analysis of X-ray texture data: an interpretation in Fortran allowing triclinic or axial specimen symmetry and most crystal symmetries. Tectonophysics, 78, 51-64. DVORAK, J. 1973. Synsedimentary tectonics of the Paleozoic of the Drahany Upland (Sudeticum, Moravia, Czechoslovakia). Tectonophysics, 17, 359-391.

PHYLLOSILICATE CONTROL OF MAGNETIC FABRIC DVORAK, J. & WOLF, M. 1979. Thermal metamorphism in the Moravian Paleozoic (Sudeticum, C.S.S.R). Neues Jahrbuch fur Geologic und Paldontogie Abhandlungen, Monatshefte, 10, 597-607. FRANCU, E., FRANCU, J. & KALVODA, J. 1999. Illite crystalinity and vitrinite reflectance in Paleozoic siliciclastics in the SE Bohemian Massif as evidence of thermal history. Geologica Carpatica, 50(5), 365-372. HARTLEY, A. J. & OTAVA, J. 2001. Sediment provenance and dispersal in a deep marine foreland basin: the Lower Carboniferous Culm Basin, Czech Republic. Journal of the Geological Society, London, 158, 137-150. HENRY, B. 1983. Interpretation quantitative de 1'anisotropie de susceptibilite magnetique. Tectonophysics, 9\, 165-177. HENRY, B. & DALY, L. 1983. From qualitative to quantitative magnetic anisotropy analysis: the prospect of finite strain calibration. Tectonophysics, 98, 327-336. HROUDA, F. 1982. Magnetic anisotropy of rocks and its application in geology and geophysics. Geophysical Surveys, 5, 37-82. HROUDA, F. 1994. A technique for the measurement of thermal changes of magnetic susceptibility of weakly magnetic rocks by the CS-2 apparatus and KLY-2 Kappabridge. Geophysical Journal International, 118, 604-612. HROUDA, F. & JELINEK, V. 1990. Resolution of ferromagnetic and paramagnetic anisotropies, using combined low-field and high-field measurements. Geophysical Journal International, 103, 75-84. HROUDA, F., JEZEK, J. 1999. Magnetic anisotropy indications of deformations associated with diagenesis. In: TARLING, D. H. & TURNER, P. (eds) Palaeomagnetism and Diagenesis in Sediments. Geological Society, London, Special Publications, 151, 127-137. HROUDA, F. & SCHULMANN, K. 1990. Conversion of magnetic susceptibility tensor into orientation tensor in some rocks. Physics of the Earth and Planetary Interiors, 63, 71-77. HROUDA, F. & ULLEMEYER, K. 2001. Quantitative correlation of biotite lattice preferred orientations and magnetic fabrics in granulites from the Southern Bohemian Massif. Zeitschrift der deutschen geologischen Gesellschaft, 152(2-4), 547-561. HROUDA, F., JELINEK, V. & ZAPLETAL, K. 1997a. Refined technique for susceptibility resolution into ferromagnetic and paramagnetic components based on susceptibility temperature-variation measurement. Geophysical Journal International, 129, 715-719. HROUDA, F., SCHULMANN, K., SUPPES, M., ULLEMEYER, K., DE WALL, H. & WEBER, K. 1997b. Quantitative relationship between low-field AMS and phyllosilicate fabric: a review. Physics and Chemistry of the Earth, 22, 153-156. HROUDA, F., SIEMES, H., HERRES, N. & HENNIGMICHAELI, C. 1985. The relation between the magnetic anisotropy and the c-axis fabric in a massive hematite ore. Journal of Geophysics, 56, 174-182.

379

JELINEK, V. 1981. Characterization of the magnetic fabric of rocks. Tectonophysics, 79, T63-T67. JELINEK, V. 1984. On a mixed quadratic invariant of the magnetic susceptibility tensor. Journal of Geophysics, 56, 58-60. JELINEK, V. & POKORNY, J. 1997. Some new concepts in technology of transformer bridges for measuring susceptibility anisotropy of rocks. Physics and Chemistry of the Earth, 22, 179-181. LAMPERT, S. 1996. Palaomagnetische und gesteinsmagnetische Beitrage zur Losung von stratigraphischen und tektonischen Problemen in den abruzzischen Apenninen und in SW-Sardinien. PhD thesis, ETH Zurich. Li, G., PEACOR, D. R., MERRIMAN, R. J., ROBERTS, B. & VAN DER PLUIJM, B. A. 1994. TEM and AEM constrains on the origin and significance of chloritemica stacks in slates: an example from Central Wales, U.K. Journal of Structural Geology, 16, 1139-1157. LUNEBURG, C. M., LAMPERT, S. A., LEBIT, H. D., HIRT, A. M., CASEY, M. & LOWRIE, W. 1999. Magnetic anisotropy, rock fabrics and finite strain in deformed sediments of SW Sardinia (Italy). Tectonophysics,3Ql, 51-74. MARTIN-HERNANDEZ, F. 2002. Determination of fundamental magnetic anisotropy parameters in rock-forming minerals and their contributions to the magnetic fabric of rocks. PhD thesis, ETH Zurich. MARTIN-HERNANDEZ, F. & HIRT, A. M. 2001. Separation of ferrimagnetic and paramagnetic anisotropies using a high-field torsion magnetometer. Tectonophysics, 337, 209-221. MARTIN-HERNANDEZ, F. & HIRT, A. M. 2003. The anisotropy of magnetic susceptibility in biotite, muscovite and chlorite single crystals. Tectonophysics, 367, 13-28. MILODOWSKI, A. E. & ZALASIEWICZ, J. A. 1991. The origin, depositional and prograde evolution of chlorite-mica stacks in Llandovery sediments of the central Wales Basin. Geological Magazine, 128, 263-278. NAGATA, T. 1961. Rock Magnetism. 2nd edition, Maruzen, Tokyo. OWENS, W. H. & BAMFORD, D. 1976. Magnetic, seismic, and other anisotropic properties of rock fabric. Philosophical Transactions of Royal Society, London, A, 283, 55-68. OWENS, W. H. & RUTTER, E. H. 1978. The development of magnetic susceptibility anisotropy through crystallographic preferred orientation in a calcite rock. Physics of the Earth and Planetary Interiors, 16,215-222. PARMA, J. & ZAPLETAL, K. 1991. CS-1 Apparatus for measuring the temperature dependence of lowfield susceptibility of minerals and rocks (in cooperation with the KLY-2 Kappabridge). Leaflet, Geofyzika, Brno. PARMA, J., HROUDA, F., POKORNY, J. & WOHLGEMUTH, J. 1993. A technique for measuring temperature dependent susceptibility in weakly magnetic rocks. Supplement to EOS Transactions, AGU, 113, Washington.

380

M. CHADIMA ET AL.

PASSCHIER, C. W. & TROUW, R. A. J. 1998. Microtectonics. 2nd edition, Berlin. RATHORE, J. S. 1979. Magnetic susceptibility anisotropy in the Cambrian Slate Belt of North Wales and correlation with strain. Tectonophysics, 53, 83-97. RICHTER, C., PLUIJM, B. A. VAN DER, HOUSEN, B. A. 1993. The quantification of crystallographic preferred orientation using magnetic anisotropy. Journal of Structural Geology, 15, 113-116. ROCHETTE, P. 1987. Magnetic susceptibility of the rock matrix related to magnetic fabric studies. Journal of Structural Geology, 9, 1015-1020. SCHAFER, W. 2002. Neutron diffraction applied to geological textures and stress analysis. European Journal of Mineralogy, 14(2), 263-289. SCHEIDEGGER, A. E. 1965. On the statistics of the orientation of bedding planes, grain axes and similar sedimentological data. US Geological Survey Professional Paper, 525C, 164-167. SlEGESMUND, S., ULLEMEYER, K. & DAHMS, M.

1995.

Control of magnetic rock fabrics by mica preferred orientation: a quantitative approach. Journal of Structural Geology, 17, 1601-1613. SlEMES, H., SCHAEBEN, H., ROSIERE, C. & QUADE, H.

2000. Crystallographic and magnetic preferred orientation of hematite in banded iron ores. Journal of Structural Geology, 22, 1747-1759.

TARLING, D. H. & HROUDA, F. 1993. The Magnetic Anisotropy of Rocks. Chapman and Hall, London. ULLEMEYER, K. & WEBER, K. 1994. Correction of phyllosilicate (002) X-ray pole figure measurements. In: BURGE, H. J., SIEGESMUND, S., SKROTZKI, W. & WEBER, K. (eds) Textures of Geological Materials. DGM Informationsgesellschaft, Oberursel, 83-91. ULLEMEYER, K. & WEBER, K. 1999. Lattice preferred orientation as an indicator of a complex deformation history of rocks. Textures and Microtextures, 33, 45-60. DE WALL, H., BESTMANN, M. & ULLEMEYER, K. 2000. Anisotropy of diamagnetic susceptibility in Thassos marble: A comparison between measured and modelled data. Journal of Structural Geology, 22, 1761-1771. WOOD, D. S., OERTEL, G., SINGH, J.& BENNET, H. G. 1976. Strain and anisotropy in rocks. Philosophical Transactions of Royal Society, London, A, 283, 27-42. WOODCOCK, N. H. 1977. Specification of fabric shapes using an eigenvalue method. Geological Society of America, Bulletin, 88, 1231-1236. ZAPLETAL, K. 1990. Low-field susceptibility anisotropy of some biotite crystals. Physics of the Earth and Planetary Interiors, 63, 85-97.

Anisotropy of magnetic susceptibility in the Montes de Toledo area (Hercynian Iberian Belt, Spain) and its petrostructural significance A. GIL-IMAZ1 & L. BARBERO2 1

Departamento de Ciencias de la Tierra, Universidad de Zaragoza, Cj Pedro Cerbuna 12. 50009-Zaragoza, Espana (e-mail: [email protected]) 2 Departamento de Geologia, Universidad de Cadiz, 11510-Puerto Real (Cadiz), Espana. Abstract: Anisotropy of magnetic susceptibility data are presented for samples from the anatectic complex of Toledo and the Mora batholith. The units located at the axial part of the Central Hercynian belt are separated by an important listric fault. Anatectic granites (Layos granite), epizonal granites (Mora granite) and, to a lesser extent, high-grade metamorphic rocks and metasediments are considered in this work in order to characterize their magnetic fabric and determine their structural significance in the framework of a large-scale extensional deformation of Late Variscan age. Samples from the Layos granites (footwall) show a magnetic lineation compatible with the stretching related to ductile shear deformation of the Toledo shear zone, which was responsible for the exhumation of the anatectic complex. During the exhumation, the folding of a previous metamorphic foliation in the anatectic complex took place, which is also recognized from the magnetic pattern of the minimum susceptibility axes of Layos granite. On the contrary, the magnetic fabric of the epizonal Mora granites (hanging wall), which intruded at the beginning of the extensional Late Variscan tectonics, mainly reflects magmatic processes.

The anisotropy of magnetic susceptibility (AMS) is a fast and non-destructive method of investigating rock fabrics widely used to solve different petrostructural problems. Since Graham (1954) proposed the use of the AMS for petrostructural interpretations, many efforts have been made to establish the structural significance of the magnetic fabric of different rock types within their tectonic framework. Structural interpretations of AMS data from granitic rocks largely concern emplacement mechanisms of epizonal intrusions and their relation to large-scale structures (Heller 1973; Guillet et al. 1983; Bouchez et al. 1990; Gleizes et al 1993; Leblanc et al 1994; Roman-Berdiel et al 1995; Aranguren et al 1997; Bouchez 1997; Ferre et al 1997; Olivier et al 1999). Magnetic fabric studies of metagranitic and anatectic rocks have focused on their relationship with metamorphic processes (Hrouda et al 1971; Heller 1973; Hrouda & Chlupacova 1980; Bascou et al 2002; Raposo et al 2003). Anatectic complexes represent very interesting petrostructural domains in which the relationships between the kinematics of the exhumation process and different petrological processes can be investigated. This paper presents a study of the AMS of both high-grade metamorphic rocks (peraluminous granitoids, the Layos granite, and migmatic granulites) and epizonal granites (calc-alkaline granitoids: the Mora granite) whose regional evolution is closely

related to the Toledo shear zone: a large-scale brittle-ductile shear zone. The main objective of this work is to analyse the AMS of rock types linked to different petrological, metamorphic and structural conditions in order to establish petrostructural interpretations. The AMS is a suitable method to address this purpose for two main reasons: (1) AMS can be applied to several rock types containing different magnetic mineralogy; and (2) the high sensitivity of the susceptibility meters (up to bulk susceptibilities of the order of 5 x 10 SI units), which permits the measurement of tectonic, magmatic or metamorphic structures even in very weakly structured rocks. Geological setting The Montes de Toledo area constitutes a part of the axial zone of the Central Iberian Zone within the Variscan Iberian Belt. Two different sectors can be recognized (Fig. 1): a northern anatectic complex and a southern epizonal unit, the Mora batholith. The anatectic complex comprises igneous rocks that intrude pre-Ordovician metamorphic series that reached granulitic conditions (Barbero et al 1990; Barbero & Villaseca 1992; Barbero 1995). These igneous rocks include synorogenic calc-alkaline granitoids, associated with the Variscan metamorphic peak (Barbero & Villaseca 1992), with associated minor basic rocks

From: MARTIN-HERNANDEZ, F., LUNEBURG, C. M., AUBOURG, C. & JACKSON, M. (eds) 2004. Magnetic Fabric: Methods and Applications. Geological Society, London, Special Publications, 238, 381-394. 0305-8719/04/ $15.00 © The Geological Society of London 2004.

Fig. 1. Geological map of the Toledo Complex. The location of AMS sites is also indicated. Modified after Barbero & Villaseca (1992). Inset map shows the two petrostructural units considered in this work.

AMS IN ANATECTIC COMPLEXES

(mainly gabbros) and several types of peraluminous granitoids: the Layos granite and the Moncloa granite (Barbero 1995). The most generalized structure is a penetrative slaty cleavage, with a 120°-170° regional strike and dips to the east and NE. This foliation is assigned to the second (D2) Variscan deformation phase (Martin Escorza & Lopez Martinez 1978; Barbero 1992, 1995). According to the isothermal decompression path deduced in this area (Barbero 1995), crustal thinning and exhumation of the Toledo anatectic complex must have taken place at the end of the Variscan orogeny, becoming the ascent and emplacement of calc-alkaline granitoids (e.g. the Mora pluton). This geodynamic model has been proposed in other parts of the Central Iberian Zone (Doblas & Martinez 1988; Macaya et al 1991; Diez-Balda et al. 1991). To the north, the anatectic complex is in contact with deformed Cenozoic sediments of the Tajo Basin. The southern boundary is a major low angle normal ductile-brittle shear zone with listric geometry, the Toledo shear zone (Hernandez Enrile 1991; Santa Teresa et al. 1983). Mylonites in the shear zone can reach up to 400 m in thickness, with a mylonitic foliation dipping 30°-40° to the south. This crustal-scale structure brings the anatectic complex in contact with low-grade Palaeozoic metasediments intruded by the Mora epizonal batholith. This igneous intrusion is an east-west elongated body composed of granodioritic rocks of calcalkaline affinity (Andonaegui 1990) (Fig. 1). It is a coarse-grained biotite-bearing granite with several disperse K-feldspar megacrysts. A detailed description of the mineral chemistry of these granites can be found in Villaseca & Barbero (1994). In some places accumulation of both these megacrysts and several types of enclaves are observed, which is indicative of important dynamic processes in the magma chamber (Andonaegui 1990; Villaseca et al. 1998). The main intrusion of the Mora pluton is limited to 320 ± 8 Ma (Andonaegui 1990; Villaseca et al. 1998). In relation to the Late Variscan tectonics, the Mora granitoids can be considered as a late-orogenic (post-D2 compressive deformation phase) or a post-metamorphic peak pluton (Villaseca et al. 1993). Research methodology Graham (1954) demonstrated the application of the AMS for structural analysis. Many efforts have been made since then to decipher the relationship between the magnetic fabric of a

383

rock and its petrofabric. AMS represents a measure of the crystallographic orientation of paramagnetic phases (mainly phyllosilicates and other tabular grains) and of the shape orientation of ferrimagnetic grains like magnetite. The magnetic susceptibility (k) of a rock is a physical parameter that relates the induced magnetization (M) with the applied magnetic field (H): H = fcM. The analysis of the magnetic fabric of rocks is based on determining the anisotropy of the magnetic susceptibility (AMS) when a low magnetic field is applied along different directions on a rock specimen. The variations of the magnetic susceptibility with spatial direction can be modelled as a second-rank tensor, which can be represented by an ellipsoid, with maximum (K\), intermediate (^2) and minimum (K3) principal axes (Ki > K2 > K3) defined by their magnitude and direction. The magnetic parameters used in this work are: (1) the 'degree of anisotropy' where QI =ln(Kl/Km)9 a2=\n(K2/Km\ a3 =\n(K3/Km) and Km is the bulk susceptibility (see definition below) (2) the 'symmetry of shape' The PJ parameter is used to quantify the intensity of the magnetic anisotropy, and 7} characterizes the shape of the susceptibility ellipsoid (Jelinek 1981). In addition to these parameters, Km = (Ki + K2 + K3)/3 represents the mean susceptibility of a single rock specimen. The AMS was measured with a KLY-1.02 susceptibility meter using a bridge at low magnetic field (±4 x 10~ 4 T and 920 Hz). The measurements are based on determining the directional susceptibilities along 15 suitable directions in the rock specimens (Jelinek 1981). In order to characterize the magnetic fabric of the two mentioned units, sampling sites were performed across three cross sections (Figs 1, 2). In the northern anatectic complex (footwall), samples were mainly collected in the Layos granite (129 specimens, 8 sites) due to two particularities: (1) its high textural and compositional homogeneity in comparison to the metamorphic rocks of the area; and (2) its close relationship with the anatectic process since these rocks have been related with a biotite dehydration melting process from granulitic migmatites. Samples of three other sites were taken from granulitic migmatites (TO 14), orthogneisses (TO 15) and leucogneisses (T003). In the epizonal

384

A. GIL-IMAZ & L. BARBERO

385

AMS IN ANATECTIC COMPLEXES

Mora granite (hanging wall) nine sites (108 specimens) were drilled across the three cross sections. Samples of one site (TO 19) were obtained from the Cambrian host metasediments. AMS results The AMS study focuses on: (1) the correlation between the bulk magnetic susceptibility (Km) and the degree of magnetic anisotropy (Pj); (2) the shape of the magnetic ellipsoids (7}); and (3) the relationship between the orientation of the principal susceptibility axes and the representative petrofabric indicators (e.g. tectonic and magmatic foliations and lineations). For the statistics, we calculated the standard deviation of the arithmetical mean of the anisotropy parameters (Table 1). Equal area, lower hemisphere, projections have been used for directional data.

AMS data in the hanging wall AMS parameters (Km, Pj and Tj) Km reflects the contribution of all the mineralogical phases present in a rock specimen. From this point of view, Km values below 300 x 10~6 SI units are related to rocks with dominant paramagnetic susceptibility (Rochette 1987). Furthermore, Km values below 300 x 10~6 and Pj < 1.35 are linked to phyllosilicate-bearing rocks (Rochette 1987). In other words, if a rock contains paramagnetic minerals as common constituents (c. 10%) and its susceptibility is below 300 x 10~6 SI units, the paramagnetic fraction controls the susceptibility and anisotropy of the rock (Tarling & Hrouda 1993). On the other hand, amounts of paramagnetic minerals >10% and susceptibility values above this value indicate a dominant ferrimagnetic control of both susceptibility and the magnetic anisotropy of the rock. According to this, the obtained Km values for samples from the Mora granite (between 100 and 300 x 10"6SI) and the P, values (<1.35) indicate that paramagnetic minerals are the main carriers of the magnetic susceptibility (Fig. 3a). Such paramagnetic behaviour is consistent with its mineralogical composition

Table 1. Mean anisotropy susceptibility data Site

N

Pj

Tj

1.049 1.021 1.022 1.029 1.029 1.023 1.023 1.027 1.024

0.27 -0.28 0.16 0.38 -0.30 -0.17 0.19 0.18 0.32

Low-grade host metasediments TO19 10 251

1.169

0.21

Footwall block Lay os Granite TO02 10 TO04 11 TO08 16 12 TO10 12 TO11 12 TO16 T017 17 TO18 20

1.070 1.021 .059 1.033 1.015 .028 1.064 1.261

0.48 0.02 0.35 0.34 -0.12 0.21 0.26 -0.07

High-grade metamorphic country-rocks TO03 14 274 1.066

0.22

Migmatite TO14

Hanging wall block Mora Granite 11 TO01 TO05 6 TO06 13 TO07 12 T009 11 TO12 14 T013 12 TO20 13 TO21 16

^m

229 121 120 188 131 254 241 141 123

256 305 197 230 359 300 329 658

12

1160

1.157

0.02

Orthogneiss TO15 12

241

1.054

-0.15

N = number of specimens; Km = bulk susceptibility (xlO~ 6 SI units); Pj, — degree of anisotropy (after Jelinek 1981)); 7} = symmetry of shape (after Jelinek 1981)

(Andonaegui 1990; Villaseca & Barbero 1994), which includes euhedral biotite as the main paramagnetic phase (with an anisotropy at room temperature from 1.1 to 1.37; Rochette 1987) with anhedral quartz, euhedral to subhedral albite-rich plagioclase, and euhedral Kfeldspar phenocrysts as major mineralogical phases. Other less abundant phases include cordierite, pinite and small garnet crystals with associated spinel as the only ferrimagnetic component. No correlation between Km and Pj is observed, so there is no strong influence of the magnetic mineralogy content on the degree of

Fig. 2. Cross sections and stereoplots with AMS and structural data, (a) Toledo cross section, (b) Galvez cross section, (c) Mora cross section. Squares, triangles and circles inside stereoplots indicate K\, K2 and K^ directions respectively. Great circles correspond to the different types of foliations measured in field. Directional data were plotted using Richard Allmendinger's program (Stereonet 4.9.5). Lower hemisphere, equal area.

386

A. GIL-IMAZ & L. BARBERO

Fig. 3. (a) Km versus Pj plots for samples from the hanging wall, (b) Km versus Pj plots for samples from the footwall. Inset plot corresponds to Layos granite.

arrangement of the magnetic minerals (Rochette 1987; Rochette at al 1992; Hrouda & Jelinek 1990). Although some dispersion in the ellipsoid shapes is observed (-0.5 < 7} < 0.5), the mean shape of the AMS ellipsoids for the Mora granites fits into the oblate ellipsoids field (Fig. 4a). The magnetic ellipsoids show a low dispersion in the degree of anisotropy with a mean value of 1.025. Nevertheless, site TO 19, corresponding to a schist (with quartz, muscovite and chlorite as major minerals and tourmaline, zircon and iron oxides as accessory minerals) from the country rocks, shows a higher degree of anisotropy, possibly due to a ferrimagnetic phase or to a different clay composition.

As shown by the plots in Figure 5, no significant variations in the Pj and Km parameters exist with the distance from the Toledo shear zone. Although only nine sites have been considered, the constant mean value of Pj (« 1.025) with respect to the distance to the shear zone suggests a weak influence of the associated deformation in the magnetic fabric of these rocks. The peak at 2.5km is related to a better preferred orientation of K-feldspar megacrysts. Directional AMS data Directional AMS data from the Mora granite samples also agree with a weak influence of the Toledo shear zone in the acquisition of their

AMS IN ANATECTIC COMPLEXES

Fig. 4. Anisotropy plots of degree of anisotropy (P,-) versus symmetry of shape (7}) for each site of the hanging wall (a) and the footwall (b) respectively. Error bars for each site represent the standard deviations of the arithmetic mean. Mean error bar (bold line) for sites of the Mora granite (hanging wall) is also reported.

magnetic fabric. Sites located far from the shear zone (sites TO01, TO06, TO 12, TO20) show a magnetic lineation consistent with the mean direction of the measured K-feldspar phenocrysts in the Mora granites (Fig. 6a), and thus, the preservation of an original feature of the igneous petrofabric is suggested. On the other hand, the parallelism between magnetic and granitic foliations (site TOO?) indicates the primary origin of the planar magnetic fabric, So that the sub-horizontal magnetic foliation defined by the K\ and K2 axes (Figs. 6b, c) could be interpreted as a mean magmatic surface, From the petrographic point of view, these rocks are characterized by an isotropic and

inequigranular subeuhedral texture indicating their non-deformed condition (Andonaegui 1990; Villaseca et al. 1998). Site TO 13, located within the mylonitic zone, shows a different directional magnetic pattern (Fig. 2b). Here the magnetic lineation shows a southernmost strike and then is compatible with the shear sense. In addition, it lies on the extensional foliation measured in field. The existence of a magmatic fabric prior to a deformation linked to an extensional ductile-brittle shear zone could explain the lack of coincidence between the tectonic and magnetic foliations. In fact, its rock texture is characterized by a set of spaced micro-shear zones that do not affect to

Fig. 5. Plots of (a) degree of anisotropy (P;) and (b) bulk susceptibility (Km) against distance (D) to the Toledo shear zone for the samples from the hanging wall block.

388

A. GIL-IMAZ & L. BARBERO

the phenocrysts but are adapted to the shear movement (Macaya et al. 1991). Finally, in site TO 19 (from Cambrian schists of host rocks affected by a penetrative slaty cleavage linked to the second (D2) Variscan compressive phase), the magnetic lineation, parallel to the intersection lineation between cleavage and bedding, indicates its tectonic origin (Fig. 2a).

AMS data in the footwall

Fig. 6. (a) Frequency stereogram of K\ axes of Mora granite. The rose diagram corresponding to the measured directions of K-feldspar phenocrysts is also reported, (b) Frequency stereogram of K3 axes for the same sites, (c) The whole of K} and K2 axes (magnetic foliation). Contour intervals are 1.4% in (a) & (b), and 1.1% in (c). See text for explanations.

AMS parameters (Km, Pj and Tj) Samples from the footwall, with Km values ranging between 70 and 1200 x 10~6 (SI), give more complex patterns of Km versus Pj than in the Mora granites (Fig. 3b). Although most of the PJ values are below 1.1, some sites (TO 18 of Layos granite, and TO 14 of migmatic granulites) show a rapid increase of Pj from Km above 300 x 10 SI. Such correlation between these magnetic parameters has been reported in rock types such as pyrrhotite-bearing schists (Hrouda 1987; Rochette 1987), magnetite-bearing granitoids (Damm 1988) and basaltic dykes and lava flows (Rochette et al. 1991; Margraves et al 1991; Bouchez 1997). In all cases, different amounts of the ferrimagnetic versus paramagnetic minerals to the whole susceptibility have been invoked. This mineralogical argument can be used to explain the magnetic pattern of Km versus Pj for site TO 14. The existence of different mineralogical phases contributing in different ways to the total anisotropy is linked to the mesosome and leucosome components of migmatites and thus could explain the high variability of Km (Fig. 3b). The mesosome, with a conspicuous granoblastic texture, includes cordierite as the main paramagnetic phase, quartz, K-feldspar, plagioclase, biotite, garnets and sillimanite as major minerals (Barbero 1992; Barbero & Villaseca 1992). Accessory spinel is the only ferrimagnetic phase. By contrast, the leucosomes essentially contain quartz, K-feldspar and plagioclase with large rounded garnets as paramagnetic minerals. Concerning site TO 18 from Layos granite (rocks where magnetite is absent, Barbero et al. 1995), the correlation pattern between Pj and Km is not explicable by means of different amounts of paramagnetic and ferrimagnetic minerals. This is a peraluminous granite with a hypidiomorphic texture (Barbero & Villaseca 1992). The major minerals are quartz, plagioclase, K-feldspar with cordierite (up to 30%) and biotite as paramagnetic minerals. Accessory

AMS IN ANATECTIC COMPLEXES

389

minerals include apatite, zircon, monazite, tourmaline, sillimanite, garnet, rutile and ferrimagnetic minerals such as iron sulphides and ilmenite. In this case (site located just over the shear zone) the sudden increase of Pj could be explained by a high intensity of deformation Similar to the Mora granite, samples in the footwall show predominantly oblate shapes (Fig. 4b) with the same range of variation (+0.5 > TJ > -0.5). Again the greatest difference is shown by site TO 18, which presents the more prolate shapes. No significant variations of Km and Pj with distance from the Toledo shear zone are observed (Fig. 7a, b). Excluding site TOM, whose magnetic parameters strongly reflect the combined contribution of both paramagnetic and ferrimagnetic minerals, site TO 18 is the only one showing a rapid increase of the magnetic parameters. Finally, no clear variation of TJ with the distance to the shear zone is observed (Fig. 7c). Considering all sites, a predominant neutral mean shape of the magnetic ellipsoids can be observed (7} « 0). Directional AMS data As can be seen in the cross sections (Fig. 2), a constant southward plunge of magnetic lineation is well defined in most of the sites (mainly in those of Layos granite). The direction of the magnetic lineation is consistent with the stretching direction from the previous structural data in the shear zone (Aparicio 1971; Martin Escorza & Lopez Martinez 1978; Hernandez Enrile 1991). The presence of a southward plunging magnetic lineation is even recorded in site TO 15, corresponding to orthogneisses, where a SSE magnetic lineation with no relation to the previous granitic foliation is defined. On the other hand, site TO 14, corresponding to granulitic migmatites, exhibits a magnetic fabric probably related to the anatectic process (e.g. magnetic lineation parallel to the migmatic fold axes). Sites from the Layos granite, which are located very near, or just over the fault, show the main magnetic lineation parallel to the mean stretching lineation measured in the mylonites (Fig. 8a) and, in addition, their magnetic foliation coincides with the tectonic foliation (Figs. 2a, b). Further (sites TO02, TO11, TO 16), the magnetic lineation is preserved with the same direction, but magnetic foliation is not yet parallel to the tectonic one (Fig. 2). Finally, for the most distant sites (e.g. TO04 and TO 10), the magnetic lineation, with a northward plunge, lies on the granitic foliation linked to a previous deformation phase. Considering all sites, the mean orientation of the magnetic lineation, however, contrasts with the pole of the girdle defined

Fig. 7. Plots of (a) degree of anisotropy (Pj), (b) bulk susceptibility (Km) and (c) symmetry of shape (7}) against distance (D) to the Toledo Shear Zone for the samples from the footwall block.

by the K^ axes, which is compatible with the folding of a previous NW-SE magnetic foliation (Fig. 8b). From these directional AMS data, it can thus be said that the magnetic fabric of

390

A. GIL-IMAZ & L. BARBERO

these samples probably records two different deformational processes. The first process explains the constant southward plunge magnetic lineation in most of the sites, and the other the dispersion of the K3 axes. Petrostructural interpretation of the AMS data The commonest application of the AMS technique is the petrostructural interpretation of the magnetic fabric of a rock from previous knowledge of the tectonic evolution of a region. In the present study, our results from the AMS data are referred to the general geodynamic scheme of the Montes de Toledo area during Late Variscan times, which provides a process of extensional detachment at greater structural depths involving a transition from ductile to brittle deformation (Hernandez Enrile 1991). According to this model, the present-day configuration of the Montes de Toledo area is the result of the exhumation of a northern domain (footwall) of high-grade metamorphic rocks (the anatectic complex) that is in contact with an epizonal batholith of granitic rocks (the Mora granite) located in the hanging wall of the extensional detachment (the Toledo shear zone) (Fig. 9).

Interpretation of AMS data in the hanging wall

Fig. 8. (a) Frequency stereogram of KI axes of Layos granite, (b) Frequency stereogram of K3 distributions axes for the same samples, (c) The whole of K\ and K^ axes (magnetic foliation). Squares on stereograms 10 (b) and (c) represent the poles of the corresponding girdles. Contour intervals are 1.5% in (a) & (b), and 1.2% in (c). See text for details.

The data expounded in the previous sections indicate that magmatic flow is the main process responsible for the magnetic fabric of the studied samples in the Mora granite. This is supported by the presence, in sites located far from the influence of the shear zone, of both a well-defined magnetic lineation and foliation parallel to the petrofabric indicators and magnetic ellipsoids with predominantly oblate shapes. Tectonic deformation linked to the Toledo shear zone is only detected in samples located near the fault (Fig. 9). For these samples, the magnetic lineation, which shows a southward dipping, lies on the mean penetrative extensional foliation measured in field. These planes (with dips between 45° and 60° to the south) are closed parallel to the main fault plane, so they must represent discrete C-planes of the shear zone. Thus, this magnetic lineation represents a true extensional magnetic fabric. Taking account of the fact that biotite crystals are the main magnetic phases of these rocks, the observed magnetic pattern can be explained in terms of

AMS IN ANATECTIC COMPLEXES

391

Fig. 9. Idealized scheme of the structural relationship between the epizonal batholith (Mora granites) and the anatectic complex (Layos granite, high-grade metamorphic rocks and granulitic migmatites). Stereoplots show the synthetic magnetic fabrics on each petrostructural domain.

'composite magnetic fabrics' from S-C mylonites. In fact, S-C mylonites display a better preferred crystallographic orientation of biotite crystals on C-planes than in S-planes that show a noticeable enrichment in quartz and K-feldspar (Aranguren et al. 1996). Concerning the samples corresponding to Cambrian schists of host rocks, its magnetic fabric (characterized by a magnetic lineation parallel to the intersection lineation between cleavage (D2) and bedding) is the tectonic origin. Interpretation of AMS data in the footwall In accordance with the results from AMS data of the samples from the footwall, two main deformational processes are suggested to explain the variation patterns of their magnetic fabric: (1) deformation linked to a large-scale ductile-brittle shear zone; and (2) simultaneous folding during the exhumation of the anatectic complex affecting the whole of the previously deformed footwall. In ductile-brittle shear zones, deformation patterns are characterized by a sudden increase of the intensity of deformation very close to discrete shear planes; and sudden variations in the petrofabric features (and in those related to

the magnetic fabric) are also very likely to occur. The first argument supporting this model is the variation pattern of Km and Pj with respect to the distance to the Toledo shear zone, which is characterized by a high constancy, except right over the fault (site TO 18) where both parameters rapidly increase. The shape parameter (7}) can be explained in the same way. Again, except for site TO 18, which is characterized by the most prolate shapes, a high constancy dominates in the ellipsoid shapes. Such elongated shapes are consistent with stretching deformation linked to shear deformation right near the fault. Directional data of the susceptibility ellipsoids also support the close relationship between the magnetic fabric and ductile deformation mechanism linked to the Toledo shear zone: a constant well-defined southward plunge magnetic lineation, parallel to the stretching lineation linked to the shear zone, in most of the analysed sites (in both those from Layos granite and the orthogneisses located far from the shear zone). On the other hand, the simultaneous folding during the exhumation of the anatectic complex is supported by both magnetic and structural data. First, it is supported by the scattering of the K3 axes along a NE-SW girdle whose pole is not coincident with the magnetic lineation, and the directional pattern shown by K\/Ki (magnetic foliation) characterized by the

392

A. GIL-IMAZ & L. BARBERO

presence of two different magnetic surfaces trending E-W and NE-SW respectively (Fig. 8c). Second, the pattern of relationship between the orientation of the magnetic lineation (which remains constant) and that of the granitic foliation in these rocks (which gradually varies from one site to another across the studied area) also supports the hypothesis of folding of a pre-extensional NW-SE foliation, probably linked to the D2 Variscan deformation phase, during the exhumation of the anatectic complex (Fig. 9).

the tectonic deformation is also reflected by a magnetic lineation compatible with an extensional top to south movement. This work was supported financially by the BTE200204168-C0301 (DOES) Project of the Spanish Ministry of Science and Education and RNM 326 Research Group of the Junta de Andalucia. We acknowledge the constructive review and comments by J. M. Tubia and an anonymous reviewer, which helped to improve the manuscript. We sincerely thank M. Garces for his suggestions and help with the AMS data processing at the Institute of Earth Sciences 'Jaume Almera', C.S.I.C. (Barcelona, Spain).

Conclusion AMS measurements in granitic rocks are used essentially to determine the magma flow dynamics and thus the emplacement mechanisms. Structural interpretations of the AMS data of this kind of rock are focused on establishing the relationship between the variation patterns of magnetic parameters and the tectonic structures at a regional scale. On the basis of such a methodology we interpret the magnetic fabric of some rock types of the Montes de Toledo area including epizonal granites, highgrade metamorphic rocks, granulitic migmatites and restite-rich granites. The study of the anisotropy of magnetic susceptibility (AMS) in the Montes de Toledo area reveals substantial differences in the magnetic fabric of rocks corresponding to both an anatectic complex and later epizonal intrusion. By means of the study of magnetic parameters (Km, Pj, TJ) and from the analysis of AMS directional data, we establish a close relationship between their magnetic and petro structural features. Rocks from the anatectic complex (footwall) mainly reflect stretching linked to the extension of a large-scale shear zone. The directional pattern of the KI axes corresponding to Layos granite is consistent with the folding of a previous NW-SE foliation during the exhumation of the anatectic complex. The patterns of relationship between the magnetic parameters (Km, Pj, TJ) and the distance to the Toledo shear zone suggest the ductile behaviour of the deformation linked to the fault. By contrast, the magnetic fabric of samples from the Mora granite (hanging wall) differs considerably from those of the footwall rocks. The most conspicuous feature is the presence of a magnetic lineation, which is consistent with the mean direction of K-feldspar phenocrysts. This fact indicates an origin probably linked to magmatic processes. Close to the shear zone

References ANDONAEGUI, P. 1990. Geoquimica y geocronologia de los granitoides del sur de Toledo. Tesis Doctoral. Universidad Complutense de Madrid, Espana. APARICIO, A. 1971. Estudio geologico del macizo cristalino de Toledo. Estudios Geologicos, 27, 369-414. ARANGUREN, A., CUEVAS, J. & TUBIA, J. M. 1996. Composite magnetic fabrics from S-C mylonites. Journal of Structural Geology, 18(7), 863-869. ARANGUREN, A., LARREA, F. J., CARRACEDO, M., CUEVAS, J. & TuBiA, J. M. 1997. The Los Pedroches Batholith (Southern Spain): polyphase interplay between shear zones in transtension and setting of granites. In: BOUCHEZ, J. L., HUTTON, D. H. W. & STEPHENS, W.E. (eds) Granite from Segregation of Melt to Emplacement Fabrics. Kluwer Academic Publishers, Dordrecht, 215-229. BARBERO, L. 1992. Plutonismo sin-orogenico en un area granulitica hercinica: El Complejo Anatectico de Toledo. Tesis Doctoral. Universidad Complutense. Madrid. BARBERO, L. 1995. Granulite-facies metamorphism in the anatectic complex of Toledo, Spain: late Hercynian tectonic evolution by crustal extension. Journal of the Geological Society, London, 152, 365-382. BARBERO, L. & VILLASECA, C. 1992. The Layos Granite, Hercynian Complex of Toledo (Spain): an example of parautochthonous restite-rich granite in a granulitic area. Transactions of the Royal Society of Edinburgh, Earth Science, 83, 127-138. BARBERO, L., VILLASECA, C. & ANDONAEGUI, P. 1990. On the origin of the gabbro-tonalite-monzogranite association from Toledo area (Hercynian Iberian belt). Schweizerische Mineralogische und Petrogaphische Mitteilungen 70, 209-221. BARBERO, L., VILLASECA, C., ROGERS, G. & BROWN, P. E. 1995. Geochemical and isotopic disequilibrium in crustal melting: An insight from the anatectic granitoids from Toledo, Spain. Journal of Geophysical Research, 100(B8), 15,745-15,765. BASCOU, J., RAPOSO, M., VAUCHEZ, A. & EGYDIO, S-M. 2002. Titanohematite lattice-preferred orientation and magnetic anisotropy in high-temperature

AMS IN ANATECTIC COMPLEXES mylonites. Earth and Planetary Science Letters, 198(1-2), 77-92. BOUCHEZ, J. L. 1997. Granite is never isotropic: an introduction to AMS studies of granitic rocks. In: BOUCHEZ, J. L., BUTTON, D. H. W. & STEPHENS, W.E. (eds) Granite from Segregation of Melt to Emplacement Fabrics. Kluwer Academic Publishers, Dordrecht, 95-112. BOUCHEZ, J. L., GLEIZES, G., DJOUADI, T. & ROCHETTE, P. 1990. Microstructure and magnetic susceptibility applied to emplacement kinematics of granites: the example of the Foix pluton (French Pyrenees). Tectonophysics, 184, 157-171. DAMM, V. 1988. Gesteinsmagnetische Abisotropien in Magmatiten und deren strukturgeologische Bedeutung. Zeitschriftfur Geologische Wissenschaften, 16, 739-752. DIEZ-BALDA, M., VEGAS, R. & GONZALEZ LODEIRO, F. 1991. Central-Iberian Zone: Structure. In: DALLMEYER, R. D. & MARTINEZ-GARCIA, E. (eds) Pre-Mesozoic Geology of Iberia. Springer-Verlag, Berlin, 172-188. DOBLAS, M. & MARTINEZ, J. 1988. Detachment faulting and late Paleozoic epithermal Ag-base-metal mineralization in the Spanish central system. Geology, 16, 800-803. FERRE, E., GLEIZES, G., DJOUADI, T., BOUCHEZ, J. L. & UGODULUNWA, F. 1997. Drainage and emplacement of magmas along an inclined transcurrent shear zone: petrophysical evidence from a granite-charnockite pluton (Rahama, Nigeria). In: BOUCHEZ, J. L., HUTTON, D. H. W. & STEPHENS, W.E. (eds) Granite from Segregation of Melt to Emplacement Fabrics. Kluwer Academic Publishers, Dordrecht, 253-273. GLEIZES, G., NEDELEC, A., BOUCHEZ, J-L., AUTRAN, A. & ROCHETTE, P. 1993. Magnetic susceptibility of the Mont Louis-Andorra ilmenite-type granite (Pyrenees): a new tool for the petrographic characterization and regional mapping of zoned granite plutons. Journal of Geophysical Research, 98, 4317-4331. GRAHAM, J. W. 1954. Magnetic anisotropy, an unexploited petrofabric element. Bulletin of the Geological Society of America, 65, 1257-1258. GUILLET, P., BOUCHEZ, J. L. & WAGNER, J. J. 1983. Anisotropy of magnetic susceptibility and magnetic structures in the Guerande Massif (France). Tectonics, 2, 419-429. HARGRAVES, R. B., JOHNSON, D. & CHAN, C. Y. 1991. Distribution anisotropy: the cause of AMS in igneous rocks? Geophysical Research Letters, 18, 2193-2196. HELLER, F. 1973. Magnetic anisotropy of granitic rocks of the Bergell Massif (Switzerland). Earth and Planetary Science Letters, 20, 180-188. HERNANDEZ ENRILE, J. L. 1991. Extensional tectonics of the Toledo ductile-brittle shear zone, central Iberian Massif. Tectonophysics, 191, 311-324. HROUDA, F. 1987. Mathematical model relationship between the paramagnetic anisotropy and strain in slates. Tectonophysics, 142, 323-327. HROUDA, F. & CHLUPACOVA, M. 1980. The magnetic fabric in the Nasavrky Massif. Casupis Pro Mineralogii a Geolgii., 25, 17-27.

393

HROUDA, F. & JELINEK, V. 1990. Resolution of ferrimagnetic and paramagnetic anisotropies in rocks, using combined low-field and high-field measurements. Geophysical Journal International 103, 75-84. HROUDA, F., JANAK, F., REJL, L. & WEISS, J. 1971. The use of magnetic susceptibility anisotropy for stimating the ferromagnetic mineral fabrics of metamorphic rocks. Geologische Rundschau, 60, 1124-42. JELINEK, V. 1981. Characterization of the magnetic fabric of rocks. Tectonophysics, 79, 63-67. LEBLANC, D., GLEIZES, G., LESPINASSE, P., OLIVIER, PH. & BOUCHEZ, J. L. 1994. The Maladeta granite polydiapir, Spanish Pyrenees: a detailed magneto-structural study. Journal of Structural Geology, 16(2), 223-235. MACAYA, J., GONZALEZ-LODEIRO, F., MARTINEZCATALAN, J. R. & ALVAREZ, F. 1991. Continuous deformation,ductile thrusting and backfolding of cover and basement in the Sierra de Guadarrama, Hercynian Orogen of Central Spain. Tectonophysics, 191, 291-309. MARTIN ESCORZA, C. & LOPEZ MARTINEZ, J. 1978. Analisis mesoestructural en la unidad migmatitica de Toledo. Estudios Geologicos, 34, 33-43. OLIVIER, P., AMEGLIO, L., RICHEN, H. & VADEBOIN, F. 1999. Emplacement of the Aya granitic pluton (Basque Pyrenees) in a dextral transcurrent regime inferred from a combined magnetostructural and gravimetric study. Journal of the Geological Society, London, 156, 991-1002. RAPOSO, M., D'AGRELLA, F. & SIQUEIRA, R. 2003. The effect of magnetic anisotropy on paleomagnetic directions in high-grade metamorphic rocks from the Juiz de Fora Complex. SE Brazil. Earth and Planetary Science Letters, 209(1-2), 131-147. ROCHETTE, P. 1987. Magnetic susceptibility of the rock matrix related to magnetic fabric studies. Journal of Structural Geology, 9, 1015-1020. ROCHETTE, P., JACKSON, M. & AUBOURG, C. 1992. Rock magnetism and the interpretation of anisotropy of magnetic susceptibility. Rev. Geophys., 30(3), 209-226. ROCHETTE, P., JENATTON, L., DUPUY, C., BOUDIER, F., REUBER, I. 1991. Diabase dykes emplacement in the Oman Ophiolite: a magnetic fabric study with reference to geochemistry. In: PETERS, T., NICOLAS, A. & COLEMAN, R. (eds) Ophiolite Gneisses and Evolution of the Oceanic Lithosphere. Kluwer Academic Publishers, Dordrecht, 55-82. ROMAN-BERDIEL, T., PUEYO MORER, E. L. & CASAS SAINZ, A. M. 1995. Granite emplacement during contemporary shortening and normal faulting: structural and magnetic study of the Veiga Massif (NW Spain). Journal of Structural Geology, 17(12), 1689-1706. SANTA TERESA, I., CARBO, A., CAPOTE, R. & CASQUET, C. 1983. Geometria en profundidad del Granito de Orgaz en base a datos gravimetricos. Studia Geologica Salamauticensia, 18, 237-260. TARLING, D. H. & HROUDA, F. 1993. The Magnetic Anisotropy of Rocks. Chapman & Hall, London.

394

A. GIL-IMAZ & L. BARBERO

VILLASECA, C. & BARBERO, L. 1994. Chemical variability of Al-Ti-Fe-Mg minerals in peraluminous granitoid rocks from Central Spain. European Journal of Mineralogy, 6, 691-710. VILLASECA, C., BARBERO, L. & ANDONAEGUI, P. 1993. Diversity of peraluminous granites types in the Hercynian Central zone of Spain. Terra Abstracts, 5, 433-434.

VILLASECA, C. & BARBERO, L. & ROGERS, G. 1998. Crustal origin of Hercynian peraluminous granitic batholiths of Central Spain: petrological, geochemical and isotopic (Sr, Nd) constrains. Lithos, 43, 55-79.

Statistical significance of magnetic fabric data in studies of paramagnetic granites E. L. PUEYO,1 2 M. T. ROMAN-BERDIEL,2 J. L. BOUCHEZ,1 A. M. CASAS2 & J. C. LARRASOANA3 1

Laboraloire des Mecanismes de Transfert en Geologic, Universite Paul-Sabatier, Toulouse, France 2 Geologia Estructural, Geodindmica Internet. Universidad de Zaragoza, c\ Pedro Cerbuna 12, 50009-Zaragoza, Spain (e-mail: [email protected]) Labor atorio de Paleomagnetismo, ICT 'Jaume Aimer a' CSIC, Barcelona, Spain Abstract: The low anisotropies of paramagnetic granites, due to magnetocrystalline anisotropy, require a statistical treatment of the anisotropy of magnetic susceptibility (AMS) data when systematic fabric studies are performed. Absence of statistical information on these data makes evaluation of their quality difficult. The statistical significance of magnetic fabric in granites is evaluated in this paper. Jelinek's elliptical confidence angles for the three principal susceptibility axes (E13, E ]2 , £23) of a specimen are used as markers of the quality of the AMS data. Comparing these markers at sample-, site- and massif-scale with the mean AMS axes that result from spherical statistical models helps clarify the reliability of the AMS data. This analysis is presented in detail for the plutons of Veiga and Trives (Spain). It is then applied to seven other massifs from the Pyrenees. We propose the following guides: (1) fabrics with E13 between 10° and 20° tend to isotropy; the directional data and the shape parameter should be considered with great care; (2) lineation is not reliable when E12 > 25°, i.e. when Km3LX is almost the same as K-mi\ (3) similarly, foliation is considered as not reliable when E23 > 25°, i.e. ^min does not easily differentiate from Kint. Errors attached to the mean Kmax and Km^n axes should always be produced, thus allowing further interpretation. In Trives and Veiga, 'perfect' triaxiality cannot be automatically assumed since foliation and lineation could be defined simultaneously in only 53% of the cases. Finally, a minimum of three cores (9 specimens) per site would considerably increase the proportion of reliable orientation data.

The anisotropy of magnetic susceptibility (AMS) or magnetic fabric technique has been applied to granites since the early 1960s (Balsey & Buddington 1960; Khan 1962; Birch 1979; Van derVoo& Klootwijk 1972). Systematic AMS studies, aimed at unravelling the internal structure and the emplacement mode of granite plutons started then (Duffa 1975; Ellwood & Whitney 1980; Guillet et al 1983; Hrouda & Lanza 1989) and have been successfully applied during the last two decades, representing an inflexion point in the understanding of igneous bodies. The classical methods to constrain the mineral preferred orientation were either too time consuming or impossible to apply in cases of apparently isotropic patterns (Bouchez 1997). Associated with detailed structural and microstructural maps (i.e. classical deformation markers), this fast, non-destructive and cheap technique is unavoidable to constrain deformation patterns in granite plutons, the more than the instruments (e.g. KLY susceptometers, AGICO Ltd) have greatly improved. This versatile technique can now

tackle a wide spectrum of problems in the earth sciences. Paramagnetic granites correspond to magnetite-free granites that belong to the ilmenite series of Ishihara (1977). They often correspond to granites that formed from continental crustal protolith (Wenner 1981; Ellwood & Wenner 1981). Their low bulk susceptibility (K), between 10~5 and a few 10"4SI, is mainly carried by biotite and possibly amphibole (paramagnetic minerals). In these rocks, K is a measurement of the rock iron content, hence is used for surface mapping of the rock types (Gleizes et al. 1993). The magnetocrystalline anisotropy of the carriers is at the origin of the susceptibility anisotropy of these granites. The magnetic fabric patterns are therefore interpreted in terms of lattice; hence of shape fabrics of the carriers (Borradaile & Henry 1997; Bouchez 1997, 2000). The long and short axes of the AMS ellipsoid (^max and ^min) are equated with the magmatic lineation (long axis of mineral fabric) and foliation pole (normal to the plane

From: MARTIN-HERNANDEZ, F., LUNEBURG, C. M., AUBOURG, C. & JACKSON, M. (eds) 2004. Magnetic Fabric: Methods and Applications. Geological Society, London, Special Publications, 238, 395-420. 0305-8719/04/ $15.00 © The Geological Society of London 2004.

396

E. L. PUEYO ET AL.

of mineral fabric) respectively. For the foliation, this results from the parallelism between the pole of the basal plane of the phyllosilicates (biotite, chlorite and iron-bearing muscovite) and the minimum susceptibility axis (Borradaile & Werner 1994; Martin-Hernandez & Hirt, 2003). The significance of the magnetic lineation is not straightforward since the micas have Kmax equal or very close to K-mi (Zapletal 1990; Martin-Hernandez & Hirt, 2003). However, KmSLX is the vorticity axis of the fabric, or zone axis, i.e. the mean axis around which the basal planes of the micas rotate (pole to the girdle distribution of ^min), and this axis is argued to represent the stretch direction of the magma that was deforming (see Bouchez 1997). The anisotropy degree (P = Kmax/Kmin) of these rocks cannot surpass the 'intrinsic' anisotropy of the carriers, which is of the order of 1.3 for the micas. On the other hand, the anisotropy magnitude is limited by the arrangement (fabric) of the magnetic carriers, which in granites is so weak that total anisotropy hardly exceeds 1.1 (10%) and is usually less than 1.04 (4%). Eventually the shape (prolate, oblate, triaxial) and eccentricity (anisotropy) of the AMS ellipsoid are interpreted in terms of the strain regime undergone by the magma. Despite the non-direct correlation between the magnetic fabric, the mineral fabric and the strain ellipsoid (Siegesmund et al. 1995), the study of the relative changes in the magnetic fabric shape-parameter helps to characterize variations in strain regimes at the massif scale, a goal difficult to achieve by other means. Ferromagnetic, or magnetite-bearing granites (magnetite series of Ishihara 1977) have susceptibilities often larger than 10~3 SI, due to the addition of magnetite. This melange of magnetic minerals having different 'intrinsic' susceptibilities prevents K being used as an iron content indicator. These granites are therefore composed of at least two subfabrics (micas and magnetite). For the directional aspect, a good correspondence is generally observed between the AMS axes and the preferred orientation axes of biotite and magnetite (Launeau 1990; Arbaret et al 1995; Archanjo et al 1995; Gregoire et al. 1998), and long axes of enclaves (Archanjo et aL 1995). Concerning the scalar aspect of the fabric, since two different fabricforming mechanisms are involved, namely magnetocrystalline anisotropy of the iron-bearing silicates and shape anisotropy of the magnetite grains (Stacey and Banerjee 1974; Gregoire et al. 1998), the anisotropy degree of the ferromagnetic granites is highly variable and often larger than 1.5 (Bouchez 1997).

We choose to examine the case of magnetitefree granites (paramagnetic), which are mineralogically simpler. The AMS data, however, may not necessarily be straightforward to interpret even in this case. The weak anisotropies of these granites, rarely exceeding 5%, require their AMS data to be handled with care, especially when large data sets are concerned, even if very homogeneous fabric patterns are currently described at different scales (Olivier et al. 1997). The AMS directional data axes require a spherical-statistics treatment that agrees with the anisotropy of the spatial scatter, the shape of the magnetic ellipsoid (Jezek et al. 2000) and the errors associated with the measurement (Jelinek 1981). Except for a few studies (Ernst & Pearce 1989; RomanBerdiel et al. 1995; Trindade et al. 1999; Yenes et al. 1999) in which the significance of the directional data is assessed, the fabric is usually considered as 'perfectly' triaxial. This means that, at any site of measurement, the magnetic foliation and the magnetic lineation are considered as 'perfectly' defined. This supposes that both the maximum and minimum axes of the fabric tensor have a uniaxial distribution (Fisher 1953). We contend that it may be incorrect. Since no error is attached to the averages of both the directional and scalar data, no further evaluation of their quality is possible. In addition, the number of analysed specimens at site scale should better agree with the statistical distribution needs, or at least with a given desired resolution. In this paper, the statistical significance of the AMS directional and scalar data is evaluated in the case of the paramagnetic granites of Veiga and Trives (Roman-Berdiel et al. 1995, 1998). A full summary of the AMS data and their statistical mean values derived from different methods is provided in Appendix 1. Seven other plutons from the Pyrenees will also support our deductions. Classes of quality for the AMS variables (anisotropy, lineation and foliation) will be defined from the errors attached to them. The low quality data (noise) should be filtered before examination of their geological significance. However, this is beyond the scope of the paper. Trives and Veiga granite massifs These massifs intrude the northern Iberian Hercynian belt (Fig. la) in an area belonging to the Olio de Sapo antiform (Barrera Morate et al. 1989; Roman-Berdiel et al. 1995; Aranguren et al. 1996; Roman-Berdiel et al. 1998; Vegas et al. 2001). The main structures of the area

STATISTICAL SIGNIFICANCE OF AMS DATA IN GRANITES

397

Fig. 1. (a) Geological sketch map (modified after Barrera Morate et al. 1989) of the area enclosing the Trives and Veiga massifs, and location in the Hercynian massif.

were formed during the first Hercynian phase marked by NE-verging cleavage-related folds (Parga-Pondal 1963; Gonzalez-Lodeiro & Iglesias 1977; Martinez Catalan 1985; Bastida et al. 1986). These were later modified into upright to NE-verging folds, with associated crenulation cleavage. In the central part of the Olio de Sapo anticlinorium, a medium-grade regional metamorphism, marked by staurolite and sillimanite, is present. Both plutons cut across the NW-SE trending isogrades, indicating that their emplacement post-dates the regional metamorphism, and develop metamorphic aureoles in the country rocks. These massifs have been subjected to a detailed AMS study, which along with an extensive geological setting, can be found in Roman-Berdiel et al. (1995, 1998) and Vegas et al. (2001). The Trives massif is an elongated body, 40 km2 in area, with an E-W trend to the west and NWSE trend to the east (Fig. la). It is a granodiorite in composition, with local variations to tonalite, except to the west where it is a two-micas granite. Its grain size ranges from 2 to 5cm for the Kfeldspar phenocrysts to 1-5 mm for the matrix. Bulk susceptibility site-average (Fig. Ib and

Appendix 1) is between 16 and 467 x 10 6 SI, falling into the paramagnetic field of Gleizes et al. (1993). Samples for the AMS measurements were collected from 51 stations, evenly distributed throughout the massif. At each site (about 2 m2), an average of 8 standard cores (25 mm in diameter) were collected, and 6 to 13 (9 in average) standard specimens (21 mm in height) were measured (Fig. Ic). The foliations and lineations of magmatic origin are characterized by low anisotropy degrees. A number of samples show solid-state deformation features that tend to increase the anisotropy degree (Fig. Id), with magnetic foliations and lineations sub-parallel to the macroscopic C-planes and stretching lineations. The Veiga massif is a 20 km by 8 km, E-W elongated body (Fig. la). It is also a granodiorite, with local variations to syenogranite and monzogranite, from which 42 evenly distributed sites were collected. Bulk susceptibility siteaverage (Fig. Ib and Appendix 1) ranges from 16 to 271 x 10"6 SI, again falling into the paramagnetic field. An average of 5 cores was obtained from each site, and 2 to 9 (6 in average) specimens were measured (Fig. Ic). The highest

398

E. L. PUEYO ET AL.

Fig. 1. (b) Histograms of the bulk susceptibility magnitudes of the Trives and Veiga massifs, (c) Distribution of number of sites against number of samples per site, (d) T-P' (shape versus anisotropy) diagram for site means, (e) Magnetic mineralogy analyses. High field versus low field essays in the Veiga massif from MPMS magnetometer measurements, (f) Remanent magnetization evolution from liquid nitrogen to room temperature.

degree of magnetic anisotropy is observed close to the base and top of the intrusion. Presence of C and S structures along the western border indicates that the cooling of the granite was coeval with the movement of the Chandoiro fault. Emplacement of this massif is concluded to be coeval with a NNE-SSW directed shortening, perpendicularly to the late fold axes (Roman-Berdiel et al 1995). The ferromagnetic contribution to susceptibility is negligible in both granites (Figs, le, f), as primarily attested by their low susceptibility magnitude (Fig. Ib). For Veiga, this is complemented by low-field versus high-field susceptibility measurements (MPMS cryogenic magnetometer) that allow the calculation Kferro below 5% (Fig. le). In Trives, remanence saturation essays at low temperature (sweep heating between 70 and 293 K) performed at

high fields (from 0.5 to 3T) do not show the Verwey transition that would attest to the presence of magnetite (Fig. If). The dominant paramagnetic behaviour of Veiga and Trives is therefore confirmed. Magnetic anisotropy and shape reliability The total eccentricity, or anisotropy degree, of the ellipsoid that characterizes the susceptibility tensor is simply the ratio, or P parameter, between the maximum and minimum axes. We use the more sophisticated parameter of Jelinek (1977) that takes Kmi into account:

STATISTICAL SIGNIFICANCE OF AMS DATA IN GRANITES

where / / j ^ l n A ^ , ^2 = lnAT int and Ms — In ^min and //m = (^ + /z2 + )U 3 )/3. These strength parameters have something to do with the degree of strain recorded by the magma. Typical values of P' in (paramagnetic) granites having magmatic microstructures range from 1.01 to 1.10 (Bouchez 2000). Higher values can be reached if the mineral fabric is overprinted by solid-state deformation. The quality of the AMS ellipsoid can be assessed at the sample scale by the threedimensional anisotropy F-test of Jelinek (1977), or by its mathematical equivalent confidence angle E31, which will be used to quantify the quality of a specimen. Specimens with E31 > 20° (or F < 4) are considered as isotropic at the 95% significance level, hence have to be rejected. They reflect the difficulty of defining the tensor when the eccentricity is extremely low. Accepting a small increase in the significance level, specimens with E31 > 10° (F < 10) might be described as pseudo-isotropic, hence have to be treated with care. The scatter of points in orientation diagrams attests to the lack of reliability of pseudo-isotropic samples. It is even more realistically revealed when the AMS axes are plotted along with their respective confidence angles (Fig. 2b). Since E31 against P; has an exponential trend (Fig. 2a), P' around 1.015 can be considered as separating the isotropic specimens from the pseudo-isotropic ones. However, the limit value of P' may vary from one granite to another. Finally, specimens with E31 < 10° (F > 10) will be considered as significant. At site scale (Fig. 2c), the low anisotropy values usually correspond to poorly defined principal susceptibility axes (Kmax9 ^min) and shape parameters, and display high E31 values. A poor eccentricity will obviously give weakly characterized directions whatever the statistical model in use. In Fisher's (1953) uniaxial and symmetrical statistics, a95 will range between 25° and 60°. However, Fisher's statistics is not recommended since the distributions are not necessarily symmetrical (Ernst & Pearce 1989; Tarling & Hrouda 1993). A great dispersion of the site mean standard errors (eE12, eE23, eE31) of the individual confidence angles will be observed when plotted against P' (Fig. 2e). The same remark is valid for the error attached to the shape parameter of Jelinek (1981),

399

high values of the standard error of T (Fig. 2d) as frequently observed in low anisotropic sites. We therefore propose the following sorting of qualities into Class I: reliable fabric (E31 < 10°); Class II: pseudo-isotropic fabric (20° > E31 > 10°) - directional data from these fabrics will probably show poorly defined means; and Class III: isotropic, non-reliable, fabric (E31 > 20°) - scattered orientations and fabric parameters will be found. At massif scale, the mean T values of sites having E31 > 10° should be discarded or considered with their associated errors, particularly if geostatistical methods were used to draw up the contour maps. Variability of the shape parameter is exemplified in Veiga (Fig. 3), where rejection of pseudo-isotropic and isotropic sites (classes II and III) simplifies the contours of the shape value in map-view and removes artificial and nonsense gradients of this parameter. This new map (Fig. 3b) could influence our interpretation of the emplacement mode. In Veiga, the general pattern of T isovalues is observed to be simpler parallel to the pluton boundaries, in contrast to the initial interpretation. Magnetic lineation reliability

Magnetic lineation (KmSLX) in paramagnetic granites is an apparent lineation due to the girdle distribution of the phyllosilicate c axes perpendicular to KmSiX. This has been proven by X-ray goniometry and other mineral fabric techniques (Liineburg et al. 1999). Provided that no inverse magnetic fabric carrier is present, Kmax has a very important kinematic significance since it points to the stretch direction of the deforming magma, at least for the time window during which the mineral fabric is recorded in the rock. Evaluation of the statistical significance of Kmax is therefore important. Evaluation at specimen scale can be done by considering the test of rotational anisotropy, F12 of Jelinek (1977), or its equivalent confidence angleE 12 - Specimens withF 12 < 5(E 12 > 25°) are considered as isotropic in the foliation plane at 95% confidence level. ^Tmax and K[nt have a rotational symmetry around Kmin and represent pure oblate ellipsoids (^max = K-mt). Thus, E12 will be used to reject non-significant lineations. Accepting some increase in the significance level, specimens with F < 8 (or E12 > 20°) might be considered as pseudo-symmetric (^max ~ K-mi}. where /^ = mAT max , //2 — hi^int and //3 = A low value of the parameter describing the lineaIn^min- Hence, the lower the strength of the tion (L = ^ m ax/^int) relates to the difficulty of fabric, the harder it will be to define an accurate defining Km.dx with respect to Kmt, hence a high shape of the ellipsoid (Fig. 2b). This agrees with value of E12 is expected. In Veiga and Trives

400

E. L. PUEYO ET AL.

Fig. 2. Reliability of the anisotropy values, (a) Jelinek's sample error between the maximum and minimum susceptibility axes (E31) against the individual (sample) anisotropy (P'). Non-reliable areas considering proposed critical values, (b) Projection of two low anisotropic sites. Confidence cones (95%) for the maximum (open square) and minimum (black circles) axes against the intermediate axis (E12 and E23 respectively; thin lines) in every individual sample are also plotted. The resultant Fisherian mean and its confidence cone («95; thick lines) for both axes are also shown. Note that the variability of individual T values can exceed 1 unit within the site scale, or the standard errors of T surpass 25%. (c) Mean anisotropy (site scale) against the a95 from the maximum and minimum axes, (d) P' vs. the standard error of the shape parameter (eT). (e) Corrected anisotropy degree vs. the standard error of the Jelinek's (1977) confidence angles.

STATISTICAL SIGNIFICANCE OF AMS DATA IN GRANITES

401

Fig. 3. Reliability of anisotropy (P') and shape (T) at (Veiga) massif scale. Maps after the filtering of the pseudo-isotropic sites. The removal (a) of low anisotropy sites (responsible of forced and strong gradients of the shape parameter) results in a significant difference (b), especially in the north of the massif, which shows a smoother pattern, (c) Original maps previously of Roman-Berdiel et al. (1995).

(Fig. 4a), L < 1.010 is associated with E12 > 20° and up to 70°. This also holds for T values larger than 0.6. The exact values of T and L depend on the case under study. Above E12 > 20°, the spatial definition (trend and plunge) of the magnetic lineation will lose its reliability, the fabric being purely planar.

At site scale, E12 vs. L (Fig. 4c) and T vs. E12 (Fig. 4d) have similar but smoother trends than at specimen scale (Figs. 4a, b). Again, the standard errors increase toward the non-reliable zone (Figs. 4c, d) and the more oblate the fabric, the greater the difficulty of differentiating Km.dx from ^int. The fabric orientation data at site

402

E. L. PUEYO ET AL.

Fig. 4. Reliability of magnetic lineation. (a) At sample scale: individual sample lineation (L = Km^/K-mi) versus the (95%) confidence angle between Kmax and K-mi axes (Ei 2 ). (b) At sample scale: E12 against the shape parameter (T). (c) At site scale: site mean lineation (L = Kmax/Kint) versus site mean E12 angles (and respective standard errors), (d) At site scale: E12 against the shape parameter (T). (e) At site scale: Statistical models. Cluster parameter (Lw) from Woodcock (1977) versus E12 mean (and error), (f) At site scale: long axis from the simulation ellipse (ft; Lienert 1991) versus magnetic lineation (L). (g) At site scale: Fishers' concentration parameter («) against the fabric confidence angle E 12 . Three quality regions have been represented in all graphs. Class I (reliable): white; Class II (suspicious): light grey; and Class III (non-reliable): dark grey. scale, namely means and errors of Kmax, K-mi and Kmin, are obtained by fitting the individual measurements to a given probability distribution. Although AMS is a tensor, for the sake of simplicity in this statistical study, the individual

susceptibility axes will be treated as vectors or lines to obtain site averages. Among the several statistical models to analyse spherical data sets, compiled by Fisher et al. (1987), the three following approaches are performed to check the

STATISTICAL SIGNIFICANCE OF AMS DATA IN GRANITES

reliability boundaries for magnetic lineations (Appendix 1). (i) The approach of Scheidegger (1965), Watson (1965) and Bingham (1964, 1974) calculates the orientation tensor as the matrix of the sum of cross products of the direction cosines of a given set of unitary lines having symmetrical bipolar or girdle distributions. The orientation tensor is characterized by its eigenvectors. Shape (clustered or girdled) and strength of this distribution is also analysed conformably to Woodcock (1977) and Woodcock & Naylor (1983) by using the eigenvectors moduli (normalized eigenvalues: Si + #2 + £3 = 1). Woodcock builds a pseudo-Flinn (1962) diagram by means of the ratios between the normalized eigenvalues: Lw = Si/S2, Fw = $2/83. The distribution is a cluster if Lw dominates Fw (Si > S2 and S3), otherwise it is a girdle (Si « S2 and S3 « 0). Pw (= Si/S^) is a measurement of the quality of the preferred orientation, the larger it is, the stronger the orientation. As expected, in Trives and Veiga (Fig. 4e), low E12 values correlate with high Lw values, i.e. strong clusters of Kmax, while oblate fabrics display high E12 angles and show poor cluster indexes L w . (ii) The method of Lienert (1991), based on Monte Carlo simulation of errors, calculates the elliptical cones of confidence of a population of lines using angles fa (maximum) measured on the great circle along the maximum scatter, and /32 measured along the perpendicular direction. As expected, fa of a population of Kmax varies inversely to the lineation parameter (L) using Lienert's method (Fig. 4f): the stronger the lineation, the lower the corresponding fa. (hi) Fisher's approach (1953) is used to fit vectors having radial distributions although data sets with non-axial symmetries are frequently observed in AMS studies. The concentration parameter (K) exhibits a clear correlation with the (^max - Kini) tolerance angle (E12): the lower E12, the larger K will be (Fig. 4g). Whatever the chosen statistics, E12 is selected as the most important variable since it does not depend on the rock peculiarities but only on the statistical significance at the measurements level. Three quality classes are defined: Class I: E12 < 20° (F12 > 8); lineations are always reliable. In the examples of Veiga and Trives, E12 < 20° correlates with L > 1.014,

403

the standard error of E12 (eE12) is less than 5°, and T is less than —0.65. Spherical statistical approaches of mean-^max give errors smaller than 40° (/3{; Lienert), smaller than 25° (a95; Fisher), or larger than 10 (K; Fisher) and larger than 5 (Lw = Si/S2; Woodcock). Class II: 25° > E12 > 20° (F12 ranges from 5 to 8); lineations of low quality have to be examined with caution. In Veiga and Trives this correlates with L between 1.014 and 1.007 (eE12 ranges from 5° to 10°), and T from -0.65 to -0.75. Spherical statistics of the mean-Xjnax have the following error ranges: 40-80° (fa; Lienert), 25-35° (a95; Fisher), 510 (K; Fisher) and 3-5 (Lw = S{/S2; Woodcock). Class III: E12 > 25° (F12 < 5): the lineations are unreliable since they cannot be distinguished statistically (pure oblate AMS ellipsoids). In Veiga and Trives, E12 > 30° correlates with L < 1.007 (eE12 > 10°), and T is larger than —0.75. The errors of the spherical statistics of the mea.n-KmSiX are larger than 80° (fa; Lienert), larger than 35° (a95; Fisher), smaller than 5 (K; Fisher) and smaller than 3 (Lw = Si/S2; Woodcock). In addition, classes II and III display intra-site maximum angular departures of Km&x in between 65° and 90°. In our case of Veiga and Trives, the percentages of samples and sites belonging to the different classes are shown in Table 1. At massif scale, since the magnetic lineation is almost constant in both our examples (Roman-Berdiel et aL 2001) the orientation tensors can be compared before and after removing the lineations of classes II and III, using the Woodcock diagram. The whole set of specimens' magnetic lineations (Fig. 5: All) shows that Trives is more anisotropic and better clustered than Veiga. After removing lineations from classes II and III, both data sets become more anisotropic and better clustered. By contrast, the remaining set (Fig. 5: Rest) has Table 1. Percentages of samples and sites in the Veiga and Trives Massifs belonging to the three established quality classes Total

Class I

Class II

Class III

Veiga Samples Sites

239 42

170(71%) 30(71%)

33 (14%) 7 (17%)

36(15%) 5 (12%)

Trives Samples Sites

443 51

378 (85%) 39 (76%)

30 (7%) 35 (8%) 8 (16%) 4 (8%)

404

E. L. PUEYO ET AL.

Fig. 5. Strength and shape of lineation distributions (Woodcock diagram). The whole dataset has been divided in two subsets using E12 less or larger than 20° as a cutting criterion between classes II and I. The analyses have been done for individual lineations (specimens). Equal-area projection with the Kamb contoured diagram (contours at 2cr) only for sample lineation data.

a lower anisotropy within the girdle field. These results illustrate the importance of the noise that low-quality lineations imprint to the data set. Noise due to low-quality data is even more critical when lineation trajectory maps are performed at massif scale. Individual lineations

should be represented along with their respective confidence angles (Fig. 6a). Their fa Lienert angles, simulation ellipses (Constable & Tauxe 1990), or eigenvectors (Bingham 1974) should be tabulated rather than using the a95 angles of Fisher, as already pointed out by Tarling &

STATISTICAL SIGNIFICANCE OF AMS DATA IN GRANITES

405

Fig. 6. Reliability and quality of lineations at the massif scale, (a) Map of magnetic lineations in which the Lienert confidence semi-cone angle (fa /2) has been plotted. Class II and III sites have been coloured in grey, (b) Lineation trajectory map of Roman-Berdiel et al. (1995). (c) Reinterpreted trajectory map considering the spatial scatter of individual sites and their significance (class), (d) Histogram of number of high and low quality lineations against the trend of Kmax.

Hrouda (1993). Accordingly, anomalous directions with low quality will be easily identified, and will not cause the lineation patterns to bend. This is illustrated in Figures 6c and 6b where the bending of the lineation trajectories in the southern part of Veiga disappears when lineation qualities are taken into account. Classes II and III sites exhibit a wider range of orientations and their mean falls far from that of class I (Fig. 6d). Seven granite plutons from the Pyrenees (Fig. 7), namely Aya (Olivier et al. 1999), Bassies (Gleizes et al. 1991), Bielsa (Roman-Berdiel et al. 2004), Cauterets-Panticosa (Gleizes et al. 1998), Maladeta (Leblanc et al. 1994) and Trois-Seigneurs (Leblanc et al. 1996), are used to check the validity of the correlation between the proposed lineation confidence angles of

Jelinek and the fabric parameters (T vs. E12). At one extreme, the Trois-Seigneurs massif does not display any single low-quality sample, and Bassies is very close to this situation. From the latter massif it can be deduced that T around 0.8 correlates with the boundary between classes II and III. Both granites exhibit welldefined trends between T and E12. In a second group, containing Panticosa, Cauterets, Maladeta and Trives, T values, from 0.75 to 0.65, have a larger scatter than has the first group. Aya, Bielsa, and Veiga to a lesser extent, represent the other extreme with large scatters of both variables. Classes II and III are separated by much lower T values (~0.5). Similar features could be found if the Jelinek errors were plotted against the other tensorial (L, P') or directional variables (axes means). Figure 7 illustrates that

406

E. L. PUEYO ET AL.

Fig. 7. Geological sketch map showing the location of magnetic bodies in the Pyrenean domain (Axial Zone). The Ei2 versus T graphic (specimen information) for 7 massifs are also presented alphabetically.

STATISTICAL SIGNIFICANCE OF AMS DATA IN GRANITES

the relationship between the confidence angles and the fabric parameters change considerably from one massif to the other, hence they must be checked to filter the unreliable data. Magnetic foliation reliability Magnetic foliation is the plane perpendicular to the minimum susceptibility axis. In a biotitebearing granite sample, it represents the mean orientation of biotite basal planes usually related to the rock fabric. Using the same approach as for the lineations, we briefly examine the statistical significance of foliation. Again, the quality of the magnetic foliation for an individual specimen can be assessed by considering the test of rotational anisotropy F23 of Jelinek (1977). Specimens with F23 values below 5 (or E23 > 25°) are considered isotropic in the plane perpendicular to Kmax. This means that K-mi and Kmin cannot be distinguished (Kini = Km{n; pure prolate ellipsoid). As for E12 and L (lineations), a relationship between E23 and the foliation parameter (F = ^int/^min) can be established at sampleand~ site-scales with an exponential relationship between them (Figs. 8a, c). The smaller the E23 angle, the larger F will be. High E23 angles correspond to highly prolate ellipsoids (T approximately below —0.5) in which Kmax dominates over K[ni ~ Km^n (Figs. 8b, d). Due to scale smoothing, at site-scale the variation of E23 is defined for a smaller interval (0-38°) than at sample-scale (0-70°). As for the lineations, F and E23 are the basic variables that help calibrating the quality boundaries with respect to the different spherical statistical models. Therefore three different categories of quality are established. Class I: E23 < 20° (F23 > 8): the foliations are always reliable. In Veiga and Trives, the standard error of E23 (eE23) was less than 5°, the F parameter greater than ~ 1.014, and T greater than -0.3. Spherical statistics as applied to the mean-Xjnin gave errors smaller than 40° (A; Lienert), smaller than 25° (a95; Fisher), or larger than 10 («; Fisher) and than 5 (Lw = Sr1/S2; Woodcock). Class II: 20° > E23 > 25° (5 < F23 < 8). These low-quality foliations must be treated with caution. In Veiga and Trives, eE23 ranged between 5° and 10°, F between 1.007 and 1.014, and T between - -0.3 and - -0.5. Spherical statistics of mean-^min gave the following error ranges: 40-80° (^; Lienert), 25-35° (a95; Fisher), 5-10 (K; Fisher) and 3-5 (Lw = S{ /S2; Woodcock).

407

Class III: E23 > 25° (F23 < 5). Foliations cannot be defined statistically, belonging to pure prolate ellipsoids. In Veiga and Trives, eE23 were larger than 10°, F less than 1.007 and T smaller than ~ -0.5. Spherical statistics of mean-^Tmin gave errors larger than 80° (fl\; Lienert), larger than 35° (a95; Fisher), or smaller than 5 («; Fisher) and 3 (Lw = Si/S2; Woodcock). Oblate and triaxial fabrics usually predominate in granite massifs (Fig. Id), illustrating the fact that the magnetic foliation is frequently better defined than lineation. For example, in our massifs (Figs. 7e, f), up to 27% of lineation means were removed from class I, while only 7% of foliation means were subtracted from the high quality set. In order to increase the representativity of sectoror massif-averages, we still recommend including angular errors (Lienert 1991; Constable & Tauxe 1991) in foliations maps and/or providing tables that allow quality evaluation. Number of specimens per site Validity of the statistical" models obviously depends on the amount of data. In magnetic fabric studies the number of samples per site varies, depending upon authors, between 4 and 12, and it is quite acceptable to use 2 cores yielding 4 to 6 standard specimens. Again, Trives and Veiga provide a practical analysis of quality versus number of samples since they have a wide variation of number of specimens per site, between 2 and 13, with averages of 6 for Veiga and 9 for Trives (Fig. Ic). At site-scale (Fig. 9a) two extreme fabrics have been chosen. T17 is a prolate-type where the maximum angular departure of ^min between individual samples reaches 124°. T90 is an oblate-type where the maximum departure of Kmsix reaches 126°. Both sites contain 4 cores yielding 11 and 8 specimens respectively. The evolution of some statistical parameters (a95, fa, Lw and Fw) is calculated as a function of the number of cores and specimens (Figs. 9b, c). Note that we have purposely selected the sequence of sample stacking, from 2 to 4 cores (and their corresponding specimens), that best shows the evolution of the statistical parameters. Unsurprisingly, Kmax from T17 (prolate) displays small angular errors independently of the number of cores and is also well grouped in the cluster field of the Woodcock diagram as soon as 2 cores are considered. £"min from T90 (oblate) displays a relative improvement in both the angular errors and clustering as soon as a third core is considered in the calculation.

408

E. L. PUEYO ET AL.

Fig. 8. Reliability of magnetic foliations, (a) At sample scale: foliation parameter (F — K[nt/Kmin) versus the (95%) confidence angle between K-mi and K^m (£23). (b) At sample scale: E23 against the shape parameter (T). (c) At site scale: site mean foliation parameter (F = K-mi/Kmin) versus site mean E23 angles (and their respective standard errors), (d) At site scale: E23 against the shape parameter (T). (e) Distribution histogram of number of sites for E12. (f) Same histogram for E23. These histograms give the percentages of sites in the three classes of qualities.

STATISTICAL SIGNIFICANCE OF AMS DATA IN GRANITES

Fig. 9. Variation of statistical parameters as a function of the number of data (n). (a) Equal-area projection of two end-member fabrics; square-^Tmax, triangle-^int and dot-^min. (b) Fisher-Lienert confidence angles evolution of orientation data as a function of n. (c) Woodcock-Flinn graphic to control shape and strength of the spatial distributions, (d) 0^95 versus n plot, (e) Percentage of high quality lineations versus n.

409

410

E. L. PUEYO ET AL.

In these examples, due to the extreme shapes of the ellipsoids, a good clustering of the orientations was more-or-less predictable, even for very few samples. This contrasts with the Kmin axes of the prolate site (T17) and the KmSLX axes of the oblate site (T90). They display a larger variability, with ^min from T17 enlarging its confidence region j3\ by about 35°, and its a95 by about 10° when a third core (3 more specimens) is included in the calculation. The shape of the ATmin distribution moves from the cluster-girdle field (2 cores) to the pure girdle field (4 cores) in the Woodcock diagram, which agrees better with the expected geometry of the minimum axes of a prolate fabric. Similar results occur with the maximum susceptibility axes in the oblate case (T90): j3\ and a95 of KmSLX increase by about 15° when jumping from 2 to 3 cores (Fig. 9b). This evolution is even clearer in the Woodcock diagram (Fig. 9c) where, by progressively adding data to the initial set, the drift of Kmax surprisingly starts in the cluster field before moving to the girdle field. In view of these results, two cores may not be sufficient to characterize a reliable fabric at a given site. Finally, the number of samples per site (n) is examined using, once more, the data set from Trives and Veiga. We use the mean lineations, which are more sensitive than the foliations to small variations of the fabric shape with the relationship between a95 and K, which depends on the number of samples. We have reported the Fisher parameters for six different subsets of n values, each subset having a reasonable number of sites (n < 5 (15), n = 6 (22), n = 1 (10), n = 8 (24), n = 9 (11), and n > 10 (11)). As expected, the portion of curve located in the class-I field increases with the increase of n (Fig. 9d). The percentages of class-I lineations for different values of n are calculated to constrain the optimum value of n (Fig. 9e). From the positive correlation of these percentages, we contend that using fewer than 5 samples per site will only display 40% of reliable lineations while more than 10 samples per site will provide 90% of reliable lineations. In view of this remarkable relationship, here calculated for lineations, our advice is that a minimum of 3 cores (>7 cm-long) per site, which would yield a minimum of 9 standard specimens, would definitely double the percentage of reliable orientation data. Conclusion A detailed analysis of reliability of AMS data in granitic rocks at different scales (sample, site and

massif) by using different statistical approaches leads to the following conclusions: (i) Susceptibility tensors with very low anisotropies (E13 between 10° and 20°) can be considered as pseudo-isotropic and their orientation data and parameters must be used with caution or rejected, (ii) Lineation is not reliable when the angle E12 is above 25°. Therefore the use of this information should be avoided. Lineation data should be used with care when the E12 ranges between 25° and 20°. (iii) Foliation is not reliable when the E23 is above 25° and should not be utilized. Foliation data should be treated carefully when the E12 ranges between 25° and 20°. (iv) It is highly recommended that error information be included along with the orientation data (in figures and/or tables). This will allow evaluation of their quality, (v) The systematic definition of foliation and lineation at any site assuming triaxiality is far from correct. In the examples of Veiga and Trives, only 53% of sites agree with this assumption. And (vi) a minimum of three cores (9 specimens) per site would considerably increase the proportion of reliable orientation data. This research has been funded by the Projects BTE2001-0634 and BTE2002-04168 of the Spanish Ministry of Science and Technology. Two postdoctoral positions from the Spanish Ministry of Education and Culture (ELP) and from the Navarra Government (JCL) are also acknowledged. We used the 'Stereonet' program of R. Allmendinger and the AplotlO software of Lienert, to whom we are indebted. I. Gil-Pena, C. Maldonado and G. Valenzuela, helped us greatly during the field campaigns. P. Lespinasse, P. Olivier and G. Gleizes from Toulouse are acknowledged for providing the original data from the Pyrenees. Comments by S. Siegesmund and an anonymous reviewer improved this paper.

References ARANGUREN, A., TUBIA, J. M., BOUCHEZ, J. L. & VIGNERESSE, J. L. 1996. The Guitiriz granite, Variscan belt of northern Spain: extension-controlled emplacement of magma during tectonic escape. Earth and Planetary Science Letters, 139, 165-176. ARBARET, L., DIOT, H., BOUCHEZ, J. L. & LAUNEAU, P. 1995. Tridimensional shape preferred orientation of particles in experimental simple shear flow; AMS fabric and image analysis. Journal of the Czech Geological Society, 40(3), 57-58. ARCHANJO, C. J., LAUNEAU, P. & BOUCHEZ, J. L. 1995. Magnetic fabric vs. magnetite and biotite shape fabrics of the magnetite-bearing granite pluton of Gameleiras (Northeast Brazil). Physics of the Earth and Planetary Interiors, 89(1-2), 63-75. BALSEY, J. R. & BUDDINGTON, A. F. 1960. Magnetic susceptibility anisotropy and fabric of some

STATISTICAL SIGNIFICANCE OF AMS DATA IN GRANITES Adirondack granites and orthogneisses. American Journal of Science, 258-A, 6-20. BARRERA-MORATE, J. L., FARIAS-ARQUER, P., ET AL. 1989. Memoria y Hqja no 17/27 (Ourense/Verin). Mapa geologico de Espafia, E. 1:200.000, ITGE, Madrid. BASTIDA, F., MARTINEZ-CATALAN, J. R. & PULGAR, J. A. 1986. Structural, metamorphic and magmatic history of the Mondofiedo Nappe (Hercynian belt, NW Spain). Journal of Structural Geology, 8, 415430. BINGHAM, C. 1964. Distributions on the sphere and on the projective plane. PhD. Thesis, Yale Univ. New Haven, CT. BINGHAM, C. 1974. An antipodally symmetric distribution on the sphere. Annales Statistical, 2, 12011225. BIRCH, F. S. 1979. Magnetic fabric of the Exeter pluton, New Hampshire. Journal of Geophysical Research, 84, 1129-1137. BORRADAILE, G. J. & HENRY, B. 1997. Tectonic applications of magnetic susceptibility and its anisotropy. Earth Science Reviews, 42, 49-93. BORRADAILE, G. J. & WERNER, T. 1994. Magnetic anisotropy of some phyllosilicates. Tectonophysics, 235, 223-248. BOUCHEZ, J. L. 1997. Granite is never isotropic: an introduction to AMS studies of granitic rocks. In: BOUCHEZ, J. L., HUTTON, D. H. W. & STEPHEN, W. E. (eds) Granite: from Segregation of Melt to Emplacement Fabrics. Kluwer Academic Publishers, Dordrecht, 95-112. BOUCHEZ, J. L. 2000 . Anisotropie de susceptibilite magnetique et fabrique des granites. Comptes Rendus Academic des Sciences Paris, Serie 2, 330, 1-14. CONSTABLE, C. & TAUXE, L. 1990. The bootstrap for magnetic susceptibility tensors. Journal of Geophysical Research, 96(6), 8383-8395. DUFFA, B. 1975. The magnetic fabric of the Moriya granite, New South Wales. Journal of the Geological Society of Australia, 22, 159-165. ELLWOOD, B. B. & WENNER, D. B. 1981. Correlation of magnetic susceptibility with O18/O16 data in late orogenic granites of the southern Appalachian Piedmont. Earth and Planetary Science Letters, 54, 200-202. ELLWOOD, B. B. & WHITNEY, J. A. 1980. Magnetic fabric of the Elberton granite, Northeast Georgia. Journal of Geophysical Research, 85, 1481-1486. ERNST, R. E. & PEARCE, G. W. 1989. Averaging of anisotropy of magnetic susceptibility data. In: AGTERBERG, F. P. & BONHAM-CARTER, G. F. G. (eds), Statistical Applications in the Environment. Geological Survey of Canada, Ottawa, 89-9, 297-305. FISHER, N. I., LEWIS, T. & EMBLETON, B. J. J. 1987. Statistical Analysis of Spherical Data. Cambridge University Press, Cambridge. FISHER, R. A. 1953. Dispersion on a sphere. Proceedings of the Royal Society of London, A, 217, 295-305. FLINN, D. 1962. On folding during three-dimensional progressive deformation. Quarterly Journal of the Geological Society of London, 118, 385-433.

411

GLEIZES, G., LEBLANC, D. & BOUCHEZ, J. L. 1991. Le pluton granitique de Bassies (Pyrenees ariegeoises): zonation, structure et mise en place. Comptes Rendus Academic des Sciences, Paris, Serie 2, 312(7), 755-762. GLEIZES, G., LEBLANC, D., SANTANA, V., OLIVIER, P. & BOUCHEZ, J. L. 1998. Sigmoidal structures featuring dextral shear during emplacement of the Hercynian granite complex of Cauterets-Panticosa (Pyrenees). Journal of Structural Geology, 20(9-10), 1229-1245. GLEIZES, G., NEDELEC, A., BOUCHEZ, J. L., AUTRAN, A. & ROCHETTE, P. 1993. Magnetic susceptibility of the Mont Louis-Andorra ilmenite-type granite (Pyrenees): a new tool for the petrographic characterization and regional mapping of zoned granite plutons. Journal of Geophysical Research, 98, 4317-4331. GoNZALEZ-LoDEiRO, F. & IGLESIAS, M. 1977. Memoria y Hoja No 156 (Monforte de Lemos). Mapa Geologico de Espana E. 1:50000. IGME. GREGOIRE, V., DARROZES, J., GAILLOT, P., NEDELEC, A. & LAUNEAU, P. 1998. Magnetite grain shape fabric and distribution anisotropy vs rock magnetic fabric: a three-dimensional case study. Journal of Structural Geology, 20(7), 937-944. GUILLET, P., BOUCHEZ, J. L. & WAGNER, J. J. 1983. Anisotropy of magnetic susceptibility and magnetic structures in the Guerande granite massif (France). Tectonics, 2(5), 419-429. HROUDA, F. & LANZA, R. 1989. Magnetic fabric in the Biella and Traversella stocks (Periadriatic Line): implications for the mode of emplacement. Physics of the Earth and Planetary Interiors, 56(3-4), 337348. IsfflHARA, S. 1977. The magnetite-series and ilmeniteseries granitic rocks. Mining Geology, 27, 293-305. JELINEK, V. 1977. The statistical theory of measuring anisotropy of magnetic susceptibility of rocks and its application. Geofyzika Brno. JELINEK, V. 1981. Characterization of the magnetic fabric of rocks. Tectonophysics, 79, 63-67. JEZEK, J. & HROUDA, F. 2000. The relationship between the Lisle orientation tensor and the susceptibility tensor. Physics Chemistry Earth (A), 25(5), 469474. KHAN, M. A. 1962. The anisotropy of magnetic susceptibility of some igneous and metamorphic rocks. Journal of Geophysical Research, 67, 28732885. LAUNEAU, P., BOUCHEZ, J. L. & BENN, K. 1990. Shape preferred orientation of object populations: automatic analysis of digitized images. Tectonophysics, 180,201-211. LEBLANC, D., GLEIZES, G., LESPINASSE, P., OLIVIER, P. & BOUCHEZ, J. L. 1994. The Maladeta granite polydiapir, Spanish Pyrenees; a detailed magnetostructural study. Journal of Structural Geology, 16(2), 223-235. 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(4), 331-345.

412

E. L. PUEYO ET AL.

LIENERT, B. R. 1991. Monte Carlo simulation of errors in the anisotropy of magnetic susceptibility: a second-rank symmetric tensor. Journal of Geophysical Research, 96, 19539-19544. LUNEBURG, C. M., LAMPERT, S. A., LEBIT, H. D., HIRT, A. M., CASEY, M. & LOWRIE, W. 1999. Magnetic anisotropy, rock fabrics and finite strain in deformed sediments of SW Sardinia (Italy). Tectonophysics, 307, 51-74. MARTINEZ-CATALAN, J. R. 1985. Estratigrafia y Estructura del Domo de Lugo (Sector Oeste de la zona Asturoccidental-leonesa). Laboratorio Geologico Laxe, Corpus Geologicum Gallaeciae. MARTIN-HERNANDEZ, F. & HIRT, A. M. 2003. The anisotropy of magnetic susceptibility in biotite, muscovite and chlorite single crystal. Tectonophysics,361, 13-28. OLIVIER, P., SAINT BLANQUAT, M., GLEIZES, G. & LEBLANC, D. 1997. Homogeneity of granite fabrics at the metre and dekametre scales. In: BOUCHEZ, J. L., HUTTON, D. H. W. & STEPHEN, W. E. (eds) Granite: from Segregation of Melt to Emplacement Fabrics. Kluwer Academic Publishers, Dordrecht, 113-127. OLIVIER, P., AMEGLIO, L., RICHEN, H. & VADEBOIN, F. 1999. Emplacement of the Aya Variscan granitic pluton (Basque Pyrenees) in a dextral transcurrent regime inferred from a combined magnetostructural and gravimetric study. Journal of the Geological Society, London, 156, 991-1002. PARGA-PONDAL, I. 1963. Rocas y minerales de interes economico del macizo galaico. Trabajos del Laboratorio Geologico de Lage (La Coruna), 14, 99-104. ROMAN-BERDIEL, T., PUEYO-MORER, E. L. & CASASSAINZ, A. M. 1995. Granite emplacement during contemporary shortening and normal faulting: structural and magnetic study of the Veiga Massif (NW Spain). Journal of Structural Geology, 17, 1689-1706. ROMAN-BERDIEL, T., ARANGUREN, A., CUEVAS, J. & TUBIA, J. M. 1998. Compressional graniteemplacement model: structural and magnetic study of the Trives Massif (NW Spain). Lithos, 44, 37-52. ROMAN-BERDIEL, T., ARANGUREN, A., CASAS, A. M., CUEVAS, J., PUEYO, E. L. & TUBIA, J. M. 2001. La lineacion magnetica en los granites hercinicos del Noroeste de la Peninsula Iberica. Proceedings MAGIBER I, 53-54. ROMAN-BERDIEL, T., CASAS, A. M., OLIVA-URCIA, B., PUEYO, E. L. & RILLO, C. 2004. The main Variscan deformation event in the Pyrenees: new data from

the structural study of the Bielsa granite. Journal of Structural Geology, 26, 659-677. SlEGESMUND, S., ULLEMEYER, K. & DAHMS, M.

1995.

Control of magnetic fabrics by mica preferred orientation: a quantitative approach. Journal of Structural Geology, 17, 1601-1613. SCHEIDEGGER, A. M. 1965. On the statistics of the orientation of bedding planes, grain axes, and similar sedimentological data. US Geological Survey Professional Papers, 525C, 164-167. STAGEY, F. D. & BANERJEE, S. H. 1974. The Physical Principles of Rock Magnetism. Elsevier, Amsterdam. TARLING, D. H. & HROUDA, F. 1993. The Magnetic Anisotropy of Rocks, Chapman and Hall, London. TRINDADE, R. I. F., RAPOSO, M. I. B., ERNESTO, M. & SIQUIERA, R. 1999. Magnetic susceptibility and partial anhysteretic remanence anisotropies in the magnetite-bearing granite pluton of Tourao, NE Brazil. Tectonophysics, 314, 443-468. VAN DER Voo, R. & KLOOTWIJK, C. T. 1972. Paleomagnetic reconnaissance study of the Flamanville granite, with special reference to the anisotropy of its susceptibility. Geologic en Mijnbouw, 51, 609617. VEGAS, N., ARANGUREN, A., CUEVAS, J. & TUBIA, J. M. 2001. Variaciones en los mecanismos de emplazamiento de los granitos del eje Sanabria-Viana do Bolo (Macizo Iberico, Espana). Bole tin Geologico y Minero, 112(3), 79-88. WATSON, G. S. 1965. Equatorial distributions on a sphere. Biometrika, 52, 193-201. WENNER, D. B. 1981. Oxigen isotopic compositions of the late erogenic granites in the southern Piedmont of the Appalachian Mountains, USA, and their relationship to subcrustal structures and lithologies. Earth and Planetary Science Letters, 54, 186-199. WOODCOCK, N. H. 1977. Specification of fabric shapes using an eigenvalue method. Bulletin of the Geological Society of America, 88, 1231-1236. WOODCOCK, N. H. & NAYLOR, M. A. 1983. Randomness testing in three-dimensional orientation data. Journal of Structural Geology, 5, 539-48. YENES, M., ALVAREZ, F. & GUTIERREZ-ALONSO, G. 1999. Granite emplacement in erogenic compressional conditions: the La Alberca-Bejar granitic area (Spanish Central System, Variscan Iberian Belt). Journal of Structural Geology, 21, 14191440. ZAPLETAL, K. 1990. Low-field susceptibility anisotropy of some biotite crystals. Physics of the Earth and Planetary Interiors, 63, 85-97.

STATISTICAL SIGNIFICANCE OF AMS DATA IN GRANITES

413

Appendix 1. Site means for the AMS information from Veiga and Trives granitic massifs Jelinek angles Site

n

N

Kbulk

eK

E12

eE12

E23

eE23

E31

eE31

Veiga VI V2 V3 V4 V5 V6 V7 V8 V9 V10 II1 V12 V13 V14 V15 V16 V17 V18 V19 V20 V21 V22 V23 V24 V25 V26 V27 V28 V29 V30 V31 V32 V33 V34 V35 V36 V37 V38 V39 V40 V41 V42

6 6 6 6 6 7 4 5 6 5 7 6 6 6 7 8 6 6 9 2 5 6 4 9 6 6 7 5 5 6 6 5 5 6 5 4 3 6 4 6 3 6

6 6 6 6 6 7 4 5 6 5 7 6 6 6 7 8 6 6 9 2 5 6 4 9 6 6 7 5 5 6 6 5 5 6 5 4 3 6 4 6 3 6

99.115 169.07 157.5 117.4 114.2 92.48 59.61 121.92 97.52 86.77 97.05 116.1 117.99 16.26 139.3 48.92 130.4 96.8 82.06 128.2 109.01 81.89 109.67 97.01 54.14 130.17 193.11 187.46 270.94 141.15 152.62 176.24 180.22 158.88 185.8 136.38 113.4 123.46 101.69 182.38 143.27 93.54

12.234 34.9 11.4 11.9 20.1 4.4 18.29 4.9 7 12.7 23.1 9.7 18.3 1.7 33.5 2.3 12.5 11.9 10.4 7 18.9 2.7 10.7 6.8 3.6 6.9 39.8 11.3 15.9 3.6 13.7 13.1 21.3 7.13 10.9 12.7 10.7 16.1 5.6 30.33 17.5 7.6

7.91 8.98 9.85 18.8 12.5 14.7 39.75 20 19.3 19.5

1.32 0.93 1.21 4.56 3.67 4.53 11.2 5.45 9.5 2.62

8.56 18.6 17.9 6.18 11.6 21.7 28.73 20.1 11.8 29.1

2.83 5.63 7.81 1.1 3.54 5.34 10.8 4.55 2.35 6.76

3.31 5.58 5.12 4.55 5.53 6.58 14.45 8.36 5.67 11.5

0.51 0.62 0.45 0.85 1.58 0.76 2.45 1.26 1.25 1.74

10.4 11.2 30 15.8 14.7 18.8 17.2 25 11.3 16.1 11 23.7 13.9 9.6 9.1 24 11.3 34.5 24.4 34.5 18.9 22 33.3 27.9 27.7 7.9 4.9 6.4 5.4 4.7 9.6

1.97 2.6 5 4 7.4 7 8.1 6.4 0.6 1.3 3.7 2.3 3.2 1.5 1.2 6 2.6 11 6.9 6.6 4.3 3.7 5.6 7.9 7.1 0.4 0.3 1.1 0.7 1.2 2.7

10.6 12 12.5 12 7 7.5 15.8 6.9 16.8 10.7 9.8 11.3 11.5 13.8 3.5 18.7 6 16 5.8 7.6 8.9 7.7 6.4 13.3 5.9 6.2 8.8 8.9 6.1 5.4 9.5

3.54 3.6 2.4 3 1 1 7.1 1 0.7 2.7 2.6 3.9 2.9 7.6 0.5 3.9 0.9 5.4 1.8 1.2 1.9 1.5 1.2 4.5 1.5 1.1 1.7 1.4 1 1.4 2.4

4.5 5.3 8.5 7 3.7 5.2 7.7 5.5 6.9 6.2 4 7 5.4 4.2 2.5 9.7 3.9 9.1 4.6 6.2 6 5.4 5.2 7.1 4.7 3.5 3.1 3.7 2.8 2.5 4.3

0.84 1.2 1 2 0.4 1 3 1 0.4 1.3 0.5 1.3 0.9 0.7 0.3 1.2 0.7 1.6 1.3 0.9 1.3 0.7 0.7 0.9 1.3 0.3 0.3 0.6 0.4 0.6 0.7

n/N: number of considered/analysed specimens. Kbulk (eK): Bulk susceptibility at room temperature and low field (standard error). E12 (eE12), E23 (eE23), E31 (eE31): Jelinek (1977) confidence angles and their respective standard errors. L (eL), F (eF), P' (eP'), T (eT): Fabric parameters; Lineation, Foliation, Corrected anisotropy degree and shape with their respective standard errors. Kmsix mean information, kl-dec, kl-inc, kl-a, kl-k: Fisher mean (declination and inclination) and a95 and concentration parameter. K1-E1, K1-E2, K1-E3: Eigenvalues of KmaK. El-dee-kl, El-inc-kl: First eigenvector orientation. E3-dec-kl, E3-inc-kl: Third eigenvector orientation. Kmin mean information: k3-dec, k3-inc, k3-a, k3-k: Fisher mean (dec, inc) and a95 and concentration parameter. K3-E1, K3-E2, K3-E3: Eigenvalues of Km[n. El-dec-k3, El-inc-k3: First eigenvector orientation. E3dec-k3, E3-inc-k3: Third eigenvector orientation. Lienert (1991) confidence ellipse angles of Kmax (/31-kl, /32-kl) and A:min (/31-k3, /32-k3)

E. L. PUEYO ET AL.

414

Appendix 1. (cont.) A:max-Fisher

Fabric parameters

eF

Site

L

eL

Veiga VI V2 V3 V4 V5 V6 V7 V8 V9 V10 Vll V12 V13 V14 V15 V16 V17 V18 V19 V20 V21 V22 V23 V24 V25 V26 V27 V28 V29 V30 V31 V32 V33 V34 V35 V36 V37 V38 V39 V40 V41 V42

1.018 1.023 1.016 1.01 1.017 1.021 1.006 1.014 1.014 1.013 1.012 1.018 1.013 1.013 1.013 1.023 1.012 1.012 1.01 1.016 1.014 1.016 1.007 1.014 1.017 1.017 1.01 1.015 1.006 1.005 1.005 1.01 1.01 1.005 1.009 1.007 1.028 1.04 1.039 1.02 1.026 1.02

1.003 1.023 1.005 1.003 1.013 1.002 1.002 1.014 1.003 1.003 1.001 1.03 1.003 1.023 1.005 1.003 1.017 1.005 1.002 1.014 1.006 1.014 1.004 1.004 1.003 1.014 1.003 1.002 1.01 1.002 1.003 1.029 1.003 1.003 1.021 1.003 1.002 1.014 1.003 1.002 1.038 1.008 1.004 1.001 1.02 1.004 1.028 1.003 1.002 1.025 1.002 1.003 1.002 1.02 1.002 1.002 1.03 1.003 1.011 1.002 1.002 1.026 1.004 1.003 1.019 1.005 1.001 1.019 1.004 1.002 1.016 1.003 1.003 1.021 1.004 1.001 1.046 1.002 1.001 1.016 1.003 1.002 1.027 1.003 1.002 1.018 1.005 1.001 1.023 1.003 1.001 1.026 1.002 1.002 1.028 1.004 1.003 1.029 1.002 1.028 1.004 1.001 1.004 1.003 1.02 1.003 1.003 1.03 1.003 1.037 1.002 1.004 1.022 1.003 1.002 1.03 1.003 1.003 1.002 1.02 1.024 1.003 1 1.002 1.021 1.004

F

P'

eP'

T

eT

kl-dec

kl-inc

kl-a

kl-K

1.042 1.037 1.031 1.042 1.041 1.04 1.022 1.029 1.029 1.023 1.043 1.04 1.027 1.054 1.034 1.052 1.038 1.033 1.042 1.026 1.041 1.037 1.027 1.031 1.038 1.065 1.027 1.042 1.026 1.029 1.033 1.04 1.042 1.035 1.031 1.039 1.065 1.063 1.068 1.04 1.051 1.042

1.003 1.002 1.002 1.003 1.005 1.003 1.007 1.003 1.003 1.003 1.005 1.003 1.004 1.009 1.005 1.005 1.001 1.005 1.003 1.004 1.004 1.004 1.004 1.002 1.003 1.002 1.003 1.004 1.006 1.004 1.003 1.004 1.002 1.004 1.002 1.002 1.001 1.004 1.005 1.004 1.003 1.003

0.071 -0.254 -0.11 0.507 0.123 -0.169 0.236 -0.01 0.051 -0.186 0.456 0.08 -0.011 0.426 0.175 0.142 0.357 0.225 0.54 -0.197 0.272 0.012 0.404 0.081 0.075 0.458 0.123 0.291 0.387 0.607 0.661 0.465 0.484 0.673 0.335 0.616 0.147 -0.303 -0.15 -0.023 -0.046 0.018

0.186 0.132 0.177 0.076 0.151 0.205 0.319 0.228 0.21 0.146 0.08 0.172 0.159 0.112 0.085 0.139 0.101 0.13 0.054 0.005 0.12 0.2 0.183 0.13 0.192 0.031 0.138 0.076 0.24 0.063 0.058 0.105 0.117 0.081 0.228 0.142 0.08 0.081 0.04 0.081 0.056 0.145

287 119 112 95 120 111 132 114 88 100 84 70 126 351 135 309 17 39 122 23 324 23 99 122 134 129 297 313 143 119 336 347 322 298 337 14 286 302 293 289 298 127

0 4 24 21 3 35 20 15 9 13 39 32 2 7 16 18 16 13 70 14 9 8 71 26 27 1 13 21 20 43 30 59 35 16 47 84 20 20 16 12 19 7

13 24 12 34 13 13 28 22 33 21 14 14 17 28 15 34 21 22 57 39 65 22 34 22 20 24 23 16 48 47 23 40 29 20 47 37 10 9 17 11 11 27

24 8 28 4 24 21 9 10 4 11 17 21 13 5 15 3 10 8 1 22 2 9 6 6 10 7 7 19 3 3 8 4 6 10 3 5 96 51 23 30 86 6

STATISTICAL SIGNIFICANCE OF AMS DATA IN GRANITES

415

Appendix 1. (cont.) Amax-Bingham Eigenvalues Site Veiga VI V2 V3 V4 V5 V6 V7 V8 V9 V10 II1 V12 V13 V14 V15 V16 V17 V18 V19 V20 V21 V22 V23 V24 V25 V26 V27 V28 V29 V30 V31 V32 V33 V34 V35 V36 V37 V38 V39 V40 V41 V42

Eigenvectors

^Tmin-Fisher

K1-E1

K1-E2

K1-E3 El-dec-kl

El-inc-kl

E3-dec-kl

E3-inc-kl

k3-dec k3-inc k3-a

k3-K

0.9439 0.8441 0.9512 0.7785 0.9445 0.9325 0.8825 0.8828 0.7666 0.8914 0.9154 0.9369 0.8998 0.7714 0.9094 0.7307 0.8641 0.8497 0.4144

0.0481 0.1504 0.042 0.2126 0.0501 0.046 0.1001 0.087 0.2214 0.0961 0.0502 0.0613 0.0796 0.2 0.0774 0.199 0.1254 0.1391 0.3737

0.008 0.0055 0.0067 0.0089 0.0054 0.0215 0.0174 0.0303 0.012 0.0125 0.0344 0.0018 0.0206 0.0286 0.0132 0.0703 0.0105 0.0113 0.2119

287 121 112 92 120 111 132 115 97 100 84 71 127 348 133 303 17 40 129

0 10 24 30 3 35 19 15 13 13 40 32 2 6 16 9 17 13 8

195 11 12 193 219 223 223 210 333 5 202 238 34 239 302 69 241 277 222

83 62 22 17 73 29 5 21 69 23 29 58 54 73 74 75 68 68 23

0.4535 0.8525 0.8336 0.7515 0.8698 0.8374 0.7963 0.9355 0.673 0.6081 0.8412 0.6967 0.8211 0.8705 0.631 0.8051

0.3729 0.1145 0.1597 0.1434 0.1087 0.159 0.1896 0.0595 0.2979 0.3861 0.1394 0.2925 0.1735 0.1025 0.3592 0.1661

0.1736 0.0331 0.0067 0.105 0.0214 0.0036 0.0141 0.005 0.0291 0.0058 0.0194 0.0108 0.0054 0.0271 0.0098 0.0287

0.973 0.9524 0.9542 0.9897 0.8018

0.0222 0.0401 0.0346 0.0088 0.1315

0.0048 0.0075 0.0112 0.0016 0.0668

314 23 99 121 134 130 298 313 137 124 337 344 321 298 342 1 286 302 293 289 298 124

1 8 71 27 26 1 15 21 34 31 30 58 33 17 50 84 20 20 16 12 19 4

223 134 307 31 232 9 200 184 236 226 231 220 215 207 201 207 166 64 55 54 66 32

24 68 17 1 16 88 30 59 13 19 24 19 23 5 34 6 54 55 62 69 60 33

196 213 210 199 214 201 244 202 352 5 188 246 51 247 234 193 242 284 210 119 228 197 354 12 236 24 192 218 235 225 235 208 211 202 193 197 177 91 152 138 76 221

4 3 12 28 11 5 2 3 4 8 6 146 3 12 3 3 32 20 4 22 5 8 2 3 4 56 11 16 5 54 29 8 25 20 33 26 133 34 4 20 192 5

30 46 9 19 32 6 35 21 39 1 17 57 46 56 53 13 66 55 21 22 28 80 23 23 18 82 42 12 28 16 21 24 28 20 38 5 42 67 74 74 65 39

33 42 19 12 19 27 61 45 35 25 25 5 41 18 37 40 11 14 30 37 32 23 70 31 34 8 17 18 32 8 11 26 14 14 10 16 9 11 46 14 7 29

E. L. PUEYO ET AL.

416

Appendix 1. (cont.) Amin-Bingham Eigenvalues Site Veiga VI V2 V3 V4 V5 V6 V7 V8 V9 V10 II1 V12 V13 V14 V15 V16 V17 V18 V19 V20 V21 V22 V23 V24 V25 V26 V27 V28 V29 V30 V31 V32 V33 V34 V35 V36 V37 V38 V39 V40 V41 V42

Eigenvectors

Lienert

K3-E1

K3-E2

K3-E3

El-dec-k3

El-inc-k3

E3-dec-k3

E3-inc-k3 /31-kl

/?2-kl

(3\-k3 (32-k3

0.7614 0.669 0.888 0.9521 0.895 0.7703 0.6415 0.6796 0.7075 0.853 0.7801 0.9905 0.753 0.8926 0.693 0.5887 0.9572 0.931 0.712

0.2305 0.2237 0.1052 0.0379 0.1019 0.2 0.3446 0.2818 0.2692 0.1164 0.1744 0.0062 0.2301 0.0799 0.2839 0.2664 0.0264 0.0355 0.1757

0.0081 0.1073 0.0068 0.01 0.0031 0.0297 0.0138 0.0385 0.0234 0.0306 0.0455 0.0032 0.017 0.0275 0.023 0.145 0.0164 0.0335 0.1124

199 217 210 198 212 204 238 202 353 6 189 246 44 247 230 212 242 284 216

38 36 9 19 34 7 38 10 36 0 18 57 61 55 34 16 66 55 30

98 103 115 93 322 105 111 293 99 96 91 85 146 340 135 101 39 148 318

14 29 26 38 28 51 38 3 18 5 25 31 7 2 8 51 22 27 19

25 18 13 23 44 38 82 31 32 44 33 27 19 60 28 19 47 51 33

15 9 11 17 20 8 70 10 7 17 22 5 8 18 2 14 10 25 25

51 48 55 73 61 53 84 69 44 64 39 19 25 28 70 55 71 26 64

11 16 10 21 23 19 7 6 23 25 20 4 9 23 12 13 10 16 29

0.8066 0.8466 0.5836 0.6229 0.7721 0.9753 0.8778 0.922 0.7904 0.9745 0.9537 0.8506 0.9501 0.9324 0.9758 0.9571 0.9933 0.9602 0.768 0.9323 0.9954 0.7636

0.1764 0.1527 0.3527 0.3501 0.1939 0.0173 0.1054 0.0475 0.1793 0.0213 0.0435 0.1342 0.0458 0.0463 0.0219 0.0395 0.0059 0.0336 0.228 0.0599 0.0035 0.1637

0.017 0.0007 0.0637 0.027 0.0339 0.0074 0.0168 0.0305 0.0303 0.0042 0.0027 0.0152 0.0041 0.0213 0.0023 0.0033 0.0008 0.0062 0.004 0.0078 0.0011 0.0727

222 188 1 13 240 25 192 218 232 225 235 211 211 202 193 197 177 92 115 138 76 216

29 81 17 21 27 82 41 12 25 16 21 22 28 20 38 6 42 67 74 74 65 41

21 359 127 130 142 244 293 42 127 323 350 330 304 331 41 105 279 306 316 289 330 343

60 9 63 49 14 7 12 78 29 27 48 51 5 60 49 25 14 19 15 15 7 34

65 38 66 29 12 44 34 19 45

23 5 57 20 9 4 20 14 19

67 44 65 34 80 19 40 24 59

40 4 58 22 6 11 18 1 15

53 60

26 10

74 29

32 7

43 58 64 59 18 45 13 71 48

24 2 19 25 7 14 8 13 24

36 27 50 43 17 31 35 53 26

12 2 18 11 5 4 9 13 20

STATISTICAL SIGNIFICANCE OF AMS DATA IN GRANITES

417

Appendix 1. (cont.) Jelmek angles Site

n

N

Kbulk

eK

E12

eE12

E23

eE23

E31

eE31

Trives T10' T104' T106 T108 Til' T112 T115 T12 T13 T134 T136 T14' T140 T141 T145 T148 T15 T151 T152 T154 T158 T159 T161 T164 T17 T172 T173 T174 T175 T176 T177 T178 T179 T19 T2 T3' T59' T6 T60' T67 T71' T78 T79' T8 T86 T89' T9 T9' T90 T94 T95

8 13 11 9 8 8 7 10 8 7 8 10 8 7 8 9 9 8 8 10 7 8 8 8 12 6 11 9 10 8 8 8 8 11 7 7 8 8 8 9 10 9 9 9 12 8 8 6 8 9 8

8 13 11 9 8 8 7 10 8 7 8 10 8 7 8 9 9 8 8 10 7 8 8 8 12 6 11 9 10 8 8 8 8 11 7 7 8 8 8 9 10 9 9 9 12 8 8 6 8 9 8

105.29 85,386 76,235 139.09 87,385 96,537 73.13 106.1 131.8 423.04 42,483 139.49 91,797 60,011 44,084 198.7 100.3 192.54 37,816 28,465 96.89 16 189.25 41.1 82,475 116.92 143.11 109.95 113.11 109.93 154.21 137.86 268.6 93.37 271.16 154.11 129.01 217.75 466.63 127.98 96,642 185.4 208.59 136.94 96,905 127.29 152.53 145.93 107.89 43,819 48.92

1,959 1,153 1,416 8,879 2,296 6,654 2,052 2,457 11,398 12,824 2,913 6,037 5,529 2,479 1,124 3,322 4,363 11,707 0.621 1,536 9,831 1,176 6.01 0.471 1,586 7,819 8,799 5,887 7,924 9,261 3,115 10,177 18,799 2,231 15,325 2,905 8,266 9,416 155.19 7,839 12,184 9,115 5.59 10,168 4,376 7,791 6,821 15,467 7,566 4,994 1,459

4.15 10.36 7,709 9,922 6.6 7,087 11,786 10.85 6,862 5,329 35,737 3.51 10,138 27,543 31,438 14,533 4,656 11,237 36.04 29.31 24,713 24,713 4,862 30,475 8.9 23,783 4.55 4,456 9.48 5.05 3,512 4,775 2.45 12,155 2.5 5,829 7.7 3,337 2.5 25,133 10.83 8,578 4,689 9,122 10,808 11.15 7,925 5.65 25,412 9,833 22.5

0.71 1,137 0.772 1,302 0.656 0.91 1,442 1,018 0.923 0.438 7,229 0.545 0.704 7,178 5,335 3,884 0.525 2,434 7.85 6,091 3,859 3,859 0.313 4.31 0.903 9,833 0.43 0.374 1,315 0.297 0.449 0.587 0.208 1,387 0.26 0.409 0.828 0.318 0.228 8.25 2,506 2,017 0.463 2,243 1,638 1,903 1,435 0.974 6,532 3,155 6,806

5,887 8,973 3.7 3,333 14.15 8,612 5,071 4.26 11.8 42,857 17.2 2.61 36,625 6,686 5,075 1,078 6,289 1,775 4,325 4.42 2,275 2,275 2,087 5,537 26,975 35,833 3,175 7,333 3.85 25.1 10,363 6,962 5,487 4,082 7,643 3,214 3,937 7.2 3,387 4,478 13.67 7,556 4,422 8,178 3,783 7,787 49,625 5,767 8.35 19,589 8,575

0.87 0.855 0.423 0.284 1,624 1,376 1,004 0.436 1,712 0.857 3,208 0.4324 5,525 0.72 0.535 0.198 0.729 0.218 0.408 0.441 0.343 0.343 0.185 0.559 5,927 0.392 0.432 1,136 1,177 5.88 1,296 1,045 0.876 0.385 1,899 0.258 0.608 1,378 0.484 0.577 2,998 1,949 0.552 2,163 0.323 1,234 0.69 0.876 1,913 3,368 2,023

2,425 4,673 2,491 2,444 4,475 3,725 3,471 2.99 4.2 2,271 9.95 1.5 8 5,057 4,337 1,011 2,633 1,512 3,812 3.8 2,075 2,075 1.45 4,737 6,075 2,967 1,867 2,611 2.51 3.92 2,562 2,812 1,625 3,036 1,714 2,071 2,575 2,187 1,412 3,511 5.42 3,444 2,256 3,722 2,775 4,512 3 2,817 6.45 5,622 5.55

0.37 0.384 0.256 0.194 0.437 0.45 0.552 0.237 0.507 0.281 1,032 0.24 0.656 0.48 0.383 0.196 0.289 0.201 0.405 0.414 0.304 0.304 0.122 0.519 0.585 0.355 0.221 0.169 0.501 0.341 0.294 0.355 0.145 0.278 0.153 0.154 0.347 0.254 0.149 0.563 10,549 0.535 0.246 0.556 0.278 0.738 0.418 0.447 1,649 1,097 1,184

E. L. PUEYO ET AL.

418

Appendix 1. (cont.}

Fabric parameters

/Tmax-Fisher

Site

L

eL

F

eF

P

eP

T

eT

kl-dec

kl-inc

kl-a

kl-K

Trives T10' T104' T106 T108 Til' T112 T115 T12 T13 T134 T136 T14' T140 T141 T145 T148 T15 T151 T152 T154 T158 T159 T161 T164 T17 T172 T173 T174 T175 T176 T177 T178 T179 T19 T2 T3, T59' T6 T60' T67 T71' T78 T79' T8 T86 T89' T9 T9' T90 T94 T95

1.0512 1.0217 1.0289 1.0158 1.0525 1.0239 1.0179 1.0142 1.0335 1.0393 1.0244 1.0299 1.0347 1.0209 1.0137 1.0124 1.0411 1.0199 1.0137 1.0203 1.0101 1.0101 1.0391 1.0131 1.0229 1.0095 1.0356 1.0398 1.0181 1.0395 1.0465 1.0532 1.0656 1.0192 1.061 1.0429 1.03 1.049 1.0556 1.0169 1.0294 1.0202 1.0314 1.0333 1.0166 1.0261 1.0281 1.0432 1.0114 1.0402 1.0184

1.0021 1.0013 1.0013 1.0016 1.0019 1.0024 1.0014 1.0013 1.0022 1.0025 1.0075 1.0016 1.0026 1.0051 1.0026 1.0013 1.0027 1.0023 1.0037 1.0037 1.0016 1.0016 1.0014 1.0018 1.0012 1.0023 1.0013 1.0012 1.002 1.0037 1.0017 1.0014 1.0024 1.0013 1.0059 1.002 1.0025 1.002 1.0054 1.0041 1.0032 1.0031 1.0007 1.0031 1.001 1.002 1.0021 .0035 .0022 1.0058 1.0034

1.036 1.0249 1.065 1.0466 1.0256 1.0206 1.0479 1.0371 1.0216 1.0584 1.0447 1.0431 1.01 1.072 1.0856 1.1754 1.0312 1.122 1.0994 1.1322 1.1113 1.1113 1.1009 1.0811 1.01 1.0453 1.056 1.0283 1.0551 1.0114 1.0164 1.0386 1.0352 1.0588 1.0293 1.0851 1.0645 1.0273 1.0466 1.0598 1.0252 1.0227 1.0357 1.0433 1.0463 1.0412 1.0444 1.0438 1.0352 1.0162 1.0449

1 0015 1 0014 1 0021 1 0033 1 0021 1 0026 1 0044 1 0027 1 003 1 007 1.006 1 0029 1 0021 1 0051 1 0059 1 0027 1 0025 1 0043 1 0049 1 .0052 1 .0026 1 0026 1 0031 1.0031 1 0019 1 .0042 1 .0025 1 .0039 1 .0044 1 .003 1 .0018 1.0028 1 .0056 1 .0018 1 .0075 1 .0017 1 .0038 1 .0041 1 .0065 1 .0038 1 .0035 1 .0026 1 .0017 1 .0052 1 .0016 1 .0053 1 .0017 1 .0042 1 .0023 1 .0022 1 .0064

] .0896 ] .0475 ] .0984 ] .0658 .0809 .0454 1.0691 L.0536 L.0566 .1011 .0736 1.0749 1.0479 1.0999 1.1096 1.2143 L.0742 1.157 .126 .1696 .1365 .0769 .1485 .1039 .0344 .0592 .0948 .0701 .078 .0546 .0661 .0948 .1054 1.083 1.0951 1.1341 1.099 1.0788 1.1055 1.0828 1.0566 1.0442 1.0686 .0793 .066 .0692 .0744 .0892 .0495 1.0596 1.0668

1.0022 1.0015 1.0021 1.0035 1.0025 1.0034 1.0051 1.0028 1.0021 1.0059 1.0052 1.004 1.0039 1.0062 1.0078 1.0034 1.0042 1.0036 1.0062 1.0071 1.0029 1.0114 1.0038 1.004 1.0093 1.0051 1.0033 1.0028 1.0043 .0037 .0019 1.0018 .0034 1.0015 1.0094 1.0022 1.0051 1.0024 1.0108 1.003 1.0025 1.0034 1.002 1.0035 1.0019 1.0062 1.0017 1.0064 1.0035 1.0053 1.0066

-0.19075 0.0564 0.3562 0.4702 -0.358 -0.0835 0.4231 0.4292 -0.2308 0.1454 0.321 0.1551 -0.5744 0.5453 0.7106 0.8454 -0.1511 0.6894 0.7495 0.7134 0.8195 0.011 0.40175 0.705 -0.4212 0.6378 0.1922 -0.2066 0.475 -0.56 -0.4889 -0.1815 -0.3379 0.4864 -0.394 0.2949 0.3369 -0.3124 -0.1184 0.548 -0.0952 0.07055 0.0407 0.088 0.45575 0.17875 0.2061 -0.0173 0.5141 -0.374 0.366

0.03422 0.05044 0.02862 0.05682 0.04179 0.08179 0.05739 0.05025 0.08211 0.08327 0.18186 0.0336 0.06492 0.1019 0.04933 0.01524 0.04509 0.03882 0.06271 0.0456 0.0284 0.10803 0.01936 0.03306 0.10228 0.07722 0.02249 0.07991 0.07167 0.11436 0.04957 0.04439 0.07878 0.03455 0.11409 0.02539 0.04168 0.07954 0.05662 0.10164 0.11085 0.11553 0.02699 0.10908 0.03028 0.06903 0.0476 0.05395 0.06298 0.12596 0.1072

331 161 283 114 128 261 159 275 287 142 143 306 315 93 120 166 326 291 269 127 141 354 123 64 9 310 301 323 168 286 320 297 315 314 302 312 324 319 306 287 310 2 299 313 203 309 316 319 238 335 71

18 16 30 14 13 41 14 16 39 39 5 13 27 24 11 12 21 10 14 16 12 40 16 14 44 27 32 16 17 6 55 50 45 50 74 31 24 41 37 14 21 46 28 27 14 21 15 30 42 24 20

3.5 3.9 5.6 6.6 9.2 11.2 4.1 9.9 9.9 7.6 22.3 4 7.9 18.9 20.3 5 3.4 23.3 23.1 10.3 28.3 32.6 3 12.2 4.9 26.2 3.8 6.6 8.5 10.8 3.3 25.1 7.5 20.3 14.1 4.2 12 7.3 21 18 7.1 15.6 3.6 7 11.4 7.5 5.6 4.2 33.6 6.2 22.7

218.1 89.2 60.6 55.8 32.5 22.2 187.8 22.3 28 55 6.2 130.1 43.9 9.6 7.4 94.8 204 5.8 5.8 20.7 4.2 3.4 291.1 18.9 71.9 6.3 123.7 54.6 29.9 23.7 242.1 5.1 48.5 5.5 16.5 181.8 20.1 51 6.9 8.1 42 10.5 180.1 49.5 14.1 48.5 86.8 212.3 3.2 61.4 6.1

STATISTICAL SIGNIFICANCE OF AMS DATA IN GRANITES

419

Appendix 1. (cont.) ^max-Bingham Eigenvalues

Kmin -Fisher

Eigenvectors

Site

K1-E1 K1-E2 K1-E3 El-dec-kl

El-inc-kl

E3-dec-kl

E3-inc-kl

k3-dec k3-inc k3-a

Trives T10' T104' T106 T108 Til' T112 Til 5 T12 T13 T134 T136 T14' T140 T141 T145 T148 T15 T151 T152 T154 T158 T159 T161 T164 T17 T172 T173 T174 T175 T176 T177 T178 T179 T19 T2 T3' T59' T6 T60' T67 T71' T78 T79' T8 T86 T89' T9 T9' T90 T94 T95

0.993 0.9808 0.973 0.9722 0.9537 0.9336 0.9922 0.931 0.947 0.9736 0.7856 0.9876 0.9655 0.8618 0.8398 0.9835 0.9923 0.8495 0.7816 0.9243 0.6874 0.6099 0.9948 0.9226 0.9769 0.8039 0.9865 0.9714 0.9474 0.9369 0.9937 0.7704 0.9687 0.7329 0.9134 0.9919 0.9288 0.9703 0.8112 0.8617 0.962 0.8603 0.9913 0.9685 0.8888 0.9689 0.9825 0.9935 0.7504 0.9745 0.7853

18 16 30 14 13 41 14 16 39 39 7 13 27 25 16 12 21 10 13 16 17 39 16 14 44 27 32 16 17 7 55 44 45 49 74 31 25 41 35 13 21 45 28 27 13 21 15 30 35 24 19

241 43 115 295 326 130 4 25 109 10 240 60 63 209 28 67 184 118 35 32 47 221 31 228 170 172 200 174 38 28 194 78 85 137 63 155 85 148 85 70 83 219 197 50 301 43 188 137 40 177 177

1 58 59 76 76 37 74 51 51 39 51 60 31 44 10 37 64 80 67 17 20 38 8 76 44 56 17 71 65 60 22 33 33 40 8 56 49 48 45 72 60 38 21 12 30 11 67 60 47 64 39

93 4 16 339 18 79 257 24 92 14 238 210 122 196 29 61 215 32 40 30 51 88 6 241 178 165 196 213 14 13 101 40 113 150 49 165 95 58 57 46 56 215 198 51 299 51 89 119 46 125 174

0.0054 0.0172 0.0179 0.0242 0.0394 0.0542 0.0043 0.0602 0.0468 0.0243 0.1619 0.0094 0.0286 0.1173 0.1533 0.0163 0.0042 0.1492 0.2168 0.0521 0.3028 0.3257 0.0042 0.0726 0.0143 0.1933 0.008 0.0269 0.0518 0.0523 0.0043 0.2284 0.0256 0.2598 0.0789 0.0063 0.0625 0.0231 0.1705 0.1222 0.0235 0.1246 0.0073 0.0264 0.0937 0.0244 0.0137 0.0039 0.2319 0.0187 0.2055

0.0016 0.0021 0.0091 0.0036 0.0069 0.0122 0.0035 0.0088 0.0062 0.0021 0.0526 0.003 0.0058 0.0209 0.0069 0.0003 0.0035 0.0014 0.0016 0.0236 0.0099 0.0644 0.001 0.0048 0.0089 0.0028 0.0056 0.0017 0.0008 0.0108 0.002 0.0012 0.0057 0.0073 0.0077 0.0018 0.0087 0.0066 0.0183 0.0162 0.0146 0.0151 0.0015 0.0051 0.0175 0.0067 0.0038 0.0026 0.0176 0.0068 0.0092

331 161 283 114 128 261 159 275 287 142 141 306 315 92 121 166 326 282 273 127 144 350 123 64 9 311 301 323 168 286 320 308 315 306 302 312 323 319 311 292 310 0 299 313 203 309 316 319 260 335 71

57 72 6 71 56 50 28 49 46 36 39 22 67 26 14 52 44 61 66 26 11 16 59 77 47 59 22 47 72 24 28 22 52 17 6 54 56 11 39 58 37 42 19 15 24 33 69 58 47 66 29

k3-K

5.5 90.9 3.7 100.7 5.9 56.2 4.6 111 6.4 66.3 16.6 10.6 6.8 67.7 6.3 54.5 32.7 3.4 9.2 32.8 20 7.6 5.4 74.2 10.4 29.8 7.9 43.5 2.2 470.1 5.6 76.2 11.5 21.1 3.6 208.8 6.6 49.5 6.3 68 43.2 2.3 3.5 220.6 4.1 158.3 18.4 6 6.6 87.9 3.7 124.1 8.4 34.6 4.8 92.9 32.1 3.4 7.9 44.3 34.2 3.1 36.7 2.8 20.3 5.5 38 3 4.1 188.7 4.7 120.7 28.4 4.2 22 6.4 8.2 36 6.6 49.2 13 15.2 2.8 298.9 11.5 18.5 6.2 46.4 19.8 7.7 4.5 135 7.3 71.3 9 33.7 36.3 2.6 9.6 29.9

E. L. PUEYO ET AL.

420

Appendix 1. (cont.} A:min-Bingham Eigenvalues

Eigenvectors

Site

K3-E1 K3-E2 K3-E3 El-dec-k3 El-inc-k3

Trives T10' T104' T106 T108 Til' T112 T115 T12 T13 T134 T136 T14' T140 T141 T145 T148 T15 T151 T152 T154 T158 T159 T161 T164 T17 T172 T173 T174 T175 T176 T177 T178 T179 T19 T2 T3' T59' T6 T60' T67 T71' T78 T79' T8 T86 T89' T9 T9' T90 T94 T95

0.9833 0.9829 0.9709 0.9858 0.9772 0.8646 0.9786 0.9707 0.6623 0.9542 0.8283 0.9784 0.5343 0.9526 0.9652 0.9966 0.9795 0.9311 0.9927 0.9677 0.9777 0.4869 0.9931 0.9904 0.7905 0.9843 0.9865 0.9557 0.9827 0.6815 0.9661 0.7476 0.54 0.7343 0.6817 0.9922 0.9874 0.6749 0.7966 0.9572 0.9674 0.9043 0.9947 0.9183 0.9644 0.8333 0.9887 0.9807 0.9554 0.5832 0.9496

0.0128 0.0139 0.0184 0.0113 0.0156 0.11 0.0179 0.0233 0.3051 0.0427 0.1034 0.0141 0.4471 0.0371 0.0238 0.0025 0.0181 0.0637 0.0059 0.0296 0.0215 0.423 0.0039 0.0066 0.2045 0.0136 0.0088 0.0393 0.0152 0.2837 0.0309 0.2449 0.4552 0.2634 0.3069 0.0073 0.0107 0.3026 0.1925 0.0365 0.0168 0.0892 0.004 0.0779 0.0237 0.1594 0.0068 0.0178 0.0271 0.4086 0.0329

0.0039 0.0032 0.0107 0.0028 0.0072 0.0255 0.0035 0.0059 0.0326 0.0031 0.0682 0.0075 0.0187 0.0104 0.011 0.0008 0.0024 0.0052 0.0014 0.0027 0.0008 0.0901 0.003 0.003 0.005 0.0021 0.0047 0.005 0.0021 0.0348 0.003 0.0075 0.0048 0.0023 0.0114 0.0005 0.002 0.0225 0.0109 0.0063 0.0157 0.0065 0.0012 0.0038 0.0119 0.0073 0.0044 0.0015 0.0175 0.0082 0.0174

93 4 16 339 18 78 257 24 72 14 236 210 210 196 29 61 215 31 40 30 51 82 6 241 185 165 196 213 14 13 101 63 120 156 59 165 95 56 64 46 56 216 198 51 299 50 89 118 46 93 174

51 72 6 71 56 50 28 49 44 36 37 22 28 26 14 52 44 60 66 26 11 19 59 77 47 59 22 47 72 10 28 15 53 17 4 54 56 10 38 58 37 42 19 15 24 31 69 58 47 48 29

Lienert E3-dec-k3

E3-inc-k3 01-kl

02-kl

01-k3 /32-k3

324 177 123 143 241 263 153 118 281 153 129 119 313 327 296 312 338 266 261 243 169 352 126 135 4 287 310 313 157 282 331 185 316 341 214 342 210 315 230 285 304 342 305 313 50 308 208 340 315 338 71

22 18 70 19 26 40 24 4 42 46 21 4 23 53 14 14 30 18 18 60 68 1 17 4 42 18 44 9 14 6 50 64 36 73 85 36 16 46 51 18 27 33 40 28 39 19 10 25 1 21 22

7 8 9 19 22 22 4 17 29 16 42 7 13 61 30 8 6 10 22 15 54 55 6 33 9

5 4 6 4 9 21 1 4 12 3 13 5 6 11 11 1 4 5 4 11 11 32 4 9 4

20 38 9 22 14 38 29 22 24 31 27 80 44 17 11 8 14 30 6 12 18 53 8 22 67

6 6 7 5 9 22 4 4 13 3 15 6 6 11 10 2 4 5 3 4 3 27 6 6 5

7 11 16 24 5 28 23 57 31 11 20 14 33 69 13 29 7 14 24 12

4 4 4 19 4 9 7 24 8 3 9 13 12 8 7 10 2 5 16 8

39 42 26 48 25 21 76 29 53 40 12 29 21 17 12 18 8 32 19 76

4 9 4 6 3 7 12 12 17 4 9 14 11 11 6 6 3 5 6 7

42 9 63

9 6 11

12 29 50

9 6 14

Magnetic fabric constraints on oroclinal bending of the Texas and Coffs Harbour blocks: New England Orogen, eastern Australia CHARLES AUBOURG,1 CHRIS KLOOTWIJK2 & RUSSELL J. KORSCH3 l

Laboratoire de Tectonique, UMR 7072, Universite de Cergy-Pontoise, 8 Avenue du Pare, Le Campus, Bat I, 95031 Cergy Cedex, France (e-mail: charly.aubourg@geol. u-cergy.fr) 2 Department of Earth and Marine Sciences, Australian National University, Canberra ACT 0200, Australia ^Australian Geodynamics Cooperative Research Centre, Geoscience Australia, GPO Box 378, Canberra ACT 2601, Australia Abstract: We have carried out a magnetic fabric study on deformed, clay-rich rocks from the Carboniferous Texas beds and CofTs Harbour Association in order to test the hypothesis of oroclinal bending of the Texas and Coffs Harbour blocks of the southern New England Orogen, eastern Australia. The magnetic susceptibility is dominated by paramagnetic phyllosilicates, with site-dependent contributions from the ferromagnetic minerals pyrrhotite, magnetite and hematite. Pyrrhotite is ubiquitous in the Texas block and in the central part of the CofTs Harbour block, and may be important as an indicator of the grade of lowtemperature metamorphism. The magnetic fabric results, in general, show good agreement with structural observations in the Texas and Coffs Harbour blocks. The magnetic foliation is related to the pervasive cleavage associated with accretionary deformation prior to oroclinal bending. The magnitudes of the anisotropy parameters and the plunge of the magnetic lineation indicate an increase in the intensity of deformation from east of the Coffs Harbour orocline towards the Texas orocline. An imprint from oroclinal deformation is suggested by a significant increase in the anisotropy parameter and by the development of steeply plunging magnetic lineations towards the hinges of the Texas and Coffs Harbour oroclines. The Terrica beds from the Early Permian (Allandale) Terrica inlier in the Texas orocline also show another, pre-tilt induced, tectonic imprint. Remanence data from volcanics in the Alum Rock inlier (293 Ma) and magnetic fabric data from this inlier tentatively constrain the onset of oroclinal bending as prior to extrusion of the Alum Rocks, and the completion of bending as post-Terrica beds deposition and pre-Illawarra Reversal (est. 265 Ma).

In low-grade metamorphic rocks, magnetic fabric studies based on the anisotropy of magnetic susceptibility (AMS) can determine rapidly and accurately the preferred orientation of the magnetic carriers that define the magnetic fabric. Magnetic fabric is described by a susceptibility ellipsoid Ky, whose shape is defined by the length and orientation of each of the principal axes Kmsai > K^ni > ATmin. It has been established (Hrouda 1982; Borradaile 1988; Rochette et al 1992; Tarling & Hrouda 1993) that the orientation and length of the principal axes of the AMS ellipsoid are generally coaxial with the structural petrofabric, and related to the X > Y > Z axes of deformation. The axes of minimum susceptibility, Kmin, often group perpendicular to a foliation plane (e.g. bedding or cleavage). The axes of maximum susceptibility, Xmax, are generally parallel to a structural lineation, which often represents a stretching (X) within the bedding or within a cleavage in moderately deformed rocks (Hrouda 1982). We therefore assume in this study that the Kmax

and ^min axes equate with the magnetic lineation and the pole to the magnetic foliation respectively. For better illustration of the shape of the AMS ellipsoid, various anisotropy parameters are commonly used (Jelinek 1981, see Table 3 for definitions). The corrected degree of anisotropy (Pf) indicates the magnitude of anisotropy. The T parameter indicates the shape of the ellipsoid, with the ellipsoid being oblate for 0 < T < 1 and prolate for -1 < T < 0. Also used in this study is the mean susceptibility Km. We applied Jelinek's tensorial statistics (Jelinek 1978) to obtain tensorial mean values of AMS data, indicated by the * symbol. In metamorphosed rocks, both the deformation history and changes in magnetic mineralogy can lead to a more complex magnetic fabric. In the case of polyphase deformation, there is often no one-to-one correspondence between deformation phases and magnetic fabric (Hirt et al, 2000). Such a complication has been described by Robion et al. (1995) who showed that an early structural magnetic lineation can

From: MARTIN-HERNANDEZ, F., LUNEBURG, C. M., AUBOURG, C. & JACKSON, M. (eds) 2004. Magnetic Fabric: Methods and Applications. Geological Society, London, Special Publications, 238, 421-445. 0305-8719/04/ $15.00 © The Geological Society of London 2004.

422

C. AUBOURG ET AL.

be preserved in rocks of moderate epimetamorphic grade, despite late development of a pervasive cleavage. In addition to the introduction of multiple petrofabric patterns, metamorphism can also result in drastic changes to the magnetic mineralogy. Lamarche & Rochette (1986) pointed out effects related to the development of pyrrhotite at temperatures close to 300 °C in metapelites. They noted an increase in magnetic susceptibility accompanied by an enhancement of anisotropy parameters that directly reflect the strongly anisotropic susceptibility of pyrrhotite. This is readily understood from a comparison with magnetite which has an intrinsic, shape-controlled, anisotropy (P = Kmax/Kmin) in natural rocks of less than 5 (Uyeda et al 1963) and generally close to 1.3 (Borradaile et al. 1986), whereas the intrinsic, magnetocrystalline, anisotropy of pyrrhotite is generally higher by two orders of magnitude (P > 100) (Rochette et al. 1992). In this paper, we apply the AMS technique to moderately metamorphosed and strained shales in the southern New England Orogen of eastern Australia. We present magnetic fabric data (546 specimens, 313 samples, 28 sites) from the Texas and Coffs Harbour blocks, with the aim to test and constrain models of oroclinal bending of the southern New England Orogen.

lower greenschist facies (Korsch 1978). The second deformation, D2, although widespread, was not as intense, forming gentle flexures, kinks and chevron folds. The mesoscopic structures produced by the D} and D2 events have quite different orientations to the north and to the south of Red Rock (Fig. 2, box c). To the south, both bedding and cleavage strike approximately east-west, with the strike changing more towards NW-SE in the southernmost part of the block; younging trends for bedding are dominantly towards the north. To the north of Red Rock, the strikes of bedding and cleavage are approximately north-south; younging trends for bedding are dominantly to the west. This led Korsch (1981) to propose that the D{ and D2 structures had been folded around a complex, regional, syncline (D3) at some stage after the D t and D2 events. Structural data such as bedding and cleavage orientations and younging directions for bedding from the Solitary Islands, located offshore of the Coffs Harbour block, confirm the presence of the synclinal structure and help constrain its geometry (Korsch 1993). Korsch (1978) demonstrated that low-grade regional metamorphism accompanied the first mesoscopic deformation event (DO, followed some time later by a regional static thermal metamorphic event of post-D3 age. These metamorphic events have been dated by Graham & Korsch (1985) using the Rb-Sr whole-rock techTectonic framework nique at 318 ± 8 Ma and 238 ± 5 Ma respectively The southern New England Orogen (Fig. 1) has (Fig. 3). In the central Coffs Harbour block (Fig. 2, box been interpreted by many workers as a Late Palaeozoic volcanic arc - forearc basin - accre- b), Fergusson (1982«,/?) also documented the tionary wedge complex related to a convergent presence of two mesoscopic deformations, plate margin (e.g. Korsch et al. 1990, 1997; which he correlated with the Dj and D2 deformaMurray 1997; Murray et al. 1987; Roberts & tions of the eastern Coffs Harbour block. In this Geeve 1999). The Texas and Coffs Harbour area, however, the structures are dominated by blocks form part of the accretionary wedge and NW-SE striking bedding and cleavage. Fergushave suffered penetrative deformation. The two son also described mesoscopic folds and macroblocks are separated by the Demon Fault, a scopic anticlines, synclines and faults associated strike-slip fault active during the Triassic with the Dj deformation. The Gundahl Complex from this area described by Fergusson (1982a, (Korsch et al. 1978). In the eastern Coffs Harbour block, Korsch 1984) represents a classic tectonic melange, typi(1973, 1981) recognized three distinct deforma- cally found in subduction-related accretionary tion events. The two early events, D{ and D2, wedges. Lucas (1960) noted that the rocks in the Texas produced mesoscopic structures, whereas the subsequent D3 event is obvious only on macro- block (Fig. 2, box a) constitute a thrust pile scopic scale, producing a complex, regional- folded into a regional anticlinal structure, with scale synclinal structure (Fig. 1, Coffs Harbour younging trends predominantly outwards from orocline). The first deformation, D b produced the core. Olgers et al. (1974), in contrast, conupright folds in bedding and an associated sidered the Texas beds folded into a large synaxial-plane cleavage. Intensity of deformation clinal structure, but Butler (1974) confirmed the increases towards the south (Korsch 1973), original, anticlinal interpretation of Lucas accompanied by an increase in grade of regional (1960). The anticlinal structure and consistent metamorphism from prehnite-pumpellyite to outward younging directions were subsequently

AMS OF THE TEXAS & COFFS HARBOUR BLOCKS

423

Fig. 1. (a) Outline of the Texas, Coffs Harbour and Manning oroclines in the southern New England Orogen, including interpretation beneath younger cover, after Korsch & Harrington (1987). Heavy solid line separates the Tamworth Belt (Terrane) from the accretionary wedge succession. In places, particularly in the south, the stipple pattern for the Woolomin Association and Sandon Association also covers areas of presently exposed Permian successions. The boxes show the study areas of the Texas block (a), central Coffs Harbour block (b) and eastern Coffs Harbour block (c), see Figures Ib and 2. (b) Australian mega-elements, representing continent-scale groups of crustal elements after Shaw et al. (1995).

424

C. AUBOURG ET AL.

Fig. 2. Geology of the Texas-Coffs Harbour region (after Fergusson 1982a) and location of the sampling sites (Table 1). See Figure la for legend to boxes. The Terrica inlier is indicated by ATTO/CTTA-H.

supported by the regional mapping of Fergusson & Flood (1984), but Lennox & Flood (1997) advocate a tight to isoclinal synclinal structure. The presence of cleavage was reported by Olgers et al. (1974) and Butler (1974), who mapped different orientations in three structural domains. Butler also noted that his Dj cleavage was deformed by a later event which produced kinks and gentle mesoscopic folds, presumably equivalent to the D2 structures recorded from the Coffs Harbour block (Korsch 1973, 1975,

1981). Lennox and Flood (1997) reported two, or possibly three, post-bedding foliations: a dominant foliation (S{) that is sub-parallel to bedding, probably of subduction-accretion origin, and folded during oroclinal deformation; a patchily developed, cross-cutting, foliation (S2), probably of early oroclinal origin and subsequently fanned around the developing orocline; and a poorly developed east-west oriented foliation (83), probably formed after the oroclinal bending event.

AMS OF THE TEXAS & COFFS HARBOUR BLOCKS

425

Fig. 3. Schematic diagram showing the time of deposition of sediments, Dj and D2 deformations, and oroclinal bending following various authors.

Oroclinal bending Korsch (1975, 1978) suggested that the southern part of the Coffs Harbour block had been deformed by oroclinal bending (Fig. 1). Flood & Fergusson (1982, 1984) and Fergusson & Flood (1984) subsequently suggested that melange and tectonostratigraphic units in the Texas block could be correlated with similar units in the central Coffs Harbour block mapped by Fergusson (1982a). They suggested that the rocks of the Texas and Coffs Harbour blocks were once continuous, and that the regional anticline described by Lucas (1960) and the regional syncline described by Korsch (1975) were produced during the same deformational event. Thus, they proposed that the rocks had been folded into a Z-shaped megafold, which they termed the Texas-Coffs Harbour

Megafold, subsequently called the Texas and Coffs Harbour oroclines by Korsch & Harrington (1987). Mesoscopic structures associated with the development of the oroclines, such as minor folds or axial-plane cleavage, have not been unambiguously recognized in the field, although Lennox & Flood (1997) propose that a patchily developed D2 foliation formed during oroclinal bending. It is, therefore, likely that this deformation was not pervasive at outcrop scale. Korsch (1975) suggested that the Coffs Harbour orocline was folded around an axis plunging steeply to the NW. Flood & Fergusson (1982) suggested that the Texas orocline was folded about a NW-striking axial-plane with a near-vertical fold axis. The model of oroclinal bending in the New England Orogen was refined by Korsch & Harrington (1987) and Murray et al (1987).

426

C. AUBOURG ET AL.

Although the bending occurred principally in rocks of the accretionary wedge, Korsch & Harrington (1987) proposed that parts of the forearc basin - the Tamworth and Yarrol Belts - were involved as well. Interpretation of aeromagnetic and gravity anomalies (Wellman & Korsch 1988; Wellman 1990; Wellman et al 1994) support the outline of the oroclines consisting of both the accretionary wedge and the forearc basin succession. Thus, there is general consensus on the geometry of the oroclines. The time of bending (Fig. 3), however, remains debated although some agreement is now developing for the Early Permian age, which was previously argued for by Korsch (see Korsch & Harrington 1987; Murray et al. 1987). Murray et al (1987) and Lennox & Roberts (1988), in contrast, proposed Late Carboniferous (310300 Ma) oroclinal bending associated with 400500km dextral strike-slip displacement. Geeve et al. (2002) recently argued on stratigraphic and geological grounds for a comparable age, mid-Namurian to latest Carboniferous, for rotation of the Manning Orocline (Fig. 1) as concluded by them from their palaeomagnetic data of the southern Tamworth Belt (Terrane). Korsch and coworkers (e.g. Korsch & Harrington 1987) have argued that oroclinal bending occurred during the Early to Middle Permian (280-265 Ma). They relate this to the intense deformation of the Nambucca block (Leitch 1978; Coney et al. 1990), and infer a north-tosouth 450-500 km relative movement above a sub-horizontal decollement (Korsch et al. 1990). Fergusson & Leitch (1993) favour a slightly earlier initiation of the orocline at about 290 Ma, with development continuing until about 280 Ma. Lennox & Flood (1997) favour development during essentially the Early Permian, with pre-Late Asselian initiation and post-middle Artinskian continuation. Alternatively, a Late Permian age for oroclinal bending has been suggested by Collins (1990, 1991, 1994) and Collins et al (1993). Preliminary palaeomagnetic data from the Texas block, CofTs Harbour block and Tamworth Belt (Klootwijk 1985, 1996a, b\ Aubourg et al 1994) indicate a pervasive reverse polarity, Kiaman-type, magnetic overprint which post-dates oroclinal bending. This overprint was originally estimated to be of Late Carboniferous to Early Permian age from comparison with the Australian pole path (e.g. Klootwijk et al 1993; Klootwijk & Giddings 1993). However, the shape of the Permian part of the Australian APWP is not well defined (e.g. Giddings et al, 1994) and a younger limit of Early or Late Permian age cannot be excluded on the basis of current

knowledge. Extensive recent palaeomagnetic studies on Carboniferous and Permian volcanic successions from the northwestern Tamworth Belt (e.g. Opdyke et al 2000; Klootwijk 1996«, b, 2002, 2003) will lead to better definition of the New England pole path and may lead to a better constrained younger age limit on oroclinal bending. In summary, in the Texas and Coffs Harbour blocks, deformation can be related to two distinctive tectonic events: subduction and development of the accretionary wedge produced pervasive deformation (Dj and D2) at mesoscopic scale, followed by oroclinal bending (D3). Studies of anisotropy of magnetic susceptibility (AMS) and remanence (not presented in this paper) have been undertaken to identify and to possibly further constrain the evolution of these deformation phases. Sampling Twenty-eight sites (313 cores) were sampled in shale-rich rocks of the Texas and Coffs Harbour blocks (Fig. 1). The sites were selected for broad coverage of strike variations around the two oroclines (Fig. 2, Table 1). Seven sites were sampled in the Texas beds. This unit was derived from a volcanic arc that is presumed to have been situated to the west of Tamworth Belt and forms part of an accretionary wedge containing melange facies and highly deformed turbidites with pervasive cleavage (Flood & Fergusson 1982; Lennox & Flood 1997). Korsch & Harrington (1987) revised Korsch's (1977) interpretation of stratigraphic associations within the southern New England Orogen, and correlated part of the Texas beds with the Coffs Harbour Association (Fig. 1). Korsch (1977) suggested a Late Devonian to Early Carboniferous depositional age (Fig. 3) and this has been confirmed by radiolarian dating (Aitchison 1988; Aitchison & Flood 1990). Four sites were sampled in the central Coffs Harbour block and seven sites at headlands in the eastern Coifs Harbour block (Fig. 2), areas studied and described respectively by Fergusson (1984) and Korsch (1971, 1981). A younger age limit on the deposition of these rocks is given by a Rb-Sr whole-rock isochron date of 318 ± 8 Ma (Graham & Korsch 1985) for a regional metamorphic event in the Coffs Harbour area. One site (ATCA) was sampled east of the Demon Fault in the Late Permian (Fauna IV; Dickins & Malone 1973; Archbold & Dickins 1996) Gilgurry Mudstone (Fig. 3) (Thomson 1976) and nine sites (ATTO, CTTACTTH) were sampled in the Terrica beds

427

AMS OF THE TEXAS & COFFS HARBOUR BLOCKS Table 1 Sampling sites and structural data Site

Long (°)E

Texas block ATTA 151.43 ATTB 151.15 ATTC 151.33 ATTD 151.48 ATTE 151.53 ATTF 151.98 ATTN 151.60

Lat OS

N*

Lithology

Outcrop

ss

V

GF§ Dip (°)

Dip az (°)

225 25 undef 180 5 188 0 218 47 50 25 180 5

undef 1 275 20 197 3 26 47 255 undef undef 93 237 74 93

52 44 0 30

193 196 182 356 262 332 130

47 30 9 25 58 8 44

silicified shale silicified shale sandstone, siltstone silicified shale mudstone sandstone, siltstone silicified shale

creek road cut road cut road cut road cut field crop gully

280 252 170 undef 252 undef 336

20 27 5

Central Coffs Harbour block ATCK 152.55 29.58 ATCL 152.58 29.75 ATCI 152.43 29.86 29.84 ATCJ 152.35

12 13 10 15

silicified shale, melange massive sandstone shale siltstone

creek creek old road cut field crop

18 242 undef 237

32 16 26

undef undef 238 10 80 10

Eastern Coffs Harbour block 30.21 ATCE 153.14 ATCD 153.19 30.18 30.08 ATCC 153.20 30.06 ATCB 153.30 29.82 ATCF 153.29 ATCH 153.30 29.78 ATCG 153.34 29.61

10 9 8 13 12 13 13

mudstone soft mudstone cracked mudstone siltstone, mudstone mudstone marly mudstone soft mudstone, siltstone

Moonee Beach Emerald Beach Mullaway Headland Arawarra Headland Diggers Camp Minnie Waters Broomes Head

222 192 192 185 280 286 110

2 14 20 26 10 36 24

undef undef 170 5 180 20 undef 350 5 undef

Terrica beds ATTO 151.48

28.51

13

creek

CTTA CTTB CTTC CTTD CTTE CTTF CTTG CTTH

28.51 28.51 28.51 28.51 28.51 28.51 28.51 28.51

10 10 10 11 10 9 7 8

mudstone 1-8 sandstone 9-16 mudstone, siltstone sandstone, mudstone sandstone mudstone mudstone mudstone mudstone mudstone

creek creek creek creek creek creek creek creek

197 237 207 198 200 203 13 undef 192 196

71 72 36 45 50 59 69 90 74 74

28.92

12

gully

40

70

151.48 151.48 151.48 151.48 151.48 151.48 151.48 151.48

Gilgurry Mudstone ATCA 152.29

Dip O

Dip 0

13 9 13 12 12 14 12

28.98 28.81 28.39 28.32 28.26 28.21 28.67

Dip az(°)

Dip az(°)

50 5

30

5

*N = number of studied samples per site. ^SQ = pole to bedding plane, field measurement. *S\ = pole to cleavage plane, field measurement. § GF = indication of pole to cleavage plane, orientation of the plane determined in the laboratory from goniometer measurements on individual specimens.

(Olgers et al. 1974) of the Terrica inlier in the Texas block. The Terrica beds are Early Permian in age (Fig. 3) (most probably Allandale, i.e. Asselian to Sakmarian according to Archbold & Dickins 1996; Briggs 1998). While the sampling sites cover the study area sparsely, we consider the sampling scheme to be ample for the purpose of this study because there is a good consistency between magnetic fabric and structural data when compared, and there are significant differences between magnetic fabrics observed throughout the three blocks and also between pre-Permian and Permian rocks.

Rock magnetism We carried out saturation IRM and Lowrie (1990) tests in order to identify the minerals that contribute to the magnetic anisotropy and the degree of their contribution, and hysteresis loop determinations. For proper interpretation of AMS data it is desirable to estimate the matrix susceptibility contribution (non-ferromagnetic grains) from the ferromagnetic susceptibility contribution: both contributions can have a fundamental impact on the AMS of sedimentary rocks

428

C. AUBOURG ET AL.

(Borradaile et al 1986; Hirt et al. 2000). We can divide the many sources of low-field magnetic susceptibility (K) into those of matrix origin (Kp), comprising paramagnetic, diamagnetic and antiferromagnetic susceptibilities, and those of ferromagnetic sensu lato origin (Kf),

which are also responsible for the remanent magnetization (Borradaile & Tarling 1981; Hrouda 1982; Henry 1983; Hounslow 1985; Rochette 1987^). The magnetic susceptibility of the clay rocks studied is of the order of 1 x 10~4 to 5 x 10~4SI (Fig. 4a), except for the Gilgurry

Fig. 4. (a) Histogram of mean magnetic susceptibility (Km) for the Texas beds and the CofTs Harbour Association (Permian sedimentary rocks are not shown). The threshold at 350 x 10~6 SI indicates the lower limit for ferromagnetic susceptibility tending to dominate the low field magnetic susceptibility in shales, (b) Representative hysteresis loop showing the strong paramagnetic trend characteristic for the Texas beds and the CofTs Harbour Association.

429

AMS OF THE TEXAS & COFFS HARBOUR BLOCKS

Mudstone with susceptibility of the order of 1 x 1(T3 SI. Various studies have shown that clay-rich rocks with weak susceptibilities (typically below 350x 1(T6SI) have Kp values of the same order as their Kf values (Borradaile et al. 1986; Rochette 19870), with anisotropies that are often dominated by the matrix component, particularly within the foliation plane (Housen & Van der Pluijm 1990). We therefore also carried out studies on the behaviour of magnetically oriented rock powders following the method proposed by Aubourg et al. (1996) in order to gain an insight into the matrix susceptibility, the nature of the ferromagnetic minerals and the relative contributions of the magnetic carriers to the susceptibility anisotropy.

Matrix susceptibility Hysteresis loops for samples from the eastern Coffs Harbour block were obtained with a Micromag 2900 at the Centre des Faibles Radioactivites (CNRS, Gif sur Yvette, France). The quasi-linear hysteresis curves (e.g. Fig.4b) confirm the suspected predominance of the matrix contribution. The relative proportion of this fraction, the Kp percentage, was determined from comparison of the slopes of the high-field hysteresis curves and the low field magnetic susceptibility measured with a Kappabridge: it represents at least 60% of the low-field magnetic susceptibility for shales from the eastern Coffs Harbour block (Table 2). This

Table 2 Rock magnetic properties Site

Py*1

Mg*1"

Hem*1

Matrix (%) *

Texas block ATTA ATTB ATTC ATTD ATTE

£p0wder§

Fpowder A § A

2.0 2.0 1.8 1.9

1.3 1.3 1.3

2.0 1.6 <1.1 <1.1 1.2 1.2 1.6

1.3 1.3 1.3 1.3 1.3 1.3 1.4

1.1

ATTF

ATTN Central Coffs Harbour block ATCJ ATCI ATCL ATCK Eastern Coffs Harbour block ATCB ATCC ATCD ATCE ATCF ATCH ATCG

69 63 67 72 71 74 70

Terrica beds CTTA CTTB CTTC CTTD CTTE CTTF CTTG CTTH Gilgurry Mudstone ATCA *Py = pyrrhotite, Mg = magnetite, Hem = hematite. ^The number of * symbols reflects relative occurrence. * Matrix % = Kp/K0, Kp represents the matrix susceptibility derived from high field susceptibility, and KQ represents the initial or low field susceptibility. ^powder (^max/^int) anc^ ^powder (^int/^min) represent the lineation and foliation parameters determined from rock powder experiments (Aubourg et al.1996).

Fig. 5. Examples of three-axes stepwise thermal demagnetization of IRM following Lowrie (1990). Dots = soft IRM acquired at 120 mT, circles = medium IRM acquired at 400 mT, full squares = hard IRM acquired at 2150mT. Inset diagrams show the progression of IRM acquisition for the same specimens.

AMS OF THE TEXAS & COFFS HARBOUR BLOCKS

indicates the minor contribution of the ferromagnetic fraction in these clay-rich rocks, as we suspected. It is likely therefore that the matrix contributes significantly to the AMS of these sites.

Nature of the ferromagnetic minerals Saturation IRM acquisition and thermal demagnetization of 3-axes IRM, using the Lowrie (1990) method, was studied systematically in one representative specimen per site. The IRM acquisition graphs indicate the presence of low, medium and high coercivity minerals (Fig. 5 insets). Step wise thermal demagnetization of hard (Z = 2150mT), medium (7-400mT) and soft (X = \20mT) components of IRM help identifying ferromagnetic minerals through analysis of their maximum unblocking temperatures. Four representative specimens are illustrated in Fig. 5 (main plots), showing: (i) a spectrum of magnetite of low coercivity and a maximum unblocking temperature (Tbmax) at 580 °C (Fig. 5a); (ii) a mixture of magnetite (Tbmax = 580 °C) and pyrrhotite indicated by predominantly low to medium coercivity and a sharp decrease of IRM intensity between 320 and 340 °C (Fig. 5b); (iii) an example of pyrrhotite with low to medium coercivity and a Tbmax of 340 °C (Fig. 5c); (iv) a complex mixture of goethite (Tbmax = 80 °C), pyrrhotite (Tbmax = 340 °C), magnetite (Tbmax = 580 °C) and hematite (Tbmax = 680 °C) also showing a spectrum of low, medium and hard coercivities (Fig. 5d). We interpret the diagrams in terms of pyrrhotite rather than Ti-rich magnetite because the sharp intensity decrease is observed always within the same temperature range (320-340 °C, the Curie temperature range of pyrrhotite) and the coercivity has low, medium and hard values. Presence of other iron sulphides, such as greigite, is unlikely in metamorphosed shales (e.g. Sagnotti & Winkler 1999). By contrast, pyrrhotite is ubiquitously present in low-grade metamorphic shales (Rochette et al. 19876; Aubourg et al. 2000) and its occurrence is thus likely in our metasediments. We list in Table 2 the relative occurrence of ferromagnetic minerals for each site. It appears that pyrrhotite is predominant throughout the Texas and central Coffs Harbour blocks. Pyrrhotite seems also well developed within the hinge of the eastern Coffs Harbour block (sites ATCB, ATCC). Magnetite is abundant in the Terrica beds and in the eastern Coffs Harbour block, whereas hematite is restricted to the hinge of the Texas orocline. The recognition of pyrrhotite

431

and hematite is significant because: (i) hematite and pyrrhotite have a strong planar intrinsic anisotropy, which is two to three orders of magnitude larger than that of magnetite - one may thus observe an enhanced anisotropy that is not necessarily due to a more complete alignment of magnetic carriers; and (ii) pyrrhotite is an index mineral for epimetamorphism (Lamarche & Rochette 1986; Robion et al. 1995).

Relative contribution of magnetic carriers to anisotropy We studied the AMS properties of rock powders from selected samples for the purpose of comparing the relative anisotropy contribution of different magnetic components. We selected 11 representative samples, one per site, from four sites in the Texas block and seven sites in the eastern Coffs Harbour block, each with an anisotropy close to the site's tensorial mean. The powders, with grain sizes under 100 urn, were made to settle under the combined action of a horizontal magnetic field (0.3T) and gravity (Aubourg et al. 1996). The magnetic fabric acquired by the powder samples showed Kmax and Kmin axes that were aligned with the applied magnetic field and gravity directions respectively. The shape of their susceptibility ellipsoids is generally prolate except for some of the eastern Coffs Harbour sites (Table 2). The foliation parameter (Fpowder) shows an approximately constant value of about 1.3, whereas the lineation parameter (Lpowder) shows variations from 1.1 up to 2.0. There is a fairly close relationship between the composition of the ferromagnetic grains and the value of the ^powder parameter. Sites with magnetite as the dominant ferromagnetic carrier show a low value for the Lpowder parameter, whereas sites with pyrrhotite as the dominant magnetic carrier show a higher Lpowder value and a prolate fabric. This observation supports Borradaille et a/.'s (1986) suggestion that magnetite has a low intrinsic anisotropy. This simple experiment demonstrates the contrasting effects of magnetite and pyrrhotite even if the ferromagnetic fraction represents less than 40% of the initial susceptibility. In summary, the magnetic susceptibility of the rocks studied throughout the area is dominated by a large matrix contribution (clay minerals) that exceeds 60% of the initial magnetic susceptibility. The main ferromagnetic minerals are pyrrhotite and magnetite, with lesser hematite.

432

C. AUBOURG ET AL.

Magnetic fabric results Measurements All AMS measurements reported here have been carried out with a KLY-3S Spinner Kappabridge. The scalar and vector parameters obtained with KLY-3S are documented in Table 3 and compared in Figures 7 and 8. All samples had been measured previously with a spinner magnetometer (DIGICO) and an impedance bridge (KLY-2) (Aubourg et al 1994). The KLY-3S results consistently show good agreement with the KLY-2 results, but also some differences with the DIGICO results, in particular for the direction of the principal axes of susceptibility and the value of the lineation parameter (L). As an example we compare KLY-3S and DIGICO AMS data for sites ATCE and ATTO in Figure 6. Both sites show DIGICO magnetic lineations that are vertical and parallel to the drilling direction during sampling. In contrast, KLY-3S magnetic lineations are nearly horizontal. DIGICO and KLY-3S magnetic foliations are similar for site ATCE, but differ for site ATTO with a smearing of ^min axes. The value of the DIGICO lineation parameter (L = Kmax/Kint) is significantly larger than the KLY-3S value, although the foliation parameter (F = Kini/Kmin) has comparable values. These differences probably stem from the DIGICO's specimen-shape sensitivity

(Veitch et al. 1985; Heller & Schultz-Krutisch 1988) and from the higher frequency of its applied magnetic field (10 kHz for DIGICO: ~lkHz for KLY-2/KLY-3S) that tends to reduce the contribution of the coarser grains (Bloemendal et al. 1985). Overall comparison of the data obtained by the two instruments indicates the DIGICO's Kmax axis to deviate towards the specimen's long axis. Coffs Harbour block Central part Fergusson (1982<2, b) presented detailed structural data for this area that can be compared directly with the AMS axes (Fig. 7a). Sites ATCJ, ATCI and ATCL lie within structural domains 3, 7 and 5 respectively as defined by Fergusson. There are no published structural data available for site ATCK that can be used for comparison. The magnetic foliations are well defined, except at site ATCL where the Km^n axes girdle along a great circle. Magnetic foliations for sites ATCK and ATCL are distinctly oblique to bedding, cleavage or (goniometer) foliation, and for sites ATCI and ATCJ are steep and parallel to the regional cleavage Si. Orientation of the magnetic lineation varies from oblique to down-dip with respect to the magnetic foliation. The magnetic lineation at site ATCL is taken as the pole of the great

Fig. 6. Comparison of magnetic fabric results for two representative sites using the DIGOCO and KLY-3S apparatus respectively. In both examples, drilling was performed close to vertical in geographic coordinates. Note the distinct magnetic lineation (Kmax) obtained by the DIGICO and the KLY-3S (site ATTO) and also the considerable distribution of Kmin axes obtained by the KLY-3S for site ATTO. Equal area projection on the lower hemisphere.

Table 3. AMS mean tensorial data Site

Mean tensorial (*) susceptibility & anisotropy 1 2 6 7 3 4 5

^min

r

r*

^max norm. Dec* (°)

Inc* (°) Dp* 0

Dm* (°)

^min* norm.

Dec*8 (°)

Inc*8 (°) Dp*9 (°) Dm*9 (°)

1.0238 1.0159 1.0066 1.0167 1.0341 1.0127 1.0117

-0.3232 0.2983 0.8051 0.8359 0.4505 0.6817 -0.0402

1.0198 1.0202 1.0245 1.0712 1.0519 1.0306 1.0113

45 350 76 277 342 181 265

63 49 64 83 39 31 5

7 8 13 16 21 5 10

4 5 5 2 5 3 5

0.9841 0.9755 0.9577 0.8753 0.9310 0.9518 0.9890

235 248 181 17 231 77 355

27 10 7 1 23 22 4

15 14 6 8 8 4 9

6 4 2 2 5 4 7

1.0251 1.0153 1.0139 1.0151 1.0565 1.0251 1.1241 1.0175

0.2402 -0.0408 0.3784 0.7419

1.0184 1.0147 1.0349 1.0500

161 302 188 355

31 43 84 35

8 8 11 7

6 4 3 2

0.9785 0.9858 0.9556 0.9180

4 203 63 85

57 9 4 1

10 12 6 3

5 5 3 1

1.020 1.0387 1.0028 1.028 1.0509 1.0096 1.024 1.0262 1.0284 1.042 1.0700 1.0233 1.022 1.0491 1.0367 1.025 1.0437 1.0104 1.011 1.0180 1.0062

0.8644 0.6780 -0.0395 0.4924 0.5018 0.6115 0.4819

1.0144 1.0228 1.0274 1.0378 1.0200 1.0211 1.0101

304 88 81 93 112 141 220

15 6 26 9 55 63 9

17 17 7 3 33 21 13

5 8 6 3 5 5 8

0.9740 0.9641 0.9735 0.9479 0.9722 0.9683 0.9861

214 178 180 2 291 300 124

1 1 17 7 35 25 34

8 13 6 3 16 13 12

4 9 3 2 4 4 8

1.006 1.002 1.006 1.006 1.005 1.003 1.005 1.005 1.005

1.0026 1.0101 -0.5848 0.5014 1.0037 1.0012 1.0032 1.0094 -0.4888 1.0020 1.0104 -0.6824 1.0029 1.0077 -0.4487 1.0003 1.0066 -0.9055 1.0027 1.0081 -0.5024 1.0047 1.0062 -0.1380 1.0035 1.0069 -0.3295

1.0076 1.0020 1.0073 1.0076 1.0061 1.0045 1.0062 1.0057 1.0057

96 291 298 303 297 283 284 284 271

4 7 6 12 1 1 2 5 3

4 22 4 4 4 6 5 5 5

2 5 2 2 2 2 3 3 5

0.9949 0.9971 0.9947 0.9952 0.9955 0.9976 0.9956 0.9948 0.9954

188 52 198 157 204 189 193 193 181

34 78 56 75 67 67 14 10 0

17 8 15 31 9 61 12 8 8

2 5 2 2 2 4 4 4 5

1.040

1.0410

1.0466

352

5

4

3

0.9571

88

53

4

3

N sp ^ (xio" ) p * Texas block ATTA 37 281 ATTB 9 200 ATTC 19 144 ATTD 14 203 ATTE 17 322 ATTF 22 233 ATTN 20 239 Central Coffs Harbour block ATCK 17 145 ATCL 25 260 ATCI 14 295 ATCJ 30 321 Eastern Coffs Harbour block ATCE 12 83 ATCD 10 106 ATCC 10 285 ATCB 30 254 ATCF 15 235 ATCH 15 211 ATCG 16 124 Terrica beds ATTO 19 172 CTTA 18 169 CTTB 19 142 CTTC 24 108 CTTD 31 181 CITE 16 211 CTTF 12 184 CTTG 12 195 CTTH 12 184 Gilgurry Mudstone ATCA 51 861

#"*

-^max

F*

1.020 1.0121 1.020 1.0296 1.033 1.0628 1.102 1.2037 1.056 1.0926 1.038 1.0691 1.010 1.0108 1.018 1.013 1.036 1.066

1.0504

-01012

8

7

8

9

9

N = number of specimens studied. K^ = mean susceptibility, (/Cax + *int + ^min)/3, in 10"6. V* = corrected degree of anisotropy, exp(v/(2((?71 - 7?m)2 + (r?2 - ry m )+ (r?3 - r/ m ) 2 ))) with r^ = In K, and ?7m = (771 + r/2 + %)/3. F* — foliation parameter, K*nt/K^in. L* = lineation parameter, K^ax/K*nt. 6 T* = shape of AMS ellipsoid parameter, 2(ln^ ax - In K*nt) / (In K*ni - In^in) - 17 ^max norm., A^in norm, — normalized values of mean tensorial eigenvalues representing maximum and minimum susceptibility. 8 Dec, Inc = mean tensorial eigendirections representing principal susceptibility axes. 9 dp, dm = half angles representing the ellipse of 95% confidence around the mean tensorial principal susceptibility axes. 2

7

434

C. AUBOURG ET AL.

circle formed by the scattered Kmin axes. In other words, the magnetic lineation of specimens from this site is aligned along the intersection of individual magnetic foliations, and is also aligned along the regional fold axes (Fergusson 1982<2, b). At sites ATCI and ATCJ, the magnetic lineation is closely parallel to the intersection of magnetic foliation and cleavage. Note that this magnetic lineation is not parallel to the fold axes. At site ATCK, the magnetic lineation is not aligned with any plane intersection. Scalar data show that the shape of the AMS ellipsoid is oblate to triaxial and that the degree of anisotropy is less than 1.2 (Fig. 8a). There is a rough positive relationship between Km and P1 (Fig. 8b, Table 3). A high magnitude of Pf is also correlated with the occurrence of pyrrhotite. The mean degree of anisotropy for all rocks is P'* = 1.078 ± 0.058. From site ATCK (P7* = 1.041, 7 = 0.24) towards site ATCJ OP'* = 1.157, r = 0.74) (Table 3), there is a southward enhancement of the degree of anisotropy P7* and a more pronounced oblateness. This southward enhancement is also correlated with an ~80° clockwise rotation of the strike of the magnetic foliation (Fig. 8c).

Eastern part Structural data (Korsch 1975) together with AMS axes are shown in Figure 7b. Magnetic fabric is well defined for all sites. The magnetic foliations are aligned either close to bedding (ATCH, ATCE), between bedding and cleavage (ATCG, ATCF, ATCB, ATCC), or oblique to bedding and cleavage (ATCD). The strike of the magnetic foliation follows the regional structural bending from east-west to NE-SW (Fig. 2). There is a general agreement between the magnetic lineation and the fold axes, although field observations for the latter show a wide spread. The magnetic lineation is horizontal for east-west trending sites (ATCB to ATCE) and for site ATCG, but the magnetic lineation is steeply dipping at sites ATCF and ATCH where the trends of bedding, cleavage and magnetic foliation are changing. In contrast to the sites from the central part of the Coffs Harbour block, the shape of the AMS ellipsoids for the sites from the eastern part is strongly oblate, apart from site ATCC, and the value of the degree of anisotropy parameters is low (P7* < 1.1) (Fig. 8a). The P7* and K^ values are roughly correlated (Fig. 8b). Low and high P7* values are found preferentially in magnetiterich or pyrrhotite-rich rocks. The overall value of P7, meaned over all specimens, is

1.057 ±0.022. A rough relationship between the shape of the megafold and the value of P7 is suggested, with the more pronounced value for site ATCB close to the hinge of the orocline (Fig. 8c).

The Texas block We use structural data from the Texas block provided by Butler (1974). He divided the region into three structural domains, with significant change in orientation of the cleavage plane between each domain. We present here AMS data from (i) the Texas beds, which we consider to pre-date oroclinal bending and to represent the D! deformation, and from (ii) the Early Permian Terrica beds and the Late Permian Gilgurry Mudstone, which we consider to be of syn-(possibly post-)oroclinal and post-oroclinal bending origin respectively. Texas beds Butler's (1974) structural data (poles to cleavage) are compared with the magnetic fabric data in Figure 7c. There is generally good agreement throughout the Texas block between the regional cleavage and the magnetic foliation. This suggests that the magnetic fabric is tectonic in origin. Only at sites ATTE and ATTF are the magnetic foliations distinctly oblique to bedding or cleavage. The magnetic lineations are well defined. Five sites (ATTE, ATTD, ATTC, ATTN, ATTA) show agreement between the magnetic lineation and the intersection between the magnetic foliation and bedding or cleavage, or both. The steep magnetic lineations (inclination > 60°) in sites ATTA, ATTC and ATTD are particularly noticeable, with site ATTA showing a prolate anisotropy (Table 3), unlike other sites in the Texas block, and site ATTD showing the strongest degree of anisotropy (P7* = 1.250). The shape of the AMS ellipsoids is oblate (five sites out of seven) to prolate (Fig. 8d). The P7 values exceed those observed in the Coffs Harbour sites (Fig. 8e). However, as previously noted for the Coffs Harbour sites, there is an apparent relationship between the occurrence of strongly anisotropic minerals such as pyrrhotite or hematite and the highest values of the degree of anisotropy parameter P7 (Fig. 8e). Likewise the lowest values of P7 are observed in sites where magnetite is the principal ferromagnetic mineral (site ATTN). The overall degree of anisotropy P7 for all Texas beds specimens is 1.094 ±0.079. Figure 8f shows that the largest degree of anisotropy is recorded for strikes of the magnetic foliation between N100-140°E,

AMS OF THE TEXAS & COFFS HARBOUR BLOCKS

435

Fig. 7. Comparison between mesoscopic structural observations (left-hand figures) and magnetic fabric data obtained with the KLY-3S Spinner Kappabridge (right-hand figures). All figures are equal area projections on the lower hemisphere. Structural data (left): pole to cleavage (Si) = full circle; fold axis (FAj) = full square; pole to bedding (So, Figure 7D only) = full circle, (a) Central Coffs Harbour block, structural data according to Fergusson (1982b). No structural data for site ATCK; (b) Eastern Coffs Harbour block, structural data according to Korsch (1975). The solid lines represent contours of poles to cleavage (number of poles and contour levels indicated, 1% counting area).

436

C. AUBOURG ET AL.

Fig. 7. (c) Texas block, structural data according to Butler (1974). These are presented as data for Domain 1 (sites ATTE, ATTB, ATTF, ATTA) and for Domain 2 (site ATTD, ATTN). Site ATTN falls within Domain 1, but it is very close to domain 2. Its Km[n axes are more akin to pole of cleavage of Domain 2. We have no structural data for site ATTC. (d) Terrica beds and Gilgurry Mudstone, structural data for the Terrica beds according to Olgers et al (1974), Lucas (1957), P. G. Flood (pers. comm. 1993), Klootwijk (unpublished), and for the Gilgurry Mudstone according to Thomson (1976). AMS data (right): Kmax, K-mi and ^min are the principal axes of susceptibility for individual specimens, and K^ax, K*ni and A^in are the principal axes of susceptibility determined from tensorial mean results for individual sites. Bedding (So) and cleavage (S\) orientations determined by us in the field are also indicated on these AMS plots (right), whereas cleavage plane orientations estimated from goniometer measurements on specimens are listed in Table 1.

AMS OF THE TEXAS & COFFS HARBOUR BLOCKS

437

Fig. 8. Overview of (a, d) the mean shape parameter (T*) versus the mean degree of anisotropy (P7*) for sites across the eastern and central Coffs Harbour block, the Texas block and the Terrica beds, (b, e) P7* versus the mean susceptibilility, with the main ferromagnetic mineral indicated: pyr: pyrrhotite, mag: magnetite, hem: hematite, (c, f) P7* versus the strike of magnetic foliation. Error bars indicate the standard deviation from the arithmetical mean of P7*.

which is perpendicular to the inferred axial zone of the orocline. Terrica beds and Gilgurry Mudstone The Terrica beds and Gilgurry Mudstone are far less deformed than the surrounding Texas beds, yet the magnetic fabric of these Permian units is tectonic in origin. For the Terrica beds and Gilgurry Mudstone, the magnetic lineations C^max axes) are well grouped parallel to the

bedding strike but the Km^n axes are not always perpendicular to the bedding (Fig. 7d, Table 3). Furthermore, the AMS ellipsoids are generally prolate (L* > P1*), indicating a tectonic fabric (e.g. Lee et al 1990; Fig. 8d, Table 3). The degree of anisotropy is weak for the Terrica beds (overall P7 = 1.019 ± 0.028), as commonly observed in magnetite-bearing clastic sedimentary rocks (e.g. Kanamatsu et al. 2001), and stronger for the Gilgurry Mudstone

438

C. AUBOURG ET AL.

Discussion

Fig. 9. Overview of Terrica beds AMS results (CTTA-H, ATTO). Density diagram (a 1% counting area was applied and contours are shown at 2% intervals) of Kmax and Kmin axes. Lower hemisphere equal area projection (a) Geographic coordinates; (b) Stratigraphic coordinates, after restoring the bedding to the horizontal.

(Pf* = ^1.1). The Terrica beds show a main and a subsidiary cluster of ^min axes (Figs. 7, 9). Correction for bedding (Fig. 9) improves the groupings, bringing the main cluster close to vertical and the subsidiary cluster close to horizontal. There are therefore occurrences of two main magnetic foliations: one is parallel to bedding, while the other is perpendicular to bedding. While magnetic foliation parallel to bedding originates from diagenesis and post-diagenesis processes (vertical compaction), the magnetic foliation perpendicular to bedding suggests a pre-tilting origin, and is possibly related to an early layer parallel shortening (horizontal compaction). The steep magnetic foliation probably corresponds to the east-west oriented incipient cleavage that Lennox & Flood (1997) observed in the Texas beds and in the Permian basins and regarded as of post-oroclinal bending origin. Examples of technically induced magnetic foliation in weakly deformed rocks are sparse. Averbuch et al. (1992) have suggested the presence of a relict technically induced magnetic foliation in weakly to moderately deformed red sandstone with obvious spaced cleavage. Development of technically induced, magnetic foliation without evidence of cleavage has been well documented by Hounslow (1990) and Housen et al. (1996), but this relates to late development of magnetic foliation during or after tilting of the sediments. Aubourg & Robion (2002) documented orthogonal magnetic foliations in elastics, that are pre-tilting in origin and without evidence of incipient cleavage. The ability to determine magnetic fabrics acquired during the early stages of deformation in weakly deformed rocks is testament to the resolving power of the AMS method as implemented in the KLY-3S Kappabridge.

Two major Late Palaeozoic tectonic events have been recognized in the New England Orogen: subduction-related accretion (Dj) and oroclinal bending (D3). Outcrop-scale structures, however, have been recognized for only one of the two events, the deformation that occurred during subduction and accretion (Korsch 1981). The geometry of the orocline is indicated by regional-scale patterns, such as outcrop geology, continuous change in regional strike patterns of bedding and cleavage, and gravity and aeromagnetic data. Outcrop-scale deformation related to oroclinal bending, however, has yet to be recognized. There is no obvious field evidence for structures that could be related to this event, such as axial-plane cleavage or a stretching lineation, although Lennox & Flood (1997) suggest that their S2 (D2) cleavage in the Texas block may have developed during the oroclinal bending. In the context of such a polyphase deformational history, what can be expected from magnetic fabric studies? We discuss below the importance of the magnetic mineralogy and try to integrate the AMS results that we have presented into a comprehensive tectonic scheme. The importance of magnetic mineralogy We have demonstrated for the clay-rich rocks of relatively low susceptibility (1 x 10~4 to 5 x 10~ 4 SI) that the magnetic susceptibility is dominated by the matrix. Results from previous studies (Aubourg et al. 1995) suggest that the magnetic fabric and the magnetic foliation in particular are likely also to be controlled by the matrix, i.e. by the texture of the paramagnetic clays. The contribution of ferromagnetic carriers to AMS is, however, not negligible as confirmed by our rock powder analysis. We noticed that the highest degrees of anisotropy are observed in pyrrhotite-bearing rocks in the Texas beds. Also, the southward increase in mean susceptibility K^ and anisotropy P7* in the central Coffs Harbour block can be related to the southward increase in development of the strongly anisotropic magnetic mineral pyrrhotite. What then is the meaning of the observed southward increase of P7* in the central CofTs Harbour block? We believe that this increase is due to both an increase in deformation and a change in magnetic mineralogy, because the enhanced magnetic fabric with better grouping of AMS axes and a steep magnetic lineation reflects an increase in deformation. Therefore we believe

AMS OF THE TEXAS & COFFS HARBOUR BLOCKS

that it is not possible to determine the degree of finite deformation solely on the basis of AMS parameters (e.g. Hrouda 1992): account must be taken of the magnetic mineralogy and we note that there is future potential to correct the anisotropy parameters for enhancement by it. The experiment with the magnetically oriented rock powders attempts to do so, although the mechanism for orientation of magnetic carriers in rock powders differs from that of natural rock. We cannot exclude the possibility of a mineralogical origin for the large values of the anisotropy parameter within the hinge zone of the Texas orocline in particular. However, the similarity, from one block to the other, of mean susceptibility (K^) values and magnetic mineralogy, the large presence of pyrrhotite in particular, suggests that the overall trend of the anisotropy parameter reflects a regional variation in the intensity of deformation. Rock magnetic studies indicate the widespread presence of pyrrhotite. It has been established from studies in the Western Alps (Lamarche & Rochette 1986; Rochette 1987^,6) and in the Ardennes (Robion et al. 1995), that pyrrhotite is an index mineral for low-grade metamorphism characterized by low temperature (~300 °C) and low-to-medium pressure. Our identification of pyrrhotite thus supports petrological arguments (Korsch 1978) for low-grade metamorphism of prehnite-pumpellyite to lower greenschist facies. The regional variation in magnetic minerals warrants further discussion: (i) hematite is conspicuous within the hinge of the Texas orocline (sites ATTC, ATTD); (ii) pyrrhotite is less predominant in the eastern Coffs Harbour block, and its occurrence seems to be restricted to the hinge area of the orocline (sites ATCB, ATCC).

AMS and the regional trend of Dj deformation The New England Orogen shows a strong D! deformation that was acquired during development of the accretionary wedge. The D3 deformation is attributed to oroclinal bending. The Carboniferous formations of the Texas beds and Coffs Harbour Association show magnetic foliation patterns that are related to the D! cleavage, suggesting that the total magnetic fabric, foliations and lineations, may be primarily controlled by the Dj deformation. The overall pattern of AMS data suggests an increase in deformation from east to west. In this context we will discuss (i) the trend of the magnitude of Pf, (ii) the geometry of the magnetic lineation within the magnetic foliation,

439

Fig. 10. The mean degree of anisotropy (P7*) and its standard deviation for the various areas studied. The westward increase is related to a stronger deformation, a better organization of phyllosilicates, and the appearance of strongly anisotropic minerals such as pyrrhotite.

and (iii) the obliquity of the magnetic foliation with respect to bedding and cleavage: (i) The average value of the degree of anisotropy parameter P', while primarily controlled by the texture of the phyllosilicates, increases westwards from the Coifs Harbour Association to the Texas beds (Fig. 10). The magnitude of ff is partly controlled by the occurrence of pyrrhotite and hematite, but the appearance of these minerals can be also related to an increase in metamorphism, and in turn, to a more complete reorganization of the phyllosilicates. This reorganization of minerals and the emergence of more anisotropic minerals results in higher P1 values; (ii) The magnetic lineation is essentially parallel to the strike of the magnetic foliation in the eastern Coffs Harbour block (5 out of 7 sites), but is oblique to down-dip in the Texas beds (5 sites out of 7). All east-west trending sites from the eastern Coffs Harbour block have a horizontal magnetic lineation, while the two down-dip magnetic lineations are located close to the hinge of the megafold. In the central Coffs Harbour block, 3 out of 4 sites have a down-dip to oblique (>30°) magnetic lineation. In the Texas megafold, down-dip magnetic lineations occur close to the hinge of megafold. Hirt et al. (2000) observed comparable behaviours in the Asturian-Iberian Arc, and proposed that down-dip magnetic lineations in cleaved slates reflect true stretching lineations and strain heterogeneity. Following their observation, we propose that the development of down-dip magnetic lineations within the Texas beds and the Coffs Harbour Association reflects higher intensities of deformation;

440

C. AUBOURG ET AL.

(iii) Regarding the obliquity of the magnetic foliation with respect to bedding and cleavage, we note that 4 out of 7 sites in the eastern CofTs Harbour block show a magnetic foliation between the bedding and the cleavage. AMS is primarily controlled by the phyllosilicates. One can interpret this behaviour therefore as incomplete reorganization of the phyllosilicates parallel to the cleavage. In the eastern CofTs Harbour block, sites with east-west trending magnetic foliations show this magnetic foliation to be located either between bedding or cleavage, or parallel to bedding in southernmost site ATCE.

Relationship of metasediment AMS with the or o dine In addition to the primary relationship between the magnetic fabric and the Dj deformation there is also evidence for a relationship with the D3 deformation: The highest values of the anisotropy parameters in the Texas block occur within the hinge zone of the Texas orocline, that is in site ATTD and less so ATTE (Fig. 8f, Table 3). These sites are, however, also characterized by higher susceptibility (up to 800 x 10"6 SI in site ATTE) and the local presence of hematite. It is possible, therefore, that the higher values of the foliation parameter partly represent differences in magnetic mineralogy. There are good indications, however, that increasing values of the foliation parameter are related, to a significant degree, to an increase in deformation as may be concluded from the prolate shape of the magnetic fabric in the centre of the Texas orocline (site ATT A). We propose, therefore, that the Texas orocline originated as a buckling fold (e.g. Nicolas 1987) where the deformation is concentrated in the hinge zone, with oblate (ATTD) and prolate (ATTA) fabrics in the outer and inner parts respectively; The relationship between the strike of the magnetic foliation and the degree of anisotropy (Fig. 8f) in the four central Coffs Harbour block sites suggests a possible relationship between the D3, oroclinal bending, deformation and increasing values of P7; The eastern CofTs Harbour block sites show different magnetic fabric trends for the sites with east-west and NNE-SSW trending structures. Sites within the hinge of the megafold show higher values of Pf and also the occurrence of down-dip magnetic lineations.

Deformation of Permian sediments The Permian Terrica beds and the Gilgurry Mudstone show a tectonic magnetic fabric, characterized by prolate ellipsoids (L* > F*), horizontal magnetic lineation patterns, and oblique magnetic foliation with respect to the bedding. The magnetic fabric of the Terrica beds is of particular significance in that some magnetic foliations are strongly oblique to bedding and likely of pre-tilting origin. In the Texas block, it is tempting to relate east-west magnetic foliations to the post-oroclinal S3 documented by Lennox & Flood (1997). However, one can relate the pre-tilting AMS imprint deformation to oroclinal deformation (D3). We argue in the next section that AMS imprint is more likely related to D3 oroclinal deformation.

Timing of oroclinal bending: magnetic constraints As discussed above, estimates of the time of bending of the Texas and Coffs Harbour oroclines range from latest Carboniferous to Late Permian. Permian inliers in the Carboniferous Texas beds potentially hold important clues to the time of bending. Two of the inliers - Alum Rock and the Terrica beds - have been studied by palaeomagnetic and magnetic fabric methods. Preliminary palaeomagnetic results from the Alum Rock volcanics (Aubourg et al 1994) of Sakmarian age (293 Ma: Roberts et al. 1996) indicate that oroclinal rotation had already proceeded over -40° at the time of their extrusion, with a further —80° of rotation afterwards. Results from the Terrica beds of most probably Allandale age (Briggs 1998), show complete magnetic overprinting by a 'Kiaman' reverse polarity component (Klootwijk 1996Z?). Directional consistency of such 'Kiaman' overprints across the Texas and Coffs Harbour oroclines (Klootwijk 19966) shows their acquisition to post-date oroclinal bending. These primary and overprint palaeomagnetic results constrain the completion of oroclinal bending to between the Sakmarian extrusion of the Alum Rock volcanics and the end (marked by the Illawarra Reversal) of the 'Kiaman' Permo-Carboniferous Reverse Superchron (PCRS), estimated at about 265 Ma, with estimates ranging from MidUfimian to Mid-Tatarian (e.g. Opdyke & Channell 1996). Field observations of structural deformation of the Texas beds and Permian inliers showed the presence of an east-west oriented foliation that was tentatively dated as post-oroclinal

442

C. AUBOURG ET AL.

AUBOURG, C. P. & ROBION, P. 2002. Composite ferromagnetic fabrics (magnetite, greigite) measured by AMS and partial AARM in weakly strained sandstones from western Makran. Geophysical Journal International, 151, 729-737. AUBOURG, C., KLOOTWIJK, C. T. & KORSCH, R. J. 1994. Magnetic fabric and palaeomagnetic studies in the Texas and Coffs Harbour blocks, New England Orogen. Australian Geological Survey Organisation Record, 1994/58, 87p. AUBOURG, C., ROCHETTE, P. & BERGMULLER, F. 1995. Composite magnetic fabric in weakly deformed black shales. Physics of the Earth and Planetary Interiors, 87, 267-278. AUBOURG, C., BAKHTARI, H. & ROCHETTE, P. 1996. Magnetic fabric of magnetically oriented rock powders. Geophysical Research Letters, 23, 19771980. AUBOURG, C., HEBERT, R. JOLIVET, L. & CARTAYRADE, G. 2000. The magnetic fabric of metasediments in a detachment zone: the example of Tinos Island. Tectonophysics, 321, 219-236. AVERBUCH, O., FRIZON DE LAMOTTE, D. & KISSEL, C. 1992. Magnetic fabric as a structural indicator of the deformation path within a fold thrust structure: a test case from the Corbieres (NE Pyrenees, France). Journal of Structural Geology, 14, 461— 474. BLOEMENDAL, J., BARTON, C. E. & RADHAKRISHNAMURTY, C. 1985. Correlation between Rayleigh loops and frequency dependent and quadrature susceptibility: application to magnetic granulometry of rocks. Journal of Geophysical Research, 90, 8789-8792. BORRADAILE, G. J. 1988. Magnetic susceptibility, petrofabrics and strain. Tectonophysics, 156, 1-20. BORRADAILE, G. & TARLING, D. H. 1981. The influence of deformation mechanisms on magnetic fabrics in weakly deformed rocks. Tectonophysics, 77, 151168. BORRADAILE, G., MOTHERSILL, J., TARLING, D. & ALFORD, C. 1986. Sources of magnetic susceptibility in a slate. Earth and Planetary Science Letters, 76, 336-340. BRIGGS, D. J. C. 1998. Permian Productidina and Strophalosiidina from the Sydney-Bowen Basin and the New England Orogen: systematics and biostratigraphic significance. Association of Australasian Palaeontologists Memoir, 19, 1-258. BUTLER, H. R. 1974. The geology of the StanthorpeLeyburn mineral district, Queensland and New South Wales. Unpublished BSc. (Rons.) thesis, Department of Geology, University of New England, Armidale. COLLINS, W. J. 1990. A reinterpretation of Permian tectonism in the Sydney Basin and southern New England Fold Belt. Proceedings of the 24th Newcastle Symposium on Advances in the Study of the Sydney Basin. Department of Geology, University of Newcastle, Publication 344, 8-16. COLLINS, W. J. 1991. A reassessment of the 'HunterBowen Orogeny': tectonic implications for the southern New England Fold Belt. Australian Journal of Earth Sciences, 38, 409-423.

COLLINS, W. J. 1994. Escape tectonics and the redistribution of Permian basins in the southeastern New England Fold Belt. Proceedings of the 28th Newcastle Symposium on Advances in the study of the Sydney Basin. Publication of the Department of Geology, University of Newcastle, Publication 606, 72-77. COLLINS, W. J., OFFLER, R., FARRELL, T. R. & LANDENBERGER, B. 1993. Recent advances in understanding the geological development of the New England Province of the New England Orogen. In: FLOOD, P. G. & AITCHISON, J. C. (eds) New England Orogen, eastern Australia. University of New England, Armidale, 69-84. CONEY, P. J., EDWARDS, A., HINE, R., MORRISON, F. & WIDRIM, D. 1990. The regional tectonics of the Tasman orogenic system, eastern Australia. Journal of Structural Geology, 12, 519-543. DICKINS, J. M. & MALONE, E. J. 1973. Geology of the Bowen Basin, Queensland. Bureau of Mineral Resources Geology and Geophysics, 130, 1-154. FERGUSSON, C. L. 19820, Structure and tectono-stratigraphy of the Central Coffs Harbour Block, eastern New England Fold Belt. Unpublished Ph.D. thesis, Department of Geology, University of New England, Armidale. FERGUSSON, C. L. 19826. Structure of the Late Palaeozoic CofTs Harbour Beds, northeastern New South Wales. Journal of the Geological Society of Australia, 29, 25-40. FERGUSSON, C. L. 1984. Tectono-stratigraphy of a Palaeozoic subduction complex in the central CofTs Harbour Block of north-eastern New South Wales. Australian Journal of Earth Sciences, 31, 217-236. FERGUSSON, C. L. & FLOOD, P. G. 1984. A Late Palaeozoic subduction complex in the Border Rivers area of southeast Queensland. Proceedings of the Royal Society of Queensland, 95, 47-55. FERGUSSON, C. L. & LEITCH, E. C. 1993. Late Carboniferous to Early Triassic tectonics of the New England Fold Belt, eastern Australia. In: FLOOD, P. G. & AITCHISON, J. C. (eds) New England Orogen, eastern Australia. University of New England, Armidale, 53-59. FLOOD, P. G. & FERGUSSON, C. L. 1982. Tectono-stratigraphic units and structure of the Texas-CofTs Harbour region. In: FLOOD, P. G. & RUNNEGAR, B. N. (eds) New England Geology: proceedings of the Symposium on the Geology of the New England Region. University of New England, Armidale, 71-78. FLOOD, P. G. & FERGUSSON, C. L. 1984. The geological development of the northern New England Province of the New England Fold Belt. In: HERBERT, H. K. & RYNN, J. M. (eds) 1984 Field Conference. Geological Society of Australia: Queensland Division, Brisbane, 1-19. GEEVE, R. J., SCHMIDT, P. W. & ROBERTS, J. 2002. Paleomagnetic results indicate pre-Permian counter-clockwise rotation of the Southern Tamworth Belt, southern New-England Orogen, Australia. Journal of Geophysical Research, 107, 2196-2210. GIDDINGS, J. W., THEVENIAUT, H. & KLOOTWIJK, C. 1994. Late Permian APWP. AGSO94, 105-106.

AMS OF THE TEXAS & COFFS HARBOUR BLOCKS GRAHAM, I. J. & KORSCH, R. J. 1985. Rb-Sr Geochronology of coarse grained grey wackes and argillites from the Coffs Harbour Block, eastern Australia. Chemical Geology (Isotope Geoscience Section), 58, 45-54. HELLER, F. & SCHULTZ-KRUTISCH, T. 1988. Comparative measurements of magnetic susceptibility anisotropy in red sandstones. Physics of the Earth and Planetary Interiors, 51, 320-325. HENRY, B. 1983. Interpretation quantitative de 1'anisotropie de susceptibilite magnetique. Tectonophysics, 91, 165-177. HlRT, A. M., JULIVERT, M. & SOLDEVILA, J. 2000.

Magnetic fabric and deformation in the NaviaAlto Sil slate belt, northwestern Spain: Tectonophysics, 320(1), 1-16. HOUNSLOW, M. W. 1985. Magnetic fabric arising from paramagnetic phyllosilicate minerals in mudrocks. Journal of the Geological Society of London, 142, 995-1006. HOUNSLOW, M. W. 1990. Grain fabric measured using magnetic susceptibility anisotropy in deformed sediments of the Barbados accretionary prism: Leg 110. In: MOORE, J. C. & MASCLE, A. (eds) Proceedings of the Ocean Drilling Program, Scientific Results. College Station Texas, 110, 257-275. HOUSEN, B. A. & VAN DER PLUIJM, B. A. 1990. Chlorite control of correlations between strain and anisotropy of magnetic susceptibility. Physics of the Earth and Planetary Interiors, 61, 315-323. HOUSEN, B. A. T., TOBIN, H. J., LABAUME P., LEITCH, E. C., MALTMAN, A. J. & LEG 156 SHIPBOARD SCIENTIFIC PARTY, 1996. Strain decoupling across the decollement of the Barbados accretionary prism. Geology, 24, 127-130. HROUDA, F. 1982. Magnetic anisotropy of rocks and its application in geology and geophysics. Geophysical Surveys, 5, 37-82. HROUDA, F. 1992. Separation of a component of tectonic deformation from a complex magnetic fabric. Journal of Structural Geology, 14, 65-71. JEL!NEK, V. 1978. Statistical processing of anisotropy of magnetic susceptibility measured on a group of specimens. Studia Geophysica et Geodynamica, 22, 50-62. JELINEK, V. 1981. Characterization of the magnetic fabric of the rocks. Tectonophysics, 79, 63-67. KANAMATSU, T., HERRERO-BERVERA, E. & TAIRA, A. 2001. Magnetic fabrics of soft-sediment folded strata within a neogene accretionary complex, the Miura group, central Japan. Earth and Planetary Science Letters, 187, 333-343. KLOOTWIJK, C. T. 1985. Palaeomagnetism of the Tasman Fold Belt: indication for Mid-Carboniferous large-scale southward displacement of the New England region. Geological Society of Australia Abstracts, 14, 124-127. KLOOTWIJK, C. T. 19960. Phanerozoic configurations of Greater Australia: evolution of the North West Shelf, Part Two - Palaeomagnetic and geologic constraints on reconstructions. Australian Geological Survey Organisation Record, 1996/52.

443

KLOOTWIJK, C. T. \996b. Phanerozoic configurations of Greater Australia: evolution of the North West Shelf, Part Three - Palaeomagnetic Data Base: Australian Geological Survey Organisation Record, 1996/53. KLOOTWIJK, C. T. 2002. Carboniferous palaeomagnetism of the Rocky Creek Block, northern Tamworth belt, and the New England pole path. Australian Journal of Earth Sciences, 49, 375-405. KLOOTWIJK, C. T. 2003. Carboniferous palaeomagnetism of the Werrie Block, northwestern Tamworth Belt, and the New England pole path. Australian Journal of Earth Sciences, 50, 865-902. KLOOTWIJK, C. T. & GIDDINGS, J. 1993. Palaeomagnetic results of Upper Palaeozoic volcanics, northeastern Queensland, and Australia's Late Palaeozoic APWP. In: FLOOD, P. G. & AITCHISON, J. C. (eds) New England Orogen, eastern Australia. University of New England, Armidale, 617-627. KLOOTWIJK, C. GIDDINGS, J. & PERCIVAL P. 1993. Palaeomagnetic reconnaissance of upper Palaeozoic volcanics, northeastern Queensland. Australian Geological Survey Organization Record, 1993/36. KORSCH, R. J. 1971. Palaeozoic sedimentology and igneous geology of the Woolgoolga district, North Coast, New South Wales. Journal and Proceedings of the Royal Society of New South Wales, 104, 63-75. KORSCH, R. J. 1973. Structural analysis of the Palaeozoic sediments in the Woolgoolga district, North Coast, New South Wales. Journal and Proceedings of the Royal Society of New South Wales, 106, 98103. KORSCH, R. J. 1975. Structural analysis and geological evolution of the Rockvale-Cojfs Harbour region, northern New South Wales. Unpublished Ph.D. thesis, University of New England, Armidale. KORSCH, R. J. 1977. A framework for the Paleozoic geology of the southern part of the New England geosyncline. Journal of the Geological Society of Australia, 23, 339-355. KORSCH, R. J. 1978. Regional-scale thermal metamophism overprinting low-grade regional metamorphism, CofTs Harbour Block, northern New South Wales. Journal and Proceedings of the Royal Society of New South Wales, 111, 89-96. KORSCH, R. J. 1981. Deformational history of the CofTs Harbour Block. Journal and Proceedings of the Royal Society of New South Wales, 114, 17-22. KORSCH, R. J. 1993. Reconnaissance geology of the Solitary Islands: constraints on the geometry of the Coffs Harbour orocline. In: FLOOD, P. G. & AITCHISON, J. C. (eds) New England Orogen, eastern Australia. University of New England, Armidale, 265-274. KORSCH, R. J. & HARRINGTON, H. J. 1987. Oroclinal bending, fragmentation and deformation of terranes in the New England Orogen, eastern Australia. In: LEITCH, E. C. & SCHEIBNER, E. (eds) Terrane Accretion and Orogenic Belts. American Geophysical Union Geodynamics Series, 19, 129-139.

444

C. AUBOURG ET AL.

KORSCH, R. J. ARCHER, N. R. & MCCONACHY, G. W. 1978. The Demon Fault. Journal and Proceedings of the Royal Society of New South Wales, 111, 101-106. KORSCH, R. J. HARRINGTON, H. J. MURRAY, C. G. FERGUSSON, C. L. & FLOOD, P. G. 1990. Tectonics of the New England Orogen. Australian Bureau of Mineral Resources Bulletin, 232, 35-52. KORSCH, R. J., JOHNSTONE, D. W. & WAKE-DYSTER, K. D. 1997. Crustal architecture of the New England Orogen based on deep seismic reflection profiling. In: ASHLEY, P. M. & FLOOD, P. G. (eds) Tectonics and Metallogenesis of the New England Orogen. Geological Society of Australia, Special Publication, 19, 29-51. LAMARCHE, G. & ROCHETTE, P. 1986. Evolution des proprietes magnetiques lors de transformations minerales dans les roches: exemples du Jurassique Dauphinois (Alpes fran9aises). Bulletin Mineralogique, 109, 687-696. LEE, T. -Q., KISSEL, C., LAJ, C., CHORN-SHERN, H. & YI-TEH, L. 1990. Magnetic fabric analysis of the Plio-Pleistocene sedimentary formations of the coastal range of Taiwan. Earth and Planetary Science Letters, 98, 23-32. LEITCH, E. C. 1978. Structural succession in a Late Paleozoic slate belt and its tectonic significance. Tectonophysics, 47, 311-323. LENNOX, P. G. & FLOOD, P. G. 1997. Age and structural characterisation of the Texas megafold, southern New England Orogen, eastern Australia. In: ASHLEY, P. M. & FLOOD, P. G. (eds) Tectonics and Metallogenesis of the New England Orogen. Geological Society of Australia, Special Publication, 19, 161-177. LENNOX, P. G. & ROBERTS, J. 1988. The Hastings Block - a key to the tectonic development of the New England Orogen. In: KLEEMAN, J. D. (ed) New England Orogen - Tectonics and Metallogenesis. Department of Geology and Geophysics, University of New England, Armidale, 68-77. LOWRIE, W. 1990. Identification of ferromagnetic minerals in rock by coercivity and unblocking temperature properties. Geophysical Research Letters, 17, 159-162. LUCAS, K. G. 1960. The Texas Area. Journal of the Geological Society of Australia, 7, 229-235. MURRAY, C. G. 1997. From geosyncline to fold belt: a personal perspective on the development of ideas regarding the tectonic evolution of the New England Orogen. In: ASHLEY, P. M. & FLOOD, P. G. (eds) Tectonics and Metallogenesis of the New England Orogen. Geological Society of Australia Special Publication, 19, 1-28. MURRAY, C. G., FERGUSSON, C. L., FLOOD, P. G., WmTAKER, W. G. & KORSCH, R. J. 1987. Plate tectonic model for the Carboniferous evolution of the New England Fold Belt. Australian Journal of Earth Sciences, 34, 213-236. NICOLAS, A. 1987. Principles of Rock Deformation. Kluwer Academic Publishers, Dordrecht. OLGERS, F., FLOOD, P. G. & ROBERTSON, A. D. 1974. Palaeozoic geology of the Warwick and Goondiwindi 1:250 000 sheet areas, Queensland and

New South Wales. Australian Bureau of Mineral Resources Report, 164. OPDYKE, N. D. & CHANNELL, J. E. T. 1996. Magnetic Stratigraphy. International Geophysics Series, 64. Academic Press, San Diego. OPDYKE, N. D., ROBERTS, J., CLAOUE-LONG, J., IRVING, E. & JONES, P. J. 2000. Base of the Kiaman: its definition and global stratigraphic significance. Geological Society of America Bulletin, 112, 1315-1341. ROBERTS, J. & GEEVE, R. 1999. Allochthonous forearc blocks and their influence on an orogenic timetable for the Southern New England Orogen. In: FLOOD, P. G. (ed) Regional Geology Tectonics and Metallogenesis, New England Orogen, Ne099. Department of Earth Sciences, University of New England, Armidale, 105-114. ROBERTS, J., CLAOUE-LONG, J. C. & FOSTER, C. B. 1996. SHRIMP zircon dating of the Permian System of eastern Australia. Australian Journal of Earth Sciences, 43, 401-421. ROBION, P., FRIZON DE LAMOTTE, D., KISSEL, C. & AUBOURG, C. 1995. Tectonic versus mineralogical contribution to the magnetic fabrics of epimetamorphic slaty rocks: an example from the Ardennes Massif (France-Belgium). Journal of Structural Geology, 17, 1111-1124. ROCHETTE, P. 1987<2. Magnetic susceptibility of the rock matrix related to magnetic fabric studies. Journal of Structural Geology, 9, 1015-1020. ROCHETTE, P. 19876. Metamorphic control of the magnetic mineralogy of black shales in the Swiss Alps: toward the use of 'magnetic isogrades'. Earth and Planetary Science Letters, 84, 446456. ROCHETTE, P., JACKSON, M. & AUBOURG, C. 1992. Rock magnetism and the interpretation of anisotopy of magnetic susceptibility. Reviews of Geophysics, 30, 209-226. SAGNOTTI, L. & WINKLER, A. 1999. Rock magnetism and paleomagnetism of greigite bearing mudstones in the Italian peninsula. Earth and Planetary Science Letters, 165, 67-80. SHAW, R. D., WELLMAN, P., GUNN, P., WHITAKER, A. J., TARLOVSKI, C. & MORSE, M. 1995. 'Australian crustal elements' map: a geophysical model for the tectonic framework of the continent. Australian Geological Survey Organisation Research Newsletter, 23, 1-3. TARLING, D. H. & HROUDA, F. 1993. The Magnetic Anisotropy of Rocks. Chapman and Hall, London. THOMSON, J. 1976. Geology of the Drake 1:100000 sheet. Geological Survey of New South Wales, Department of Mines. UYEDA, S., FULLER, M. D., BELSHE, J. C. & GIRDLER, R. W. 1963. Anisotropy of magnetic susceptibility of rocks and minerals. Journal of Geophysical Research, 68, 279-291. VEITCH, R., HEDLEY, I. G. & WAGNER, J. J. 1985. Magnetic anisotropy delineator calibration error. Geophysical Journal of the Royal Astronomical Society, 75, 407-409. WELLMAN, P. 1990. A tectonic interpretation of the gravity and magnetic anomalies in southern

AMS OF THE TEXAS & COFFS HARBOUR BLOCKS Queensland. Australian Bureau of Mineral Resources Bulletin, 232, 21-34. WELLMAN, P. & KORSCH, R. J. 1988. Oroclinal bending of the New England Orogen - a refinement using magnetic and gravity anomalies. Geological Society of Australia Abstracts, 21, 417. WELLMAN, P., WILLIAMS, J. W. & MAHER, A. R. 1994.

445

Interpretation of gravity and magnetic anomalies in the Clarence-Moreton Basin region. In: WELLS, A. T. & O'BRIEN, P. E. (eds) Geology and Petroleum Potential of the Clarence-Moreton Basin, New South Wales and Queensland. AustraHan Geological Survey Organisation Bulletin, 241, 217-229.

This page intentionally left blank

Application of magnetic fabrics to the emplacement and tectonic history of Devonian granitoids in central Argentina M. G. LOPEZ DE LUCHI,1 A. E. RAPALINI,2 S. SIEGESMUND3 & A. STEENKEN3 1

Institute de Geocronologia y Geologia Isotopica, Pabellon INGEIS, Ciudad Universitaria, C1428EHA, Buenos Aires, Argentina 2

Institute de Geofisica Daniel Valencia (INGEODAV), Departamento de Ciencias

Geologicas, Facultad de Ciencias Exactas y Naturales, Universidad de Buenos Aires, CONICET, Pabellon 2, Ciudad Universitaria, C1428EHA, Buenos Aires, Argentina ^Geoscience Centre of the University ofGottingen (GZG), Department of Structural Geology and Geo dynamics, University ofGottingen, 37077 Gottingen, Goldschmidtstrasse 3, Germany Abstract: Regional considerations on the tectonic regime during the emplacement of the Early Devonian magmatic units in the Sierra de San Luis are inferred from combined field, petrographic and AMS observations. Devonian granitoids of the Sierra de San Luis, in central Argentina, constitute elliptical composite batholiths and make up the most voluminous magmatism that appears in the Sierra. Detailed fabric studies have been carried out on the La Totora batholith (33°09/ S, 65°42' W), which complement previous studies on two of the largest plutons in the Sierra de San Luis: The Renca and Las Chacras-Potrerillos batholiths. The studies comprised systematic field surveys, petrographic observations and anisotropy of magnetic susceptibility (AMS) measurements. Microstructural studies indicate that the batholith rocks are mainly characterized by magmatic microstructures with limited sub-magmatic to high-temperature solid-state deformation. All three batholiths possess concentric foliation patterns. The average magnetic foliation patterns in the studied plutons agree well with the macroscopic fabrics measured in the field indicating that the AMS-data can be used to study the orientation of fabric elements. Bulk susceptibility indicates a predominance of ferromagnetic contributions, although some paramagnetic sub-units are also present. Most foliations and lineations reflect magmatic flow and their attitude is linked to the interference between regional deformation and batholith inflation, i.e. fabrics may be due to regional strain in combination with the internal dynamics of the magma bodies. Rock fabrics are mainly described by oblate magnetic fabric ellipsoids. Magnetic lineations generally show a NNE-SSW trend that is interpreted to be controlled by the opening transtensional pull-apart structures during batholith inflation. It turns out that the Devonian batholiths intruded the basement syn-kinematically with respect to the Achalian deformational cycle.

Petrofabric studies are a fundamental step in order to map the fabrics in plutons accurately and to constrain the kinematics of their emplacement(e.g. Archanjo er a/. 1994, 1997; Neves et al. 1996, 2003; Benn et al. 1997, 1998, 1999, 2001; Djouadi et al. 1997; Becker et al. 2000; Siegesmund & Becker 2000; Steenken et al. 2000; Lopez de Luchi et al. 2002c; Siegesmund et al. 2004; Talbot et al. 2004). Fabric studies can be performed combining traditional field structural measurements with mesoscopic to microscopic studies. However, rock fabrics are sometimes difficult to determine in the field even where a well-developed foliation is present. In particular, the linear fabric is often hard to determine in the field. The anisotropy of magnetic susceptibility (AMS) is a powerful tool to constrain fabrics in granitic rocks, particularly if it is accompanied

by microstructural studies in order to establish a connection between the measured magnetic anisotropies, the minerals that contribute to the magnetic signal, and their relation with the deformation history of the pluton (e.g. Tarling & Hrouda 1993; Bouchez 1997, 2000; Hutton & Siegesmund 2001). At the same time, scalar magnetic magnitudes such as bulk susceptibility (k) have been used successfully as a lithological indicator or mapping tool in igneous rocks (e.g. Bouchez et al. 1990; Gleizes et al. 1993; Cruden et al. 1999; Becker et al. 2000; Steenken et al. 2000). The anisotropy of low-field magnetic susceptibility (AMS) is used to measure the preferred orientation of Mn- and Fe-rich minerals. The magnetic susceptibility magnitude (k) of granitic rocks is controlled by two types of magnetic

From: MARTIN-HERNANDEZ, F., LUNEBURG, C. M., AUBOURG, C. & JACKSON, M. (eds) 2004. Magnetic Fabric: Methods and Applications. Geological Society, London, Special Publications, 238, 447-474. 0305-8719/04/ $15.00 © The Geological Society of London 2004.

448

M. G. LOPEZ DE LUCHI ET AL.

minerals (Tarling & Hrouda 1993; Rochette 1987): (i) ferromagnetic (sensu lato) minerals, mainly represented by magnetite, and (ii) paramagnetic ferromagnesian minerals. Due to its high intrinsic susceptibility (k\ magnetite tends to control the magnetic parameters even when present in minor amounts. Tarling & Hrouda (1993) have proposed that k < 0.5 x 1(T3 SI indicates the lack of ferromagnetic influence and that the magnetic fabric can be interpreted as controlled by the paramagnetic fraction. In these cases, the magnetic fabric is defined by main rock forming ferromagnesian silica minerals, which makes the interpretation more straightforward. In paramagnetic granitoids, the magnetic fabric generally results from the crystallographic preferred orientations of biotite and amphibole. In ferromagnetic granitoids (k > 5 x 1(T3 SI), AMS results from the shape fabric or distribution anisotropy of the magnetite grains (Gregoire et al 1998), which are generally present in a few percentages. This is a general source of ambiguity in the interpretation of magnetic fabrics in these granitoids. Magnetic fabrics may record a deformational event that must finally be connected with detailed observations on the microstructures to constrain the rheological properties and thermal state of the plutons when a particular deformation affected them. Fabric patterns in a granitoid have been interpreted as a record of either the internal dynamics of the magma chamber or the regional tectonics (i.e. Saint Blanquat & Tikoff 1997; Benn et al. 1997, 2001; Cruden et al. 1999; Becker et al. 2000; Neves et al. 1996, 2003). Thus, if kinematic information can be obtained from their fabric, granitic plutons can be used as markers of regional deformation. Magmatic microstructures and crystal shape fabrics are preserved if no pervasive deformation post-dates solidification, even though solid-state overprints are usually concentrated on pluton margins and localized in intra-pluton shear domains. Although magmatic fabrics in plutons may form in response to magma chamber dynamics and/or the regional stress field (Paterson et al. 1998), major scale features like main axes and alignment of the sub-intrusives, will be essentially controlled by the regional stress field (Tikoff & Teysier 1992; Vigneresse et al. 1999). Granite emplacement can change from a passive mode, i.e. without deforming the country rock, to a forceful mode at the same crustal level depending on the interaction of several variables that affect crustal rheology (Cruden et al. 1995). Devonian magmatism in Sierra de San Luis is mainly represented by elliptical batholiths in which polyphase granodioritic to monzogranitic

intrusions show a close association between mafic and felsic magmas. The major axes of the batholiths cross-cut the regional NNE penetrative structures. Clear-cut contacts against the country rock, concentric magmatic foliation, microgranular enclaves, external coarse-grained porphyric facies and minor lamprophyric dykes are diagnostic features. Their particular tectonic setting offers the opportunity of studying internal structures in plutons emplaced in a crustal section in which no younger pervasive ductile deformation is present. This study is based on a systematic multidisciplinary project that encompassed field measurements, microstructural analyses and AMS studies of Devonian plutons from the Sierra de San Luis (Eastern Sierras Pampeanas, central Argentina; Fig. 1) in order to understand their emplacement kinematics and relationship with country the rock deformation. This paper combines the results of rock magnetic and petrofabric studies in the La Totora batholith (Fig. 1) with those obtained from the nearby Renca and Las Chacras-Potrerillos batholiths (Lopez de Luchi et al. 2002c; Siegesmund et al. 2004). Evidences are presented that document the syn-deformational emplacement of the Devonian batholiths instead of the post-orogenic emplacement so far considered. Geological setting The Sierra de San Luis is part of the Eastern Sierras Pampeanas geologic province of central Argentina (Criado Roque et al. 1981). It is characterized by an Early Palaeozoic metamorphic basement intruded by Ordovician to Devonian magmatic units. The end of ductile deformation is marked by the deposition of Upper Carboniferous continental sediments (Salfity & Gorustovich 1983). The most important map scale feature of the Sierra de San Luis is the NNE arrangement of parallel belts of different metamorphic grade and/or structural level (Fig. 1). Several authors have published regional or detailed studies on the magmatic, metamorphic, structural and tectonic evolution or metallogenesis of the Sierra de San Luis (cf. Criado Roque et al. 1981; Lopez de Luchi 1986, 1993, 1996; Ortiz Suarez et al. 1992; Dalla Salda et al. 1998; Llambias et al. 1998; Sims et al. 1997, 1998; von Gosen 1998; von Gosen & Prozzi 1996, 1998; Quenardelle & Ramos 1999; Lopez de Luchi et al. 2000, 20010,6, 2002a,6,c, 2003; Hauzenberger et al. 2001; Steenken et al. 2002, 2003; Siegesmund et al. 2004 and references therein). Three main polyphase deformed

AMS & EMPLACEMENT OF DEVONIAN GRANITOIDS

449

Fig. 1. Schematic map showing the Palaeozoic metamorphic belts and granitoids units of the Sierra de San Luis modified after Lopez de Luchi et al. (200\b). Crystallization and cooling ages were taken from Sims et al. (1997), Lopez de Luchi et al. (20026, 2003), Steenken et al. (2003) and Siegesmund et al. (2004).

450

M. G. LOPEZ DE LUCHI ET AL.

metamorphic belts (phyllites, schists, gneisses, amphibolites and migmatites) of Precambrian(?) to Early Palaeozoic age make up the Sierra de San Luis (Fig. 1). The eastern Conlara Metamorphic Complex (Fig. 1) was considered to be the result of a Cambrian metamorphic event (Sims et al 1997). The central Pringles Metamorphic Complex showed high T-moderate P peak metamorphic conditions at c.460Ma (Sims et al 1997, 1998); whereas the western Nogoli Metamorphic Complex is considered as being either Proterozoic, Cambrian or Ordovician. (Gonzalez et al. 2002 and references therein). The metamorphic overprint within the phyllite belts (in part San Luis Fm.) is considered as Ordovician or Devonian (Steenken et al. 2004submitted; von Gosen & Prozzi, 1998; Sims et al. 1997). Peak metamorphic conditions are equivalent to the amphibolite facies, except for the phyllites that were equilibrated at greenschist facies, and restricted areas in the central west sector of the Pringles Metamorphic Complex where granulite facies is present. The metamorphic rocks are the country rocks for granitoids, pegmatoids and mafic rocks that show coupling or decoupling of their internal structure in relation to the pervasive regional deformational fabrics. The composition of the granitoids of the Sierra de San Luis ranges from tonalites to high-silica alkali-granites. The few available crystallization ages for the plutons are either Ordovician or Devonian (Varela et al. 1994; Llambias et al. 1998; Lopez de Luchi et al. 2QQla,b, 2003; Sims et al. 1997, 1998; Siegesmund et al. 2004). On the contrary, the number of K-Ar cooling ages for the granites, both in biotite and muscovite, is higher, with most of them falling in the Devonian (Steenken et al. 2003, 2004-in review). Based on the petrographic, geochemical, geochronological and structural data, granitoids were separated into two series: (i) the Ordovician intrusives of tonalitic (OTS) or granodioritic to granitic (OGGS) composition, and (ii) the Devonian granites (DGS) and monzogranite/granite (DMGS) association (Lopez de Luchi et al. 2004). Ordovician stocks are broadly parallel to the regional penetrative NNE foliation. On the other hand, Devonian granitoids (Table 1) appear either as sheet batholiths parallel to the tectonic(?) contact between the Phyllite and Pringles/Nogoli Metamorphic Complex or as elliptical batholiths that are discordant to the regional foliation (Fig. 1). These elliptical batholiths appear in all different basement domains (Siegesmund et al. 2004). The tectonic evolution of the Sierras Pampeanas took place by lateral crustal growth through

processes of arc and terrane accretion acting on the Proto-Andean margin of Gondwana during Cambrian to Devonian times (see for instance Pankhurst & Rapela 1998; Ramos & Keppie 1999 and references therein). Relationships between magmatism, metamorphism and deformation during the Early to Middle Palaeozoic tectonic evolution of the SW Gondwana margin are still controversial. Several hypotheses were proposed to explain the timing and evolution of this major collisional orogen. Dalla Salda et al. (1998 and references therein) considered that the Eastern Sierras Pampeanas belong the late Proterozoic to Devonian Famatinian Orogenic Belt, in which two main metamorphic and igneous cycles were recognized: The Pampean (5807-540 Ma) and Famatinian (540-330 Ma) orogenic cycles. Based on geochemical characteristics and their relation (?) to deformational events, three main groups of granitoids were proposed: A pre-Famatinian, a Famatinian, and a Late- to Post-Famatinian group (Lopez de Luchi & Dalla Salda 1997; Lopez de Luchi et al. 1998). In contrast Stuart-Smith et al. (1999 and references therein) proposed three main compressive events: The Pampean (5307-510 Ma) and the Famatinian (490/470-330 Ma) orogenies and the final Achalian cycle (Sims et al. 1997), which roughly correspond to the Late to Post-Famatinian events of the Famatinian cycle of Dalla Salda et al. 1998). It has been proposed that the Achalian orogeny resulted from the collision of the allochthonous Chilenia terrane with Gondwana (Ramos et al. 1986; Ramos 1988; Sims et al. 1997, 1998; Davis et al. 1999; Quenardelle & Ramos 1999). The Achalian cycle is a period of heterogeneous deformation along crustal scale fault lines that may have resulted from the resumption of the convergence on the western margin of Gondwana (Siegesmund et al. 2004). During the Achalian orogeny a series of extensive Devonian granites, that were first considered as post-tectonic in relation to the Famatinian cycle (Llambias et al. 1998; Dalla Salda et al. 1998) intruded the older basement and are especially widespread in the Sierra de San Luis and the Sierras de Cordoba (Sims et al. 1997; Siegesmund et al. 2004). The La Totora batholith

Composition and structures The La Totora batholith is located at around 10km to the SW of La Toma town (Fig. 2). The pluton intrudes with clear-cut contacts metasedimentary rocks, schists and related

Table 1. Summary of the main geological features and ages for the Devonian batholith of the Sierra de San Luis. Location of the plutons is shown in Figure 1 Cooling ages

Contact

Deformation

References

Bt 397 ± 10

Tectonic contact Phyllites/ Pringles Eastern border mylonitized

Magmatic foliation partially obliterated by well-developed solid-state fol. along the borders

Sato et al. 1996; Sims et al. 1997; Lopez de Luchi et al. 20026

Bt-Ms-Ep granodioritesmonzogranites

Bt 380 ± 7

Nogoli Complex Structural contact aureole// internal foliation Thermal aureole FibroliteCord SE Contact of Barroso is tectonic

Magmatic foliation Shear zones that affect Barroso CRE

Gonzalez & Sato 2000

Bt-Amph-Ep Gabbromonzodiorites monzonites Amph-Bt monzogranite

Bt 364 ± 7

Conlara Met. Complex Structural contact aureole// internal foliation Discordant to regional foliation Conlara Met. Complex Structural contact aureole// internal foliation Discordant to regional foliation Conlara Met. Complex Structural contact aureole// internal foliation Discordant to regional foliation Conlara Met. Complex Discordant to regional foliation Conlara Met. Complex Structural contact aureole// internal foliation Discordant to regional foliation

Magmatic foliation MME

Steenken et al. 2002

Magmatic foliation Locally solid state overprint MME

Brogioni 1991, 1993; Lopez de Luchi et al. 200 1/?; Siegesmund et al. 2004

Magmatic foliation Locally solid state overprint MME

Lopez de Luchi et al. 20026 Steenken et al. 2002

Magmatic foliation and some solid state overprint

Quenardelle 1993

Magmatic foliation Locally HT solid state overprint MME

Lopez de Luchi 1986,1987, 1996; Lopez de Luchi et al. 2001^,20026; Sims et al. 1997, 1998; Steenken et al. 2002

Pluton

Principal rock types

U-Pb

DOS

La Escalerilla

Bt monzogranite

403 ±5

DMGS

El Molle

Barroso El Hornito

Las ChacrasPotrerillos

Bt-granodiorite-monzogranite Amph-Bt monzogranite Amph-Bt monzonite

La Totora

Bt-granodiorite-monzogranite Bt(Amph)-granodiorite monzogranite Amph-Bt monzonite

San Jose del Morro

Amph-Bt monzogranite Amph-Bt monzonite

Renca

Bt-granodiorite-monzogranite Amph-Bt monzogranite Amph-Bt monzonite

Bt, biotite; Ms, muscovite; Ep, epidote.

Rb-Sr

Bt 342 ± 7

Bt 352 ± 7

382 ±5

Bt 358 ± 8 Bt 371 ± 8

382 ±4

393 ±5

Ms 368 ± 9 Bt 346 ± 8

452

M. G. LOPEZ DE LUCHI ET AL.

Fig. 2. Geological map of La Totora granite with field data for foliation.

migmatites of the Conlara Metamorphic complex (Fig. 3a). An exposed semicircular area of 40km 2 constitutes roughly the northwestern half of the elliptical pluton with a major axis trending nearly NNW-SSE as inferred from aeromagnetic data (Sims et al 1997; Figs. 1, 2). The batholith is composite and exhibits an external very coarse-grained biotite porphyric monzogranite facies, the La Portena granite, that is locally amphibole bearing; and an internal medium- to coarse-grained light grey biotite

monzogranite, the Gobelli granite, that varies from equigranular to porphyric (Fig. 2). Contacts between the two monzogranites are mainly transitional although some local sharp changes from one facies into the other have been observed. Available K-Ar biotite cooling ages for the pluton are 371 ± 8 Ma for a sample of the northern border (La Portena granite) and 358 ± 8 Ma for the central Gobelli granite (Lopez de Luchi et al. 2004).

AMS & EMPLACEMENT OF DEVONIAN GRANITOIDS

La Portena granite is a porphyritic monzogranite that consists of 20-30% quartz, 2040% oligoclase, 28-46% microcline, up to 10% biotite, up to 2% amphibole and a mean 0.5% of magnetite that appears either associated with the mafic minerals or isolated in (within crystals or as crystals?) tiny crystals included in the light minerals. The most prominent accessory minerals are apatite, zircon, titanite, epidote, allanite and scarce tourmaline. The central

453

Gobelli granite is a biotite granodiorite-monzogranite, that is composed of 10-25% quartz, 35-50% oligoclase, 8-30% microcline, up to 22% biotite and a mean 0.5% of magnetite. Accessory minerals are zircon, muscovite and scarce apatite. The macroscopic fabrics range from foliated to weakly foliated or locally unfoliated (Fig. 3b). A dipping inwards magmatic foliation is defined in the La Portena granite by the shape-preferred

Fig. 3. (a-c) Outcrop scale features of the La Totora batholith: (a) clear-cut contact against the host Conlara Metamorphic Complex; (b) parallelism between the magmatic foliation and the major axis of the enclave; (c) section of a tubular schlieren in the Gobelli granite, (d-f) Photomicrographs in XPL of the microstructures in the La Portena granite of the La Totora batholith; (d) chessboard pattern in quartz which is characteristic of high-temperature deformation, base of the photo: 6mm; (e) melt migration textures, myrmekitic plagioclasequartz form an interstitial aggregate with an incipient polygonization (recovery), base of the photo: 0.5mm; (f) grain boundary migration. Peninsulas of the dark quartz remain joined to the main body of the parent crystal. Some incipient new grains have already detached, base of the photo: 0.5mm

454

M. G. LOPEZ DE LUCHI ET AL.

orientation (SPO) of undeformed microcline megacrysts set in a coarse-grained matrix in which the SPO of plagioclase and some biotite aggregate alignments are recognized. Close to the borders and especially along the western sector, the foliation is of solid-state nature and parallel to the foliation in the host. In the Gobelli granite, foliation is weakly defined at outcrop scale and in some areas the granite looks isotropic. The major axes of the microgranular enclaves and country rock septa that are observed in the La Portena granite are parallel to the trace of the foliation (Fig. 3b) planes. Thin (up to 20cm) aplitic leucogranite dykes cross-cut the foliation at high angle and seem to define a radial system in the La Portena granite. Well-developed mafic schlieren are observed in the Gobelli granite (Fig. 3c). They are commonly curved, either individually or in groups. Common features of this pluton are schlierenrimmed mushroom or elongated hook shapes, typically up to 2m in a single outcrop. Similar shapes were considered as schlieren-rimmed magma blobs (Barriere 1977, 1981; Weinberg et al. 2001). In the Gobelli granite, the blobs are composed of slightly finer-grained granite than that of the surrounding facies and the mafic schlieren marking the borders of the blobs are characterized by a sharp external boundary, where mainly biotite is concentrated. In some sectors, the magmatic foliation and schlieren define metre-scale two-dimensional elliptical shapes in which the K-feldspar megacrysts are parallel to a concentric elliptical foliation. Where vertical sections are observed, some of the ellipses are the head of the mushrooms or hook-shaped blobs but others exhibit an elliptical vertical section indicating in this case a three-dimensional ellipsoidal structure. Structures similar to the latter case were described for the Renca batholith (Lopez de Luchi 1986) and in many others granites (see Weinberg et al. 2001 for references).

Micro structures A systematic microstructural study on thin sections of the oriented core samples collected for the magnetic fabric study indicates that the batholith rocks are mainly characterized by magmatic microstructures with limited sub-magmatic to high-temperature solid-state deformation. The criteria for identifying igneous textures have been extensively described by Paterson et at. (1989, 1998), Bouchez et al (1990, 1992) and Vernon (2000). The macroscopic shape-preferred

orientation of minerals in the La Portena granite is associated in the majority of the sections with magmatic microstructures that are characterized by porphyric K-feldspars set in an equigranular to slightly porphyric matrix with slightly bowed contacts between grains. At some sites, quartz grains contain chessboard pattern subgrains (Fig. 3d) and melt pockets are observed in interstitial sites (Fig. 3e). Close to the borders, and especially along the western sector, a submagmatic to incipient high-temperature solidstate overprint is recognized by the chessboard pattern in quartz, late stage melt pods, submagmatic fractures filled with fine-grained feldspar-quartz aggregates and the local development of myrmekites. Subgrains in biotite and in some feldspar crystals are sporadically seen. High-temperature recrystallization is inferred from grain-boundary migration between adjacent quartz grains in some sections (Fig. 3f). In the Gobelli granite, foliation is weakly defined and these rocks only show some subgrain development in quartz.

The Renca and Las Chacras-Potrerillos batholiths These plutons appear as discordant zoned elliptical batholiths. The larger is the Las ChacrasPotrerillos batholith that is situated at the northern end of the transpressive, sinistral Rio Guzman shear zone (Table 1; Siegesmund et al. 2004). The Las Chacras-Potrerillos batholith consists of three sub-domains, each of them comprising two granitoid sub-units. A biotite porphyritic granite (PG), a biotite equigranular granite (EG), a giant amphibole-biotite porphyritic granite (GPG), a biotite porphyritic granite (BPG), and a red granite (RG) that was subdivided into a biotite-bearing (BRG) and a muscovite-bearing red granite (MRG). The modal compositions of the entire batholith indicate monzogranitic, granodioritic and less abundant syenogranitic compositions (Lopez de Luchi et at. 200\b). Contacts with the country rock are clear-cut. The magmatic contacts are sharp but locally irregular. Country rock septa from tenths up to one hundred metres width are interlayered with the BPG. K-feldspar megacrysts tend to be smaller towards the outer borders of the pluton. Two main groups of enclaves can be distinguished in the batholith: (i) metamorphic xenoliths, and (ii) microgranular enclaves of variable size, ranging from monzonites to syenites. They appear either isolated

AMS & EMPLACEMENT OF DEVONIAN GRANITOIDS

or as enclave swarms in the BPG, the GPG, and the PG. In the western sector of the PG, finegrained syenitic stocks or syn-plutonic dykes occur together with circular aplitic to pegmatitic dykes. The Renca batholith (Lopez de Luchi et al. 2002&, c) is normally zoned and comprises two main units: an external grey to slightly pinkish porphyroid biotite ± hornblende granodiorite/ monzogranite (Unit 1) and a central pink to grey equigranular biotite-muscovite monzo/ leucomonzogranite (Unit 2). Grey to greenish grey monzonitic rocks and quartz syenites appear as microgranular enclaves, stocks and syn-plutonic dykes within Unit 1 (Lopez de Luchi 1996). The stocks and syn-plutonic dykes of monzonitic rocks are mainly melanocratic to normal amphibole-biotite quartz-monzonites, monzonites and granodiorites. Microgranular enclaves are porphyroid or equigranular and medium- to fine-grained amphibole-biotite monzonite, syenites or monzodiorites. Contact between Units 1 and 2 are almost everywhere sharp, with some discontinuous fine-grained borders of the Unit 2 against the porphyric granites. External and internal contacts mostly follow the foliation trajectories in the country rock of the narrow ductile contact aureole that developed by meso- and microfolding of the regional NNE-SSW striking pervasive foliation. Septa of country rocks within the pluton are in clear structural continuity with the surrounding country rocks in the Renca (Lopez de Luchi et al. 2002c) and Las Chacras-Potrerillos batholiths (Lopez de Luchi et al. 200\b; Siegesmund et al. 2004). The steep to vertical foliation that is recognized in the porphyroid granitoids parallels the X—Y plane of the enclaves and the synplutonic dykes attitude but frequently cross-cut the internal contacts of the different sub units at acute angles. Microstructures A systematic microstructural study on thin sections of the oriented core samples collected for the magnetic fabric study indicates that the batholith rocks are mainly characterized by magmatic microstructures with less developed sub-magmatic to high-temperature solid-state deformation. However, since euhedral crystal shapes and igneous zoning are preserved in the feldspars the rocks did not undergo extensive sub-solidus deformation. In the Las Chacras-Potrerillos batholith, magmatic microstructures are characteristic for

455

the EG and the MRG. In contrast, microstructures indicating a sub-magmatic to hightemperature solid-state overprint are observed in the BRG, the PG, the BPG and the GPG. This deformation increases towards the contacts between each sub-unit or to the country rock. The coexistence of sub-magmatic and hightemperature solid-state deformation within the marginal parts of the BPG and the GPG implies a continuous deformation history (Siegesmund et al. 2004). Lower-temperature solid-state microstructures are most pronounced in the SE part of the pluton affecting both BPG and BRG. Those feature lozenge-shaped feldspar phenocrysts, quartz ribbons that wrap around the megacrysts, elongated pockets of microaplite, and abundant myrmekite and flame pertithes in the interstitial microcline (Siegesmund et al. 2004). In the Renca batholith, magmatic, sub-magmatic, and locally high-temperature sub-solidus deformation microstructures are present in Unit 1 and subordinate in Unit 2. High-temperature solid-state deformation, although moderate, is present mainly along the contact with the country rock, at the western half of the batholith within amphibole-biotite granodiorites or monzogranites (Unit 1), and it is sporadically more intense towards the internal contact between the Units 1 and 2. In all the plutons, solid-state deformation is common along contact zones between major units or with the metamorphic host. This suggests that each granite pulse may have intruded the still incompletely or the just solidified outer granite shell. The gradual microstructural transition along the margins of the plutons indicates that deformation occurred in a continuum from the magmatic to the solid state during the emplacement. Microstructures at high solid fraction, i.e. at the sub-magmatic stage as defined by Bouchez et al. (1992) are mainly evidenced by late-stage melt migration into intra-crystal fractures (Fig. 3c) or interstices. Microaplite (quartz, plagioclase, microcline ± myrmekite) could have formed from the crystallization of a residual melt that was channelled and subsequently deformed along high-strain zones. Microaplites form 'pockets' of various shapes and sizes filling interstices between larger grains or fractures inside the feldspars. The microaplite pockets have no particular shape, except for a slight shape-preferred orientation parallel to the magmatic foliation. They become more abundant and elongate toward the pluton margins (Lopez de Luchi et al. 2002c; Siegesmund et al. 2004).

M. G. LOPEZ DE LUCHI ET AL.

456

Anisotropy of magnetic susceptibilty (AMS) of the studied batholiths

Sampling and measurements A systematic sampling of the La Totora granite was undertaken. This comprised 44 sites (230 samples) distributed as evenly as possible on

the granite. Restrictions to access impeded collection of samples in the northwestern area of the batholith. Generally, five samples of 2.5cm in diameter, separated around 1m from each other, were collected at each site with a portable drill. Orientation was done with both sun and magnetic compasses whenever possible. One to three 2.2cm high standard specimens

Table 2. Mean site AMS data for the La Totora batholith Unit

Site

N

^ vol (io- 6 si )

^max

^min

*int

Dec (°)

Inc (°)

Dec (°)

Inc (°)

Dec (°)

Inc (°)

P*

T

La Portena

T4 T6 T7 T8 T9 T10 Til T12 T14 T15 T17 T18 T19 T20 T23 T24 T25 T26 T30 T31 T32 T33 T34 T35 T36 T37 T39 T40 T41 T42 T45 T48

5 3 5 5 5 6 5 5 5 5 5 5 5 5 5 5 7 5 5 5 5 5 5 5 5 5 5 5 5 6 5 5

7258.6 7504.9 3632.7 11280.5 11538 9883.4 8882.8 9702.6 3 525.9 10701.7 5732.3 8069.9 6131.6 5809.5 6872.9 9932.1 4710.7 3360.5 2137.3 5337.4 1 776.9 1200.1 1401.3 5361.5 9543.3 3458.7 1 924.3 1965 436.8 747 1845.1 3384.3

135.2 251.7 262.7 267.9 65.7 223.1 258.4 115.4 222.3 230.3 166.9 63.1 351.4 247.0 280.9 57.5 138.1 275.1 195.8 236.3 211.0 12.2 268.1 38.0 69.8 229.5 24.3 30.6 342.8 209.2 140.4 183.4

39.7 1.8 11.3 30.4 52.4 49.0 1.3 50.5 20.7 22.7 35.3 39.1 31.0 1.9 32.0 45.4 14.5 13.9 29.2 64.3 14.6 4.3 65.6 67.1 12.3 34.3 36.3 13.1 20.9 11.4 69.8 34.8

289.5 354.8 78.0 56.0 232.8 92.2 162.0 220.9 19.0 64.6 35.3 168.6 176.8 341.0 54.5 314.2 331.1 165.8 76.8 13.5 104.9 278.8 3.0 170.7 262.0 95.8 184.1 151.0 105.6 105.4 11.4 11.6

47.4 82.0 78.7 55.4 36.9 29.7 78.8 12.5 31.4 66.6 43.2 18.2 58.9 64.5 47.8 12.8 75.2 53.2 40.9 19.5 46.7 38.5 2.2 16.0 77.5 45.4 52.0 65.3 54.8 49.8 13.0 54.9

34.1 161.4 172.5 168.9 327.5 346.4 348.7 320.4 340.1 322.5 277.5 277.9 82.9 156.1 174.5 212.4 228.9 14.5 309.1 109.4 213.4 107.6 94.0 275.4 160.3 337.9 287.0 295.6 241.7 308.2 277.8 276.1

13.1 7.7 0.9 15.0 6.3 25.5 11.1 36.8 51.0 5.2 26.5 45.3 2.4 25.5 24.3 41.8 3.2 33.2 35.2 16.2 39.6 51.2 24.3 16.0 2.6 24.8 9.9 20.5 26.9 38.0 15.1 3.9

1.052 1.035 1.034 1.039 1.108 1.065 1.031 1.041 1.050 1.052 1.054 1.079 1.064 1.043 1.019 1.044 1.027 1.014 1.090 1.076 1.024 1.055 1.050 1.060 1.053 1.081 1.057 1.085 1.025 1.021 1.018 1.071

0.49 -0.40 -0.45 0.09 0.86 0.44 0.17 0.69 0.05 0.30 0.69 0.49 0.61 0.52 -0.88 0.53 0.76 -0.28 0.63 0.69 0.01 0.14 0.28 0.55 0.46 0.36 0.88 0.50 -0.27 -0.50 0.12 0.43

Gobelli

Tl T2 T16 T21 T22 T27 T28 T29 T44 T46 T47

5 5 5 5 5 5 5 5 5 5 5

5366.1 966.7 1 360.7 1781 1 752.6 956.7 4026.8 3151.1 8810.3 1081.7 1 804.4

16.6 328.5 232.8 231.5 89.1 46.1 41.3 347.8 194.9 268.5 14.4

28.8 43.1 3.1 37.8 7.3 22.4 22.2 38.2 3.5 37.2 46.7

221.0 139.9 138.3 56.5 189.6 277.3 289.6 213.2 306.6 85.7 239.9

58.9 46.6 55.5 52.1 54.9 56.7 42.1 41.8 80.5 52.8 33.4

112.6 234.5 324.9 323.4 354.2 146.3 151.1 99.0 14.4 177.5 132.8

10.8 4.3 34.3 2.5 34.1 23.3 39.6 24.7 8.8 1.3 24.1

1.027 1.040 1.022 1.027 1.020 1.021 1.034 1.028 1.060 1.017 1.023

0.26 0.26 0.19 -0.03 0.00 -0.14 0.24 -0.06 0.43 -0.50 0.46

N: number of independent cores per site used to compute the mean, k is a geometric mean of the magnetic susceptibility. P1'. anisotropy degree, T: shape anisotropy factor as defined by Jelinek (1981).

AMS & EMPLACEMENT OF DEVONIAN GRANITOIDS

were sliced from each sample. Anisotropy of magnetic susceptibility (AMS) was measured for one specimen per sample with a KLY-2 (Kappabridge) at the University of Gottingen, Germany. This allowed the calculation of the principal susceptibilities axes for each sample. The AMS ellipsoid for each site was computed according to Jelinek (1977, 1978). Geometric means of bulk susceptibility (k) were computed for each site. Mean site data are presented in Table 2.

457

thermal treatments, indicate a ferrimagnetic fraction as the carrier of the remanence. NRM values as high as lOA/m were commonly found. AF demagnetization (Fig. 4) indicates medium destructive fields generally below 10 mT. On the other hand, thermal treatment (Fig. 4) shows a broad spectrum of unblocking temperature with virtually no remanence remaining at 600 °C. This, together with the scattered remanence directions observed (Lopez de Luchi et al. 2002a; Rapalini et al. 2003) indicates that multidomain (MD) magnetite is likely the main magnetic carrier in the granitoids of La Totora.

Magnetic mineralogy Identification of principal magnetic carriers is of great importance when analysing AMS data (Rochette et al. 1992). Magnetic carriers in the La Totora granitoids were identified as part of a palaeomagnetic reconnaissance study carried out on these rocks (Lopez de Luchi et al. 2002<2; Rapalini et al. 2003). Standard demagnetization techniques, both by alternating field and

Bulk susceptibility (k) Both the La Portena and Gobelli units present k values mostly falling into the ferromagnetic field (Tarling & Hrouda 1993). Values range between 0.9 and 11.2 x 10~3 SI units (Table 2). Geometric means of site values yield an average of 4.17 and 2 . 1 5 x l O ~ 3 S I units for the La Portena and

Fig. 4. Typical demagnetization behaviour of samples from the La Totora batholith represented by AsZijderveld diagrams and normalised demagnetization curves. Note the very low coercive forces as shown by rapid decay of remanence intensity under AF treatment, and unblocking temperatures under 600 °C. Open (solid) symbols represent projections on the vertical (horizontal) plane.

458

M. G. LOPEZ DE LUCHI ET AL.

Gobelli granites respectively. This suggests that the susceptibility parameters are governed, or at least influenced, by the ferromagnetic fraction. Thin section analysis and rock magnetic properties (see below) indicate that magnetite is the dominant ferromagnetic mineral in both units. The lower value of the central Gobelli granite can explain the lower magnetic relief shown by the aeromagnetic survey (Sims et al. 1997). Lower magnetite content in the Gobelli unit is consistent with normal compositional evolution of the magma source (Fig. 5a). In order to analyse this parameter, a subjective subdivision has been made. Low k values (<2 x 10~3 SI units) are predominant in the Gobelli domain with few exceptions. The La Portena unit, however,

shows a greater range of values with low k values only found in the western sector close to the margin. On the other hand, the highest values (k > 8 x 1(T3 SI) are mainly found in the northern area approximately corresponding to a facies of porphyric coarse-grained granodiorite-monzodiorite. The Renca granitoids (Lopez de Luchi et al. 2002c) were sampled in a similar fashion to those of La Totora, with 60 sites evenly distributed over the entire intrusive (Figs. 5c, d). In contrast, the Las Chacras-Potrerillos batholith (Siegesmund et al. 2004) was not sampled for AMS in such detail, with only 49 sites widely distributed on all six major units. This less dense sampling in the Las Chacras-Potrerillos

Fig. 5. Distribution of geometric mean site k values on the La Totora granite (a) and comparison with Renca batholith (b). Similar maps for P1 are shown in (c) and (d) respectively. Data is also presented in Table 2. Note that scales of values for each batholith differ. Data of Renca is taken from Lopez de Luchi et al. (2002&).

AMS & EMPLACEMENT OF DEVONIAN GRANITOIDS

459

Fig. 6. (a) Pf vs. K plot for each site of the La Totora granite. Note the lack of any significant correlation between the two parameters, (b) P' vs. K for the Renca batholith. For comparison, fields for the La Totora (in grey) and for the different magmatic sub-units of the Las Chacras-Potrerillos batholith are indicated. Values are site means (see also Table 2). Note that for Unit 1 of the Renca batholith and BPG of the Las ChacrasPotrerillos batholith a strong correlation between both parameters is observed, (c) T vs. P1 plot for each site of La Totora granite. Note the predominance of oblate shapes for the La Portena granite (d) T vs. Pf plot for the Renca batholith. Oblate fabrics are dominant for high anisotropy degrees. Note the different scales for (c) and (d). For comparison fields for La Totora (in grey) and for the different magmatic sub-domains of Las ChacrasPotrerillos are indicated. The area for the central domain of the Las Chacras-Potrerillos batholith is open because values up to 1.7 for Pf are present. Data of the Renca batholith is from Lopez de Luchi et al. (20026), those for the Las Chacras-Potrerillos batholith from Siegesmund et al. (2004).

batholith makes comparison of AMS results from this batholith less certain. However, the structural field observations from this pluton are of high quality. Comparable to La Totora, the Renca and Las Chacras-Potrerillos batholiths are largely made up by ferromagnetic granites (Figs. 6a, b). Exceptions to this are the internal Unit 2 of Renca with k < 3 x 10~4 SI and the muscovite-bearing red granite (MRG) of the Las Chacras-Potrerillos intrusion with k < 10~4 SI. The internal Unit 2 of Renca with significantly lower k values (Fig. 6b) is similar to the pattern of the La Totora pluton but much more conspicuous. All other units mainly correspond to the ferromagnetic field with values of k occasionally greater than 10~2SI. The La Totora granite (both the La Portena and Gobelli units) tends to fall into intermediate values with much more reduced variation, comparable with the susceptibility range observed in the EG of the Las ChacrasPotrerillos batholith (Fig. 6b). The much more

reduced dispersion of values is consistent with a lesser petrographic and geochemical differentiation of the units composing the La Totora batholith. On the other hand, the extreme values reached at both the Renca and Las ChacrasPotrerillos batholiths are nearly three orders of magnitude higher, indicating the larger variability and differentiation of these more extensive and complex bodies.

AMS scalar parameters The anisotropy degree (Pf) and shape parameter (T) were computed following Jelinek (1978). Pf values in La Totora are generally low, with a single site showing values over 1.1. Distribution of P' is plotted in Figure 5c. Values have been subjectively classified in very low (P7 < 1.03), low (1.03 < P1 < 1.05), moderate (1.05 < P7 < 1.07) and high (P7 > 1.07) anisotropies. Although the correlation between the

460

M. G. LOPEZ DE LUCHI ET AL.

intensity of a solid-state overprint and the degree in anisotropy of a magnetic fabric influenced by ferromagnetic fractions may not be straightforward, a lack of significant correlation between Pf and k is observed in the La Totora batholith (Fig. 6a). In the northern area, where the high k values have been observed P' is characterized by intermediate values (Fig. 5c). In contrast, low bulk susceptibilities along the western contact of the pluton come along with high magnetic anisotropies (see Figs. 5a, c). Thus, a correlation between the solid-state overprint and P' is inferred to exist in the La Totora batholith. Then, very low and low anisotropies suggest a lack of any sub-solidus deformational fabric, while moderate and especially high anisotropies suggest that the magnetic fabric may be influenced by a solid-state overprint. All except one site in the Gobelli domain show very low or low P/ values, indicating that the magnetic fabric is likely governed by magmatic processes only. This is consistent with field and thin sections observations (see above). The La Portena unit presents variable degrees of anisotropy, with high values concentrated on the margins, which suggests that deformation occurred there, possibly during emplacement of the external body itself or the internal one. Again, this is consistent with field and microscopic observations of signs of sub-magmatic to incipient high-temperature solid-state deformation at those sites. However, no significant increase in Pf is seen close to the boundary of both units suggesting that rheological conditions during emplacement were quite similar between both units. This is coherent with the general transitional boundary between the La Portena and Gobelli granites. Figure 6c shows the relationship between P1 and T. For very low P1', the AMS ellipsoid shows no shape preference. However, for values over 1.04, all magnetic fabric ellipsoids correspond to the oblate field with some correlation between T and P1'. Due to its generally low P'', the Gobelli granite has no preferred magnetic fabric ellipsoid shape, while oblate fabrics dominantly characterize the La Portena granites. This is also similar to AMS results gained for the granitoids of Renca and assigned to radial compression associated with the intrusion of the central body (Figs. 5d, 6d; Lopez de Luchi et al. 2002c). However, the dominant ferromagnetic character of most units from the Renca and Las Chacras-Potrerillos batholiths suggests that the AMS interpretation, particularly its anisotropy degree, may be problematic. Furthermore, some units (Fig. 6b) show a strong correlation between Pf and k. Taking this into account,

correlation of P/ values with microstructural analysis, correspondence of directional AMS values with field foliations and lineations as well as biotite texture measurements (Las Chacras-Potrerillos batholith) have been used to test confidence of AMS results as representative of the granites petrofabric. In particular, the Unit 1 of Renca and the BPG body of the central domain of the Las ChacrasPotrerillos batholith show a strong correlation between Pf and k (Fig. 6b), which makes any interpretation of anisotropy degree as indicator of deformational degree more dubious. However, such distinct correlation is not observed in the La Totora pluton, as already mentioned, as well as in the central Unit 2 of Renca, and the PG and EG units from the northern domain, and the BRG and MRG granites from the southern domain of the Las Chacras-Potrerillos batholith. In all these latter cases Pf is very seldom higher than 1.1 (Fig. 6b). Within the GPG unit of the Las Chacras-Potrerillos batholith high Pf values of >1.09 show no variation with the very high bulk susceptibility (Fig. 6b).

AMS directional data Distribution of the magnetic foliation planes in La Totora batholith is presented in Figure 7a and magnetic lineations in Figure 8a. The magnetic foliation planes are generally sub-vertical. Since in most areas the fabric is interpreted to be magmatic in origin, sub-vertical foliation planes may suggest that the exposure level of both the La Portena and Gobelli granites are far from the top of the bodies. Planes tend to follow a concentric pattern, especially in the La Portena granite, parallel to both the external and internal boundaries. Sub-horizontal to shallowly inclined lineations plunge roughly towards the NNE or SSW and may reflect magma flow patterns (Fig. 8a). This seems to be a systematic pattern in the granites of the Sierra de San Luis and is discussed later. In the southern area, a few lineations are sub-vertical or of moderate plunge. This may reflect zones of vertical magma flow in relation with a possible feeder zone for the entire batholith. However, since the southern half of the batholith is not exposed and given the lack of gravimetric survey in La Totora, this has to be taken as highly speculative. Both lineations and foliation planes show a systematic angularity with the border of the body in the western exposures of the La Portena granite. This is inconsistent with deformational fabric controlled solely by radial compression, and may be reflecting a

Fig. 7. Distribution of magnetic foliation planes for the La Totora, Renca and Las Chacras-Potrerillos batholiths. Due to the low density of AMS data for the Las Chacras-Potrerillos batholith foliation trajectories based on magnetic fabrics and field investigation are presented also. Note concentric sub-vertical foliations in the La Totora and Unit 1 of the Renca batholiths. The Unit 2 of the Renca batholith shows a 'dome' like distribution. Foliations tend to cross-cut internal limits in the Las Chacras-Potrerillos batholith. Note the different scale of each geological sketch. Data for the Renca batholith are from Lopez de Luchi et al. (20026), those for Las Chacras-Potrerillos batholith from Siegesmund et al. (2004).

Fig. 8. Distribution of magnetic lineations for the La Totora, Renca and Las Chacras-Potrerillos batholiths. Note that except for the Unit 1 of the Renca batholith, most lineations tend to be sub-horizontal or shallowly plunging. Note the different scale of each geological sketch. Data for the Renca batholith are from Lopez de Luchi et al. (20026), those for Las Chacras-Potrerillos batholith from Siegesmund et al. (2004).

AMS & EMPLACEMENT OF DEVONIAN GRANITOIDS

more complex stress pattern that produced shear in that part of the magmatic unit. It is noteworthy that the border of the pluton is particularly straight in that area, which may suggest that a previous NNW-SSE fault was significantly involved in the emplacement of the granite. Distribution of foliation planes and lineations for the Renca and Las Chacras-Potrerillos batholiths is shown in Figures 7c and 8c. In the latter case, due to the relatively scarce database, foliation trajectories have been interpreted additionally based on field observations (Siegesmund et al. 2004). In all cases, AMS foliation planes tend to be parallel to macroscopic foliations and generally follow the external border of the batholith. As in the case of La Totora, in the outer Unit 1 of the Renca and in most units of the Las Chacras-Potrerillos batholiths, foliations are sub-vertical to steeply inward inclined. Exceptions to this are the central Unit 2 of Renca and the eastern area of the central domain of the Las Chacras-Potrerillos batholith. The first one shows moderately dipping foliations in the centre that become increasingly inclined towards the internal contact with Unit 1. This is interpreted as a possible expression of magmatic fabric near the top of the body. In the latter batholith, although mainly concordant to the external border, foliation planes cross-cut most internal boundaries, suggesting that the foliation development post-dates the juxtaposition of all main units. As already mentioned, the plunge of the magnetic lineation in La Totora is generally of low to moderate angle and is interpreted as reflecting the magmatic flow direction. A similar interpretation is suggested for the internal Unit 2 of Renca, in which lineations show generally subhorizontal to shallow inclination. Except for a small area in the SW corner of the central domain, in the Las Chacras-Potrerillos batholith, the magnetic lineation is also of low to moderate inclination. Only the external Unit 1 of Renca has dominant sub-vertical lineations (Fig. 8b). This was interpreted as governed also by magma flow, although many sites were located very close to the external border of the batholith and may be affected by important flattening due to radial compression. Lineations are also discordant to most internal boundaries of the three batholiths. Figure 9 shows density plots (equal area projections, lower hemisphere) of the magnetic lineations (&max) for the three batholiths. In all of them, with perhaps the exception of the Las Chacras-Potrerillos central domain and the external Unit 1 of Renca, magnetic lineations tend to follow a NNE-SSW trend, which is more

463

evident in the northern and southern domain of the Las Chacras-Potrerillos and the La Totora batholiths. This systematic orientation of the magnetic lineations in several different granitic bodies suggests a regional control, either related to the far field stress pattern during the emplacement of the batholiths or the direction of magmatic flow following the direction of transtensional extension. Discussion

Significance of the magnetic fabrics Microfabrics mainly result from magmatic to sub-magmatic deformation except for the border zones. Therefore, measured fabrics mostly represent emplacement related fabrics and indicate that deformation of the magmas took place while they were not entirely crystallized and that there was no significant postemplacement deformation. Most of the units are characterized by a magmatic foliation defined by the preferred orientation of primary igneous minerals, predominantly K-feldspar megacrystals, plagioclase, amphibole and biotite. In the Las ChacrasPotrerillos batholith, the foliation is variably developed in the centre of the pluton and increases in intensity towards the western margin of the BPG and GPG. This intensification is locally accompanied by a minor hightemperature sub-solidus overprint. In the equigranular granitoids of the southern domain, the foliation is hard to recognize, apart from a shape-preferred orientation of quartz aggregates frequently noticed in the BRG. A first survey on the planar magnetic fabric anisotropy exhibits a marked conformity with the foliation obtained from field measurements. A deviation of macroscopic from magnetic fabrics has to be assigned to the low AMS sample density. Hence, it has to be proved if the AMS is able to reflect magmatic flow fabrics. To perform this a comparison of the AMS data with a calculated AMS tensor based on the biotite texture was made. With the exception of the GPG where amphibole with up to 6% contributes to the paramagnetic fabric, biotite is the only paramagnetic carrier. A set of samples from the sub-units of the Las Chacras-Potrerillos batholith was selected for U-stage biotite texture determination. The obtained data set was subsequently processed by a series of computer programs leading to the theoretical AMS tensor assuming a monomineralic rock composition (see Siegesmund et al. 1995). The obtained biotite

464

M. G. LOPEZ DE LUCHI ET AL.

Fig. 9. (a) Comparison of the biotite texture (density plot) based theoretical AMS tensor with the AMS experimental data (squares: ATmax, circles: A^ nt , triangles: K^^n). The P1 versus T diagram shows the calculated and experimental magnetic parameters. It is indicated that the measured biotite fabrics of samples SM 38, SM24, SM 104 and SM27 correspond to the magnetic directions. Differences between the calculated and experimental orientation of the magnetic fabric (SM 23 and SM 40) have to be addressed to weakly defined biotite textures and the limited sample volume acquired by biotite texture measurements. (Isolines within the theoretical AMS tensor plots represent the normalised bulk susceptibility) (b) P1 versus T diagram showing the discrepancy of calculated and experimental magnetic fabric parameters for most of the studied samples.

AMS & EMPLACEMENT OF DEVONIAN GRANITOIDS

textures of the samples SM 27, SM 38, SM 104, and SM 24 (Fig. 9a) are fairly able to explain the directional data of the measured AMS fabric. In contrast, for the samples SM 23 and SM 40 a large discrepancy occurs between the experimental and calculated AMS data. In case of the sample SM 23 from the EG domain this deficiency might be a result of the poorly constrained biotite texture, even though a large number of biotite crystals orientations were estimated. Similarly, the method failed for the porphyric fabric of the sample SM 40, which exhibits a highly undulating planar fabric with biotite crystals huddled against the large microcline phenocrysts. However, for the porphyric samples SM 104 and SM 27 from the BPG and GPG respectively, the orientation of the calculated AMS tensor coincides quite well with the experimental data (Fig. 9a). Therefore, the difference in orientation between the calculated and the experimental AMS tensor will be founded in the limited sample volume subjected to the biotite texture measurements. It is argued that the experimental AMS data is able to represent the fabric orientation of the sampled granitoids, due to the acquired larger sample volume. Comparing the calculated with the experimental degree of eccentricity and shape of the magnetic fabric ellipsoid, large differences for most of the samples are evident (Fig. 9b). Only in case of the samples SM 38 and SM 27 is a rough agreement of the two data sets obtained. However, numerical modelling approaches (Siegesmund & Becker 2000) have shown that due to a textured amphibole subfabric very different values for P1 and T are possible. This might account for sample SM27 of the GPG only. Nonetheless, exceptionally high bulk susceptibilities for most of the units of the Las Chacras-Potrerillos batholith will not be founded in the paramagnetic carriers of the magnetic properties, but have to be attributed to the ferromagnetic ore minerals that obliterate the paramagnetic fabric. Thermomagnetic measurements indicate that those ferromagnetic (sensu lato) ore minerals are dominated by magnetite and minor amounts of hematite (Mosch 2002).

Interpretation of batholith structures Foliations in the La Totora batholith (Fig. 7a) are parallel to the external border of the batholith but tend to cross-cut the internal boundaries. They are mostly sub-vertical, which may indicate a position close to the equatorial plane of the batholith.

465

Foliations in the Renca batholith (Fig. 7b) are steep to vertical and inward dipping in the porphyroid granitoids of the Unit 1. They are parallel to the X-Y flattening plane of the enclaves and the syn-plutonic dykes attitude but cross-cut locally the contact between the two major units of the batholith (Lopez de Luchi et al. 2002c). These internal features are concordant with the axial planes of the mesoand microscale folding of the regional NNE striking pervasive foliation in the ductile aureole that developed in the medium grade metamorphic Ordovician country rocks. Unit 2 shows a planar magnetic fabric shallowly dipping outwards. The foliation trajectories in the Las ChacrasPotrerillos batholith (Fig. 7c) define roughly three areas at the plutons scale: NW, central and SE. In the first, a concentric funnel-shaped foliation pattern is roughly concordant to the contact with the country rock. In the central area, the foliation is either parallel or cut across the internal contacts at different angles despite maintaining a concordant relationship with the plutons western margin. At the eastern contact, the magmatic foliation lies at moderate to high angles with the contact and bends into parallelism when traced towards the northernmost sector. The foliation is steep to inward dipping except for some areas along the eastern border where an outward dipping foliation is observed. Country rock septa are parallel to both the foliation and the contacts. In the southernmost part of the central sector, sharp inflections of the foliation trajectories are locally associated with magmatic shear zones. In the SE sector, most of the foliation planes are steeply inclined. Patterns of the foliation trajectories are elliptic with a major NNE axis. They tend to be parallel to the external contact but cross-cut internal contacts. Within a 3km wide NNE trending belt along the border between the central and southern domain the heterogeneous high to low temperature solid-state deformation overprints the magmatic fabrics and leads to a more penetrative NNE fabric. The lineation in the La Totora intrusion is mainly shallowly plunging either to the NNE or SSW. The local appearance of highly inclined lineations in the centre of the batholith might reflect a vertical magmatic flow direction (Fig. 10). In Unit 1 of Renca, lineations close to the eastern and western contacts are shallowly dipping, whereas they are steeper at the north and south boundaries. Inward plunging lineations at the north and south represent zones of limited ballooning (Lopez de Luchi et al. 2002c). The linea-

466

M. G. LOPEZ DE LUCHI ET AL.

Fig. 10. Stereographic projections (equal area, lower hemisphere) for the magnetic lineations of the La Totora, Renca and Las Chacras-Potrerillos batholiths. Data for La Totora and Renca batholiths correspond to mean site lineations whereas data for the Las Chacras-Potrerillos correspond to sample directions. Note that N-NE lineations are dominant in most units. Data for the Renca batholith are from Lopez de Luchi et al. (2002&), those for Las Chacras-Potrerillos batholith from Siegesmund et al. (2004).

tions in the central Unit 2 strike mainly NNESSW (Figs. 8b, 10). The obtained linear fabric within the Las Chacras-Potrerillos batholith generally shows a NNE-SSW trend. This trend is most conspicuous in the northern and southern domain, where £max axes fit along a NNE-SSW striking great circle that indicates the prevalence of shallow plunging lineations. In contrast &max axes within the central part of the batholith exhibit a scattered arrangement along a NESW striking, slightly NW inclined great circle, depicting a broad variation from shallowly to steeply plunging lineations (Fig. 10). Due to the complex processes in the formation of magmatic fabrics, the question arises at which stage between magma ascent and final emplacement the fabric pattern was developed. From all available data, it is apparent that the foliation post-dates the compositional zoning because a general structural continuity is found across all

the subunits. Therefore, it can be inferred that magmas were deformed contemporaneously during their emplacement and the fabric was acquired during a late stage of the emplacement history.

Emplacement Emplacement models applied to the Devonian batholiths should account for the predominance of magmatic fabric, the ellipsoidal outcrop shape, the stresses related with magma input, narrow ductile aureoles and the influence of a regional stress field related with active regional scale shear zones. The three analysed batholiths are characterized by dominant oblate shape of the magnetic fabric ellipsoid, weak linear fabric independently of its plunge, concentric nature of the planar fabric and the deflection or folding of the country rock

AMS & EMPLACEMENT OF DEVONIAN GRANITOIDS

parallel to the pluton contacts (Fig. 7), which suggest predominantly flattening strain. These characteristics define in part a concentrically expanded pluton (Paterson & Vernon 1995) for which multiple processes for ascent and emplacement were proposed. Exceptional are localized belts where a higher deformation, as indicated by high-temperature solid-state microstructures, correlates with higher Pf values and shallow plunging lineations (Figs 5, 8). In order to generate a batholith, melts are needed that could migrate up to a level where there is a switch from upward to sub-horizontal flow (Petford et al. 2000). Then, any spacecreating mechanism that could accommodate the magma volumes is required. Most of the controversy over the space problem refers to the role of the pluton as either the space maker through expansion or the passive marker of processes operating outside its aureole. Diapirism vs. dyke-controlled ascent are two extreme ascent mechanisms that were proposed for concentric expanded plutons (see Paterson & Miller 1998; Miller & Paterson 1999; Petford et al. 2000; Dietl & Koyi 2002 and references therein). In between these extreme behaviours of the host, i.e. elastic for the dykes and viscous flow in the diapirs, the visco-elastic diapirs represent a type associated with a complex rheology of the host that ranges from elastic to viscous (Paterson & Miller 1998; Miller & Paterson 1999). Petrological studies on the Devonian batholiths of the Sierra de San Luis demonstrate that the monzonite rocks represent mantle-derived magmas, whereas the granodiorite-granite formed by mixing of mantle and crustal derived melts (Lopez de Luchi 1986, 1996; Lopez de Luchi et al. 2004). This suggests multiple batches of magma that could have mixed due to the lower viscosity contrast that can occur along preheated pathways that would favour a viscoelastic diapir model (Paterson & Miller 1998; Miller & Paterson 1999). On the other hand, dyke-controlled ascent is favoured by tectonic overpressure but limited mixing is expected. The depth of emplacement is constrained by hornblende thermobarometry to approximately 12-14 km (Lopez de Luchi et al. 2004) The lack of a widespread contact metamorphic overprint in the host of the Devonian batholiths together with the preservation of muscovite cooling ages in the contact aureole older than the pluton crystallization ages (Table 1; Lopez de Luchi et al. 2002a; Siegesmund et al. 2004; Steenken et al. 2004-in review) suggest that temperatures in the contact aureole were below 350 °C. Contact metamorphic minerals are represented by scarce biotite flakes that display a discontinuous

467

foliation parallel to the axial plane of the microfolds that deformed the NNE regional pervasive foliation (Lopez de Luchi 1986, 1996). Narrow ductile aureoles led to a general concordance between country rock foliation and contacts. This narrow deformation aureole together with the mainly magmatic fabrics suggest that the former more likely record some degree of lateral spreading of a laccolith intrusion (Benn et al. 1999). On the other hand, narrow ductile aureoles were considered as the result of the behaviour of the country rocks as a power law material that together with a pre-heated pathway for magma ascent would enable magmas to ascend as visco-elastic diapirs (Miller & Paterson 1999). In both the Renca and La Totora batholiths a first generation of aplitic dykes define a radial pattern, whereas the latest stage epidote tourmaline veins are preferentially oriented in the WNW direction, i.e. perpendicular to the regional NNE stretching direction (Rossello et al. 1999). Concentric foliation patterns together with radial dykes that cross-cut the foliation trend may indicate flattening of the magma due to a balloon-like inflation (cf. Hutton & Siegesmund 2001) and subsequent horizontal stretching. In order to control the more qualitative problem of the far-field controls on the emplacement of the plutons the regional framework has to be considered. A main set of NNW and NNE trending ductile shear zones of Cambrian, Ordovician and Devonian ages are recognized in the Eastern Sierras Pampeanas (Whitmeyer & Simpson 2003; von Gosen & Prozzi 1998; Sims et al. 1998). Mid-Palaeozoic resumption of convergence on the western margin of Gondwana is evidenced by a widespread compressive deformation in the Sierras Pampeanas (Sims et al. 1997). This Devonian cycle is referred to as the Achalian and would be defined by the collision of the Chilenia terrane with the Precordillera terrane (Astini 1996; Sims et al. 1998). Deformation is partitioned between domains of west-directed thrusting and strike-slip with extension beyond the western Sierras Pampeanas. In the Sierras de Cordoba, Sims et al. (1998) reported Ar/Ar mica ages ranging from 358 to 451 Ma for the NNW trending Guamanes shear zone. In addition, a Devonian reactivation of the Ordovician Los Tuneles shear zone (Martino et al. 2002) based on K-Ar amphibole ages of 365-373 Ma. In the Sierras de Comechigones, Devonian ages are also assigned to the NW trending Las Lajas shear zone and to Las Albahacas shear zone. The NW oriented shear zones are also well documented by strong linear trends of moderate to highly magnetic, strike

468

M. G. LOPEZ DE LUCHI ET AL.

parallel anomalies (Sims et al. 1997). In the Sierra de San Luis, the NNE trending shear zone along the eastern contact of the La Escalerilla pluton (Ar/Ar muscovite age of 344.7 ± 0.5 Ma) and the NNE striking Rio Guzman shear zone (Ar/

Ar muscovite ages of 375-351 Ma) were described (Sims et al. 1997, 1998). WNW-ESE oriented lamprophyric dykes and tourmalineepidote veins appear in the basement of Conlara complex (Rossello et al. 1999).

Fig. 11. Satellite image showing the Sierra de San Luis. The major Devonian batholiths are well accentuated by their smooth relief. Depicted are the major regional scale shear zones. Extensional step over structures point to the sinistral character of the NNE-SSW striking shear zones. A secondary set of NNW-SSE trending shear zones is less well documented. On site findings indicate the sinistral character of those structures. They were interpreted to facilitate the counter clockwise step over of the sinistral displacement in a transtensional regime.

AMS & EMPLACEMENT OF DEVONIAN GRANITOIDS

As indicated from microstructures the plutons are locally affected by a solid-state overprint. A continuum from magmatic to high-temperature solid-state microstructural features indicates the involvement of regional shear zones in the emplacement process of the batholiths. This is in accordance with the obtained age spectra for the activity of the Rio Guzman shear zone (Sims et al. 1998). A detailed satellite image survey clearly depicts this NNW-SSE striking shear zones that is prominent from field observations also. Additional fault lines (Fig. 11) parallel the Rio Guzman shear zone towards the east and west. Extensional overstep structures manifest the sinistral transtensional character of these shear zones. A second set of NNE-SSW striking shear zones is less well documented from field investigations, but clearly visible in the satellite image (Fig. 11). Where observed in the field, structures indicate a left lateral displacement as indicated from the deflection of the pervasive foliation. These fault lines are interpreted to accommodate the anticlockwise step over of the sinistral displacement along the Rio Guzman and related fault lines. In fact, no continuation of this lineament to the north of the Las Chacras-Potrerillos batholith has been reported yet. The contemporaneous activity of both strikeslip systems led to a NNE-SSW directed extension allowing the formation of magma conduits (Fig. 12). Subsequent inflation of the batholiths follows the direction of space creation. The latter mainly accounts for the Las ChacrasPotrerillos and La Totora batholiths, with a marked NNE-SSW lineation trend. Contrasting is the steep marginal lineation of the Renca batholith. However, the same mechanism for magma ascent is suggested. The deep erosion level of the body of Renca, as indicated from the funnel arrangement of the marginal foliation, denotes a position below the equatorial plane of the pluton. Steep lineations in this level will reflect the ascent path of the magma, rather than the direction of expansion. However, the observation of shallowly plunging roughly NNE-SSW trending lineations in the central Unit 2 argues for the typically observed NNE directed inflation of a balloon. Crystallization ages of the Devonian batholiths (Lopez de Luchi et al. 2003 and references therein), structural features, and geochronological constraints on the activity of the crustal scale shear zones of the Sierra de San Luis indicate that at a regional scale the emplacement of the Devonian batholiths is coeval with the activity of the shear zones. Therefore, the Devonian magmatism is not post-tectonic regarding the Famatinian cycle in the Ordovician but results

469

from another compressive event that reflects the protracted convergence along the southwestern margin of Gondwana. Conclusions The Sierra de San Luis is characterized by the generation of voluminous Devonian granitoids. Most of them exhibit ellipsoidal outcrop shapes. Field observations, geochemical and structural data suggest that they are pulsed, composite intrusions. The available ages reveal crystallization between 400 and 380 Ma, while the cooling below 350 °C took place between 380 Ma and 350 Ma. Observed fabrics of the investigated batholiths are largely related to magmatic flow. Sub-magmatic and minor amount of high-temperature solid-state deformation along the internal contacts of the different sub-units is in agreement with emplacement kinematics related to the opposed forces of magma buoyancy and the regional strain field. Moreover, localized belts of intensified high (to low) temperature solid-state deformation, overprinting the pre-existing (sub-)magmatic fabrics reflect the activity of regional scale shear zones during the emplacement of the magmas. Magnetic data clearly show that the three investigated plutons (La Totora, Renca and Las Chacras-Potrerillos) are largely made up by ferromagnetic granitoids (k occasionally greater than 10~2 SI). The determined magnetic foliations tend to parallel the macroscopic foliations. In the La Totora, Renca and Las ChacrasPotrerillos batholiths, foliations are sub-vertical to steeply inward inclined. In most of the different units of the plutons, the magnetic lineation is of low to moderate inclination with a mainly NNE-SSW trend, except for the external Unit 1 of Renca where a more steeply to locally subvertical lineation was observed. Devonian age determinations on the fault rocks and granitoids point to a syn-tectonic emplacement of the batholiths. The plutons are interpreted to be positioned at the crossover of sinistral shear zones. The origin of this NNE directed extensional setting in a transpressive regime seems to be related to the transfer of emplacement along a secondary set of NNW trending sinistral faults. The final emplacement is due to a subsequent balloon-like inflation of the batholith following the direction of space creation. This model is in accordance with the relative timing of the emplacement sequence and macroscopically visible planar fabrics in the field as well as magnetic fabric data. Therefore, the Devonian batholiths are syn-kinematic with respect to the

470

M. G. LOPEZ DE LUCHI ET AL.

Fig. 12. Schematic model showing the transtensional stress field during the emplacement of the Devonian granitoids of the Sierra de San Luis, (a) Schematic block diagrams illustrating the inferred shape of the batholiths based on magnetic and macroscopic fabrics, (b) Cartoon/sketch illustrating the syn-tectonic emplacement of the major Devonian granitoids in the Sierra de San Luis. Schematic depiction of the major tectonic strike slip faults connected with the 'space-creation' for the magma ascent. The development of a secondary set of NNW trending sinistral strike slip faults leads to the counter clockwise step over of the sinistral displacement along the Rio Guzman and related NNE-SSW trending shear zones. Consequently, NNE directed crustal extension would allow the formation of magma conduits and subsequent magma accommodation in some kind of pull-apart structure.

Achalian deformational event. The active tectonics could also favour the ascent of mantle-derived magmas that are associated with the granitoids in all the plutons. Our study suggests that even granitoids with mainly magmatic fabric that cross-cut the host rock fabric at a regional scale might show their

syn-kinematic character after a careful structural study. The authors are thankful for the constructive and helpful review by J. Y. Talbot, which helped to improve the manuscript. For the careful editorial handling of our manuscript, we are in debt to F. Hernandez. Her additional comments were an advantage to the contents

AMS & EMPLACEMENT OF DEVONIAN GRANITOIDS of the paper. Geochronological constraints were kindly enabled by K. Wemmer and his K/Ar laboratory staff at the Geoscience Centre of the University of Gottingen. Moreover, constructive discussion with K. Wemmer was essential for the understanding of the Devonian history in the Sierra de San Luis. Both should be highly appreciated. Financial support for the study was granted by the DAAD-ANTORCHAS program and the DFG (Si 438/16-1). A. St. is grateful for the DFG research scholarship STE 1036/1-1.

References ARCHANJO, C. J. & BOUCHEZ, J. L. 1997. Magnetic fabrics and microstructures of the post-collisional aegirine-augite syenite Triunfo pluton, northeast Brazil. Journal of Structural Geology, 19, 849-860. ARCHANJO, C. J., BOUCHEZ, J. L., CORSINI, M. & VAUCHEZ, A. A. 1994. The Pombal granite pluton: magnetic fabric, emplacement and relationship with the Brasiliano strike-slip setting of NE Brazil (Paraiba State). Journal of Structural Geology, 16, 323-335. ASTINI R. A. 1996. Las fases diastroficas del Paleozoico Medio en la Precordillera del oeste argentinoevidencias estratlgraficas. XII Congreso Geologico Argentine y IV Congreso de Exploration de Hidrocarburos, Actas V, 509-526. BARRIERE, M. 1977. Deformation associated with the Ploumanac'h intrusive complex, Brittany. Journal of the Geological Society, London, 134, 311-324. 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, 21^-244. BECKER J. K., Siegesmund S. & Jelsma H. A. 2000. The Chinamora batholith, Zimbabwe: structure and emplacement-related magnetic rock fabric. Journal of Structural Geology, 22, 1837-1853. BENN, K., HAM. N. & PIGNOTTA, G. S. 1998. Emplacement and deformation of granites during transpression: magnetic fabrics of the Archean Sparrow Pluton, Slave Province, Canada. Journal of Structural Geology, 20, 1247-1259. BENN, K., HORNE, R. J., KONTAK, D. J., PIGNOTTA, G. S. & EVANS, N. G. 1997. Syn-Acadian emplacement model for the South Mountain batholith, Meguma Terrane, Nova Scotia: magnetic fabric and structural analyses. Geological Society of America, Bulletin, 109, 1279-1293. BENN, K., ROEST, W. R., ROCHETTE, P., EVANS, N. G. & PIGNOTTA, G. S. 1999. Geophysical and structural signatures of syntectonic batholith construction: the South Mountain Batholith, Meguma Terrane, Nova Scotia. Geophysical Journal International, 136, 144-158. BENN, K., PATERSON, S. R., LUND, S. P., PIGNOTTA, G. S. & KRUSE, S. 2001. Magmatic fabric in batholiths as markers of regional strain and plate kinematics: example of the Cretaceous Mt. Stuart batholith. Physics and Chemistry of the Earth (A), 26, 343-354.

471

BOUCHEZ J. L. 1997. Granite is never isotropic: An introduction to AMS studies of granite rocks. In: BOUCHEZ, J. L., HUTTON, D. W. H. & STEPHENS, W. E. (eds) Granite: From Segregation of Melt to Emplacement Fabrics. Kluwer Academic Publishers, Amsterdam, 95-112. BOUCHEZ, J. L. 2000. Anisotropie de susceptibilite magnetique et fabrique des granites. Comptes Rendus de rAcademic des Sciences, Series IIA, Paris, 330, 1-14. BOUCHEZ, J. L., GLEIZES, G., DJOUADI, T. & ROCHETTE, P. 1990. Microstructures and magnetic susceptibility applied to emplacement kynematics of granites: the example of the Foix pluton (French Pyrenees). Tectonophysics, 184, 157-171. BOUCHEZ, J. L., DEALS, C., GLEIZES, G., NEDELEC, A. & LUNEY, M. 1992. Submagmatic microfractures in granites. Geology, 20, 35-38. CRIADO ROQUE, P., MOMBRU, C. & RAMOS, V. A. 1981. Estructura e interpretation tectonica: In: IRIGOYEN, M. (ed) Geologia y Recursos Miner ales de la provincia de San Luis. VIII Congreso Geologico Argentine, Relatorio, 155-192. CRUDEN, A. R., KOYI, H. & SCHMELING, H. 1995. Diapiric basal entrainment of mafic into felsic magma. Earth and Planetary Science Letters, 131, 321-340. CRUDEN, A. R., TOBISCH, O. T. & LAUNEAU, P. 1999. Magmatic fabric evidence for conduit-fed emplacement of a tabular intrusion: Dinkey Creek Pluton, central Sierra Nevada batholith, California. Journal of Geophysical Research (B5), 104, 10511-10530. DALLA SALDA, L. H., LOPEZ DE LUCHI, M. G., CINGOLANI, C. & VARELA, R. 1998. LaurentiaGondwana collision: the origin of the Famatinian-Appalachians Orogenic Belt. In: PANKHURST, R. J. & RAPELA, C. W. (eds) The Proto-Andean Margin of Gondwana. Geological Society, London, Special Publications, 142, 219-234. DAVIS, J. S., ROESKE, S. M., MCCLELLAND, W. C. & SNEE, L. W. 1999. Closing the ocean between the Precordillera Terrane and Chilenia: early Devonian ophiolite emplacement and deformation in the south west Precordillera. In: RAMOS, V. A. & KEPPIE, J. D. (eds) Laurentia Gondwana connections before Pangea. Geological Society of America, Special Paper, 336, 115-138. DIETL, C. & KOYI, H. A. 2002. Emplacement of nested diapers: results of centrifuge modeling. Journal of the Virtual Explorer, 6, 81-88. DJOUADI, M. T., GLEIZES, G., FERRE, E., BOUCHEZ, J. L., CABY, R. & LESQUER, A. 1997. Oblique magmatic structures of two epizonal granite plutons, Hoggar, Algeria: late-orogenic emplacement in a transcurrent orogen. Tectonophysics, 279, 351374. GLEIZES, G. L., NEDELEC, A., BOUCHEZ, J. L., AUTRAN, A. & ROCHETTE, P. 1993. Magnetic susceptibility of the Mount-Louis Andorra Ilmenite-Type Granite (Pyrenees): A new tool for the petrographic characterisation and regional mapping of zone granite plutons. Journal of Geophysical Research, 98,4317-4331.

472

M. G. LOPEZ DE LUCHI ET AL.

GONZALEZ, P. D. & SATO, A. M. 2000. Los plutones monzoniticos cizallados El Molle y Barroso: dos nuevos intrusivos pos-orogenicos en el oeste de las sierras de San Luis, Argentina. IX Congreso Geologico Chileno, Puerto Varas, 1 (sesion tematica 4), 621-625. GONZALEZ, P. D., SATO, A.M., BASEI, M. A. S., VLACH, S. R. F. & LLAMBIAS, E. 2002. Structure metamorphism and age of the Pampean Famatinian orogenies in the western Sierra de San Luis. XV Congreso Geologico Argentine, Calafate, Actas II, 51-56. GOSEN, W. VON 1998. Transpressive deformation in the southwestern part of the Sierra de San Luis (Sierras Pampeanas, Argentina). Journal of South American Earth Sciences 11, 233-264. GOSEN, W. VON & PROZZI, C. R. 1996. Geology, Structure and metamorphism in the area south of La Carolina. XIII Congreso Geologico Argentine y III Congreso de Exploracion de Hidrocarburos, Buenos Aires, Actas II, 301-314. GOSEN, W. VON & PROZZI, C. R. 1998. Structural evolution of the Sierra de San Luis (Eastern Sierras Pampeanas, Argentina): implications for the Proto-Andean margin of Gondwana. In: PANKHURST, R. J. & RAPELA, C. W. (eds) The ProtoAndean Margin of Gondwana. Geological Society, London, Special Publications, 142, 235-258. GREGOIRE, V., DARROZES, P., GAILLOT, P., NEDELEC, A. & LAUNEAU, P. 1998. Magnetite grain shape fabric and distribution anisotropy vs. rock magnetic fabric: a three-dimensional case study. Journal of Structural Geology, 20, 937-944. HAUZENBERGER, C. A., MOGESSIE, A., HOINKES, G., FELFERNIG, A., BJERG, E. A., KOSTADINOFF, J., DELPINO, S. & DIMIERI, L. 2001. Metamorphic evolution of the Sierras de San Luis, Argentina: granulite facies metamorphism related to mafic intrusions. Mineralogy and Petrology, 71, 95-126. MUTTON, D. H. W. & SIEGESMUND, S. 2001. The Ardara Granite: Reinflating the Balloon Hypothesis. Zeitschrift der Deutschen Geologischen Gesellschaft 152, 309-323. JELINEK, V. 1977. The statistical theory of measuring anisotropy of magnetic susceptibility of rock and its application. Geofyzika Brno, Czechoslovakia. JELINEK, V. 1978. Statistical processing of of magnetic susceptibility measured in groups of specimens. Studia Geophysica et Geodaetica, 22, 50-62. JELINEK, V. 1981. Characterization of magnetic fabric of rocks. Tectonophysics, 79, 563-567. LLAMBIAS E. J., SATO A. M., ORTIZ SUAREZ A. E. & PROZZI C. 1998. The granitoids of the Sierra de San Luis. In: PANKHURST, R. J. & RAPELA, C. W. (eds) The Proto-Andean Margin of Gondwana. Geological Society, London, Special Publications, 142, 325-341. LOPEZ DE LUCHI, M. G. 1986. Geologia y Petrologia del basamento de la Sierra de San Luis, Region del Batolito de Renca. PhD thesis, Departamento de Geologia. Facultad de Ciencias Exactas, Universidad de Buenos Aires. LOPEZ DE LUCHI, M. G. 1993. Caracterizacion geologica y emplazamiento del Batolito de Renca. XII

Congreso Geologico Argentino y II Congreso de Exploracion de Hidrocarburos, Mendoza, Argentina, Actas IV, 42-53. LOPEZ DE LUCHI, M. G. 1996. Enclaves en un batolito postectonico: petrologia de los enclaves microgranulares del batolito de Renca. Revista de la Asociacion Geologica Argentina, 51, 131-146. LOPEZ DE LUCHI, M. G. & DALLA SALDA, L. H. 1997. Late Famatinian Granitoids: a progressive stabilizing crust in SW South America. Terrane Dynamics 97, International Conference on Terrane Geology, University of Christchurch, New Zealand. Conference Abstracts, 103-106. LOPEZ DE LUCHI, M. G., DALLA SALDA, L. H., CINGOLANI, C. & VARELA, R. 1998. An overview of the granitoids of the Famatinian Orogenic Belt in Pampean Ranges. 14th International Conference on Basement Tectonics, Departmento Geologia, Escola de Minas, Univ. Federal de Ouro Preto, Brazil and International Basement Tectonics Association, Abstracts, 89-92. LOPEZ DE LUCHI, M. G., ROSSELLO, E. A. & LE CORRE, C. 2000. Emplacement related structures of the Renca Batholith, Sierra de San Luis, Argentina. Geophysical Research Abstracts, 2 (CD edition). LOPEZ DE LUCHI, M. G., OSTERA, H. A., LINARES, E. & ROSSELLO, E. A. 200la. Preliminary calculations on the cooling rate of the Renca batholith, Sierra de San luis, Argentina. In: TOMLINSON, A. (ed) Edicion Especial HI South American Symposium on Isotope Geology, Pucon, Chile, Abbreviated Abstracts Volume, Revista Comunicaciones, 52, 15. LOPEZ DE LUCHI, M. G., SIEGESMUND, S., HOFMANN, A., HUBNER, H., HULKA, C. & MOSCH, S. 200\b. Geological setting and composition of the Las Chacras-Potrerillos Batholith, Sierras Pampeanas, Argentina: First results. Zeitschrift der Deutschen Geologischen Gesellschaft, 152, 325-350. LOPEZ DE LUCHI, M. G., FANTIN, M. & RAPALINI, A. E. 2002<2. Granito La Totora, Sierras de San Luis, Argentina: Primeros Resultados. XV Congreso Geologico Argentino, Calafate, Actas I, 269-274. LOPEZ DE LUCHI, M. G., HOFFMANN, A., SIEGESMUND, S., WEMMER, K. & STEENKEN, A. 20026. Temporal constraints on the polyphase evolution of the Sierra de San Luis. Preliminary report based on biotite and muscovite cooling ages. XV Congreso Geologico Argentino, Calafate, Actas I, 309-315. LOPEZ DE LUCHI, M. G., RAPALINI, A. E., ROSSELLO, E. & GEUNA, S. E. 2002c. Rock and magnetic fabric of the Renca batholith (Sierra de San Luis, Argentina): constraints on its emplacement. Lithos, 61, 161-186. LOPEZ DE LUCHI, M. G., CERREDO, M. E., SIEGESMUND, S., STEENKEN, A. & WEMMER, K. 2003. Provenance and tectonic setting of the protoliths of the metamorphic complexes of Sierra de San Luis. Revista de la Asociacion Geologica Argentina, 58, 525-540. LOPEZ DE LUCHI, M. G., SIEGESMUND, S., WEMMER, K., STEENKEN, A. & NAUMANN, R. 2004 - in review. Paleozoic tectonomagmatic evolution of the Sierra de San Luis (Sierras Pampeanas, Argentina). Lithos.

AMS & EMPLACEMENT OF DEVONIAN GRANITOIDS MARTINO, R. D., GUERRESCHI, A., SFRAGULLA, J. A. 2002. Deformation fragil y realciones regionales de la faja de deformation Los Tuneles en las Sierras de Pocho y Guasapampa, Cordoba, Argentina. XV Congreso Geologico Argentine, Calafate, CD edition. MILLER, R. B. & PATERSON, S.R. 1999. In defense of magmatic diapirs. Journal of Structural Geology, 21, 1161-1173. MOSCH, S. 2002. Strukturelle und Gesteinsmagnetische Untersuchungen an den Granitoiden des Las Chacras-Potrerillos Batholith, Sierra de San Luis, Argentinien. Unpublished Diploma thesis, Geoscience Centre of the University of Gottingen, Department of Structural Geology and Geodynamics, University of Gottingen. NEVES, S. P., VAUCHEZ, A. & ARCHANJO, C. J. 1996. Shear zone-controlled emplacement or magma assisted nucleation of shear zones? Insights from northeast Brazil. Tectonophysics, 262, 349-364. NEVES, S. P., ARAUJO, A. M. B., CORREIA, P. B. & MARIANO, G. 2003. Magnetic fabrics in the Cabanas Granite (NE Brazil): interplay between emplacement and regional fabrics in a dextral transpressive regime. Journal of Structural Geology, 25, 441-453. ORTIZ SUAREZ, A. E., PROZZI, C. & LLAMB!AS, E. 1992. Geologia de la parte sur de la Sierra de San Luis y granitoides asociados. Estudios Geologicos, 48, 269-277. PATERSON, S. R. & MILLER, R. 1998. Mid-crustal magmatic sheets in the Cascades Mountains, Washington: implications for magma ascent. Journal of Structural Geology, 20, 1345-1363. PANKHURST, R.J. & RAPELA, C.W. (eds) 1998. The Proto-Andean Margin of Gondwana. Geological Society, London, Special Publications, 142. PATERSON, S. R. & VERNON, R. H. 1995. Bursting the 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. Journal of Structural Geology, 11, 349-363. PATERSON, S. R., FOWLER, T. K., SCHMIDT, K. L., YOSHINOBU, A. S., YUAN E. S. & MILLER, R. B. 1998 Interpreting magmatic fabric patterns in plutons. Lithos, 44, 53-82. PETFORD, N., CRUDEN, A. R., MCCAFFREY, K. J. W. & VIGNERESSE, J. L. 2000. Granite Formation, transport and emplacement in the crust. Nature, 408, 669-673. QUENARDELLE, S. 1993. Caracterization geologicapetrologica del Granito San Jose del Morro, provincia de San Luis. XII Congreso Geologico Argentine y II Congreso de Exploration de Hidrocarburos, Mendoza, Actas IV, 61-67. QUENARDELLE, S. & RAMOS, V. A. 1999. Ordovician western Sierras Pampeanas magmatic belt: Record of Precordillera accretion in Argentina. In: RAMOS, V. A. & KEPPIE, J. D. (eds) Laurentia Gondwana Connections before Pangea. Geological Society of America, Special Paper, 336, 63-86.

473

RAMOS, V. A. 1988. Late Proterozoic-Early Paleozoic of South America - a collisional history. Episodes, 11, 168-174. RAMOS V. A. & KEPPIE, J. D. (eds) 1999. Laurentia Gondwana Connections before Pangea. Geological Society of America, Special Paper, 336. RAMOS, V. A., JORDAN, T. E., ALLMENDINGER, R. W., MPODOZIS, C., KAY, S. M., CORTES, J. M. & PALMA, M. 1986. Paleozoic terranes of the Central Argentine - Chilean Andes. Tectonics, 5, 855-880. RAPALINI, A. E., LOPEZ DE LUCHI, M. G. & FANTIN, M. 2003. Magnetic Characterization of the Devonian La Totora Granite, Sierras de San Luis, Argentina. X Congreso Geol. Chileno, Conception, CD edition. ROCHETTE, P. 1987. Magnetic susceptibility of the rock matrix related to magnetic fabric studies. Journal of Structural Geology, 9, 1015-1020. ROCHETTE, P., JACKSON, M. & AUBOURG, C. 1992. Rock magnetism and the interpretation of anisotropy of magnetic susceptibility. Reviews of Geophysics, 30, 209-226. ROSSELLO, E. A., LOPEZ DE LUCHI, M. G., MASSABIE, A., CERREDO, M. E., ALONSO, M. S. & SPIKERMANN, J. P. 1999. Control estructural de los filones tardio-magmaticos emplazados en el Batolito de Renca (San Luis, Argentina). XIV Congreso Geologico Argentine, Actas I, 145148. SAINT BLANQUAT, M. & TIKOFF, B. 1997. Development of magmatic to solid-state fabrics during syntectonic emplacement of the Mono Creek Granite, Sierra Nevada Batholith. In: BOUCHEZ, J. L., HUTTON, D. W. H. & STEPHENS, W. E. (eds) Granite: From segregation of melt to emplacement fabrics. Kluwer Academic Publishers, Amsterdam, 231-252. SALFITY, J. A. & GORUSTOVICH, S. 1983. Paleogeographia de la cuenca del Grupo Paganzo (Paleozoico Superior). Revista Asociacion Geologica Argentina, 38, 437-453. SIEGESMUND, S. & BECKER, J. 2000. Emplacement of the Ardara pluton (Ireland): new constraints from magnetic fabrics, rock fabrics and age dating. International Journal of Earth Sciences, 89, 307327. SIEGESMUND, S., STEENKEN, A., LOPEZ DE LUCHI, M. G. & WEMMER, K. 2004. The Las Chacras-Potrerillos Batholith: Structural evidences on its emplacement and timing of the Intrusion. International Journal of Earth Sciences, 93, 23-43. SIEGESMUND, S., ULLEMEYER, K. & DAHMS, M. 1995. Control of magnetic rock fabrics by mica preferred orientation: a quantitative approach. Journal of Structural Geology, 17, 1601-1613. SIMS, J. P., SKIRROW, R. G., STUART-SMITH, P. G. & LYONS, P. 1997. Informe geologico y metalogenetico de las Sierras de San Luis y Comechingones (provincias de San Luis y Cordoba), 1:250000. Anales, XXVIII, Institute de Geologia y Recursos Minerales, SEGEMAR, Buenos Aires. SIMS, J. P., IRELAND, T. R., ET AL. 1998. U-Pb, Th-Pb and Ar/Ar geochronology from the southern

474

M. G. LOPEZ DE LUCHI ET AL.

Sierras Pampeanas, Argentina: implications for the Paleozoic evolution of the western Gondwana margin. In: PANKHURST, R. J. & RAPELA, C. W. (eds) The Proto-Andean Margin of Gondwana. Geological Society, London, Special Publications, 142,259-281. STEENKEN, A., SIEGESMUND, S. & HEINRICHS, T. 2000. The emplacement of the Rieserferner pluton (eastern Alps, Tyrol): Constraints from field observations, magnetic fabrics and microstructures. Journal of Structural Geology, 22, 18551873. STEENKEN, A., LOPEZ DE LUCHI, M. G., SIEGESMUND, S. & WEMMER, K. 2002. An insight into the structural evolution of the Sierra de San Luis (southeastern Sierras Pampeanas, Argentina); a progress report. XV Congreso Geologico Argentine, Calafate, Actas I, 321-325. STEENKEN, A., WEMMER, K., LOPEZ DE LUCHI, M. G. & SIEGESMUND, S. 2003. The cooling history of the basement complex of the Sierra de San Luis (Eastern Sierras Pampeanas/Argentina) and its implication for Famatinian vs. Achalian events based on K/Ar-geochronology. 4th Symposium on Isotope Geology, Bahia, Brazil, Short Papers I, 117-120. STEENKEN, A., LOPEZ DE LUCHI, M. G., SIEGESMUND, S., WEMMER, K. & PAWLIG, S. 2004 - in review. Crustal provenance and cooling of the basement complexes of the Sierra de San Luis: An insight into the tectonic history of the proto-Andean margin of Gondwana. Gondwana Research. STUART-SMITH, P. G., CAMACHO, A., SIMS, J. P., SKIRROW, R. G., LYONS, P., PIETERS, P. E. & BLACK, L. P. 1999. Uranium lead dating of felsic magmatic cycles in the southern Sierras Pampeanas, Argentina: Implications for the tectonic development of the Proto-Andean Gondwana

margin. In: RAMOS, V. A. & KEPPIE, J. D. (eds) Laurentia Gondwana Connections before Pangea. Geological Society of America, Special Paper, 336,87-114. TALBOT, J. Y., MARTELET, G., COURRIOUX, G., CHEN, Y. & FAURE, M. 2004. Emplacement in an extensional setting of the Mont Lozere-Borne granitic complex (SE France) inferred from comprehensive AMS, structural and gravity studies. Journal of Structural Geology, 26, 11-28. TARLING, D. H. & HROUDA, F. 1993. The Magnetic Anisotropy of Rocks. Chapman and Hall, London. TIKOFF, B. & TEYSSIER, C. 1992. Crustal scale enechelon 'P-shear' tensional bridges: A possible solution to the batholith room problem. Geology, 20, 927-930. VARELA, R., LLAMBIAS, E., CINGOLANI, C. A. & SATO, A. M. 1994. Datacion de algunos granitoides de la Sierra de San Luis (Argentina) e interpretation evolutiva. VII Congreso Geologico Chileno, Conception, Actas II, 1249-1253. VERNON, R. 2000. Review of microstructural evidence of magmatic and solid state flow. Electronic Geosciences, 5, DOI 10.1007/sl0069-000-1010-z. VlGNERESSE, J. L., TlKOFF, B. & AMEGLIO, L.

1999.

Modification of the regional stress field by magma intrusion and formation of tabular granitic plutons. Tectonophysics, 302, 203-224. WEINBERG, R. F., SIAL, A. N. & PESSOA, R. R. 2001. Magma flow within the Tavares pluton, northeastern Brazil: Compositional and thermal convection. Geological Society of America, Bulletin, 113, 508-520. WHITMEYER, S. J. & SIMPSON, C. 2003. High strain-rate deformation fabrics characterize a kilometersthick Paleozoic fault zone in the Eastern Sierras Pampeanas, central Argentina. Journal of Structural Geology, 25, 909-922.

Relationships between magnetic and structural fabrics revealed by Variscan basement rocks subjected to heterogeneous deformation—a case study from the Klodzko Metamorphic Complex, Central Sudetes, Poland M. KADZIALKO-HOFMOKL,1 S. MAZUR,2 T. WERNER 1 & J. KRUCZYK1 1

Institute of Geophysics, Polish Academy of Sciences, 01-452 Warsaw ul. Ks. Janusza 64, Poland, (e-mail:[email protected]) 2Institute of Geological Sciences, Wroclaw University, 50-204 Wroclaw, pi M. Borna 38, Poland Abstract: The Klodzko Metamorphic Complex comprises a number of thrust units, consisting of meta-igneous and metasedimentary rocks of the Variscan basement of the Sudetes, NE Bohemian Massif. The thrust sheet stack rests upon unmetamorphosed Nowa Ruda ophiolite and is unconformably overlain by Frasnian-Fammenian sediments. The studied rocks underwent six deformation events, scattered in time between the Middle Devonian and Late Carboniferous. The multiphase deformation produced a steep foliation that consistently trends WNW-ESE and a mineral lineation plunging to the ESE at a shallow to moderate angle. The results of the AMS study show that, despite the complex structural evolution, all the studied rocks bear a similar magnetic fabric mirroring the principal structural directions. This relationship between magnetic and structural fabrics is apparent in all tectonic units irrespective of the considerable variations in strain rate and metamorphic grade. This suggests a rather low dependence of magnetic anisotropy on the changing conditions and intensity of deformation.

The correlation between magnetic anisotropy and structural fabric in rocks belonging to the metamorphosed Variscan basement of the Sudetes, NE Bohemian Massif, has been a subject of considerable interest in recent years (Werner et al 2000; Werner 2002; Jelenska et al. 2002). The results of previous studies revealed a clear, if not perfect, concordance between magnetic and structural directions in rocks subjected to penetrative syn-metamorphic deformation. In all cases, however, the earlier investigations concentrated on selected domains within the Sudetes that suffered high but fairly uniform strain and metamorphism. Here we present data collected in an area characterized by highly heterogeneous deformation and variable metamorphic overprint. Our case study was carried out in the Klodzko Metamorphic Complex (KMC) which consists of six individual thrust units. All these units underwent a similar structural evolution, although equivalent deformation events in individual thrust sheets are characterized by different strain intensities and contrasting metamorphic conditions (Mazur 2003). Consequently, such a situation provides an opportunity to examine the potential influence of strain rate and metamorphic gradient on the relationship between the magnetic and structural fabrics produced under the same tectonic regime.

Outline of geology and sampling The Klodzko Metamorphic Complex (KMC) crops out in the Central Sudetes between the Gory Sowie Massif and the Orlica-Snieznik Dome. The KMC is comprised of metasedimentary, meta-igneous rocks predominantly basic in composition, and orthogneisses metamorphosed under greenschist-to-amphibolite facies conditions (Wojciechowska 1990; Kryza & Mazur 2001). These crystalline rocks form a NW-SE elongated outcrop approximately 100km2, bounded to the south and west by Upper Carboniferous and Lower Permian clastic sedimentary sequences of the Intra-Sudetic Basin. In the NE, the KMC is unconformably overlain by unmetamorphosed Upper Devonian conglomerates and limestones, which form the basal part of the Carboniferous flysch succession that fills the Bardo Basin (Haydukiewicz 1990). The SE boundary of the KMC is defined by an intrusive contact with the Klodzko-Zloty Stok granitoids of still uncertain Early Carboniferous origin (Fig. 1). The KMC comprises a number of separate tectonic units (Mazur & Kryza 1999) labelled 1-6 in Figure 2. These are respectively, from base to top, as follows: (1) the Mary Bozkow Unit, comprising the Givetian metasedimentary shelf sequence of 370-380 Ma (Hladil et al.

From: MARTIN-HERNANDEZ, F., LUNEBURG, C. M., AUBOURG, C. & JACKSON, M. (eds) 2004. Magnetic Fabric: Methods and Applications. Geological Society, London, Special Publications, 238, 475-491. 0305-8719/04/ $15.00 © The Geological Society of London 2004.

476

M. KADZIALKO-HOFMOKL ET AL.

Fig. 1. (a) Geological setting of the Klodzko Metamorphic Complex (denoted by rectangle) in the Bohemian Massif. EFZ: Elbe Fault Zone, ISF: Intra-Sudetic Fault, MGH: Mid-German Crystalline High, MO: Moldanubian Zone, MS: Moravo-Silesian Zone, NP: Northern Phyllite Zone, OFZ: Odra Fault Zone, OSD: Orlica-Snieznik Dome, RH: Rhenohercynian Zone, SX: Saxothuringian Zone, (b) Geological setting of the Klodzko Metamorphic Complex within the Gory Sowie - Klodzko region of Central Sudetes. (c) Geological sketch map of the Klodzko Metamorphic Complex.

1999); (2) the L^czna Unit, a melange body (Mazur & Kryza 1999); (3) the Bierkowice Unit, consisting of mafic metavolcanics; (4) the Scinawka Unit, metagabbros; (5) the OrlaGologfowy Unit, metagabbros and mafic volcanics and their sub volcanic equivalents, intruded by granitoids and accompanied by deep marine sediments; and (6) the Klodzko Fortress Unit,

metasediments and metabasalts, associated with volcanoclastic sandstones and dacitic-andesitic tuffs (Fig. 2). These tectonic elements are interpreted as thrust sheets within a nappe pile assembled at the turn of the Middle and Late Devonian due to the WNW-directed tectonic transport. A nappe structure of the KMC is characterized by

AMS OF VARISCAN METAMORPHIC ROCKS FROM SUDETES

477

Fig. 2. A generalized geological map of the Ktodzko Metamorphic Complex showing its tectonic subdivision and mean orientations of tectonic foliation and lineation. In the legend: lithotectonic log of the Klodzko Metamorphic Complex. The numbers denote studied units, letters: sampled lithotypes, as in Tables 1 and 2.

478

M. KADZIALKO-HOFMOKL ET AL.

Fig. 3. Synoptic stereoplots showing the orientation of different structural fabrics in the Klodzko Metamorphic Complex: foliation Si (a) and lineation LI (b) in the Orla-Gologlowy, Scinawka and Bierkowice units; foliation ^2 + 3 (c) and lineation L2 + 3 (d) in the Maly Bozkow, La_czna and Klodzko Fortress units, foliation S3 (e) and lineation L3 (f) in the Orla-Gologlowy Unit. Equal-area net, lower hemisphere. Girdle scatter of foliation S\ is brought about by folds F2 and foliation S2 + 3 by folds F6 (mostly in the Klodzko Fortress Unit). 0.5%, 2%, 4%, 8%: contours for counting method.

AMS OF VARISCAN METAMORPHIC ROCKS FROM SUDETES

the occurrence of structurally higher tectonic units from the west to east (Fig. 2) and by a tectonic inversion of the metamorphic P-T conditions (Kryza & Mazur 2001). The metamorphic grade increases from greenschist facies in the Maty Bozkow Unit, through epidote-amphibolite facies in the Lqczna and Bierkowice units, to upper amphibolite facies in the Scinawka and Orla-Gologlowy units. Only the structurally uppermost Klodzko Fortress Unit contradicts this trend, revealing epidote-amphibolite facies metamorphism. The differences in the geochemical characteristics of the meta-igneous rocks suggest derivation of the individual units from various plate tectonic settings (Kryza et al. 2003). This conclusion is in accord with the contrasting protolith ages associated with different parts of the KMC. The lowermost Maly Bozkow Unit contains a fossil-bearing crystalline limestone with coralline fauna of Middle Devonian age (Hladil et al. 1999). By contrast, U-Pb zircon dating from the two uppermost units indicates an Upper Proterozoic age in the range of 590-600 Ma (Mazur et al. 2003). The whole KMC rests upon Nowa Ruda Massif pillow lavas, labelled 0 in Figure 2 (Mazur 2003). The latter are considered part of a dismembered ophiolite (Dubinska & Gunia 1997; K^dzialkoHofmokl 2001) and are unconformably overlain, together with the KMC, by the Frasnian-Fammenian Upper Devonian limestones (360370 Ma), labelled 7 in Figure 2 (Bederke 1924). Compilation of structural data from different tectonic units of the KMC indicates superposition of six successive deformation events D\~D6. Effects of the DI stage are clearly evident in the Orla-Gologlowy, Scinawka and Bierkowice units with their mostly hard meta-igneous rock composition. The DI event was associated with the emplacement of thrust units comprised in the KMC and resulted in foliation Si and stretching lineation L l 9 developed as penetrative deformational structures. The Si planes, together with associated top-to-the WNW kinematic indicators, are reoriented on the limbs of younger F2 folds with axes parallel to lineation LI. The F2 folds cause Si to dip, typically to the SSW and NNE, whereas lineation LI gently plunges to the ESE at the angle of 20-40° (Figs. 2, 3). The D2 deformation related to NNE-SSW shortening occurred under greenschist facies conditions throughout the KMC. This event resulted in the folding of previously juxtaposed tectonic units into regional-scale F2 folds with axes 9riented WNW-ESE. In the Orla-Gologlowy, Scinawka and Bierkowice units, D2 produced little if any penetrative structures. A more significant effect associated with Z>9 occurred in the

479

metasediments of the Maly Bozkow, La_czna and Klodzko Fortress units, where £2 planes developing as the main foliation. An almost vertical foliation, S2 developed as an axial cleavage within F2 folds and largely obliterated the older foliation Si. The L2 lineation is inclined to the ESE at the angle of 20-30° (Fig. 3), forming parallel to the F2 axes due to the intersection of Si and S2 planes. The effects of the D3 deformation are mostly concentrated within the metagabbros and gneisses of the Orla-Gologlowy unit, which were subjected to intense mylonitization. The S3 mylonitic foliation is almost vertical and strikes consistently WNW-ESE, whereas stretching lineation L3 is gently inclined to the ENE (Fig. 3). Common asymmetric fabrics associated with the Z>3 shear provide evidence for mostly strike-slip dextral kinematics. Outwith the OrlaGologlowy Unit, the D3 event also resulted in significant alteration of structures within the L^czna and Klodzko Fortress units. D3 caused the reactivation of the S2 foliation under a dextral shear regime and related modification of the symmetrical D2 fabric. Deformation D4 is apparently related to the retrogression of metamorphic conditions and their final cessation. A D4 brittle-ductile shear zone at the contact of the Bierkowice unit against the L^czna and Maly Bozkow units (Fig. 2) is cross-cut by the erosional surface of the preUpper Devonian unconformity (Fig. 2). Furthermore, D4 seems to represent the last in a succession of events recorded in the KMC before its exhumation in the Late Devonian. The deformation D5 took place along the contact of the Orla-Gologlowy unit with the Klodzko-Zloty Stok Granitoid Massif and is closely associated in that area with the effects of Carboniferous contact metamorphism. Deformation D5 led to the reactivation of the Si foliation due to rotational sinistral shear in local shear zones. The final D6 event is only recorded in the Maly Bozkow and Klodzko Fortress units, adjacent to the Bardo Basin (Fig. Ic). Deformation D6 produced medium-to-open angle folds, F6, with sharp hinges and axes plunging to the W or NW. The axial planes are accentuated by either single joints or a regular fracture cleavage S$. A study of the KMC magnetic fabric A total of 149 hand samples were collected from the Klodzko Metamorphic Complex for the study of anisotropy of magnetic susceptibility (AMS). An earlier investigation performed on

M. KADZIALKO-HOFMOKL ET AL.

480

these samples by K^dzialko-Hofmokl et al. (2003) concluded that the natural magnetic remanence is carried by minerals of secondary origin comprising pyrrhotite, post-pyrite goethite, secondary magnetite, maghemite and hematite. Magnetomineralogical experiments included four methods: thermal demagnetization of IRM, three-axis Lowrie method, changes of bulk susceptibility at low and high temperatures and changes of induced magnetization during heating. Details of these studies are shown in Ka_dzialko-Hofmokl et al (2003). AMS parameters were measured for between two and eight specimens drilled from each hand sample with the KLY2/KLY3 susceptibility bridges of Agico (former Geofyzika-Brno). The results were analysed with the ANISO 11 program of Jelinek (1977). The following set of data was obtained for each specimen: mean susceptibility, Km; degree of anisotropy, Pf; shape parameter, T\ and directions of the axes of the anisotropy ellipsoid, KmSLX, K-mi, Km^n (Jelinek 1977; Tarling & Hrouda 1993). Ranges for the Km values are summarized in Table 1 together with the number of hand samples and specimens used for the AMS study. The parameters, Pf and T, together with the directions of Kmax and Kmin are presented for the respective mapped units (Figs. 4—8). Appropriate tectonic parameters, i.e. mean orientations of lineation and foliation, are also compiled in the same diagrams. Mean

directions of KmSLX and j^min axes and tectonic parameters are summarized in Table 2. Annealing experiments were performed on 29 specimens of crystalline limestones KLM and KBW, 9 rhyolites and 9 meta-rhyolites specimens that revealed a high scatter of anisotropy directions and inverse magnetic fabric. Specimens used for experiments were heated stepwise in a non-magnetic furnace to 650 °C for palaeomagnetic purposes, their AMS properties were measured after the final heating step and once again after 40 minutes of additional annealing at 650°C in the same furnace. The annealing technique may in some cases result in an enhanced pre-existing magnetic fabric due to the formation of new ferrimagnetic minerals from previously non-magnetic phases, aligned to the original preferred orientation (e.g. Jelenska & Ka_dzialko-Hofmokl 1987; Borradaile & Henry 1997; El-Hemaly & Ka_dzialkoHofmokl 2000; Borradaile & Lagroix 2000; Souque et al. 2002; Henry et al. 2003). The ^max and ^min directions after annealing are shown in Table 2. The anisotropy of anhysteretic remanent magnetization, AARM, was used to investigate several KLM limestones specimens. These anisotropy of remanence measurements were performed in fields of 100mT a.c. and 50 uT d.c., using an AGICO-LDA3 device. A description of the method can be found in Werner (2002).

Table 1. The statistics of collected samples and their susceptibilities Unit

Lithology and label

7

limestone KP1 limestone KLW

6

N/n

Km x 10"6SI

8/43 5/12

120-210 50-160

phyllite KP2, greenstone KP3 metarhyolite KW rhyolite KSC albite-chlorite schist KT

7/18 10/20 12/34 10/16

500-800 60-125 50-100 180^40

5

metagabbro KOG amphibolite KK meta-andesite KOA metagranite KSG

9/31 11/36 11/23 10/25

30-360 800-410000 210-300 150-400

4

metagabbro KSA amphibolite KSS

8/25 14/41

140-600 500-700

greenstone KB

10/32

3 2

crystalline limestone KLM 1,2 sedimentary metabreccia KLM 3

1

metasandstone KMB crystalline limestone KBW

0

pillow lavas NRC

A^xlO-SI

50-170 50-270

1000-70000

8/30 5/8

-6-60 300-600

250-1 000

10/37 10/46

150-270 -4-12

80-1 500

9/18

530-730

N/n: number of hand samples/number of specimens measured Km: mean susceptibility for fresh rocks, Kmim: mean susceptibility after annealing

AMS OF VARISCAN METAMORPHIC ROCKS FROM SUDETES

481

Table 2. The mean orientations of directions of mineral and magnetic fabric Unit Label

Rock type

Foliation/ bedding

Lineation

7

KP 1 KLWs.l KLW s.2 KLW s.3

limestone limestone limestone limestone

205/85 (S0) 96/55 (S0) 100/45 (S0) 108/65 (S0)

6

KW KSC KP2 KP3 KT

metarhyolite rhyolite phyllite greenstone (ab-chl) schist

5

KOG KK KOA KSG

4

-Kmax

^min

299/7 68/65 97/47 158/63

33/32 262/29 285/44 295/23

80/20 (S3?) 200/25 (S2+3) 208/58 (S2 +3) 200/60 (S2 + 3)

_ 130/10 (L2+3) 125/10 (L2 + 3) 120/15 (L2 + 3)

metagabbro amphibolite meta-andesite metagranite

200/75 (SO 20/80 (Si) 190/70 (SO 20/85 (SO

115/15 110/15 11 5/35 105/35

(LO (LO (LO (LO

scat. 112/15 117/28 104/56

scat. 203/7 10/30 204/8

KSA KSS

amphibolite metagabbro

25/80 (SO 45/50 (SO

1 10/30 (LO 120/20 (LO

112/40 95/23

204/2 scat.

3 2

KB

greenstone

340/84 (SO

70/35 (LO

scat.

scat.

KLMs. 1,2 cryst.limestone KLM s. 3 metabreccia

170/65 (S2 + 3) 180/80(S2 + 3)

90/15 (L2 + 3) 90/15 (L2 + 3)

scat., ann. 257/7 73/18

scat., ann.: 341/18 343/1

1

KMB KBW

metasandstone cryst.limestone

180/65 (S2) 230/75 (SO

100/20 (L2) 145/25 (L2)

97/18 358/24 scat., ann.: 135/27 52/18, ann.: 41/7

0

NRC

pillow lava

210/65 (S0)

130/25 (LO

121/17

The resultant directions of ARM anisotropy axes obtained during these additional experiments are shown along with the respective AMS results (Fig. 5c).

The basaltic pillow lavas of the Nowa Ruda Massif (NRC—unit 0) Susceptibility of the pillow lava is principally controlled by magnetite partly surface-oxidized to maghemite, small amounts of goethite and hematite, as well as the paramagnetics, chlorite and epidote. The observed anisotropy is generally low and typically characterized by oblate or prolate shaped ellipsoids in individual samples. The magnetic foliation correlates well with palaeo-horizon plane S0, delineated by elongation of the volcanic pillows. Magnetic lineation shows a moderate scatter within the foliation plane but is roughly concordant with the direction of mineral lineation L\ (Fig. 4a).

KMB met a-sands tones and KBW crystalline limestones of the Mafy Bozkow (unit 1) The susceptibility of KMB specimens is associated with ferrimagnetic minerals, mostly

251/67, ann .: 237/63 scat. scat., ann. 139/34; 220/16, ann .: 246/26 24/67 scat. 36/34 141/18 20/32 122/15

10/26

magnetite, pyrrhotite and goethite accompanied by the paramagnetic and diamagnetic minerals, muscovite and quartz. The ellipsoid of anisotropy is oblate whereas the anisotropy parameter ranges between 1.15 and 1.2. Kmax directions form a tight cluster consistent with an L2 mineral lineation. The orientation of magnetic foliation corresponds to that revealed by the foliation, S2 (Fig. 4b). The very low susceptibility of the KBW crystalline limestone is mainly a function of the diamagnetic quartz and an admixture of paramagnetic muscovite which is concentrated in distinct streaks. The presence of low amounts of goethite and hematite as well as magnetite, maghemite and pyrite was revealed by rockmagnetism analytical methods. The shape parameter T is rather variable depending on the sample, although in the majority of cases it points to an oblate-shaped magnetic ellipsoid. For samples with clearly paramagnetic AMS (Km > 3 x 10~6 SI vol.) magnetic foliations are moderately dispersed but remain approximately consistent with the attitude of the structural foliation, Si. The magnetic lineation shows a wide scatter within the plane of mean foliation (Fig. 4c). Several KBW specimens (some with diamagnetic AMS) were selected for the heating experiments (Fig. 4d). After step wise heating to

Fig. 4. AMS results: directions of KmaK (full squares) and Kmin (full circles) and P1 - T plots for lithotypes. (a) NRC, (b) KMB, (c) KBW (only for specimens with mean susceptibility >3 x 10~ SI), (d) results of annealing experiments on KBW selected specimens (low susceptible samples are included). Open circles and squares mark the mean Kmin and ATmax axes with confidence cones calculated by the method of principal component analysis, crosses: tectonic lineations L, S0: bedding, S: planes of tectonic foliations. Numbers against tectonic parameters denote number of respective deformational episode. Same notations in Figures 5-8.

AMS OF VARISCAN METAMORPHIC ROCKS FROM SUDETES

650 °C and a subsequent 40 minutes of annealing, Kmin axes form a tighter maximum, the position of which indicates that the magnetic foliation is slightly oblique to S\ . The attitude of magnetic fabric seems to reflect in that case a finite strain acquired due to the superposition of deformation D2 on the pre-existing foliation Si. KmSLX directions group into two distinct populations, one of which is parallel to the lineation, L2- This effect is probably due to the inversion of maximum and intermediate axes in strongly oblate AMS ellipsoids.

Metabreccias and KLM crystalline limestones of the Lqczna (unit 2) The magnetic susceptibility of specimens of the KLM metabreccia is controlled by a mixture of magnetite and the paramagnetic mineral, chlorite, accompanied by small amounts of diamagnetic quartz and calcite. The ellipsoid of anisotropy is oblate; directions of Km&K and ^min form tight groups indicating slightly oblique attitude of the magnetic fabric to mineral lineation £2+3 and foliation £2 + 3 (Fig- 5a). This suggests that the magnetic fabric is more ready to adopt strain increments induced by the D3 event in comparison to foliation and lineation partly inherited after the deformation D2. Consequently, the magnetic foliation probably corresponds to an XY section of the D3 strain ellipsoid inclined to foliation S2 reused as the Z>3 shear plane. The magnetic susceptibility of the KLM crystalline limestone is principally controlled by the diamagnetics, quartz and calcite, and the paramagnetics, muscovite and chlorite. Small amounts of ferrimagnetics, comprising goethite, hematite, magnetite, maghemite and pyrrhotite are also present. Obtained values of the anisotropy parameter, Pf, are variable and associated with both oblate and prolate anisotropy ellipsoids. -Kmin directions are widely scattered within the plane parallel to tectonic foliation whereas Kmax clusters in a maximum around the pole to this plane (Fig. 5b). Such a situation clearly suggests the inversion of maximum and minimum anisotropy axes. In order to test this hypothesis, AARM was deployed to investigate the remanence anisotropy for a number of specimens. As suspected, the directions of AARM axes show clear inversion of max and min axes of ellipsoids acquired for ferromagnetic grains comparing to the AMS data (Fig. 5c). A similar change of the anisotropy axes was also achieved during the annealing experiment for several specimens (Fig. 5d).

483

KB greenstones of the Bierkowice (unit 3) The greenstones have a generally high, if somewhat variable magnetic susceptibility dominated by magnetite of various generations, sometimes surface-oxidized to maghemite, with lesser amounts of goethite and hematite. Anisotropy ellipsoids are either oblate or prolate for lower values of anisotropy parameter, whereas solely oblate for higher anisotropy (Fig. 5e). Kmax and ^min are widely scattered, the directions of which are only loosely coincident with the orientations of the foliation Si and lineation LI (Fig. 5e). This may be due to the presence of several generations of magnetite grains of various shapes and sizes. There are significant differences in the direction of respective axes of anisotropy as encountered amongst specimens cut from the same hand sample.

KSA amphibolites and KSS metagabbros of the Scinawka (unit 4) The KSA laminated amphibolites display low susceptibility associated with hematite and sparse magnetite being identified by reflected light microscopy. Paramagnetic hornblende and ilmenite are also present. P1 remains in a narrow range of high values associated with a distinctly oblate anisotropy ellipsoid. The relatively tight maxima of A"max and ^min perfectly mirror the attitude of tectonic foliation and lineation in the KSA amphibolites (Fig. 6a). The KSS metagabbros are relatively rich in hematite, goethite, pyrrhotite and ilmenite. These minerals result in a generally high value for the anisotropy parameter Pf, though significant variations do occur, whereas the anisotropy ellipsoids are usually oblate. Kmax and ATmin directions are broadly scattered, while about 50% of Kmax measurements group close to the pole of foliation Si (Fig. 6b).

KOG metagabbros, KK amphibolites, KOA meta-andesites and KSG metagranites of the Orla-Gologlowy (unit 5) The KOG metagabbros reveal a low magnetic susceptibility associated with small amounts of fine grained Fe-oxides. Pf values are low and do not appear to be correlated with any particular shape of anisotropy ellipsoid. The directions of Km3LX and Kmin are widely dispersed, though partly coherent with respect to the orientations of the foliation Si and lineation L{ (Fig. 6c).

Fig. 5. AMS results: (a) KLM metabreccia, (b) KLM crystalline limestones (only for specimens with susceptibility >3 x 10 6 SI), (c) directions of AMS (full symbols) and AARM (open symbols) for four specimens of KLM limestones (d) results of annealing experiments for selected specimens of KLM limestones (e) KB.

Fig. 6. AMS results, (a) KSA, (b) KSS, (c) KOG, (d) KK, (e) KOA, (f) KSG.

486

M. KADZIALKO-HOFMOKL ET AL.

The KK amphibolites display variable susceptibility, with high values clearly associated with the occurrence of magnetite. The shape parameter T remains low, despite highly differentiated P1 values corresponding to oblate or prolate anisotropy ellipsoids. Directions of KmaK and ^min are weH constrained and closely coincide with the orientations of foliation S\ and lineation LI respectively (Fig. 6d). The susceptibility of the KOA meta-andesites is mostly influenced by iron hydroxides, paramagnetic ilmenite and diamagnetic quartz and calcite. Low to intermediate values of P' are coupled with distinctly oblate-shaped anisotropy ellipsoids. Kmin directions moderately cluster around a pole to foliation S\ (Fig. 6e), ^max directions are slightly dispersed due to highly oblate fabric. Nevertheless, a noticeable Kmax cluster corresponds to the orientation of mineral lineation L{ (Fig. 6e). The KSG metagranites contain ferrimagnetic minerals as represented by very small amounts of goethite and unindentified Fe-oxides, possibly magnetite and maghemite, whereas paramagnetic phases comprise chlorite and biotite. The susceptibility of these rocks is low to intermediate with distinctly differentiated values of P1'. With the exception of one specimen, all anisotropy ellipsoids are oblate and characterized by low to high shape parameter values. Km^n directions are confined to a tight maximum around a pole to the foliation, Si. ^max are mostly consistent with the orientation of lineation LI although some measurements are distinctly steeper (Fig. 6f).

contribution from paramagnetic minerals like muscovite and chlorite in the KW, and biotite in the KSC. Ferrimagnetics play a minor part, and are represented in the samples by fine grains of goethite and hematite in the KSC and hematite accompanied with small magnetite inclusions within silicates in the KW. Pf parameters for these felsic volcanics are low and associated with predominantly oblate-shaped anisotropy ellipsoids. In the KW, ^max directions are broadly scattered along the plane parallel to foliation, 53, whereas ^min axes cluster around a pole to foliation. The broad dispersion of Kmax axes is in agreement with the poor development of tectonic lineation (Fig. 7c). In the KSC, ^min directions are relatively concentrated despite the fact that the rock is practically devoid of deformational structures. Magnetic lineations are widely dispersed along the girdle confirming a pervasive magnetic foliation (Fig. 7e). Samples of the KSC rhyolite, when heated to 650 °C, show Kmax directions that have shifted to form one maximum (Fig. 7f). Annealing of the KW meta-rhyolite for 40 minutes at 650 °C resulted in a decreasing scatter of Kmax directions (Fig. 7d). The ferromagnetic fraction of KT albitechlorite schists is dominated by pyrrhotite locally associated with small amounts of hematite. The intermediate values of susceptibility for KT schists are usually associated with a moderate anisotropy and sparse, high values of P7. Anisotropy ellipsoids are oblate with variable degrees of eccentricity. The orientations of magnetic foliation and lineation mirror those displayed by foliation S2+ 3 and lineation L2+ 3 (Fig. 7g).

KP3 greenstones, KP2 phyllites, KW metarhyolite, KSC rhyolite and KT albitechlorite schists of the Klodzko Fortress (unit 6)

KP1 and KL W Upper Devonian limestones (unit 7)

The KP3 greenstones and KP2 phyllites contain pyrrhotite as the principal magnetic mineral. The greenstones yield intermediate values of Pf that correlate with distinctly oblate anisotropy ellipsoids. The magnetic foliation and lineation orientations closely resemble those demonstrated by foliation S2+ 3 andlineation L 2+ 3 (Fig. 7a). The phyllites show intermediate-to-low anisotropy and oblate shaped anisotropy ellipsoids. Kmin and J^max directions are rather scattered although the majority display attitudes approximately corresponding to the orientations of foliation S2+ 3 and lineation L2+i (Fig. 7b). The low susceptibility values obtained from KW and KSC felsic volcanics imply a major

Magnetic susceptibility in KP1 limestones is associated with pyrrhotite, accompanied by small amounts of magnetite. The paramagnetic fraction is represented by an admixture of detrital muscovite. Values of Pf vary from low to intermediate, whereas anisotropy ellipsoids are oblate in the majority of specimens. Directions of ATmax and Kmin are moderately concentrated, the latter forming a maximum around the pole to sedimentary bedding (Fig. 8a). The KLW limestone, as sampled in three sites, contains goethite and magnetite. These principal ferrimagnetic minerals are accompanied by two paramagnetic species: detrital muscovite and chlorite. The P' parameters for these samples is low with a neutral shape of anisotropy ellipsoid. For sites 1 and 2, ^max and ^min directions are

AMS OF VARISCAN METAMORPHIC ROCKS FROM SUDETES

487

Fig. 7. AMS results, (a) KP3, (b) KP2, (c) KW, (d) annealing experiments for selected specimens of KW, (e) KSC, (f) annealing experiments for selected specimens of KSC, (g) KT.

weakly grouped; Kmin is observed to be poorly coincident with the orientation of sedimentary bedding, 50 (Figs. 8b, c). At site 3, magnetic foliation and lineation are more coherently oriented. Kmin directions cluster around the pole to bedding whereas the maximum for Km.dK is not related to the attitude of lineation in the adjacent metamorphic rocks (Fig. 8d). Discussion Our data show that three first deformation events penetrative at the regional scale in the KMC were effectively mimicking the same structural plan

(Fig. 3). Consequently, foliations and lineations produced during these events are practically indistinguishable in terms of geometrical criteria. Their discrimination is mainly possible based on the composition of syn-kinematic mineral assemblages, which grew in a wide range of metamorphic conditions from upper amphibolite to greenschist facies (Mazur 2003). Nevertheless, AMS obviously preserves little 'memory' of early metamorphic events. Instead of overprinted patterns, the AMS fabric reproduces the finite strain, even if its latest increments (D2-D3) were acquired at much lower temperatures than the initial deformation (Di). On the other hand, the last three deformations (D4-D6) had very

Fig. 8. AMS results, (a) KP1, (b) KLW site 1, (c) KLW site 2, (d) KLW site 3.

AMS OF VARISCAN METAMORPHIC ROCKS FROM SUDETES

limited extent and, thus, their effects were unable to largely interfere with the AMS fabric developed during the first D\-D^ events. Besides the general resemblance of the finite strain pattern and the AMS fabric, our results demonstrate several more subtle relationships, which to a certain extent influence the development of the AMS pattern. For instance, the best defined AMS directions, corresponding closely to the orientation of structural foliation and lineation associated with the DI event, were found in rocks that had experienced relatively fast cooling without a retrogressive metamorphic overprint. These are best represented in the KSA and KK amphibolites, which reveal AMS axes perfectly consistent with the attitude of deformational structures (Figs. 6a, 6d). Furthermore, the KSA samples demonstrate the high uniformity of anisotropy ellipsoids, which appear to be comparable with the finite strain ellipsoid in these rocks (Mazur 2003). By contrast, KK specimens show a wide spectrum of shapes and anisotropy parameters. The difference is probably due to the varying concentration of magnetite grains in KK, as evident in the values of Km (Table 1), and attributable to the dissimilar geological histories of the KSA and KK amphibolites. The latter were subjected to a temperature-dominated static metamorphism in the contact aureole of the Klodzko-Ztoty Stok pluton. Interestingly, other rock samples devoid of significant retrogressive overprinting revealed much worse groupings of AMS directions relative to the KK and KSA amphibolites. This observation primarily concerns KOG and KSS metagabbros characterized by a coarsegrained texture. It seems apparent in this case that grain size is a significant factor, preventing consistent AMS directions in coarse-grained rocks despite a similar deformation and metamorphic history to the finer grained KSA amphibolites. Our results show that the distribution of AMS directions is largely independent of the assemblage of magnetic minerals responsible for susceptibility and its anisotropy in the investigated samples. This also accounts for similar AMS patterns in different rock types comprised in the KMC. The least scatter of AMS axes is observed in rocks where ferrimagnetics dominate, e.g. KK, and also in rocks where both ferrimagnetic and paramagnetic minerals equally contribute to anisotropy, e.g. KOA, KMB meta-sandstones, KLM meta-breccia, KSA, KSG and KT. A high scatter of anisotropy axes was observed in the metagabbro, KSS, where AMS readings depend on both ferrimagnetics and paramagnetics. Scattered axes were also observed in the

489

greenstone, KB, where AMS is carried solely by a large amount of ferromagnetic minerals. In the latter case the presented data confirms that late retrogressive metamorphic overprints have a devastating influence on AMS directions. The KB greenstones were subjected to brittle-ductile faulting and cataclasis, attributable to deformation D4. A previously published microscopic investigation by K^dzialko-Hofmokl et al. (2003) has shown that deformation and retrograde metamorphism brought about several mineral transformations, which allowed growth of new ferrimagnetic phases. Such grains, typically magnetite, concentrate in the randomly oriented micro-cracks and veins that occur with brittle-ductile deformation. Consequently, Kmin and Kmax axes define the weakly developed magnetic lineation and foliation. A similarly scattered distribution of AMS directions is not detectable in the KLM metabreccia despite the fact that this rock still preserves features inherited from the protolith, a melange body. Subsequent phases of regional metamorphism and deformation, namely D2 and Z>3, overprint with a uniform AMS pattern. Therefore, KLM samples show consistently oriented axes of AMS ellipsoids closely corresponding to the attitude of the structural grain. Our results show that even a relatively lowgrade metamorphic overprint is entirely sufficient to produce a consistent AMS pattern. Accordingly, the KT albite-chlorite schists, KMB meta-sandstones and KP3 greenstones demonstrate well-defined AMS directions intimately corresponding to the orientation of foliation and lineation. Also the NRC pillow lavas, despite their very low metamorphic grade, display AMS directions, which can be easily compared to the orientation of deformational structures. An interesting suite of AMS information is provided by rock samples, which have apparently avoided the influence of regional metamorphism. For instance, the KP1 Upper Devonian limestones crop out on an inverted limb of a fold developed during the last deformation, £>6. The orientation of Km^n is closely associated with the attitude of bedding planes parallel to the foliation in adjacent metamorphic rocks. Furthermore, Kmax directions correspond to a regional trend of lineation, though linear structures are virtually absent in the KP1 limestones. Consequently, the AMS fabric of these rocks must have been sensitive enough to record a deformation event that did not produce any structural imprint. The role of contact metamorphism in remagnetization of the limestones remains a moot point. The boundary of the neighbouring Klodzko-Zloty Stok pluton was

490

M. KADZIALKO-HOFMOKL ET AL.

sufficiently close to potentially produce elevated temperatures. However, a cursory inspection of thin sections does not reveal any clear sign of recrystallization typically associated with a contact metamorphic aureole. The same horizon of Upper Devonian limestones, sampled much farther from the Kiodzko-Zloty Stok pluton in the KLW locality, shows only moderate steepening of sedimentary bedding due to tectonic deformation. For the specimens collected from KLW, only a limited evidence of the sedimentary bedding orientation seems to have been preserved and, therefore, may have been partly erased by later hydrothermal or diagenetic processes. However, orientations for some ^min axes appear to be approximately related to the attitude of sedimentary planes, whereas the majority of Kmax directions plunge moderately to the east or SE. A surprisingly consistent AMS pattern is demonstrated by the KSC rhyolites, which are remarkable for being virtually devoid of any recognizable macroscopic structures besides random jointing. The geological affiliation of these rocks is still unclear since they belong to the Klodzko Fortress Unit of the KMC, or alternatively, represent one of the rhyolite subvolcanic intrusions into sediments of the adjacent Bardo Basin. On the evidence of a magnetic foliation corresponding to the NNW-SSE elongation of the rhyolite body rather than the attitude of foliation in neighbouring rocks, the origin of the AMS fabric can be attributed tentatively to the process of magmatic emplacement. Well-developed magnetic foliation is also documented in the KW meta-rhyolite. In this case, however, the association is presumably deformation-related since the KW metarhyolites bear a macroscopic tectonic foliation of similar orientation. Annealing experiments have proven, in some cases, the effectiveness of this method in enhancing AMS fabrics. The pattern of AMS directions appears to be considerably clarified after annealing in both the KLM and KBW crystalline limestones, characterized by very low ferrimagnetic minerals concentrations. In the KLM samples with inverse AMS fabric annealing and AARM experiments revealed the real orientation of magnetic foliation and lineation. Annealing experiments on felsic volcanics such as the KSC rhyolites and KW meta-rhyolites were not so successful. Here the method produces only slight improvements in the grouping of AMS directions and correspondence with structural data. For the remaining units annealing experiments produced a higher scatter in the AMS data and these results are not shown in the paper in Table 2.

Conclusions The regional correlation of polyphase fold and fabric sequences in metamorphic terrains is generally hindered due to continuous reworking, overprinting and transposition of local fabrics during subsequent deformations. Therefore, in such terrains analysis of the finite fabric, which is the result of the bulk deformation, becomes the most useful method of recognizing significant patterns in regional deformation (Piazolo et al. 2004). Our results show that AMS study provides a useful tool for the generalization of a structural pattern. In our case study a relative uniformity of the AMS fabric is acquired in consequence of two independent factors: (1) the reproducibility of the main structural trends during the subsequent deformation events, and (2) the ability of AMS, modified by superimposed deformations, to reflect the bulk finite strain. In the latter case, the similarity of the magnetic and structural fabrics in deformed rocks revealed by a number of earlier studies (e.g. Pares et al. 1999, 2003) is validated for the fabrics produced by several overprinted events and, thus, reflecting the bulk finite strain. Since AMS strain-related fabrics are highly sensitive to changes in temperature and related deformation conditions, in some cases such fabrics are present in rocks apparently devoid of a tectono-metamorphic imprint, e.g. the limestones, KP1 and KLW. A similar conclusion can be drawn from annealing experiments on the rhyolites, KSC and KW. Annealing experiments as well as AARM analysis revealed the true magnetic orientation in samples where inverse magnetic fabrics were observed, e.g. the KLM limestone. Ferrimagnetic phases, although major contributors to the total susceptibility of studied rocks, were not always in accord with the trend of tectonic structures. This was clearly evident where ferrimagnetic minerals had grown during late brittle-ductile phases of deformation and retrograde metamorphism e.g. KB greenstones. We acknowledge the support of the Polish State Committee for Scientific Research, grant No 6P04D 06617. We are deeply indebted to A. Cavanagh for his efforts to improve the linguistic aspect of the paper.

References BEDERKE, E. 1924. Das Devon in Schlesien und das Alter Sudetenfaltung. Fortschritte der Geologic und Paldeontologie. 7, 1-50.

AMS OF VARISCAN METAMORPHIC ROCKS FROM SUDETES BORRADAILE, G. J. & HENRY B. 1997. Tectonic applications of magnetic susceptibility and its anisotropy. Earth Science Reviews, 42, 49-93. BORRADAILE, G. J. & LAGROIX, F. 2000. Thermal enhancement of magnetic fabrics in high grade gneisses. Geophysical Research Letters, 27(16), 2413-2416. DUBINSKA, E. & GUNIA, P. 1997. The Sudetic ophiolite; current view on its geodynamic model. Geological Quarterly, 1, 1-28. EL-HEMALY, I. & KADZIALKO-HOFMOKL, M. 2000. Paleomagnetic investigations of the Carboniferous sediments from the Intra-Sudetic Basin, southern Poland. Part II: Namurian-Westphalian deposits, Acta Geophysica Polonica, 48(1), 93-122. HAYDUKIEWICZ, J. 1990. Stratigraphy of Paleozoic rocks of the Gory Bardzkie and some remarks on their sedimentation. Neues Jahrbuch fur Geologie und Palaontologie, 179, 275-284. HENRY, B., JORDANOVA, D., JORDANOVA, N., SOUQUE., C. & ROBION, P. 2003. Anisotropy of magnetic susceptibility of heated rocks. Tectonophysics, 366, 241-258. HLADIL, J., MAZUR, S., GALLE, A. & EBERT, J. 1999. Revised age of the Maly Bozkow limestone in the Klodzko metamorphic unit (Early Givetian, late Middle Devonian): implications for the geology of the Sudetes. Neues Jahrbuch fur Geologic und Palaontologie, 211(3), 329-353. JELENSKA, M. & KADZIALKO-HOFMOKL, M. 1987. Dependence of anisotropy of magnetic susceptibility of rocks on temperature. Physics of the Earth and Planetary Interiors, 62, 19-31. JELENSKA, M., WERNER, T. & MAZUR, S. 2002. Paleomagnetic and rock magnetic properties of the Lower Paleozoic metamorphic complex of the Rudawy Janowickie (West Sudetes, Poland). Geologica Carpathica, 53(5), 283-294 JELINEK, V. 1977. The statistical theory of measuring anisotropy of magnetic susceptibility of rocks and its application, Brno, Geofyzika, 1-88. KADZIALKO-HOFMOKL, M. 2001. Rock-magnetic study of the Gogolow-Jordanow serpentinite unit of the Paleozoic ophiolite (South Poland). Ofioliti, 26(2b), 425-432. KADZIALKO-HOFMOKL, M., KRUCZYK, J., MAZUR, S. & SIEMIATKOWSKI, J. 2003. Paleomagnetism of the Upper Proterozoic and Devonian rocks from the Klodzko Metamorphic Complex in the West Sudetes (SW Poland): tectonic implications for the Variscan belt of Central Europe. Tectonophysics, 377, 83-99. KRYZA, R. & MAZUR, S. 2001. Contrasting metamorphic paths in the Klodzko Metamorphic Unit, Central Sudetes. Polskie Towarzystwo Mineralogiczne Prace Specjalne. Polish Mineralogical Society, Special Papers. 19, 97-99. KRYZA, R., MAZUR, S. & PIN, C. 2003. Subductionand non-subduction-related igneous rocks in the

491

Central-European Variscides: geochemical and Nd isotope evidence for a composite origin of the Klodzko Metamorphic Complex, Polish Sudetes. Geodinamica Acta, 16(1), 39-57. MAZUR, S. 2003. Structural evolution of the Klodzko Metamorphic Complex and its significance for Variscan tectonics of the Sudetes. Acta Universitatis Wratislaviensis no 2581, Prace GeologicznoMineralogiczne, 44, 1-197. MAZUR, S. & KRYZA, R. 1999. Preliminary report on the metamorphosed melange in the Klodzko metamorphic complex (West Sudetes, SW Poland). Mineralogical Society of Poland—Special Papers, 14, 102-104. MAZUR, S., KRYZA, R., TURNIAK, K., BROCKER, M. & PIN, C. 2003 Pre-Variscan metaigneous rocks of the Ktodzko Metamorphic Complex—a vestige of cadomian subduction in the Central Sudetes, SW Poland. Journal of the Czech Geological Society, 4%(\-2\ 90-91. PARES, J. M., VAN DER PLUIJM, B. A. & DINARES, T. J. 1999. Evolution of magnetic fabrics during incipient deformation of mudrocks (Pyrenees, northern Spain). Tectonophysics, 307(1-2), 1-14. PARES, J. M. & VAN DER PLUIJM, B. A. 2003. Magnetic fabrics and strain in pencil structures of the Knobs Formation, Valley and Ridge Province, US Appalachians. Journal of Structural Geology, 25(9), 1349-1358. PlAZOLO, S., ALSO?, G. I., VAN GOOL, J. & M0LLER

NIELSON, B. 2004. Using GIS to unravel high strain patterns in high grade terranes: a case study of indentor tectonics from West Greenland. In: ALSOP, G. I., HOLDSWORTH, R. E., MCCAFFREY, K. J. W. & HAND, M. (eds) Flow Processes in Faults and Shear Zones. Geological Society, London, Special Publications, 224, 63-78. TARLING, D. H. & HROUDA, F. 1993. The Magnetic Anisotropy of Rocks, Chapman & Hall, 1-217. SOUQUE, C., ROBION, P. & FRIZON DE LAMOTTE, D. 2002. Cryptic fabric of tectonic origin revealed by heating of sedimentary samples from the Corbieres, France. Physics and Chemistry of the Earth, 27, 1253-1262. WERNER, T. 2002. The correlation of magnetic anisotropies (AMS and AARM) with tectonic fabrics of the Niemcza shear zone (SW Poland). Acta Geophysica Polonica, 50(1), 79-107. WERNER, T., MAZUR, S. & JELENSKA, M. 2000. Changing Direction of magnetic Fabric in a Thrust Unit: an Example from the Karkonosze - Izera Massif (SW Poland). Physics and Chemistry of the Earth (A), 25(5), 511-517. WOJCIECHOWSKA, I. 1990. Geology of the Klodzko metamorphic unit (Sudetes, Poland). Neues Jahrbuch fur Geologic und Palaontologie. 179, 189-195.

This page intentionally left blank

Oblique magnetic fabric in siderite-bearing pelitic rocks of the Upper Carboniferous Culm Basin, SW England: an indicator for palaeo-fluid migration? HELGA DE WALL1 & LAURENCE N. WARR2 l

lnstitut fur Geologic der Julius-Maximilians-Universitat Wilrzburg, Pleicherwall 1, D-97070 Wurzburg, Germany 2 Centre de Geochimie de la Surface (CNRS-ULP), 1 rue Blessig, 67084-Strasbourg, France Abstract: A zone of siderite dominated magnetic fabrics is recognized within clastic argillaceous rocks of the southern part of the Upper Carboniferous Culm foreland basin of SW England. This zone was identified by measuring the anisotropy of magnetic susceptibility (AMS) before and after heat treatment of samples. A detailed investigation of a recumbent fold structure within this zone (at the well-known Crackington Haven locality) reveals the pre-folding nature of siderite formation. The restored ttmax axes of AMS-ellipsoids plot on a segment of a small circle, with a mean inclination of c. 45° to the pole of the sedimentary bedding planes. This oblique magnetic fabric geometry is considered to reflect substrate-controlled siderite growth within a migrating fluid medium, which crystallized during diagenesis and the early stages of Variscan compression. The regional distribution of siderite growth, in combination with the directional information from the AMS, is discussed as an indicator for the palaeoflow direction of diagenetic fluids within a foreland basin setting.

A large number of investigations have applied the anisotropy of magnetic susceptibility (AMS) approach to resolve strain histories, ranging from the characterization of sedimentary fabrics formed during deposition and compaction to the study of tectonic deformation associated with basin inversion, folding and thrusting in orogenic belts (e.g. Hrouda & Potfaj 1993; Sagnotti et al 1998; Hrouda & Jezek 1999; Aubourg et al 2000; Carrapa et al. 2003). The nature and evolution of rock fabrics developed in such settings is now well established. Increasingly oblate fabrics are developed in sedimentary sequences due to progressive compaction (Hirt et al. 1995). However, during rock deformation a variety of magnetic fabrics may develop, ranging from strongly uniaxial (oblate, prolate) to girdle geometries (Pares et al. 1999). Particularly complex fabrics result during folding and faulting in multilayered rock sequences characterized by strain partitioning, where strain histories are often influenced by varying contributions of flexural slip and flexural flow. Under conditions of increasing strain there is a tendency for fabrics to become simpler. Stress-induced solution transfer and oriented grain growth can result in a near perfect oblate cleavage fabric (Liineburg et al. 1999). Whereas phyllosilicates are the main constituents of pelitic rocks (>40%), these lithologies often contain other paramagnetic minerals, such as Fe-carbonates, -oxides, -hydroxides and -sulphides as well as ferrimagnetic components,

particularly magnetite. A further aspect is the various origins of component minerals, which may be detritally derived or newly formed as authigenic or metamorphic phases at various stages during the rock history. Mineral growth and recrystallization can occur during shallow and deep diagenetic cementation (Schneider et al. 2004), during the circulation of hydrothermal fluids or during prograde and retrograde metamorphic reactions (Rochette 1987). Even rock weathering at the Earth's surface is capable of precipitating and modifying both paramagnetic and ferrimagnetic minerals (Kontny et al 1997). In pelitic sediments the AMS is typically controlled by paramagnetic phyllosilicate minerals, especially chlorite, which may occur as detrital or authigenic mineral phases (Liineburg et al. 1999; Hrouda & Potfaj 1993; Pares et al 1999). In such settings, the magnetic fabric reflects the crystallographic preferred orientation of phyllosilicates with magnetic foliations lying parallel to planar sedimentary (bedding) or tectonic (cleavage) fabric elements. As a result magnetic lineations commonly reflect the intersection direction between these fabric elements and are oriented parallel to macro- or mesoscale fold axes. In this study we investigate a geological setting where the occurrence of diagenetic siderite influences the magnetic fabric. The predominance of siderite is recognized by a deviation from the typical phyllosilicate-related fabric geometry

From: MARTIN-HERNANDEZ, F., LUNEBURG, C. M., AUBOURG, C. & JACKSON, M. (eds) 2004. Magnetic Fabric: Methods and Applications. Geological Society, London, Special Publications, 238, 493-507. 0305-8719/04/ $15.00 © The Geological Society of London 2004.

494

H. DE WALL & L. N. WARR

(normal fabrics of Rochette 1988) where ftmax parallels the structural lineation and ftmin lies perpendicular to the structural foliation. In these siderite-bearing samples the magnetic lineations lie not within but oblique to the planar rock fabrics (bedding planes). Abnormal magnetic fabrics related to the predominance of carbonates have been noted in earlier studies in both sedimentary and metamorphic rocks (Rochette 1988; Ihmle et al. 1989; Hirt & Gehring 1991; de Wall et al. 2000; Hounslow 2001) whereby the maximum susceptibility corresponds with the orientation of the carbonate c axis and is generally oriented perpendicular to the foliation or bedding plane. This type of fabric has been named inverse magnetic fabric (Rochette 1988). In the study presented here, the maximum susceptibility is shown to reflect the orientation of carbonate c axes characterized by a more complex pattern of inverse fabrics that lie not perpendicular but oblique to bedding planes. It will also be

demonstrated that heat treatment of pelitic samples is an important procedure for the effective interpretation of these fabric elements, as recently shown by Henry et al. (2003). Our investigation was conducted in the Upper Carboniferous Culm Basin sequence of SW England, a classical rock section for studying the varying degrees of deformation and metamorphism. The well-exposed coastal section has been a site of intense sedimentological, structural and mineralogical investigations (e.g. Sanderson 1979; Thomas 1988; Warr et al. 1991; Hecht 1992). One of the first systematic studies of AMS in pelitic rock sequences was also conducted in this area (Singh et al. 1975) . Geological setting The Upper Carboniferous rocks and underlying sequences of the Culm Basin of SW England (Fig. 1) form part of the Rhenohercynian Zone

Fig. 1. Simplified geological map showing the Upper Carboniferous Culm Basin in SW England. The sample sites (1 to 18) along the coastal section of Devon and North Cornwall are indicated by black dots. Areas A, B, C refer to the subdivision of the Culm section based on magnetic anisotropy data (see Fig. 2). THSZ: Tintagel High Strain Zone. Numbering refers to sample locations: 1 Westward Ho!; 2 Greencliff; 3 Bucks Mill; 4 Hartland Point, 5 Hartland Quay, 6 Welcombe Mouth, 7 Duckpool, 8 Maer Cliff, 9 Widemouth Bay, 10 Millock Haven, 11 Crackington Haven, 12 Samphire rock, 13 The Strangles, 14 Voter Run, 15 Rusey Beach, 16 Rusey Fault, 17 Rusey Fault south, 18 Pentargon.

AMS IN SIDERITE-BEARING PELITIC ROCKS

of the Variscan orogenic belt of Europe. The Culm Basin is considered to have formed as a foreland basin (Dewey 1982; Hartley & Warr 1990; Warr 1993), characterized by a c.3km thick basin fill, ranging from distal turbidites (Crackington Formation) to more proximal turbidites and storm deposits (Bude Formation), as well as deltaics of the Bideford Formation (Melvin 1986; Higgs 1984). The clastic sedimentary sequence is dominated by an alternation of sandstones, siltstones and mudstones, the latter forming a dominant component of the background sedimentation. An intimate relationship has been recognized between deposition, burial and early rock deformation, with the occurrence of slump horizons, which are considered to have developed into early thrust planes following deposition of the rock sequences (Whalley & Lloyd 1986). An important feature of the Culm Basin is the well-known transition from upright and open, chevron-style folding in the central part of the basin to more strongly folded and overturned, south-vergent fold structures in the south (Sanderson 1979). This structural asymmetry is attributed to southward-directed backthrusting induced by the north-directed underthrusting of a Variscan metamorphic wedge below the southern margin of the Culm sedimentary foreland basin sequence (Warr 1993). The changes in structural style documented across the basin are also accompanied by variations in lowtemperature mineral reactions, which range from diagenetic (^zeolite facies), through anchizone (~pumpellyite-prehnite), to epizone (~greenschist facies) grades of metamorphism (Warr et al. 1991). A notable prograde increase in metamorphic grade accompanies the progressively higher strains, passing from the centre of the Culm Basin southward towards its southern margin. This gradient continues to increase into the Devonian and Carboniferous rocks lying to the south of the Upper Carboniferous basin, where greenschist facies grades of metamorphism characterize the 'Tintagel High Strain Zone' (Primmer 1985). A distinct jump in metamorphic grade occurs across the Rusey Fault zone, which marks the southern boundary of the Culm Basin rocks. Here, the middle anchizone grade of metamorphism is missing, indicating significant displacement along this structure (Warr et al. 1991). The mineralogy of the pelites, which consists of shales, compacted and cleaved mudstones, slates and phyllites, has been documented by a number of X-ray diffraction and electron microscopy studies (Warr et al. 1991; Warr & Nieto 1998).

495

Culm pelites typically consist of assemblages containing quartz, feldspar, detrital muscovite with authigenic growth of either illite-kaolinite or illite-chlorite assemblages. Minor amounts of diagenetic pyrite and carbonate cements also occur as concretionary nodules or vein fillings. Sampling and analytical methods A total of 35 oriented mudstones and slate samples were selected from 18 well-documented localities along the Devon and north Cornish coastal section, from Westward Ho! in the north to Pentargon south of the Rusey Fault zone (Fig. 1). For AMS analysis, rock cylinders (up to 5 specimens per sample) were prepared in the laboratory. The directional magnetic susceptibility was determined for 15 positions with a Kappabridge (KLY-2, AGICO). The AMS tensor and the resulting magnitudes and orientation of the principal axes of the AMS-ellipsoids were calculated using the software package Ani20. From the magnitudes of the AMS-ellipsoids, the mean susceptibility value (volume susceptibility in SI units), and the shape (T) and anisotropy factors (Pf)9 are calculated as described by Jelinek (1981) and Tarling & Hrouda (1993). For heat treatment, a selection of AMS cylinders was heated for 24 hours at 500 °C in an air oven and remeasured. The mineralogy of samples was determined by X-ray diffraction (XRD) analyses (using a Siemens D500 diffractometer) on both randomly oriented whole rock powders and textured <2um clay mineral separates. Qualitative analysis of relative mineral abundance was undertaken using homogeneous, random powder mounts following the methodology outlined by Moore & Reynolds (1997). Here, the intensity of single phase reflections measured under constant analytical and preparations conditions primarily reflects the abundance of that phase, as long as compositional variations are limited. Both natural and heattreated samples were analysed in order to characterize the thermally induced changes in the mineral assemblage. In addition, a number of polished thin sections were prepared for important samples. These were first studied optically by transmission light microscopy in order to observe rock fabric and textural relationships of both sedimentary to deformation-related fabrics. Furthermore, some samples were studied by scanning transmission electron microscopy (TEM) using a Philips CM 20 at the University of Granada, in order to obtain high resolution images (up to 200 000 x magnification) and electron diffraction patterns

496

H. DE WALL & L. N. WARR

of single crystals. The methods applied are described in more detail in Warr & Nieto (1998).

Results Regional variations in the AMS fabrics Variations in the shape of AMS fabrics across the region are described as follows, from north to south along the studied profile (Fig. 2). In the Westward Ho! to Hartland Point section (area A in Figs. 1, 2a) the AMS fabric is oblate with a relatively constant shape factor (7") and anisotropy factor (Pf) of around 1.1. Such fabrics imply a strong compactional component, in accordance with the presence of a distinct bedding fabric and the absence of a cleavage defined by preferred mineral orientation. The rock cleavages observed are localized fracture or pressure solution fabrics, associated with the higher strain regions of fold hinges. Here, two samples from such folded and cleaved rocks (locality 4 in Fig. 2a) are characterized by distinct prolate shapes. In area A the dominant fabric element observed by light microscopy is defined by the preferred orientation of detrital grains, in particular white micas. Using reflected light microscopy, opaque phases were seen to consist of Fe-oxides, Fe-sulphides and vitrinite grains aligned parallel to the bedding. At higher resolutions using TEM, various generations of authigenic illite/smectite aggregates can be seen to have grown parallel with the compactional fabric (Warr & Nieto 1998). From Hartland Quay to Crackington Haven (area B in Figs. 1, 2a) the AMS ellipsoids show a strong variation, ranging from moderately oblate (T ~ +0.7) to moderately prolate (T ~ -0.6) shapes. The anisotropy, ranging from 1.01 to 1.10, is generally lower than in area A. However, locally higher anisotropies of up to 1.2 occur. In this section of the profile a south-dipping cleavage is developed, which increases in intensity southwards towards the margin of the Culm Basin (Sanderson 1979). Two principle fabrics can be observed in hand specimens and thin sections. The first is the bedding, marked by the orientation of detrital components, as described in area A. The second is defined by a strong pressure solution cleavage with local rotation of grains into the direction of maximum compression. TEM investigations of this material show authigenic growth of illite/ smectite and kaolinite grains in both bedding and cleavage directions (Warr & Nieto 1998). From the Strangles, across the Rusey fault zone and into the Tintagel High Strain Zone

(area C in Figs. 1, 2a), the magnetic fabric shows an increasingly oblate shape (T > 0.7) which is accompanied by a notable increase in anisotropy (Pf = 1.2-1.3). This region is an area defined by a well-developed slaty cleavage, characterized by polyphase deformation and metamorphism (Primmer 1985; Warr et al. 1991; Warr & Nieto 1998). The distinct jump in metamorphic grade from lower to upper anchizonal conditions across the Rusey fault zone (Warr et al, 1991, fig. 3), and the increased temperature gradient towards the fault documented by Andrews et al. (1996), is accompanied by an abrupt increase in anisotropy and flattening of the AMS fabric across this structure. However, south of the Rusey fault, in the highly deformed rocks of the Tintagel High Strain Zone, the AMS fabrics show little variation. Multiphase deformation is homoaxial in this high strain region (Andrews et al. 1988; Warr 1989). The three areas defined (A, B, C) also show differences in both the orientation of the AMS principle axes and the rock fabric (Fig. 2b). In area A /^max and «int lie within the bedding fabric, with ^mjn aligned to the bedding pole and ftmax in the direction of fold axes. Such a fabric is referred to as a normal magnetic fabric (Rochette et al. 1992, 1999). This relationship is also observed in the higher metamorphic area C, but here the «min axes are oriented perpendicular to the main slaty cleavage. In contrast to A and C, area B is characterized by the occurrence of oblique (inverse) fabrics with a stronger scattering of axes where ^min and ftmax are oriented along a great circle and define a girdle distribution normal to the main ENE-WSE trend of fold axes in this area. The K-mi axes form a well-defined, sub-horizontal cluster in the direction of fold axes. This geometry cannot be directly related to magnetic fabrics commonly observed in folded phyllosilicatebearing rocks (e.g. Liineburg et al. 1999; Pares et al 1999).

Changes due to heat treatment After heat treatment a change in the orientation of the AMS elements is evident across the regional profile (Fig. 2c). Whereas no consistent change in the orientation of the AMS axes is observed in areas A and C, the samples from area B indicate a transformation from abnormal to normal fabrics after heating (Fig. 2c). The magnetic foliation poles of the oblique, inverse fabrics, which define a N-S trending girdle distribution (7r-plane), are transformed after heating into a slightly N-S elongated cluster distribution,

AMS IN SIDERITE-BEARING PELITIC ROCKS

497

Fig. 2. AMS data obtained from along the Culm section for the subdivided areas A, B, C; (a) shows variations in anisotropy (P7) and shape (T) factors for localities 1 to 18. RFZ: Rusey fault zone; (b) The orientation of AMS principal axes before heat treatment (stars indicate the average orientation of fold axes in the areas A and B); and (c) The orientation of AMS principal axes after heating treatment. Projections are into the lower hemisphere (ftmax: squares, « int : triangles, ftmin: dots) using a Schmidt net.

498

H. DE WALL & L. N. WARR

indicating shallow dipping magnetic foliations. Before heating, the magnetic lineations of these oblique fabrics scatter along the vertical, N-S trending ?r-plane. During heating this lineation trend migrates to lie horizontally and defines a new, general E-W and NE-SW trend lying sub-parallel to the orientation of fold axes in this section of the Culm Basin. Bulk magnetic susceptibilities before and after heat treatment also show variations for the subdivided areas A, B, C. Before heating, little variation in the susceptibilities is observed

across the regional profile, with mean values in the range of 1(T4SI. Exceptions to this pattern are two samples from Maer Cliff (locality 8) and one of the nine rock samples taken from fold structures at Crackington Haven (locality 9). The higher values recorded from these samples can be related to a coarser, sandy pelitic lithology that contains detrital magnetite. These observations are in accordance with the regular occurrence of ferrimagnetic minerals in the sandstone Ethologies of this region (pers. com. A. Morris 2003).

Fig. 3. Thermomagnetic analysis of magnetic susceptibility for a siderite-bearing sample from locality 11 (Crackington Haven): (a) the low-temperature course before heating indicates paramagnetic behaviour; (b) the high-temperature course reflects oxidation of the sample at elevated temperatures; (c) full course of thermomagnetic behaviour for a sample after initial heat treatment.

AMS IN SIDERITE-BEARING PELITIC ROCKS

After heat treatment all bulk magnetic susceptibilities show a significant increase, which varies in character for each area defined. Area A shows strong variation of values between 1 x 1(T3 and 2.5 x 1(T2 (SI units); area B has a significantly higher range of susceptibilities between 2 x 10 and 2 x 10"1; and area C displays a general tendency towards decreasing

499

values southwards with values in the range of 2.5 x 10"3 to 5 x 10"2. In order to establish the cause of these variations a combination of temperature-dependent susceptibility measurements and X-ray diffraction analyses was undertaken. The results, presented in Figures 3 and 4, reveal the thermal alteration of siderite into magnetite as the

Fig. 4. (a) Random powder X-ray diffraction pattern of a sample from zone B (Maer Cliff, the locality 8 in Fig. 1) before and after heating in an oven for 24 hours at 500 °C. Both d-values and Miller indices (in brackets) are given for siderite (the reactant) and magnetite (the product), (b) The X-ray diffraction intensity of the 2.7 A siderite peak for unheated samples (in counts per second) is plotted against the volume susceptibility after heat treatment. The peak intensity of siderite can be used as a qualitative measure of mineral abundance by following methodology outlined by Moore & Reynolds (1997).

500

H. DE WALL & L. N. WARR

primary cause of the susceptibility increase. Siderite is paramagnetic above the Neel temperature of 38K, indicated by a linear I/K relationship with T (Fig. 3a), as predicted by the Curie law. Heating above 350°C causes a transformation of siderite into a ferrimagnetic phase with a Curie point of 590 °C, characteristic of pure magnetite (Fig. 3b, c). Transformation of siderite to magnetite or maghemite during heating of samples has previously been recognized by Ellwood et al (1986), Hirt & Gehring (1991) and Pan et al. (2000). The siderite to magnetite conversion is also evident in powder XRD patterns when comparing the unheated and heat treated specimens (Fig. 4). Some hematite and maghemite can also be identified in XRD patterns, but is generally only present in small amounts. Hence, during the 24-hour heating at 500 °C in the oven experiments most of the siderite was

thermally altered to magnetite. The antiferromagnetic nature of hematite in contrast to ferrimagnetic magnetite and maghemite does not contribute significantly to the bulk increase in susceptibility observed. A clear relationship can also be observed between the intensity of the main siderite reflection at about 2.7 A and the volume susceptibility after heating (Fig. 4b), indicating this mineral to be the primary source. Conversion of other Fe-bearing minerals (such as pyrite) may also be contributing to the production of magnetite. However, comparison of natural and heated XRD patterns show these minerals to be of minor importance. High-resolution TEM images of selected pelite samples also show fine-grained siderite crystals in micro-veins that have grown sub-parallel to illite basal planes (Fig. 5a). Such micro-vein material is estimated to form c. 5-10% of the rock

Fig. 5. High resolution TEM images (x 88 000) of a mudstone sample from Crackington Haven, (a) Thin siderite veins, commonly between 10 and 30 nm in thickness, occur along the margins of illite crystallites. Note the granular texture of the vein fillings, which is considered to reflect nano-sized siderite crystals of c. lOnm. (b) shows a single siderite crystal with a rhombohedral crystal face oriented parallel to basal planes of illite and the bedding plane fabric; (c) siderite grain with areas of high defect concentration and lattice distortion.

AMS IN SIDERITE-BEARING PELITIC ROCKS

501

Fig. 6. The Crackington Haven fold structure with a flat upper limb and overturned lower limb, (a) field photo with sample locations (I, II); (b) structural field data plotted in a Schmidt net.

volume and is commonly detected in XRD patterns. The mottled texture of carbonate veins is considered to reflect a very fine particle size, with crystals of c. 10 nm in size. Larger single micron-scale euhedral crystals are also observed (Fig. 5b), showing variable concentrations of lattice defects, indicative of some intracrystalline strain (Fig. 5c). AMS fabrics at Crackington Haven In order to resolve the complexities observed in area B, which has overlapping contributions of compactional and deformational fabrics, as well as a mixture of paramagnetic phyllosilicate and siderite subfabrics contributing to the bulk AMS fabric, a detailed study was conducted at the Crackington Haven locality (grid reference

SX 1425 9695). Here, a south-verging recumbent fold structure (Fig. 6), with an inter limb angle of 60° and an axial plane dipping 30° to the north was selected. The upper limb of the fold is relatively flat lying and has been notably thinned, whereas the lower overturned limb dips more steeply and has been thickened. The flatter upper limb also contains syn-folding extensional quartz-dolomite veins documented by Beach (1977). Samples were selected from a cleaved mudstone horizon, positioned between two distal turbidite sandstone beds (Fig. 6a). This mudstone horizon shows thickening in the hinge zone, typical of folded multilayered sequences of varying competence, with a mixture of fold-classes 1 and 3 (Ramsay 1967). Two positions around the fold were sampled, from the upper and lower limb, as shown in Figure 6. In order to evaluate the relationship between the

502

H. DE WALL & L. N. WARR

Fig. 7. Orientation of the long axes of AMS ellipsoids (ft max , magnetic lineation, shown as triangles) from the Crackington fold structure plotted in Schmidt nets: (a) samples from the upper limb, and (b) the lower limb, (c) Compilation of both data sets after reorientation of the bedding planes to the horizontal position. Dots show the mean orientation of bedding poles calculated from the distribution density of elements shown in Fig. 6.

AMS fabrics and the bedding planes both the fold limbs and their respective magnetic fabric axes have been restored to a horizontal bedding position; a routine referred to as the 'fold test' in palaeomagnetic studies. Before heat treatment and prior to unfolding, magnetic fabrics from samples in the upper sub-horizontal limb show a SE-plunging ^max direction, whereas the steeper lower limb is mostly characterized by a sub-horizontal, NNW trending Kmax (Fig. 7a, b). After rotation of the bedding to the horizontal position, the Kmax axes measured in both limbs lie along part of a small circle inclined 42° to the bedding pole (Fig. 7c). This implies the magnetic fabric has a distinct relationship to the bedding and therefore can be considered as pre-folding in age. After heat treatment of specimens from the Crackington Haven fold, changes in the geometry and orientation of the AMS axes are observed (Fig. 8a, b). Before heating, the AMS ellipsoids show variations in shape from oblate to prolate, whereas after heating all samples have strongly oblate ellipsoids. The Km^n axes plot within a cluster roughly normal to the bedding plane and the Kmax axes are oriented sub-horizontally in a NE-SW direction. Interpretation and discussion Three regional areas (A, B, C) have been recognized in the Culm Basin section, each of which is characterized by distinct magnetic fabrics. The untypical magnetic fabrics of area B have been shown to be dominated by a paramagnetic siderite fabric, which based on our fold test was found to be pre-folding in origin. In the following section we discuss the mechanism of siderite formation and address the regional implications

of this fabric as a possible indicator of the palaeo-fluid flow direction in this forelandbasin setting.

Control on the AMS fabrics: indications of substrate-controlled growth The natural AMS fabric related to the siderite growth shows a distinct angular relationship to the bedding planes and is characterized by a small-circle geometry (Fig. 7). Models for different siderite growth fabrics and the resulting AMS fabrics have been discussed by Hounslow (2001) to explain the textures observed in sedimentary siderite concretions in Lower Jurassic and Carboniferous marine and non-marine mudstones. Substratecontrolled growth of carbonate can result in various magnetic fabric geometries, with differences in the direction of maximum susceptibility controlled by the preferred orientation of c axes in respect to the substrate plane. Hounslow (2001) suggested that c axes lying at a defined angle (the angle between rhomb crystal faces and c axes) to the substrate surface result in distinct small-circle distributions. The geometry we found in the Culm pelites (area B) is very similar (Fig. 9). Samples from the Crackington fold structure revealed distinct inclinations between the Kmax axes and the bedding planes with angles in the range of 34 to 55°. After application of the fold test, a distribution on a defined segment of a small circle is indicated. We interpret this geometry as substrate-controlled growth fabrics of rhomb-shaped crystals that preferentially grew epitaxially on the basal surfaces of clay substrates. Further indications of such substrate-controlled growth can be inferred from the HRTEM images, where the rhomb surfaces of siderite crystals are observed to lie

AMS IN SIDERITE-BEARING PELITIC ROCKS

503

Fig. 8. (a) Shape of AMS ellispoids for samples from the upper limb (diamonds) and lower limb (squares) before (open) and after (filled) heat treatment, (b) Orientation of ellipsoid axes plotted in a Schmidt net projection following the heat treatment. Note Kmax axes are close to horizontal and Avmin correspond roughly with the bedding poles of the respective fold limbs. UL: upper limb, LL: lower limb.

directly along the basal (001) planes of illite particles (Fig. 5b). Substrate-related growth could theoretically result in a 360° small-circle distribution, if the c axes of rhombohedral crystals have a full rotational degree of freedom around the pole to the substrate. Such a scenario is shown in Figure 9 with an isotropic bulk geometry of the resulting AMS fabric (Hounslow 2001, p. 552, fig. 14c). The samples from our study, however, plot on a defined segment (c. 130-220°) of the small circle, indicating a preferred unidirectional orientation of siderite c axes, dipping c.45° in a northerly direction. We interpret this preferred orientation of siderite crystals to reflect growth from a migrating fluid that circulated along the bedding parallel anisotropy during burial diagenesis. Based on the

c axes orientation of the siderite, a north-south axis of flow can be inferred, and assuming that the c axes are oriented away from the direction of fluid transport (rhombs facing downstream), we also speculate the fluid transport direction was from south to north. The primary occurrence of siderite veins parallel and sub-parallel to the bedding fabric (Fig. 5a), in addition to the pre-folding crystallization history, implies a low-temperature diagenetic origin for these Fe-carbonate minerals. This stage of diagenetic growth was probably related to the circulation of burial formational waters within this Upper Carboniferous sedimentary basin, although precise formation temperatures cannot be specified. On the basis of clay mineral assemblages characteristic of diagenetic/anchizone

504

H. DE WALL & L. N. WARR

Fig. 9. Model for the interpretation of magnetic fabric geometry in siderite-dominated pelites of the Culm Basin, (a) Sketch of substrate-controlled growth showing siderite crystals along the bedding plane and the inferred geometry of AMS grain ellipsoids, (b) Schmidt net projection of Kmax directions from upper and lower limbs of the Crackington fold after unfolding. The grey-shaded small circle represents the schematic oaxis pole density distribution (increasing density with darker colour) as used by Hounslow (2001) for growth fabric interpretation of inclined siderite c axes.

conditions, maximum regional rock temperatures at the Crackington Haven locality did not significantly exceed 150 °C (Kisch 1987; Warr et al. 1991). These were post-dated by higher fluid temperatures (up to c. 300 °C) that have been documented on the basis of vitrinite reflectance and fluid inclusions studies (Cornford et al. 1987; unpublished diploma thesis of Sappok 1996). These hydrothermal fluids occurred synchronous with folding and were responsible for the extensional quartz-dolomite mineralization in the thinned upper limbs of the south-vergent folds. No siderite mineralization occurred during this hydrothermal activity but pre-existing diagenetic crystals did suffer some intracrystalline deformation (Fig. 5c). After heat treatment, the AMS reflects the geometry and orientation of the ferrimagnetic fabric. Replacement of siderite (+pyrite) by magnetite (+maghemite) requires a different approach in the magnetic fabric interpretation.

Whereas the AMS before heating is controlled by the crystallographic preferred orientation of paramagnetic Fe-carbonates, the ferrimagnetic AMS after heating is related to grain shape anisometry and the distribution anisotropy of magnetite clusters (e.g. Gregoire et al. 1998). Therefore, the initial paramagnetic fabric reflects the internal structure of these siderite veins and the heated fabric can be taken to reflect the orientation of siderite veins. In the Culm area B the distinct oblate shape of the AMS after heating (Fig. 8a) and the orientation of ttmin axes normal to the bedding suggests planar magnetite fabrics parallel to the bedding planes. This fabric can be explained by the bedding-parallel siderite micro-veins observed in the high resolution TEM images. Such micro-veins could have been developed within a deviatoric stress field in the presence of high fluid pressures, enabling the veins to dilate parallel to the planar fabric anisotropy.

AMS IN SIDERITE-BEARING PELITIC ROCKS

The clustering of ttmax axes in a NE-SW direction (Fig. 8b) may be explained by a component of layer parallel shortening occurring during the initial stages of deformation, which provides a cryptic rotational fabric of the bedding planes around axes perpendicular to the maximum compressional stresses. Regional implications of the siderite fabrics Abnormal, Fe-carbonate related magnetic fabrics are not commonly observed in other foreland basin settings. Typically, phyllosilicate fabrics (predominately chlorite) dominate in such areas (e.g. Pares et al 1999). In the Culm Basin the predominance of siderite in the magnetic fabrics is limited to the southern part of the Upper Carboniferous sequence, which includes both the Crackington and the Bude formations and therefore appears to be independent of the environments of deposition. As deformed anchizonal rocks occur also north of this zone (e.g. Hartland Quay section), no clear relationship can be related to the grade of low-temperature metamorphism and intensity of deformation in these sequences. Based on the early, low-temperature formation of the siderite it therefore appears that Zone B marks an area of enhanced and directed diagenetic fluid flow. The inferred south to north direction does, however, imply a genetic relationship with the migrating fold and thrust belt of the Variscan orogeny. Based on K-Ar cooling ages of the Palaeozoic slates to the south, this suggests the proximity of an uplifting thrust wedge (Warr et al. 1991). The intense rock deformation within the thrust wedge, synchronous with sedimentation and burial in the foreland basin, provides a mechanism for fluid migration towards the foreland. In such a scenario, deeper source basinal fluids are squeezed out into the freshly deposited, less dense foreland basin sediments, before themselves being incorporated into the migrating thrust wedge. The mobility of material may have also been enhanced by the relatively high geothermal gradients recorded in these Upper Palaeozoic coal-bearing basins of Europe (Gayer et al. 1998). Conclusions AMS studies on mud- and siltstones from the Upper Carboniferous Culm Basin show regional deviations from the phyllosilicate-controlled magnetic fabrics that are typically recorded for diagenetic to low-grade metamorphic rocks. The AMS geometries observed before heating are related to the formation of diagenetic siderite

505

prior to the onset of folding and thrusting of this foreland basin rock sequence. Restored magnetic lineations define a segment of a small circle distribution, which is considered to reflect substratecontrolled siderite growth within a migrating fluid medium. The regional distribution of siderite, in combination with the directional information from the AMS, is suggested to indicate a foreland-directed fluid-flow driven by the advancing Variscan thrust wedge. We thank F. Nieto (Granada) for use of the TEM facilities and L. Nano (Wurzburg) for his technical assistance. We also acknowledge the helpful reviews provided by B. Ellwood and M. Hounslow.

References ANDREWS, J. R., BARKER, A. J. & PAMPLIN, C. F. 1988. A reappraisal of the facing confrontation in north Cornwall; fold- or thrust-dominated tectonics? Journal of the Geological Society, London, 139, 493-504. ANDREWS, J. R., DAY, J. & MARSHALL, J. E. A. 1996. A thermal anomaly associated with the Rusey Fault and its implications for fluid movements. Proceedings of the Ussher Society, 9, 68-71. AUBOURG, C., HEBERT, R., JOLIVET, L. & CARTAYRADE, G. 2000. The magnetic fabric of metasediments in a detachment shear zone: The example of Tinos Island (Greece). Tectonophysics, 321, 219-236. BEACH, A. 1977. Vein arrays, hydraulic fractures and pressure solution structures in a deformed flysch sequence, SW England. Tectonophysics, 40, 201225. CARRAPA, B., BERTOTTI, G. & KRIJGSMAN, W. 2003. Subsidence, stress regime and rotation(s) of a technically active sedimentary basin within the western Alpine Orogen: The Tertiary Piedmont Basin (Alpine domain, NW Italy). Geological Society Special Publications, 208, 205-227. CORNFORD, C., YARNELL, L. & MURCHINSON, D. G. 1987. Initial vitrinite reflectance results from the Carboniferous of north Devon and Cornwall. Proceedings of the Ussher Society, 6, 461-467. DE WALL, H., BESTMANN, M. & ULLEMEYER, K. 2000. Anisotropy of diamagnetic susceptibility in Thassos marble: a comparison between measured and modelled data. Journal of Structural Geology, 22, 1761-1771. DEWEY, J. F. 1982. Plate tectonics and the evolution of the British Isles. Journal of the Geological Society, London, 139, 371-412. ELLWOOD, B. B., BALSAM, W., BURKART, B., LONG, G. J. & BUHL, M. L. 1986. Anomalous magnetic properties in rocks containing the mineral siderite: Palaeomagnetic implications. Journal oj Geophysical Research, 91, 12779-12790. GAYER, R., GARVEN, G. & RICKARD, D. 1998. Fluid migration and coal-rank development in foreland basins. Geology, 26, 679-682.

506

H. DE WALL & L. N. WARR

GREGOIRE, V., DARROZES, J., GAILLOT, P., NEDELEC, A. & LAUNEAU, P. 1998. Magnetite grain shape fabric and distribution anisotropy of vs. rock magnetic fabric: A three-dimensional case study. Journal of Structural Geology, 20, 937-944. HARTLEY, A. & WARR, L. N. 1990. Upper Carboniferous foreland basin evolution in SW Britain. Proceedings of the Ussher Society, 7, 212-216. HECHT, C. 1992. The Variscan evolution of the Culm Basin, south-west England. Proceedings of the Ussher Society, 8, 33-38. HENRY, B., JARDANOVA, D., JARDANOVA, N., SOUQUE, C. & ROBION, P. 2003. Anisotropy of magnetic susceptibility of heated rocks. Tectonophysics, 366, 241-258. HIGGS, R. 1984. 'Lake Bude' (early Westphalian, SW England): storm dominated siliclastic shelf sedimentation in an equatorial lake. Proceedings of of the Ussher Society, 6, 417-418. HIRT, A. & GEHRING, A. U. 1991. Thermal alteration of the magnetic mineralogy in ferruginous rocks. Journal of Geophysical Research, 96, B6, 99479953. HIRT, A. M., EVANS, K. F. & ENGELDER, T. 1995. Correlation between magnetic anisotropy and fabric for Devonian shales on the Appalachian Plateau. Tectonophysics, 247, 121-132. HOUNSLOW, M. W. 2001. The crystallographic fabric and texture of siderite in concretions: implications for siderite nucleation and growth processes. Sedimentology, 48, 533-557. HROUDA, F. & JEZEK, J. 1999. Magnetic anisotropy indications of deformation associated with diagenesis. Geological Society Special Publication, 151, 127-137. HROUDA, F. & POTFAJ, M. 1993. Deformation of sediments in the post-orogenic Intra-Carpathian Paleogene Basin as indicated by magnetic susceptibility anisotropy. Tectonophysics, 224, 425-434. IHMLE, P. F., HIRT, A. M., LOWRIE, W. & DIETRICH, D. 1989. Inverse magnetic fabric in deformed limestones of the Morcles nappe, Switzerland. Geophysical Research Letters, 16, 1383-1386. JELINEK, V. 1981. Characterization of the magnetic fabrics of rocks. Tectonophysics, 79, 63-67. KISCH, H. J. 1987. Correlation between indicators of very low-grade metamorphism. In: FREY, M. (ed.) Low Temperature Metamorphism, Blackie, Chapman and Hall, 227-300. KONTNY, A., FRIEDRICH, G., BEHR, H. J., DE WALL, H., HORN, E. E., MOLLER, P. & ZULAUF, G. 1997. Formation of ore minerals in metamorphic rocks of the German Continental Deep Drilling Site (KTB). Journal of Geophysical Research, 102, B8, 18323-18336. LUNEBURG, C. M., LAMPERT, S. A., LEBIT, H. D., HIRT, A. M., CASEY, M. & LOWRIE, W. 1999. Magnetic anisotropy, rock fabric and finite strain in deformed sediments of SW Sardinia (Italy). Tectonophysics, 307, 51-74. MELVIN, J. 1986. Upper Carboniferous fine grained turbiditic sandstones from Southwest England: a model for growth in an ancient, delta-fed subsea fan. Journal of Sedimentary Petrology, 56, 19-34.

MOORE, D. M & REYNOLDS, R. C. 1997. X-Ray Diffraction and the Identification and Analysis of Clay Minerals. Oxford University Press, Oxford (2nd edition). PAN, Y., ZHU, R., BANERJEE, S. K., GILL, J. & WILLIAMS, Q. 2000. Rock magnetic properties related to thermal treatment of siderite: behavior and interpretation. Journal of Geophysical Research, 105, Bl, 783-794. PARES, J. M., VAN DER PLUIJM, B. A. & DINARESTURELL, J. 1999. Evolution of fabrics during incipient deformation of mudrocks (Pyrenees, northern Spain). Tectonophysics, 307, 1-14. PRIMMER, T. J. 1985. A transition from diagenesis to greenschist facies within a Variscan fold/thrust complex in southwest England. Mineralogical Magazine, 49, 365-374. RAMSAY, J. G. 1967. Folding and Fracturing of Rocks. McGraw-Hill, New York & London. ROCHETTE, P. 1987. Metamorphic control of the magnetic mineralogy of black shales in the Swiss Alps: towards the use of 'magnetic isogrades'. Earth and Planetary Science Letters, 84, 446^56. ROCHETTE, P. 1988. Inverse magnetic fabrics in carbonate-bearing rocks. Earth and Planetary Science Letters, 90, 229-237. ROCHETTE, P., JACKSON, M. & AUBOURG, C. 1992. Rock magnetism and the interpretation of anisotropy of magnetic susceptibility. Reviews in Geophysics, 30, 209-226. ROCHETTE, P., AUBOURG, C. & PERRIN, M. 1999. Is this fabric normal? A review and case studies in volcanic formations. Tectonophysics, 307, 219-234. SAGNOTTI, L., SPERANZA, F., WINKLER, A., MATTEI, M. & FUNICIELLO, R. 1998. Magnetic fabric of clay sediments from the external northern Apennines (Italy). Physics of the Earth and Planetary Interior, 105, 73-93. SANDERSON, D. J. 1979. The transition from upright to recumbent folding in the Variscan fold belt of southwest England: a model based on the kinematics of simple shear. Journal of Structural Geology, \, 171-180. SCHNEIDER, J., DE WALL, H., KONTY, A. & BECHSTADE, T. 2004. Magnetic susceptibility variations in carbonates of the La Vid Group (Cantabrian Zone, NW-Spain) related to burial diagenesis. Sedimentary Geology, 166, 73-88. SAPPOK, N. 1996. Mikrothermometrie an Fluideinschliissen in Quartz-/Karbonateadern des variskischen Kulm Beckens, SW England. Unpublished Diploma thesis, University of Heidelberg, p.57. SINGH, J., SANDERSON, D. J. & TARLING, D. H. 1975. The magnetic susceptibility anisotropy of deformed rocks from North Cornwall, England. Tectonophysics, 27, 141-153. TARLING, D. H. & HROUDA, F. 1993. The Magnetic Anisotropy of Rocks. Chapman and Hall, London. THOMAS, J. M. 1988. Basin history of the Culm Trough of Southwest England. In: BESLY, B. & KELLING, G. (eds) Sedimentation in a Synorogenic Basin Complex: the Upper Carboniferous of Northwest Europe. Blackie, Glasgow and London, 24-37.

AMS IN SIDERITE-BEARING PELITIC ROCKS WARR, L. N. 1989. The structural evolution of the Davidstow Anticline and its relationship to the Southern Culm Overfold, north Cornwall, north Cornwall. Proceedings of the Ussher Society, 7, 136-140. WARR, L. N. 1993. Basin inversion and foreland basin development in the Rhenohercynian of south-west England. In: GAYER, R. A. & GREILING, R. O. (eds) Rhenohercynian and Sub-Variscan Fold Belts. Earth Evolution Series, Vieweg, 197-224. WARR, L. N. & NIETO, F. 1998. Crystallite size distributions in very low-grade metamorphic pelites:

507

A HRTEM and XRD study of clay mineral crystallinity index standards. The Canadian Mineralogist, 36, 1393-1414. WARR, L. N., PRIMMER, T. J. & ROBINSON, D. 1991. Variscan very low-grade metamorphism in southwest England: a diasthermal and thrust-related origin. Journal of Metamorphic Geology, 9, 751764. WHALLEY, J. S. & LLOYD, G. E. 1986. Tectonics of the Bude Formation, north Cornwall—the recognition of northerly directed decollement. Journal of the Geological Society, London, 143, 83-88.

This page intentionally left blank

Development of magnetic fabrics during hydrothermal alteration in the Soultz-sous-Forets granite from the EPS-1 borehole, Upper Rhine Graben JANA JUST,1 AGNES KONTNY,1 HELGA DE WALL,2 ANN M. HIRT3 & FATIMA MARTIN-HERNANDEZ4 l

Geologisch-Palaontologisches Institut, Ruprecht-Karls Universitat, INF 234, 69120 Heidelberg, Germany (e-mail: [email protected]) 2 Institutfur Geologie, University in Wurzburg, Pleicherwall 1, 97070 Wurzburg, Germany 3 Institut ftir Geophysik, ETH Honggerberg, HPP 01.2, 8093 Zurich, Switzerland 4 Faculty of Geosciences, Utrecht University, Budapestlaan 17, 3584 CD Utrecht, The Netherlands Abstract: The Variscan, magnetite-bearing Soultz-sous-Forets granite is found between 1420 and 2230m of the EPS-1 borehole situated in the Upper Rhine Graben (France). Our study focuses on the changes of magnetic properties that occur during the progressive hydrothermal alteration and fracturing of the Soultz granite after emplacement. The magnetic susceptibility («) of the granite is between 10 and 80 x 10"3 SI, and suggests that ferrimagnetic minerals are the primary carrier. During cooling and later tectonic and hydrothermal overprints, including the formation of the Rhine Graben, the granite was deformed under brittle conditions and partially altered by hydrothermal fluids. Along with this fluid activity, oxidation of magnetite to hematite occurred and reduced K (< 1 x 10"3 SI). AMS analysis on oriented samples documents the history of progressive transformation from primary magmatic fabric to tectonic fabric during hydrothermal alteration and faulting. The fresh granite with multidomain magnetite grains shows sub-horizontal magnetic foliations and randomly oriented magnetic lineations within the foliation plane. This fabric is similar to the magmatic fabric reflected by biotite. Transformation of the magnetic fabric started with localized magnetite oxidation along NW-SE oriented micro-cracks, which are probably associated with a late-magmatic alteration (stage I). Elongated and co-aligned magnetite relics within the newly formed hematite caused a well-defined NW-SE trending magnetic lineation and steeper magnetic foliation. Later alteration associated with intense brittle deformation (stage II) initially adopted this magnetic fabric, but intense cataclasis destroyed it. The geometry and orientation of magnetic fabric clearly indicate a hydrothermal alteration, which relates to the acting tectonic stresses in the post-emplacement history of the Soultz granite.

The EPS-1 borehole near Soultz-sous-Forets is part of the Hot Dry Rock (HDR) project, which drilled through magnetite-bearing, Variscan basement, the Soultz granite, in the depth interval between 1420 and 2230m. During the late-magmatic cooling of the pluton and the later formation of the Upper Rhine Graben, the intrusive rocks were partially altered by hydrothermal fluids. This alteration has been monitored along the profile by magnetic susceptibility measurements, which indicated a transition from ferrimagnetic to paramagnetic susceptibilities (Rummel & Konig 1991). Magnetic fabric analysis using the anisotropy of magnetic susceptibility (AMS) is often used to define and quantify the magnetic fabric (e.g. Bouchez et al. 1987; Tarling & Hrouda 1993; Ferre & Ameglio 2000). This method has been applied to a large number of granitoid rocks from different tectonic settings in order to correlate the shape and orientation of the magnetic

ellipsoid with tectonic strain and emplacement processes (examples from the Odenwald crystalline complex are given, for example, by Dietl & Stein 2001; Greiling & Verma 2001). While the AMS method has often been used to evaluate the tectonic strain acquired during ductile microstructure development in granitoid rocks (e.g. Ferre & Ameglio 2000; Hrouda et al. 2002), very few investigations focused on the influence of hydrothermal alteration and brittle deformation on magnetic fabrics. Changes in magnetic and fractal properties of fractured granites were reported by Nakamura & Nagahama (2001) from a drill core towards the Nojima fault, Japan. In weakly fractured granites, the magnetic fabric can be related to fracturing, while in strongly fractured rocks the AMS became nearly isotropic. In contrast to the Soultz granite, where alteration reduced magnetic susceptibility (Rummel & Konig 1991), in the Nojima Fault an increase in susceptibility was observed

From: MARTIN-HERNANDEZ, F., LUNEBURG, C. M., AUBOURG, C. & JACKSON, M. (eds) 2004. Magnetic Fabric: Methods and Applications. Geological Society, London, Special Publications, 238, 509-526. 0305-8719/04/ $15.00 © The Geological Society of London 2004.

510

J. JUST ET AL.

Fig. 1. Location map of the central Upper Rhine Graben (dashed line) and a schematic profile through the rift at the EPS-1 drill site at Soultz-sous-Forets.

during alteration, indicating a new formation of magnetite. Our study focuses on the magnetomineralogical changes and the related magnetic fabric development in the fresh and hydrothermally altered and fractured Soultz granite. We use a combination of low-field and high-field magnetic methods together with electron backscatter images on polished sections oriented parallel to the magnetic fabric. These results should lead to a better understanding of the processes controlling fabric development that reflect tectonic stresses during the brittle, post-emplacement history of an originally magnetite-bearing granite. Geological setting The NNW-SSE trending Rhine Rift is one of the most prominent extensional structures in central Europe, which developed from Middle to Late Eocene times by reactivation of a complex set of crastal discontinuities, formed as late to postVariscan (Permo-Carboniferous) relaxation structures (e.g. Ziegler 1996; van Wees et al 2000; Schumacher 2002). The reactivation of the Variscan discontinuities was induced by the Alpine Africa-Eurasia crustal convergence. Hereby, the prevailing tectonic stresses influenced the structural development of the localized alteration and fault zones within the Soultz pluton. Enhanced seismicity, abundant mineral water springs and distinct heat flow anomalies are the present-day features of this tectonic structure.

The EPS-1 borehole is located on the western side of the Upper Rhine Graben within a geothermal anomaly (Fig. 1). The borehole was cored from 980 to 2230 m and consists of sedimentary cover rocks down to 1420m underlain by granitic basement. The Soultz granite belongs to the late- to post-tectonic Variscan granitoids, ubiquitous in the European Variscan basement. Alexandrov et al. (2001) dated the magma emplacement at 331 ± 9 Ma (U-Pb ion-probe, zircon) and classified the granite as a monzogranite with high-K calc-alkaline composition. The Variscan age of the Soultz granite is within the range found for several granitoid intrusions from adjacent basement areas of the Mid German Crystalline Rise (Flottman & Oncken 1992). The granitoids were emplaced during the Permo-Carboniferous phase of wrench tectonics (Ziegler 1986). Petrological investigations of the Soultz granite from Stussi et al. (2002) have shown that pluton formation occurred in two different crystallization stages. The first was deep-seated at 11-12 km depth and characterized by the crystallization from the liquidus phase (hornblende, rare titanite, allanite, biotite, orthoclase, plagioclase) in a temperature range of 755790 °C at 3.5kbar. The second crystallization stage took place at 4.5-5.5 km depth, which led to a consolidation of the granitic body at temperatures of 665-715 °C and pressures of 1.5-2.0 kbar during ascent; this is characterized by the final crystallization of hornblende and magnetite. Genter & Traineau (1996) observed a weak magmatic foliation with sub-horizontal

DEVELOPMENT OF AMS DURING ALTERATION

511

inclinations, which was correlated to primary biotite and hornblende. EPS-1 borehole - magnetic susceptibility and petrography Based on the magnetic susceptibility log of the drill cores (Rummel & Konig 1991) (Fig. 2) the granite can be subdivided into an upper (14201550m) and a lower (1550-2230 m) section. The upper section consists of strongly altered granite with low magnetic susceptibility (K) values (K < I x 10~3 SI). The homogenous distribution of the low susceptibilities indicates a persistent alteration. The lower section consists mainly of fresh, unaltered granites with high ^-values in general (K < 10 x 10~3 SI). Three main, discrete fault zones with strongly decreased K values dissect the lower section; this is also reflected in the susceptibility log (Fig. 2). Microscopic examinations showed that oxidation of magnetite into hematite (martitization) during hydrothermal alteration was responsible for the reduction of K. Traineau et al. (1991) described two main types of hydrothermal alteration within the Soultz granite. The first type (stage I) is propylitic and pervasive. It took place under retrograde conditions during cooling of the magma and affected the whole plutonic body with varying intensities. The second type (stage II) is localized to fault zones and has an argillitic character. It is related to hydrothermally altered cataclastic granites and their hydrothermalized wall rocks. Bulk magnetic susceptibility was measured on cylindrical specimens (c. 11 cm3), selected from different localities within the profile (Fig. 2). Frequency distribution allows for the definition of susceptibility intervals for the fresh granite, the hydrothermally altered granite of stage I (pervasive alteration) and hydrothermally altered granite of stage II (localized alteration) (Fig. 3). The altered granites of stage II are subdivided into cataclastic granites and their hydrothermally altered wall rocks. Although there is no fresh, unaltered granite sensu stricto within the drill profile, in this study, the term 'fresh granite' is used for granites without any evidence of magnetite oxidation. The fresh granite shows a K interval between 10 and 80 x 10"3 SI, whereby almost 70% of the specimens have K values between 20 and 40 x 10~~3 SI. The K values indicate a magnetite concentration of approximately 1 wt% in the fresh granite. The hydrothermally altered granites of stage I show decreased K values and two maxima in their frequency distribution, one in the upper part (0.4-0.5 x 10~3 SI)

Fig. 2. Lithological section (modified after Genter & Traineau 1996) and magnetic susceptibility log with data (Rummel & Konig 1991) on drill cores at the 1414-2230 m depth interval in the EPS-1 borehole. The hydrothermal alteration observed in the geological section is classified in two stages. Black dots and numbers indicate sample position for this study, according to the nomenclature of the core index (Genter & Traineau 1991).

and the second one in the central and lower part (10-15 x 10"3 SI). This bimodal distribution indicates different alteration and magnetite oxidation intensities during the pervasive alteration. Different alteration intensities are also observed for the later, localized hydrothermal alteration. The cataclastic granites have K values between 0.8 and 10 x 10~3 SI and their hydrothermally

512

J. JUST ET AL.

Fig. 3. Frequency histogram of the mean magnetic susceptibility measured on standard specimens (~10cm3) from cores of the EPS-1 borehole.

altered wall rocks have the lowest susceptibilities: between 0.1 and 0.5 x 10~3 SI. These hydrothermalized wall rocks have a maximum frequency distribution between 0.2 and 0.3 x 10~3 SI, which is characteristic of the most intense hydrothermal alteration in the Soultz granite. In the following, the magnetic minerals and fabrics of the unaltered fresh granite and the subsequent stages of the hydrothermal alteration are described in more detail. Fresh granite The high magnetic susceptibilities in the fresh granite, which does not show any evidence of oxidation, are controlled by magnetite with grain sizes between 300 and 500 urn (Fig. 4a, b). The magnetite grains are mostly euhedral or subhedral, and are frequently found in magnetite clusters associated with unaltered biotite and

titanite (Fig. 4b). The primary mineral assemblage is comprised of K-feldspar megacrysts (up to 5cm diameter) in a coarse-grained matrix of diamagnetic quartz, feldspar, paramagnetic biotite and hornblende, as well as accessory diamagnetic apatite and titanite. The primary minerals are partly affected by the early pervasive alteration stage I. This alteration started during the emplacement of the granite and persisted under retrograde conditions during cooling (Jacquemont 2002). During a late magmatic stage (700 °C and 3^kbar), magmatic epidote was formed due to the decomposition of hornblende. Continued cooling to c. 300 °C caused the formation of paramagnetic hydrogarnet, prehnite and epidote in biotite. An early hematitization of AT-feldspar megacrysts (Dubois et al. 1996), observed also in fresh granites, indicates an increase of oxygen fugacity (fO2) during the cooling history. In addition, small hematite platelets and rods

Fig. 4. Micrographs (SEM) from thin sections parallel to the magnetic foliation (ft max / K int) showing typical microfabrics of the fresh and hydrothermally altered Soultz granite. The sample locations are shown in Fig. 2. Mag: magnetite; hem: hematite; bio: biotite; sph: titanite; rut: rutile; pig: plagioclase; kfs: A^-feldspar; qtz: quartz; car: carbonate; cal: calcite. (a) Single grain of magnetite associated with biotite and titanite surrounded by partly altered plagioclase within fresh granite. Micro-crack in magnetite is parallel to ^max. (b) Clusters of magnetite grains are parallel to the shape preferred orientation of biotite associated with titanite within weakly hydrothermally altered granite, (c) Magnetite oxidized into hematite (heml) with elongated magnetite relics with (111) long axis oriented parallel ^max in hydrothermally altered granite, (d) Decomposition of plagioclase into Na-rich (dark grey) and Ca-rich (bright grey) plagioclase components. Growth of Ca-Fe-Mg carbonates and illite parallel to crystallographic axis of plagioclase within hydrothermally altered granite, (e) Cataclased heml grains, (f) Heml is cataclased and dissolved. Precipitation of fine-grained hematite (hem2) grains occurred along cracks oriented parallel to the magnetic axis ^max and /cint or (g) hematite (hem2) is associated with illite. (h) Progressive illitization in hydrothermalized wall rock accompanied by abundant Ca-Fe-Mg carbonates.

DEVELOPMENT OF AMS DURING ALTERATION

513

514

J. JUST ET AL.

(<10jim) grew along the biotite cleavage. The oxidation of magnetite to hematite (martite) occurred during the latest phase of pervasive alteration stage I.

Hydro thermally altered granite In the Soultz granite, two main hydrothermal alteration stages were previously described by Traineau et al. (1991). In this study, we focused on alteration processes, which affect the magnetic mineralogy. The onset of a gradual decomposition of magnetite into martite was accompanied by a chloritization (Fe-Mg chlorite) of biotite. These alteration processes took place during the latest phase of stage I alteration, where retrograde conditions during the pluton emplacement caused decreased temperatures of c. 200 °C and increased oxygen fugacity, which most likely initiated the magnetite oxidation. The magnetite decomposition is visible in a progressive decrease of ferromagnetic into susceptibility values dominated by paramagnetic phases (K < 1 x 10~3 SI). Hematite replacement started in grain boundaries and micro-cracks, which further evolved in general along the crystallographic axis (111) of magnetite (Fig. 4c). Depending on the alteration intensity magnetite relics of varying grain sizes were retained within the martite grain (Fig. 4c). Shape and orientation of the retaining magnetite relics were controlled by previously existent microcracks within magnetite grains. With advanced hydrothermal alteration of the granite hematite was pseudomorphed after magnetite grains. But, in hydrothermally altered granites of stage I, no completely oxidized magnetite grains were observed. Plagioclase and biotite, which were minerals of the primary magmatic assemblage, were transformed into illite 4calcite -f chlorite. Secondary hematite precipitated probably from residual fluids, which circulated during the latest cooling stage of the plutonic body. Like hematite, the newly formed minerals pseudomorphed the primary minerals (Fig. 4d). Granites with significantly oxidized magnetite, assigned to the stage I alteration, were found only in the depth interval 1550 to 1580m. This section is referred to as the central part of the borehole profile in this study (Fig. 2). The hydrothermal alteration stage II occurs within faulted and fractured zones showing the strongest alteration intensity of the granite with the strongest oxidation of magnetite. These granites contain inhomogeneously distributed networks of fine veins, mostly filled by secondary

minerals such as quartz, illite, hematite and carbonates, which were precipitated from fluids. With increasing distance from a fault zone, cataclasis vanished and wall rock halos of hydrothermally altered minerals developed. Hematite pseudomorphisms after magnetite were often fractured and a second hematite generation was formed along cracks (Fig. 4e, f)- The size of magnetite relics depends on the intensity of the stage II alteration as well as on the intensity of the earlier stage I alteration. Hereby, the magnetite relics can be completely hematitized representing the strongest hydrothermal alteration. However, a complete hematitization of magnetite is rarely observed in the Soultz granite. In the altered wall rock halos, the primary minerals, particularly feldspars, can be fully dissolved (Fig. 4g, h) and replaced mainly by illite, which pseudomorphed after the primary minerals. In contrast to stage I alteration, where chlorite replaced the Fe-Mg silicates (biotite, hornblende, chlorite), these phases are also replaced mainly by illite. From the petrographic view, the most obvious difference, compared to stage I, is the higher content of illite. Results

Rock magnetic properties In order to evaluate the decreasing influence of magnetite on magnetic susceptibility and fabric during the hydrothermal alteration and the associated increasing content of hematite, the rock magnetic behaviour was investigated on representative samples. Fresh and hydrothermally altered granite samples of different alteration intensities with magnetite relics and without any petrographic evidence of magnetite were analysed using the temperature dependence of low-field magnetic susceptibility (K (T)), isothermal remanent magnetization (IRM), and stepwise thermal demagnetization (TH) experiments of a multi-component IRM (Fig. 5). The K (T) measurements were done at GPI Heidelberg using a KLY-2 Kappabridge, combined with the CS-2/CS-L furnace apparatus of AGICO (Hrouda 1994). IRM acquisition curves were obtained with a maximum field of 1T using an ASC Scientific Model IM-10-30 Impulse Magnetometer at the ETH Zurich. Determination of unblocking temperatures was done by thermal demagnetization (TH) with progressive heating in zero-field. Before thermal demagnetization, the samples were magnetized in three orthogonal directions with different field intensities (x = 200 mT; y = 600 mT; z = 2.6 T) in order to

Fig. 5. Rock magnetic properties of fresh and hydrothermally altered granite from stage I and II. (a) K (T) curves with susceptibility normalized to room temperature, (b) Acquisition of IRM with stepwise magnetization up to 2.6T. (c) Before thermal demagnetization (TH) the samples were magnetized in three directions: x (•) = 200mT, y (A) = 600 mT and z(•) = 2.6 T.

516

J. JUST ET AL.

separate the different coercivity components (Lowrie 1990).

Fresh granite The K (T) measurements for the fresh granite show a curve typical for multidomain (MD) magnetite with a reversible heating and cooling run (e.g. Kontny & de Wall 2000; Muxworthy & McClelland 2000). Characteristic features of the MD magnetite are the Verwey transition at — 150°C, no distinct Hopkinson peak and a Curie temperature (rc) of 585 °C, which indicate the predominance of MD grains of magnetite (e.g. Dunlop & Ozdemir 1997). This interpretation is also supported by the remanence experiments. IRM acquisition curves are typical for a soft magnetic component, which can be saturated at fields well below lOOmT. The thermal demagnetization curves are typical for low coercivity material (x axis), which is demagnetized at 575 °C. indicating magnetite as the magnetic carrier (Fig. 5c). The unblocking of magnetite is sharp and suggests a limited grain size fraction and compositional range.

Hydro thermally altered granite In the pervasively altered granite of stage I with magnetite relics, the K (T) measurements still show the TC of magnetite at 585 °C but, in contrast to the fresh granite, an irreversible heating and cooling run. Although observed microscopically, hematite was not identified by its Tc. We assume that in the Soultz granite hematite content is too low for identification using the K (T) measurements, because the magnetic susceptibility of hematite is three orders smaller than that of magnetite. Additionally, hematite was probably transformed into magnetite during heating before reaching its Tc and thus only the Tc of magnetite was observed. However, the strong increase of K, a single Tc of magnetite as well as the Verwey transition at -150°C during a second cooling run indicate the formation of new magnetite phase during the heating run (Fig. 5). The IRM acquisition curve documents the magnetization of soft and hard magnetic components. Nearly 50% of the bulk magnetization is acquired at fields below lOOmT, characteristic of magnetite, but saturation of the second component requires magnetic fields higher than the peak field of 2.6 T. This behaviour is typical for hematite. Thermal demagnetization also

indicates the dominance of the low coercivity component demagnetized at 575 °C. However, in contrast to the fresh granite, this hydrothermally altered sample shows a wider temperature range of unblocking. This indicates the presence of a wide grain-size spectrum of magnetite. Magnetic behaviour of this altered sample is therefore still dominated by the magnetite component while the influence of hematite is subordinate. A grain-size reduction of magnetite due to a partial transformation into hematite during hydrothermal alteration as observed by optical microscopy (Fig. 4c) is clearly evident in the magnetic signature. K (T) measurements of granite with the most intensive hydrothermal alteration do not show the Verwey transition in the low-temperature run. During the heating run, an increase in magnetic susceptibility is observed at temperatures above 450 °C, which is interpreted as the formation of a strong magnetic phase. This newly formed phase shows a Tc of 590 °C, indicating the formation of cation-deficient magnetite. From the K (T) magnetite can be excluded as an original phase. A strong irreversibility of the heating and cooling run is observed along with a strong increase in magnetic susceptibility during cooling. In contrast to the less altered sample, a second low-temperature run reveals no Verwey transition, and magnetic susceptibility decreases with decreasing temperature. Therefore, the formation of an unstable, maghemite phase is favoured and the formation of pure magnetite during the heating experiment can be excluded. In the IRM acquisition curve the hard magnetic component dominates the magnetization behaviour. However, the IRM curve still indicates the presence of a softer magnetic component. An initially concave IRM curve (see inlay in Fig. 5b) suggests that MD hematite rather than magnetite is the magnetic carrier. This agrees with microscopic observations showing no magnetite but large grains of hematite (100300 um) in addition to fine hematite grains < 10 urn. This interpretation is in agreement with studies of Kletetschka & Wasilewski (2001) on the transition between truly multidomain (>100um) and single domain (SD; <50 jim) behaviour in hematite. The SD hematite grains possess high coercivities and are not saturated at 2.6T during IRM acquisition. The TH experiments show that the high and intermediate coercivity components are demagnetized at 700 °C, indicating hematite as the magnetic carrier. In contrast to the high coercivity component, the intermediate coercivity component shows a wider temperature range of unblocking

DEVELOPMENT OF AMS DURING ALTERATION

and probably can be related to the presence of a wide MD hematite grain-size spectrum. Magnetic fabric Low field magnetic fabric measurements reflect the bulk, preferred orientation of all ferromagnetic, paramagnetic and diamagnetic minerals. The magnetic susceptibility measurements and rock magnetic characteristics presented in the previous sections have revealed a progressive change in the magneto-mineralogy of the Soultz granite from ferrimagnetic behaviour in the fresh parts to paramagnetic behaviour in strongly altered parts. This change is also documented in the geometry and orientation of magnetic fabrics. Thirty-nine oriented samples were taken from the 1420-2230 m granite interval of the EPS-1 borehole. A Kappameter KT5 (GEOFYZIKA) with a sensitivity of 1 x 10"5 SI and an operating frequency of 10 kHz has been used for sample selection, since susceptibility was diagnostic for the fresh and altered granites. Up to 6 cylindrical specimens (\" standard cylinders) per core were prepared with a sample volume of ~10.06cm3. Determination of the directional dependence of the anisotropy of magnetic susceptibility (AMS) was done using the KLY-2 Kappabridge (AGICO) with a detection limit of 4 x 10~8 SI. Several parameters are used to describe the AMS ellipsoids. The AMS parameters used in this study are the shape factor (T = (InF - lnL)/(lnF + InL)) and the degree of AMS defined by the corrected anisotropy factor (P7 = exp(2(ln^ max - ln«) 2 + 2(ln Kint - In ft)2 4- 2(ln ftmin - Inft)2)).Cylinders were reoriented into geographic coordinates to interpret the AMS, using the formation microscanner data given by Center & Traineau (1991). After AMS measurements, thin sections of oriented specimens parallel to the magnetic foliation plane (ft max ft int -plane) were prepared in order to compare magnetic fabrics with petrofabrics using a scanning electron microscope (SEM).

Anisotropy of AMS The shape and degree of anisotropy of AMS ellipsoids are displayed in a Jelinek-diagram (Jelinek 1977) for fresh and altered granites (Fig. 6a). The fresh granites show strongly oblate to neutral (T > 1) ellipsoids with high anisotropies (1.08 < P1 > 1.28). These magnetic fabrics are related to unoxidized magnetite (Fig. 4a, b), which is always associated with

517

biotite in the Soultz granite. The onset of the magnetite oxidation during the hydrothermal alteration caused a decrease in the anisotropy and a breakout of the strictly oblate geometry, so that the ellipsoid shapes became more prolate. With further increase of hydrothermal alteration the anisotropies become nearly isotropic and the ellipsoid shapes cover the whole spectrum from strongly oblate (7" ~ +1) to strongly prolate (T ~ — 1) geometries. Figure 6b shows decreasing anisotropy factors with decreasing mean susceptibility for the fresh and hydrothermally altered granites. The fresh granite with significant amounts of magnetite shows a clear linear correlation between the susceptibility and degree of anisotropy. With further decreasing susceptibility, due to hydrothermal alteration, the anisotropy becomes weak and the linear trend changes from a steep slope in fresh to a flat slope in altered granites of stage I. Samples with hydrothermal alteration of stage II scatter mainly in the field of low anisotropy, and the bulk susceptibility is mainly within the range of typical paramagnetic-dominated values. The low magnetic susceptibility values of localized stage II alteration imply that magnetite oxidation was more intense within the fault zones.

Orientation of AMS The magnetic foliations of fresh and hydrothermally altered Soultz granite can be characterized by a general change from sub-horizontal orientation in the fresh granite to steeper dipping in the altered granite of stage I, and to a random orientation in the strongly altered granites within discrete fault zones (stage II). This trend is accompanied by decreasing bulk susceptibilities (Fig. 7). In Figure 8, the orientation of the low-field magnetic fabric is presented in a stereogram subdivided into fresh and hydrothermally altered granites of stage I and II. The stage I alteration is presented for the upper (1400-1550 m), central (1550-1580 m) and lower (1580-2220 m) parts of the Soultz granite section respectively (for definition, see Fig. 2). In the fresh granite the magnetic fabric is carried by magnetite. The magnetic foliation plane (ftmin) generally shows sub-horizontal orientation with a magnetic lineation (ftmax) that is strongly scattered within the foliation plane. However, in the central and in the lower part (below 2220m) the foliations are slightly steeper and the trend is more distinct, with NW-direction and SSW-direction respectively. The sub-horizontal orientation of the magnetic

518

J. JUST ET AL.

Fig. 6. (a) Variation of the shape parameter T with the anisotropy P' (Jelinek 1977) showing oblate shapes for the fresh granite and a strong variation from oblate to prolate shapes in the hydrothermal samples, (b) Variation of Pf with mean susceptibility. Hydrothermally altered samples of stage I overlap with fresh and as vein altered samples of stage II, for further explanation see text.

foliation planes in fresh granites is in agreement with observations from Center & Traineau (1996), who evaluated magmatic foliations from primary biotite and hornblende grain alignment. The agreement of magnetic and magmatic foliation planes and the microscopically observed association of magnetite and biotite imply that magnetite mimics the magmatic flow regime in the Soultz granite. Hydrothermally altered granites of stage I reveal distinct differences in the magnetic fabric orientation within the subdivided sections (upper, central, lower part) of the Soultz granite. The general pattern of preferred orientations in the upper and central parts is similar. However, an inversion of the ttmjn and ftmax axes is observed. The magnetic foliations from the central part show a NW-SE trending orientation with varying sub-horizontal to steep dip values and sub-horizontal lineations. In the upper part

the magnetic lineations scatter within a steeply NE-SW oriented plane, while in the central part, this geometry is occupied by the foliation poles. In the lowest part, magnetic fabrics of stage I strongly resemble the pattern evaluated for the fresh granites of the same section. The foliation is generally sub-horizontal and the magnetic lineation scattered within the foliation plane. The orientation of AMS ellipsoids in altered granites of stage II within fault zones can be subdivided into cataclastic granites and their hydrothermally altered wall rocks without cataclasis. The cataclastic granites show no preferred orientations of Kmax and ftmin, even in the majority of adjacent specimens (Fig. 8). The random distribution of magnetic foliation poles and magnetic lineations is associated with decreased anisotropy, which can be correlated with the increased fracture density within fault zones.

DEVELOPMENT OF AMS DURING ALTERATION

519

Fig. 7. Mean magnetic susceptibility versus dip of magnetic foliation for fresh (black diamonds) and hydrothermally altered granite of stages I (grey diamonds) and II (open triangles). The solid line represents a changing trend from flat to steep foliation during hydrothermal alteration.

The magnetic fabrics in cataclastic granites show a clear trend of the magnetic foliations only when a limited number of phases dominate the magneto-mineralogy within well-oriented fractures. An example is shown in a localized fault zone at 1640m depth (sample 2807, Fig. 8), where fractures are filled with secondary, finegrained hematite and acicular illite (Fig. 4g). Magnetite content was almost completely oxidized, as evidenced by low susceptibility values (< 0.362 x 10"3 SI) and K (T) experiments, where no primary magnetite was observed. The quantitatively determined ferromagnetic susceptibility was assigned to hematite with 35% and to paramagnetic components with 65% (separation of para- and ferromagnetic components based on susceptibility versus temperature measurement, Hrouda et al 1997). Here, an E-W trend of magnetic foliation planes in fractured granite has been observed indicating an E-W trending fault zone. Hematite formed from magnetite (martite) was strongly fractured and partially replaced by a secondary hematite generation precipitated from fluids (Fig. 4e, f). Unlike other cataclastic samples the magnetic foliations in adjacent samples show constant strike directions, because the

secondary hematite associated with paramagnetic illite follows the fracture pattern of the EW striking fault zone. The magnetic ellipsoids are oblate (Fig. 6a) but show a lineation caused by intersection of secondary fracture planes with small angles to the main fault plane. The small anisotropy factor P7 (1.01-1.03) can be explained by hematite and illite within the secondary fracture planes, which are not perfectly aligned. In contrast to the random distribution of magnetic fabric elements within the cataclastic granites, the hydrothermally altered wall rocks reveal a well-defined magnetic fabric orientation. The foliation planes are preferentially NW-SE oriented and the magnetic lineation has a subhorizontal NW-SE trend (Fig. 8). This geometry is comparable to altered granites of stage I in the central part of the pluton.

High-field separation offerri- and paramagnetic sub fabric The orientation of low-field AMS fabrics measured with a KLY-2 Kappabridge is

520

J. JUST ET AL.

Fig. 8. Lower hemisphere, equal area projection of magnetic lineations (squares) and foliations (circles) in fresh and altered granites with stage I divided into upper part (1420-1550m), central part (1550-1580m) and lower part (> 1580m). Grey symbols and corresponding magnetic foliation planes with E-W trend in cataclastic granites of stage II represent the core sample 2807.

Fig. 9. Separation of ferrimagnetic and paramagnetic fabric by high-field AMS measurements in comparison with low-field KLY-2 AMS data of different hydro thermally altered granite of stage I. Squares represent ftmax, triangles «int and circles ftmin. (a) Weak hydrothermal alteration (sample 2476). (b) Strong hydrothermal alteration (sample 2540; sample location see Fig. 2).

DEVELOPMENT OF AMS DURING ALTERATION

compared with the subfabrics for ferri- and paramagnetic components, separated using a highfield torque magnetometer (Martin-Hernandez & Hirt 2001). This separation (Fig. 9) has been carried out for two specimens from hydrothermally altered granites of stage I, which have different magnetite content according to different alteration degrees. Sample 2476 is from a weakly altered section with bulk susceptibility of K = 9.130 x 10"3 SI and still shows the primary magmatic fabric of the granite with a sub-horizontal foliation. Sample 2540 is from a strongly altered granite with K = 0.791 x 10~3 SI and represents the altered fabric with a steeply inclined magnetic foliation. These two samples are marked in Figure 7, where a steeper inclination for the strongly altered sample is shown. The high-field separation for the weakly altered sample 2476 (Fig. 9a) shows similar orientations of para- and ferrimagnetic subfabrics with sub-horizontal magnetic foliations according to the primary magmatic geometry. The directional properties of high-field measurements indicate that 31% of the anisotropy is carried by the paramagnetic phases and 69% by the ferromagnetic phases. Both are in agreement with the low-field AMS fabric (compare also Fig. 8). In addition, the torque signal shows a dependence on the applied field (B), which indicates that hematite also contributes to the torque signal. Although this contribution can be separated from the para- and ferrimagnetic contributions, it is not large enough to compute the hematite subfabric, even in the strongly altered sample 2540. Sample 2540 shows a distinct deviation of para- and ferrimagnetic fabrics (Fig. 9b), whereby 46% of the anisotropy is carried by paramagnetic and 54% by ferrimagnetic signal. The paramagnetic fabric still shows a sub-horizontal orientation of the magnetic foliation according to the primary magmatic fabric. In contrast, the ferrimagnetic subfabric shows a sub-vertical orientation of the magnetic foliation, which is in agreement with the lowfield AMS fabric. The high-field AMS measurements clearly highlight the transformation of magnetic fabric during magnetite decomposition in the course of hydrothermal granite alteration. The lowfield AMS fabric of the altered granites of stage I is carried by the ferrimagnetic component, which reflects the preferred alignment of the magnetite relics. The separated paramagnetic fabric is clearly different and is carried by several paramagnetic components mentioned above. The paramagnetic fabric does not contribute to the low-field anisotropy alongside ferromagnetic magnetite relics.

521

Discussion Primary magnetic fabric in the Soultz-sousForets granite AMS fabric studies can be used to define the internal structure of granitoid rocks (e.g. Bouchez et al 1990; Archanjo el al 1995). In the undeformed state the magmatic fabric reflects the preferred orientation of grains with crystalline and/or shape anisotropy (e.g. Tarling & Hrouda 1993). The fresh Soultz monzogranite has a high-K calc-alkaline composition and high magnetic susceptibilities typical for granites of the 'magnetite series' defined by Ishihara (1977) and characteristic for /-type granites. Together with other clear subduction-related /-type intrusions from the adjacent Northern Vosges, Windstein and Kaiserbach, these rocks indicate a magmatic arc setting in this area (Flottmann & Oncken 1992). The ferrimagnetic phase is shown in the Soultz granite to be pure, multidomain magnetite, which is the main carrier of the AMS. In advanced hydrothermally altered granites the grain sizes of magnetite were reduced. In the studied 700m granite section of the EPS-1 borehole, the AMS ellipsoids of fresh granites reveal a general anisotropy factor P' in the range from 1.08 to 1.28, which is within typical values for flow fabrics observed in other granitic magnetite-series (e.g. Ferre et al. 1999; Greiling & Verma 2001). AMS ellipsoids are all oblate but have variable degrees of oblateness (Fig. 6a). The high degree of oblateness is probably due to magnetite forming clusters associated with biotite. The clear positive linear correlation between anisotropy factor and magnetic susceptibility in Figure 6b implicates increased magnetite content. The higher magnetite content allows for better mimicking of biotite crystal and the resulting shape of the magnetite cluster reflects more oblate ellipsoid shapes. Therefore, the strongly oblate shapes (T < 0.7) in fresh granites resemble AMS ellipsoids of single biotite crystals (Martin-Hernandez & Hirt 2003). The authors defined a highly oblate shape for biotite (0.75 < T < 0.99) and a consistent paramagnetic ellipsoid orientation with the crystallographic structure of the phyllosilicate. In the Soultz granite, only a weak magmatic foliation is observed macroscopically and this reflects the alignment of biotite (Center & Traineau 1996). The low-field AMS measurements in the fresh granites define a strong foliation since the fabric is carried by magnetite as the amplification factor. The magnetic foliations show relatively flat dips and a flat random

522

J. JUST ET AL.

distribution of magnetic lineation (ftmax). This flat, oblate geometry is consistent throughout the fresh granite sections (700m), and is therefore regarded as the significant and primary magmatic fabric of the Soultz granite. However, the orientations of the AMS fabrics in other Variscan late-orogenic granite intrusions differ from those in the Soultz granite. In the Mid-European Variscides, e.g. in the Odenwald magmatic complex (Flottmann & Oncken 1992; Greiling & Verma 2001; Dietl & Stein 2001) and in the Oberpfalz at the western border of the Bohemian Massif (Leonhard et al. 2001), the observed magmatic fabrics were related to the orientation of NW-SE shortening Variscan tectonic stresses (Ziegler 1986). In the primary magmatic fabric of the fresh Soultz granite, such a relation could not be detected. The distinct flat and planar orientation of magnetic fabrics in the studied granite may indicate a magma emplacement in the centre of a batholith with a deceleration flow regime and horizontal spread out of magma. Deceleration flow occurs where channelized magma spreads out into a broader region and thus slows down, resulting in the flattening or oblate strains with high angles to the flow direction (Cruden 1990). This distinct magmatic fabric would imply that the 331 ± 9 Ma Soultz granite (Alexandrov et al. 2001) was not affected in the same way as the adjacent intrusions, which were influenced by the transtensional-transpressional stress regime during the late Variscan orogeny (Ziegler 1986).

Magnetic fabric in hydrothermally altered granite In the Soultz granite, the two hydrothermal alteration stages (pervasive stage I and localized stage II) appear in the strongly altered upper part and in the lower part within fault zones, while the central part is only characterized by stage I alteration. The reduction of magnetic susceptibility due to the transformation of magnetite into hematite affects the progressive transformation of magnetic fabrics in both stage I and II alterations. With increasing alteration, the bulk magnetic susceptibility is reduced as well as the anisotropy factor Pf < 1.1, typical for paramagnetic granitoids (Bouchez et al. 1987; Ferre & Ameglio 2000). This study shows that magnetite relics in martite strongly influence the rock magnetic behaviour during the progressive hydrothermal alteration. This influence ceases with the strongest degree of hydrothermal alteration when nearly all magnetite is oxidized.

The decreased magnetic susceptibility and anisotropy in the hydrothermally altered granites of stage I and II can be related to the decreased magnetite content as a consequence of oxidation and associated grain-size reduction. The formation of hematite parallel to (111) long axis of magnetite caused an increased distance between magnetite domains, and therefore the magnetic interactions weakened (Stephenson 1994). With advanced oxidation magnetite disappears and the contribution of paramagnetic minerals to the AMS dominates. The fresh Soultz granite is composed of 40vol% plagioclase, 19vol% of paramagnetic, hematitized A^-feldspar and 8vol% of biotite + chlorite (Jacquemont 2002). Along with hydrothermal alteration (stage I and II) paramagnetic illite pseudomorphed after plagioclase and therefore mimics the plagioclase alignment. In contrast to platy biotite associated with magnetite, feldspars are weakly oriented into the magmatic fabric. Thus, anisotropy is strongly decreased in altered granites without magnetite content. Hydrothermal alteration without cataclasis causes no changes of structural fabric because the paramagnetic alteration minerals mimic the lattice preferred orientation of primary minerals (Fig. 4d). The most frequent alteration mineral is acicular and fine-grained illite (~10|im), which is often accompanied by fine-grained, antiferromagnetic hematite from a second population (Fig. 4f, g). Variable contributions of ferri-, antiferro- and paramagnetic components on the AMS cause a change from primary, strictly oblate AMS ellipsoids and result in more oblate as well as prolate AMS ellipsoids in the hydrothermally altered granites. Along with this transformation of AMS ellipsoid shape, the orientation of primary magnetic foliation poles in the fresh granites evolves from a horizontal cluster distribution to a general, secondary girdle distribution with NE-SW trending great circle in the altered granites. At the same time, the primary scattered distribution of the magnetic lineation changes to a well-defined sub-horizontally NW-SE trending distribution pattern (Fig. 10). Based on the alteration mineralogy within the alteration zone of the central part of the pluton (1550-1580m), we assign this characteristic fabric development to stage I alteration according to Traineau et al. (1991). In cataclastic-deformed granites, the above mentioned AMS fabric associated with magnetite is destroyed during the stage II alteration. The development of fault zones is characterized by almost randomly oriented micro-cracks causing randomly oriented magnetite relics. The previous preferred orientation of principal axes

DEVELOPMENT OF AMS DURING ALTERATION

523

Fig. 10. Development of magnetic fabric during progressive hydrothermal alteration with typical magnetic mineralogy and susceptibility (x 10~3 SI). The primary magnetic fabric with sub-horizontal magnetic foliation changes to tectonic fabric with steepened NW-SE trends. Intensive cataclasis causes randomly distributed fabrics. Dotted areas (squares) represent magnetic lineations and striped areas (circles) magnetic foliation poles. Further explanations see text.

of the AMS changes into a random distribution of both foliation poles and magnetic lineations (Fig. 10). In contrast, hydrothermalized wall rocks of stage II without cataclasis reveal a pattern of preferred orientation corresponding to the stage I alteration (Fig. 8), although the alteration mineralogy can be assigned to the stage II alteration. A special development during the stage I alteration is observed in the upper part of the granite section (1420-1550 m). Here, the altered rocks (stage 1 alteration) show an inverse magnetic fabric with inverted orientation of magnetic foliation as compared with the central part of the

pluton (Fig. 8). This behaviour may indicate an inverse magnetic fabric due to the very small magnetite relics in hematite, which behave as single domain grains and show the lowest magnetic susceptibility for the stage I alteration (Fig. 3). Inverse magnetic fabrics related to single domain magnetite have been described by Potter & Stephenson (1988), Rochette et al (1999) and Ferre (2002). In single domain magnetite, the easy magnetic axis fixes the direction of spontaneous magnetization and thus results in zero susceptibility along the long axis. The evolution from primary to altered magnetic fabric can be discussed in the context of

524

J. JUST ET AL.

micro structural observations assigned to alteration stages I and II. Two different processes are considered to act as controlling factors during the fabric transformation: these are a microstructural and a stress-controlled alteration. Micro structural and fluid inclusion studies in the Soultz granite were done by Schild et al. (1998), who analysed three successive, orthogonal micro-crack directions in healed quartz and feldspars. One of the micro-crack directions is in agreement with the NW-SE trend of ftmax in the altered granites and is often accompanied by co-oriented micro-cracks in magnetite (Fig. 4a). We assume that a preferred micro-crack formation in magnetite with dominantly NW-SE directions was sub-parallel to one of the four (111) crystallographic axis of magnetite with the smallest angle to the micro-cracks. These planes acted as weakened zones that favoured fluid circulation in magnetite. Magnetite oxidation started along these weak zones and caused elongated and NW-SE oriented magnetite relics. Hydrothermalized wall rocks of the later stage II alteration show magnetic fabric, which corresponds to the geometry of stage I alteration (Fig. 8). The alteration minerals of stage II pseudomorphed after the available mineralogy, as already observed for stage I. This is also the case if the granite was already altered during the former stage I alteration. We assume, therefore, that the ubiquitous NW-SE trending fabrics in the hydrothermally altered granite of stage II are due to the same tectonic setting that was responsible for the micro-crack formation within magnetite during the late magma tic stage. Conclusions The magnetic fabrics reflect the primary magmatic flow fabric in the fresh Soultz granite, which had been altered during hydrothermal alteration stages I and II (Fig. 10). In the fresh granite, MD and subeuhedral magnetite were identified as the main magnetic carrier of the AMS. Strictly oblate fabrics and enhanced anisotropies were found, which are related to magnetite grains forming clusters. These magnetite clusters mimic the alignment of biotite within the magmatic foliation plane. The sub-horizontal magmatic foliation indicates a magma emplacement due to either a decelerating flow regime, which provokes stronger internal than external tectonic stresses, or internal vertical stresses during the two-stage crystallization of the magma. Therefore, the primary magmatic fabrics of the Soultz granite seem not to have been affected by the NW-SE wrench tectonics during

the Variscan orogeny, in contrast to adjacent intrusions of similar age at the eastern and western flanks of the Rhine Graben. With further cooling and solidification, the external stress conditions became stronger relative to the internal stress. Now, the solid body could absorb the external stresses (c^ = NW-SE), which caused micro-crack formation in magnetite with preferred NW-SE orientation. The micro-crack formation in magnetite acted as the controlling factor for the development of the alteration fabric during the subsequent hydrothermal alteration stages I and II. During these hydrothermal alteration stages magnetite was oxidized to hematite, and illite as well as Fe-carbonates replacing the rock forming minerals became the most important paramagnetic minerals. Therefore, the orientation of steep micro-cracks was responsible for the steepening of the magnetic foliation. These cracks allowed for circulation of residual fluids during the magma cooling, which caused the pervasive alteration (stage I alteration) of the pluton. The later alteration stage II retained this trend, which is well documented in the hydrothermalized wall rocks within fault zones. This trend is only destroyed by cataclastic deformation, which results in random orientation of the principal axes of the anisotropy ellipsoid. This faulting proceeded during the multiphase formation of the Upper Rhine Graben with changing stress directions and resulted in chaotic micro-crack orientations. It could be shown that magnetite controls the magnetic fabric during the hydrothermal alteration, until it completely oxidizes to hematite. In rocks where the magnetite was completely oxidized, the AMS is controlled by paramagnetic minerals, which have a low degree of anisotropy. In such rocks, hematite formation during the hydrothermal alteration stages is not high enough to dominate the AMS fabric. Financial support was provided by the DFG-Graduiertenkolleg 273 (Fluid-Rock Interaction) to the first author. We are grateful to J. Baumgartner from Socomine for enabling sampling of the EPS-1 cores. Furthermore, we want to thank the Soultz-group at Heidelberg, especially L. Warr, A. Schleicher, B. Kober and D. Aubert for their co-operation and constructive discussions. The constructive reviews of F. Hrouda and M. Hounslow are gratefully acknowledged. This is also an ETH publication no. 1319.

References ALEXANDROV, P., ROYER, J.-J. & DELOULE, E. 2001. 331 ± 9 Ma emplacement age of the Soultz monzogranite (Rhine Graben basement) by U/Pb

DEVELOPMENT OF AMS DURING ALTERATION ion-probe zircon dating of samples from 5km depth. Earth and Planetary Sciences, 332, 747-754. ARCHANJO, C. J., LAUNEAU, P. & BOUCHEZ, J. L. 1995. Magnetic fabric vs. Magnetite and biotite shape fabrics of the magnetite-bearing granite pluton of Gameleiras (Northeast Brazil). Physics of the Earth and Planetary Interiors, 89, 63-75. BOUCHEZ, J. L., BERNIER, S., ROCHETTE, P. & GUINEBERTEAU, B. 1987. Log des susceptibilites magnetiques et anisotropies de susceptibilites dans le granite de Beauvoir: consequences pour sa mise an place. Geologic de la France, 2-3, 223-232. BOUCHEZ, J. L., GLEIZES, G., DJOUADI, T. & ROCHETTE, P. 1990. Microstructure and magnetic susceptibility applied to emplacement kinematics of granites: the example of the Foix pluton (French Pyrenees). Tectonophysics, 184, 157-171. CRUDEN, A. R. 1990. Flow and fabric development during the diapiric rise of magma. Journal of Geology, 98, 681-698. DIETL, C. & STEIN, E. 2001. The diapiric emplacement and related magmatic fabrics of the porphyritic Ludwigshohe granite, Central Odenwald (Germany). Mineralogy and Petrology, 72, 145-164. DUBOIS, M., AYT OUGOUGDAL, M., MEERE, P., ROYER, J.-J., BOIRON, M.-CH. & CATHELINEAU, M. 1996. Temperature of paleo- to modern self-sealing within a continental rift basin: The fluid inclusion data (Soultz-sous-Forets, Rhine graben, France). European Journal of Mineralogy, 8, 1065-1080. DUNLOP, D. & OZDEMIR, 6. 1997. Rock Magnetism— Fundamentals and frontiers. Cambridge University Press, Cambridge. FERRE, E. 2002. Theoretical models of intermediate and inverse AMS fabrics. Geophysical Research Letters, 29(7), 31/1-31/4. FERRE, E. C. & AMEGLIO, L. 2000. Preserved magnetic fabrics vs. Annealed microstructures in the syntectonic recrystallised George granite, South Africa. Journal of Structural Geology, 22, 11991219. FERRE, E. C., WILSON, J. & GLEIZES, G. 1999. Magnetic susceptibility and AMS of the Bushveld alkaline granites, South Africa. Tectonophysics, 307, 113133. FLOTTMANN, T. & ONCKEN, O. 1992. Contraints on the evolution of the Mid Crystalline Rise—a study of outcrops west of the river Rhine. Geologische Rundschau, 81(2), 515-543. GENTER, A. & TRAINEAU, H. 1991. Geological survey of the HDR borehole EPS1, Soultz-sous-Forets, Alsace, France. BRGM, R 32433. GENTER, A. & TRAINEAU, H. 1996. Analysis of macroscopic fractures in granite in the HDR geothermal well EPS-1, Soultz-sous-Forets, France. Journal of Volcanology and Geothermal Research, 72, 121-141. GREILING, R. O. & VERMA, P. K. 2001. Strike-slip tectonics and granitoid emplacement: an AMS fabric study from the Odenwald Crystalline Complex, SW Germany. Mineralogy and Petrology, 12, 165-184. HROUDA, F. 1994. A technique for the measurement of thermal changes of magnetic susceptibility of weakly magnetic rocks by the CS-2 apparatus

525

and KLY-2 Kappabridge. Geophysical Journal International, 118(3), 604-612. HROUDA, F., JELINEK, V. & ZAPLETAL, K. 1997. Refined technique for susceptibility resolution into ferromagnetic and paramagnetic components based on susceptibility temperature-variation measurement. Geophysical Journal International, 129(3), 715-719. HROUDA, F., Puns, M. & MADARAS, J. 2002. The Alpine overprints of the magnetic fabrics in the basement and cover rocks of the Veporic Unit (Western Carpathians, Slovakia). Tectonophysics, 359(3-4), 271-288. ISHIHARA, S. 1977. The magnetite series and ilmenite series granitic rocks. Mining Geology, 27, 293-305. JACQUEMONT, B. 2002. Etude des interactions eauxroches dans le granite de Soultz-sous-Forets. Quantification et moderation des transferts de matiere par les fluides. Ecole et Observatoire des Sciences de la Terre, Centre de Geochemie de la France (UMR 7517). JELINEK, V. 1977. Statistical processing of magnetic susceptibility measured in groups of specimens. Studia Geophysica et Geodaetica, 22, 50-62. KLETETSCHKA, G & WASILEWSKI, P. 2001. Grain size limit for SD hematite. Physics of Earth and Planetary Interiors, 129(1-2), 173-179. KONTNY, A. & DE WALL, H. 2000. Case studies on the use of temperature-dependent susceptibility for the characterisation of magneto-mineralogical changes during metamorphism. Physics and Chemistry of the Earth, A25(5), 421-429. LEONHARD, W., DE WALL H., STEIN, E. & GREILING, R. O. 2001. AMS-Messungen am Leuchtenberger Granit/Oberpfalz.- Exhursionsfuhrer und Verojfentlichungen der GGW, 212, 64-66. LOWRIE, W. 1990. Identification of ferromagnetic minerals in a rock by coercivity and unblocking temperature properties. Geophysical Research Letters, 17(2), 159-162. MARTIN-HERNANDEZ, F. & HIRT, A. M. 2001. Separation of ferrimagnetic and paramagnetic anisotropies using a high-field torsion magnetometer. Tectonophysics, 337, 209-221. MARTIN-HERNANDEZ, F. & HIRT, A. M. 2003. The anisotropy of magnetic susceptibility in biotite, muscovite and chlorite single crystals. Tectonophysics, 367, 13-28. MUXWORTHY, A. R. & MCCLELLAND, E. 2000. Review of the low-temperature magnetic properties of magnetite from a rock magnetic perspective. Geophysical Journal International, 140, 101-114. NAKAMURA, N. & NAGAHAMA, H. 2001. Changes in magnetic and fractal properties of fractured granites near the Nojima Fault, Japan. The Island Arc, 10,486^194. POTTER, D. & STEPHENSON, A. 1988. Single-domain particles in rocks and magnetic fabric analysis. Geophysical Research Letters, 15(10), 1097-1100. ROCHETTE, P., AUBOURG, C. & PERRIN, M. 1999. Is this magnetic fabric normal? A review and case studies in volcanics formations. Tectonophysics, 307, 219234. RUMMEL, F. & KONIG, E. 1991. Physical properties of core samples, borehole EPS-1, Soultz-sous-Forets,

J. JUST ET AL.

526

velocity, density and magnetic susceptibility-logs, depth interval 933-2227 m. Internet Bericht, RuhrUniversitat Bochum. SCHILD, M., VOLLBRECHT, A., SlEGESMUND,

S.

&

REUTEL, C. 1998. Microcracks in granite cores from the EPS-1 geothermal borehole, Soultzsous-Forets (France): paleostress directions, paleofluids and crack-related Vp-anisotropies. Geologische Rundschau, 86, 775-785. SCHUMACHER, M. E. 2002. Upper Rhine Graben: Role of preexisting structures during rift evolution. Tectonics, 21(1), 6. STEPHENSON, A. 1994. Distribution anisotropy: two simple models for magnetic lineation and foliation. Physics of the Earth and Planetary Interiors, 82, 49-53. STUSSI, L. M., CHEILLETZ, A., ROYER, J. J., CHEVREMONT, P. & FERAUD, G. 2002. The hidden monzogranite of Soultz-sous-Forets (Rhine Graben, France). Mineralogy, petrology and genesis. Geologic de la France, 45-64.

TARLING, D. H. & HROUDA, F. 1993. The Magnetic Anisotropy of Rocks. Chapman & Hall, London. TRAINEAU, H., CENTER, A., CAUTRU, J. P., FABRIOL, H. & CHENREMONT, P. 1991. Petrography of the granite massif from drill cutting analysis and well log interpretation in the geothermal HDR borehole GPK1 (Soultz, Alsace, France). Geothermal Science and Technology, 3(1-4), 1-29. VAN WEES, J.-D., McCANN, T., DADLEZ, R., GAUPP, R., NARKIEWICZ, M., BITZER, F. & SCHECK, M. 2000. On the origin of the Southern Permian Basin, central Europe. Marine and Petroleum Geology, 17, 43-59. ZIEGLER, P. A. 1986. Geodynamic model for the Paleozoic Crustal Consolidation of Western and Central Europe. Tectonophysics, 126, 303-328. ZIEGLER, P. A. 1996. Geodynamic processes governing development of rifted basins. In: ROURE, F., Elluoz, N., Shein, V. S. & Sidorov, V. A. (eds) Geodynamic Evolution of Sedimentary Basins. Ed. TECHNIP, Paris, 19-67.

Sub-fabric identification by standardization of AMS: an example of inferred neotectonic structures from Cyprus THOMAS D. HAMILTON,1 GRAHAM J. BORRADAILE1 & FRANCE LAGROIX2 1

Geology Department, Lakehead University, Thunder Bay, Ontario P7B 5El, Canada (e-mail: [email protected], [email protected]) 2 Institute for Rock Magnetism, University of Minnesota, 100 Union Street S.E., Minneapolis, MN 55455-01, USA Abstract: Calcite petrofabrics are sensitive to weak strains, possibly being the most sensitive classical petrofabric indicator. Thus, calcareous sediments may reveal stress trajectories in neotectonic environments. Calcite aligns by crystal-plastic deformation and pressure solution produce corresponding alignments in accessory clay minerals and magnetite (possibly fossilbacterial). Their alignments are rapidly and precisely detected by anisotropy of low field magnetic susceptibility (AMS) with net magnetic fabrics, which blend diamagnetic contributions from matrix calcite (diamagnetic bulk susceptibility K ~ -14uSI), accessory clay minerals (« = 100 to 500 uSI) and sometimes trace magnetite (K > 2 SI). Their relative abundances and different anisotropies must be considered in interpreting AMS orientations, nevertheless our study reveals orientation distributions of AMS axes in sub-areas and regions that are sensibly interpreted as palaeostress trajectories in Neogene and Quaternary strata. The AMS axes may be correlated with the orientation of faults, plate-motion vectors and seismic solutions. Large samples (1090 specimens from 419 sites) are treated by different statistical approaches ('standardization') to emphasize or suppress the contribution of subfabrics with anomalous mean susceptibility. A sub-sample of 254 specimens from 219 sites, from different sub-areas was investigated by anisotropy of anhysteretic remanence (AARM), which isolates the orientation distributions of magnetite. Magnetic fabrics are mostly of the L-S kind with the magnetic lineations compatible with gravitational stretching of the sedimentary cover away from the Troodos massif and orthogonal to the principal faults and graben. The L-direction (fcmax) shows a smooth variation in orientation, through the sub-areas, directed radially from the Troodos massif and the S-components of the magnetic fabrics are inclined gently to the bedding, compatible with vergence toward the Cyprean Arc to the S and SW of Cyprus.

Magnetic fabric analysis, most commonly using anisotropy of low-field magnetic susceptibility (AMS), is now well established as a non-destructive technique to isolate the mean orientationdistribution of crystals in an anisotropic rock. AMS blends contributions from different minerals that have different magnetic responses, i.e. paramagnetic, diamagnetic or 'ferro'magnetic (Rochette et al. 1992). In turn, the principal AMS axes may proxy for the principal axes of the orientation distribution ellipsoid of crystals (Henry 1989; Borradaile 2001) which, in turn, may reveal the finite strain axes in technically deformed rock (Tarling & Hrouda 1993; Borradaile & Henry 1997). In deformed calcite matrices, interpretation is somewhat complicated by an intrinsic counterintuitive arrangement of crystallographic and AMS axes, a socalled 'inverse fabric' (Rochette 1988), which nevertheless still permits sensible interpretations of their finite strain axes (Ihmle et al. 1989; de Wall et al. 2000).

However, recent studies show that in neotectonic environments, with imperceptible penetrative strain, AMS axial orientations may be sensibly related to the orientations of joints or faults, for which stress axes are reliably inferred (Sagnotti et al. 1994, 1998; Mattel et al. 1999; Cifelli et al. 2004). In some instances, AMS axes may correlate with modern seismic solutions and plate movements (Kissel et al. 1986; Mattel et al. 1999; Borradaile & Hamilton 2004). Thus, AMS may proxy for stress trajectories under limited circumstances. Our goal is to examine a large sample of AMS and AARM data from post-Palaeogene strata in Cyprus that correlate with the orientations of neotectonic structures or events (faults, rifts, Tertiary uplift). It will be shown that their orientations are consistent with stress trajectories inferred from those structures and recent plate motion trajectories. Sampling at two density levels verifies the homogeneity of the regional domains and the validity of the regional conclusions. Local

From: MARTIN-HERNANDEZ, F., LUNEBURG, C. M., AUBOURG, C. & JACKSON, M. (eds) 2004. Magnetic Fabric: Methods and Applications. Geological Society, London, Special Publications, 238, 527-540. 0305-8719/04/ $15.00 © The Geological Society of London 2004.

528

T. D. HAMILTON ET AL.

sub-areas (I-VI) were structurally homogenous and sampled with as little lithological variation as possible over areas <30km 2 . The regionalscale domains (I-V) have an average area of -400 km2. Tectonic background Our study concentrates on the limestone and marl sedimentary cover to the Cretaceous Troodos ophiolite. We exclude from this discussion the juxtaposed older allochthonous Mamonia and Kyrenia terranes, which include exotic formations as old as Triassic and perhaps even Carboniferous (Robertson 1990) (Fig. la). The Troodos terrane exposes an integral ophiolite sequence with mantle harzburgite and Iherzolite at the base overlain by layered gabbros and dunite, sheeted dykes, a transitional sheeted-dyke to pillowed unit and a thick pillow basalt sequence (Malpas et al 1990). Supra-Troodos sedimentation commenced in the Maastrichtian (~68Ma) with a pelagic chalk blanket, to form a relatively continuous sequence to the present (Lord et al. 2000). Northwards subduction stalled in the Eocene, with southward thrusting of the Kyrenia Range. Subduction then retreated southwards to its present location off the south shore of Cyprus in the Miocene, initiating the present stress regime (Fig. Ib). Miocene extensional basins, most prominently the Polis graben, formed in response to changing stress trajectories during the retreat of the subduction zone's hinge away from the Troodos micropiate. Subduction continued during the Pliocene, until the Eratosthenes Seamount reached the trench in the Pleistocene when serpentitization-driven diapirism continued to uplift the Troodos dome (Robertson 2000). These events drastically changed the neotectonic regime in the last 5 Ma, which is evident from studies of structure (Robertson et al. 1995), earthquakes (Ben-Avraham et al. 1988; Arvidsson et al. 1998; Papazachos & Papaioannou 1999; Borradaile & Hamilton 2004) and global plate motion (Reilinger et al. 1997). The tectonic events, their principal plategeometrical consequences and the reasonably inferred principal tensile or compressive stress trajectories are summarized in Figure 2. Our previous studies indicated the potential for AMS (anisotropy of low field magnetic susceptibility) to indicate very weak strains and weak orientation distributions of minerals, like those accompanying calcite-crystal plastic mechanisms in neotectonic environments. AMS shows neotectonic potential in the limestone cover from the Palaeogene Lefkara Formation through the Neogene Pakhna Formation to the

localized Pleistocene cover. Without regard to detailed stratigraphic level within the Troodos Cover sequence, Lagroix and Borradaile (2000) showed that the AMS principal axes correspond to kinematic patterns. fcmax defines down-slope gravitational stretching away from the Troodos dome and locally into tectonic/depositional basins. In the Polis Rift Valley, AMS and AARM (anisotropy of anhysteretic remanent magnetization) axes correlate with principal stress trajectories inferred from Miocene and younger fault-orientations and by modern seismic fault-plane solutions (Borradaile & Hamilton 2004, Fig. 6). Neotectonic interpretation potential of AMS? The presence of consistently oriented AMS axes in young sedimentary rocks that are not related to depositional fabrics, itself argues for some subtle tectonic imprint (e.g. Sagnotti et al. 1994). In such environments, AMS axes may be oriented consistently and simply with respect to faults or joints. These structures have orientations uniquely associated with their cogentic stress trajectories and thus by induction, the AMS axes may proxy as stress trajectories. Of course, AMS axes may proxy for stress trajectories in ancient rocks too, but their recognition requires fortunate circumstances (e.g. Borradaile & Kehlenbeck 1996). Usually, in ancient or severely deformed rocks, the association of AMS with finite strain axes with earlier tectonic events will mask any young and feeble AMS overprint associated with late stress increments. Note that we use Flinrfs L-S scheme (1965) equally to describe tensor magnitude ellipsoid shapes, whether they are for strain, AMS, stress or mineral ODs (petrofabrics). Calcite twinning and other crystal-plastic deformation have long been used in petrofabrics as sensitive indicators of incremental strain axes, which are essentially parallel to the causative palaeostress axes. Indeed, the association of calcite petrofabrics, their causative stress axial orientations and magnetic fabrics have been shown in some other studies (Owens & Bamford 1976; Owens & Rutter 1978; Jackson et al. 1989; Borradaile et al. 1989). The Neogene and Quaternary limestone and marl that covers the Troodos microplate shows evidence of weak to moderate strain, expressed by calcite twinning and, rarely, a feeble stylolitic cleavage. Calcite, accessory clay and magnetite traces are aligned and readily detectable in the AMS signal (Lagroix & Borradaile 2000). These minerals

SUB-FABRIC IDENTIFICATION BY STANDARDIZATION OF AMS

529

Fig. 1. Geology of Cyprus, (a) Major lithological units of Cyprus (Geological Survey Department, Cyprus, 1979). (b) Major Tectonic features of Cyprus (Robertson 1990; Arvidsson et al 1998; Borradaile & Hamilton 2004).

530

T. D. HAMILTON ET AL.

Fig. 2. Simplified stratigraphic and tectonic events for the Troodos terrane, and its sedimentary cover rocks.

contribute quite differently to AMS and to bulk susceptibility. Calcite is diamagnetic (usually quoted as K ~ — 14uSI) but it has a large anisotropy and comprises >97% of the rocks' volumes. Its crystal-plastic deformation aligns the paramagnetic accessory clay minerals (100500|iSI, <3% by rock volume) and the scarce traces of magnetite. However, the latter may contribute significantly to AMS due to magnetite's high bulk susceptibility (~ 1.0-2.7 SI) (Hunt et al 1995); and the grains may align either as overgrowths or inclusions associated with clay grains or as independent grains. Magnetite grains may be of clastic or bacterial origins; their hysteresis properties are compatible with the latter hypothesis (Borradaile & Lagroix 2000; Borradaile & Hamilton 2003). The specimen's bulk susceptibilities

(« = \/(fcmaxfcintfcmin)) are a first guide to the

mineralogical controls on AMS. We present these for rather homogenous sub-areas (Fig. 3a), which justifies the subsequent interpretation of larger sampling schemes in regional domains (I-V) (Fig. 3b). The positive susceptibility sections of the histograms are scaled logarithmically and, to a first approximation, it appears that the weakly positive K specimens have susceptibilities with a lognormal distribution, with most subareas and regional domains having modal K in

the range +15 uSI < K < +35 j^SI. Only very low concentrations of accessory clay (<1000|iSI) and trace magnetite (~2.0SI) are required to raise the specimens' K from the level of a pure calcite diamagnetic matrix (assumed to be ~ -14uSI) to the levels of the positive K specimens shown in Figure 3 a, b. In the complete absence of magnetite, only 3% by volume of clay would be required to justify the positive K values. Many specimens are truly diamagnetic and the linear scale for the diamagnetic part of the histogram clarifies the frequency distributions. For calcite K is commonly quoted at ~ — 14jiSI (Voight & Kinoshita 1907) or -7.5 to -39^iSI. (Hunt et al. 1995). Whereas there are fairly modern high precision torque measurements for anisotropy (i.e. &max - A;min; Hellwege & Hellwege 1967; Owens & Rutter 1978) there are apparently no recent high-precision measurements for the bulk value («), which hinders modelling and interpretation of AMS. Values for synthetic calcite would be preferable for modelling work (e.g. in Borradaile 1988) whereas measurements from natural calcite are contaminated by non-diamagnetic impurities. Our instrument has a low-drift environment (<0.05jiSI) and is calibrated with MnO2 (~1654 jaSI). Due to the small range of diamagnetic susceptibilities there may be some slight bin boundary bias in that part of the histogram.

SUB-FABRIC IDENTIFICATION BY STANDARDIZATION OF AMS

531

Fig. 3. Frequency distribution of bulk magnetic susceptibility (K) for (a) sub-areas (~30 km ) and (b) regional domains (~400km ). Data in the shaded box on the left represent diamagnetic specimens and their scale is linear. Data with K > 0 represent paramagnetic specimens and are shown with a logarithmic scale.

The susceptibility frequency-distribution indicates that paramagnetic limestone is more common and it confirms that the rocks' susceptibilities, whether paramagnetic or diamagnetic are mostly bimineralic in origin. Interpreting AMS axes in such rocks necessitates the evaluation of the competition between weak paramagnetism and weak diamagnetism in the same specimens, and between weakly dia/paramagnetic specimens within a sample of several different specimens, for example from a sub-area.

A new plot of fabric anisotropy parameters, introduced by Borradaile & Jackson (this volume), simplifies the interpretation in this context. Jelinek's (1981) anisotropy parameters (Pj9 T) are usually presented on Cartesian axes. Consequently, for weak degree of anisotropy (low-eccentricity ellipsoids with P, ~ 1.0) the difference between T-values that describe ellipsoid shape is exaggerated. In the extreme case, the sphere (isotropic case, Pj = 1.0), which should plot at one point on the diagram actually

532

T. D. HAMILTON ET AL.

spread along the entire axis from T = +\ (prolate) to T = — 1 (oblate). This biases the presentation and interpretation of all low-P7 anisotropies that are essentially near-isotropic, a matter of considerable relevance here where specimens straddle the diamagnetic-paramagnetic boundary. The new polar plot presents Pj radially and T along arcs; thus, all spheres plot

uniquely at the origin and the difference in shape between weak anisotropies is not exaggerated. Moreover, the plot facilitates the simultaneous presentation and comparison of positive and diamagnetic susceptibilities (Fig. 4). When working with anisotropies of diamagnetic rocks and weakly paramagnetic rocks it is also important not to overlook some non-trivial issues that

Fig. 4. Polar plots (see Borradaile & Jackson, this volume) of four select sub-areas (a)-(d) and two select regional domains (e, f). Shaded area represents diamagnetic specimens and white area represents paramagnetic specimens.

SUB-FABRIC IDENTIFICATION BY STANDARDIZATION OF AMS

software must manage carefully: (1) Anisotropy is indeterminate for a specimen if some axes have positive susceptibility and others have negative susceptibility (Borradaile 2003). This occurs in some limestone specimens. (2) &max is obviously the largest absolute value for a paramagnetic material but for a diamagnetic material the most negative value represents the long axis of the magnitude ellipsoid. (3) Regardless of the convention for defining the ellipsoid in (2) above, the common diamagnetic minerals, quartz and calcite have the idiosyncrasy that under most metamorphic conditions their c axes tend to align parallel to the shortening axis (but see Borradaile & Jackson, this volume). Thus, c axes are parallel to the most negative susceptibility, which may be perpendicular to the rock's ^-fabric. Calcite exhibits a strong prolate diamagnetic anisotropy with (&max - kmin) estimated at 1.39jiSI (Krishnan et al. 1933; Hellwege & Hellwege 1967, p. 143) or 1.172 ± 0.028 uSI (Owens & Rutter 1978). If we accept the much less certain but commonly quoted bulk susceptibility K ~ — 14uSI, these precise anisotropy determinations indicate a large anisotropy of -10%. (4) Since most deformation mechanisms align calcite with its c axis (most negative susceptibility) parallel to shortening it produces an 'inverse fabric' (Rochette 1988; Ihmle et al. 1989). Examples of the anisotropies of both local-scale samples (Fig. 4a-d) and regional domains (Fig. 4e, f) indicate that the anisotropy degree (Pj) is similar for diamagnetic and paramagnetic limestones and that ellipsoid shape tends slightly toward the oblate case (7* > 0) for both diamagnetic and paramagnetic specimens. Anisotropies are remarkably similar for the diamagnetic and paramagnetic fields and diamagnetic limestones are only absent in marl Formations in the Omodhos and Lefkara sub-areas (Fig. 4c, d). Interpreting AMS AMS orientations and magnitudes would reflect the orientation distribution of a monomineralic rock with a unique petrofabric. However, for rocks with multiple minerals with similar contributions to the net susceptibility or with multiple subfabrics some care is required in evaluating the influence of mineral abundances and subfabrics

533

(Borradaile 1988; Borradaile & Henry 1997). For rocks of high mean susceptibility (K), AMS axes may be controlled by one or two or a few minerals (Henry 1983, 1990, 1992; Borradaile et al. 1986; Borradaile & Kehlenbeck 1996; Borradaile & Lagroix 2001; Nakamura & Borradaile 2001). In our limestones however, sub-equal competition from the diamagnetic matrix and paramagnetic clay complicate the issue. To illustrate the argument with hypothetical but not unrealistic values chosen for arithmetic simplicity, disperse a concentration of 5% clay (assume ft = 266|iSI) in a calcite matrix (assume tt = -14jiSI). The limestone's net susceptibility is approximately zero and the specimen may also be isotropic. It is for this reason that some AMS data from limestone must be examined carefully because the AMS directions are simply too unstable; their anisotropy is too low to define stable orientations of principal axes as itemized in the list above (Borradaile & Jackson, this volume). In worse cases, the principal susceptibilities may not all be of the same sign so that the AMS tensor may not be represented by any magnitude ellipsoid. Although it is now realized that the magnitude ellipsoids of individual specimen tensors have little kinematic significance due to the complex blend of multiple mineralogical responses, specimen magnitudes do affect the calculation of the orientation of the mean tensor. The mean tensor's orientation may be strongly deflected toward the orientation of a few large-magnitude specimen tensors. However, high susceptibility specimens need not be considered as outliers in the disparaging statistical sense. They may provide useful information from a subfabric or second population of grains with a different orientation distribution. That is the case in this study where specimens with larger concentrations of paramagnetic accessory clay minerals may strongly influence the orientation of the mean tensor for a sub-sample. By recalculating the mean tensor using standardized specimen tensors, the influence of high susceptibility specimens and kinematically distinct subfabrics may be isolated, or emphasized (Borradaile 2001, 2003). Specimen tensors are standardized by dividing each of the principal magnitudes for a specimen (fcmax, &int, fcmin) by K; this weights all specimens equally, regardless of their bulk susceptibility. Consequently the contribution high-ft subfabrics to the orientation distribution will be subdued in the net-AMS. On the other hand, the mean tensor for non-standardized specimens emphasizes the role of such subfabrics (Borradaile 2001,2003).

534

T. D. HAMILTON ET AL.

Fig. 5. Idealized symmetry of confidence cones about the mean orientations of the principal axes of the mean tensor for a petrofabrically homogeneous group of specimens. The shapes of confidence cones for tensor-mean principal axes may be used for other petrofabrically significant axes (as in orientation distributions of crystals), for finite strain axes and for principal stresses. The shape of the confidence cones may also be related to the relative magnitudes of the three principal mean axes: (a) elongate confidence ellipses in the maximumintermediate plane define an ellipsoid with oblate symmetry (S > L) whereas (c) elongate confidence ellipses in the plane perpendicular to the maximum axis are associated with prolate symmetry (L > S). Flinn's (1965) L—S notation was initially introduced to describe finite strain ellipsoids or fabric ellipsoids but it is a useful shorthand for any anisotropy.

The following comparison of mean tensors for standardized versus non-standardized specimens in regional-scale and local-scale samples, display different AMS axes that reveal a rapidly changing tectonic regime, especially in the last 5 Ma. Sample mean tensors for standardized and non-standardized specimens are also compared profitably to previously determined AARM tensors. Processing AMS data from IOW-K specimens The tectonic significance of AMS data in limestone having weak susceptibilities depends on a careful assessment of the mineralogical origins of AMS and the effects of specimen variation within the sample suite. Jelinek (1978) showed that the orientation distribution (OD) of a suite of tensors ('AMS ellipsoids') must be treated by a special statistical procedure so that the sample's mean tensor retains orthogonal axes, just like individual specimen tensors. Applying Jelinek statistics permits us to characterize a sample of AMS tensors much more effectively than with density contours, although the latter still have their use (Borradaile 2001, 2003). The shape and symmetry of the 95% confidence regions around the mean tensor's principal axes define the orientation distribution of specimen tensors in the L-S fabric scheme (Flinn 1965); thus regional suites of AMS ellipsoids define

the regional variation from S through L tectonites (Fig. 5). Confidence cones for the mean tensor's principal axes reflect the shape of the orientationdistribution (OD) ellipsoid for the sample, not of individual tensors. For example, individual prolate ellipsoids scattered with their long axes in a plane define a sample with an orientation distribution described by an oblate ellipsoid (e.g. Borradaile 2003, pp. 286-287). Thus in AMS studies, if the mean 7} > 0 for specimens, it does not necessarily mean that their orientation distribution, described by the mean tensor, is also oblate (Borradaile 2001). The sample's mean tensor and its confidence cones may be biased toward the orientation of specimens of anomalous K. In this study, the anomalous specimen may be a diamagnetic one or a paramagnetic one in a contrasting matrix. The standardization technique described (dividing ^max etc. by K, etc.) suppresses this bias. Whereas it may change the orientation of the mean tensor, it may also change the shape of the confidence regions about the mean tensor's axes. Thus one may re-evaluate the OD of the mean tensor as an L > S rather than S > L fabric. AMS measurements AMS was determined using a Sapphire Instruments SI2B device operating at 19 200 Hz and

SUTB-FABRIC IDENTIFICATION BY STANDARDIZATION OF AMS

~0.1mT. The anisotropy measurement utilized the seven-orientation system (Borradaile & Stupavsky 1995), which includes four body-diagonal measurements, improving precision over the more commonly used orientations confined to coordinate axes and their symmetry planes (e.g. Girdler 1961). Instrument control and real-time data processing were performed using the SI2B01 software package developed by G. J. Borradaile. A complete AMS determination and analysis requires less than four minutes per core. The low-field AMS was determined on 360 cores from 121 specimens of the local sampling campaign and 730 cores from 298 specimens of the regional sampling campaign of Lagroix and Borradaile (2000). AARM: Anisotropy of anhysteretic remanent magnetization AARM was determined using the same seven-axis anisotropy scheme as for AMS measurements. The specimen is fully demagnetized and then the seven differently oriented ARMs are applied and measured, in a sequence such that each ARM is inclined at 45° or 35.3° to the previous one. For many rock types, this effectively cleans the ARM acquired in each preceding treatment. This tested technique greatly reduces the measurement time because 3-axis AF cleaning between each ARM application is no longer necessary (Werner & Borradaile 1996). However, the success of this short-cut must be verified with each new measurement campaign and each new lithology. ARMs were imposed with a Sapphire Instruments three-axis Alternating Field (AF) demagnetizer outfitted with a supplementary d.c. coil applying a bias field of 0.05mT. The AF decayed from a peak field of 80 or lOOmT to zero while the d.c. bias field was turned on from 60 mT to zero. The imposed ARM was measured in a JR5a automatic spinner magnetometer (sensitivity 0.03 mA/m). Instrument control, measurement and simultaneous data-reductions were performed using the SPIN01 software package developed by G. J. Borradaile. Complete AARM determination and analysis requires 18 minutes per core. We have only included AARM results from regional-scale sampling suite in this study. Of the original sample (n = 1170, Lagroix & Borradaile 2000), 201 cores were rejected for AARM determination. The criteria for accepting or rejecting the AARM results occur during measurement; results were rejected if intensities were < 1 mA/m when measured in orientation 1 of the AARM scheme or if the AMS ellipsoid for the sample had a negative K. In the present

535

study, for AARM we selected 254 cores from 229 specimens from the original suite published by Lagroix and Borradaile (2000). Sampling strategy The first level of sampling focused on structurally and lithologically homogenous sub-areas <30km 2 that were densely sampled (Fig. 6). They include the Polis Rift Margins, Polis Basin, Galataria, Omodhos-Pakhna, Lefkara and Lymbia, designated sub-areas I-VI, respectively, yielding n = 360 cores from N = 121 sites. These areas include formations from the Lefkara up to the Athalassa Formation. The homogeneity of fabrics in these sub-areas justified the larger scale interpretation from the next level of sampling. The second level of sampling was a diluted regional-scale campaign. The limestone specimens covered a wide age-range of the sedimentary cover, from the Lefkara Formation up to the Nicosia Formation (Lagroix & Borradaile 2000). Of the original sample suite (cores n = \\ 70, sites TV = 434), a sub-sample (n = 730, TV — 298) was analysed in five regional domains of similar tectonic style and trend (Fig. 7), each with an area of ~400km 2 . All hand specimens were oriented in the field and three to six cores (25mm in diameter and 22 mm high) were drilled in the laboratory from each specimen, restored to geographic coordinates. Localized sampling: Polis Rift Region (sub-areas I, II, III) Non-standardized AMS in this, the Polis Rift region, differs from the other areas (Fig. 6a). From the faulted margins they exhibit an S > L fabric as shown by the elongate confidence cones for /rmax and /qnt, and a small &min confidence cone (Fig. 6a, sub-area I). The AMS axes are symmetrically compatible with stress directions of the last 5 Ma verified from fault orientations and from seismic data (Payne & Robertson 2000; Borradaile & Hamilton 2004). In the basin, non-standardized data permit straightforward fabric interpretations (Fig. 6a, sub-area II). These younger strata have S > L depositional fabric as shown by the elongate confidence cones for A:max and /qnt in the foliation and a tight, near-vertical A;mjn confidence cone. The magnetic foliation is parallel to the bedding. ^max and fcmt ^Q within the bedding plane and ^max is possibly a flow-alignment, since it is

536

T. D. HAMILTON ET AL.

Fig. 6. Sub-areas of Cyprus (I-VI) and their magnetic fabrics as discriminated by standardization (specimen tensors standardized dividing by each specimen's K). (a) Non-standardized AMS orientation distributions are biased by anomalously oriented subfabrics, especially with anomalously high-*;, (b) Standardized AMS fabrics usually suppress the contribution of specimens with anomalous orientations.

SUB-FABRIC IDENTIFICATION BY STANDARDIZATION OF AMS

537

Fig. 7. Regional domains of Cyprus (I—V) and their magnetic fabrics, (a) Non-standardized AMS. (b) Standardized AMS. (c) Standardized A ARM. A ARM isolates the magnetite subfabric and when standardized, the contribution of anomalously oriented specimens is subdued.

parallel to the rift axis (cf. Sagnotti et al 1994). Raw, non-standardized AMS data demonstrates the effects of higher susceptibility specimens in which depositional-controlled AMS fabrics dominate. In the Galataria region, non-standardized specimens exhibit an L«S fabric, which appears nearly isotropic from its large, overlapping confidence cones (Fig. 6a, sub-area III). Nevertheless, the mean tensor axes are similar

to the other two preceding localities in this region. How are these IOW-K samples affected by relatively high-AC specimens? Standardizing the specimen AMS to K reveals similar susceptibility directions overall but demonstrate progressively different confidence cone shapes (Fig. 6b, subareas I-III). All show L > S ODs. The progressive increase in size of the confidence cones from west to east may be due to the smaller

538

T. D. HAMILTON ET AL.

sample sizes (n = 152 to n = 25). The NEtrending &maxL > S fabric agrees with the crustal tension axis required for the paired normal faults defining the Polis Graben and the normal fault zone defining the Pegia fault system (Payne & Robertson 2000). The L > S fabric of the AMS data are also symmetrically compatible with the current tectonic stress regime deduced from earthquake data (Borradaile & Hamilton 2004, Fig. 2). Localized sampling: Southern slopes of Troodos (sub-areas IV, V, VI) The three easternmost detailed sub-areas show little difference between standardized and nonstandardized samples (Fig. 6, sub-areas IV—VI). The Omodhos-Pakhna sub-area (IV) displays an L w S OD fabric with a shallow northward dipping 'foliation'. The Lefkara and Lymbia (sub-areas V and VI) localities display an S > L OD fabric as shown by the elongate confidence cones £max and hmi and a tight fcmin confidence cone with a shallow NNE dipping foliation. These OD fabrics in the eastern localities represent a dominant sedimentary fabric with some secondary tectonic control. However, all are compatible with a south-directed movement (southerly vergence) within the limestone cover. Regional-scale AMS Non-standardized AMS data of the five regionalscale domains (Fig. 7a, sub-areas I-V) exhibit similar ODs for all areas. All five groups exhibit an S > L fabric as shown by the elongate confidence cones for /cmax and fcint and a tight fcmin confidence cone near vertical (Borradaile 2001). This OD is similar to the results exhibited in the density-contoured data in Lagroix and Borradaile (2000). The S-component of the fabric is bedding controlled and fcmax is kinematically compatible with N-S extension, parallel to the aligned phyllosilicates caused by the stretching of the sedimentary cover (Lagroix & Borradaile 2000). But how are these low susceptibility samples affected by relatively high susceptibility samples that may be considered as statistical outliers? Standardizing the specimen ellipsoids to «, AMS data of the five regional-scale groups (Fig. 7b) reveals similar overall AMS axes but different confidence-cone shapes and, thus, different fabric ODs in the L-S range. These differences change progressively from L « S in the east to

L > S in the west. This corresponds to the increasing tectonic strain from east to west, inferred from earthquake data (Ben-Avraham et al. 1988; Arvidsson et al 1998; Papazachos & Papaioannou 1999; Borradaile & Hamilton 2004) and overall structural trends (Robertson 2000; Borradaile & Hamilton 2004). In contrast, raw data (non-standardized) are weighted by higher susceptibility depositional subfabrics rather than the weak tectonic ones compatible with supra-subduction SW-directed shortening (Lagroix & Borradaile 2000). Regional scale AARM Only standardized data are presented for AARM to avoid the spurious effects due to the large variance in remanence intensity. Domains I, II & V (Fig. 7c) show similar overall AARM axes to standardized AMS (Fig. 7b). This verifies the important contribution of magnetite to some AMS fabrics. They show an L K, S OD similar to the easternmost standardized AMS, perhaps due to a near-uniaxial shape of magnetite, which when arranged in a weak planar fabric has a strong Am[n zone-axis girdle. Thus, the N-S extensional trend (seen in the AMS) is less effectively expressed but the fabric axes are kinematically compatible with movement directed toward the SW (or subduction to the NE). AARM fabrics for the other two sub-areas (Fig. 7c, sub-areas III & IV) exhibit S > L fabrics; AARM axes indicate a more southerly directed movement. The AARM fabric may be a composite of the stylolitic cleavage incompletely overprinting the bedding fabric (Lagroix & Borradaile 2000). Conclusions LOW-K; rocks such as limestone commonly show unstable AMS axes due to sub-equal competition from subfabrics of diamagnetic calcite and paramagnetic clays, with traces of magnetite. These subfabrics commonly represent weak tectonic overprints on depositional fabrics, as in the limestone cover of the Troodos terrane. Standardizing AMS principal values of specimens to their K equalizes the contributions of subfabrics/specimens with different K in the overall sample mean tensor. Thus, one subfabric may be emphasized or neutralized. This has a similar effect to determining a PSD magnetite subfabric with AARM (Lagroix & Borradaile 2000), although it is not a satisfactory substitute for AARM.

SUB-FABRIC IDENTIFICATION BY STANDARDIZATION OF AMS Applying this methodology to the limestone cover of Troodos, standardized and nonstandardized AMS reveal depositional fabrics, neotectonic fabrics associated with NNE subduction, neotectonic fabrics associated with rifting and composite fabrics from feeble stylolitic foliation imprints on bedding. Standardizing AMS data for petrofabrically homogenous sub-areas or domains reveals fabrics compatible with neotectonic stress trajectories in post-Paleogene strata. This is well shown in the Polis basin (<5Ma) and the Galataria sub-area (43-62 Ma) (Figure 6b, sub-areas II and III respectively), which show nearly identical fabrics in rocks with 40 million years age difference, but relatively little geographic difference. Examination of AMS tensors for sub-areas, and even larger regions may reveal tectonically significant axes, related to stress rather than finite strain. The orientation distribution of the mean tensor has an ellipsoid shape that may reveal the relative magnitudes of the principal stresses. The Natural Sciences and Engineering Research Council of Canada (NSERC) funded G. J. Borradaile for this work at the Lakehead University Rock Magnetism Laboratory. The authors were greatly assisted and encouraged by the Geological Survey Department of Cyprus, in particular through G. Petrides and I. Panayides. D. Czeck, P. Kelso and M. Jackson provided helpful, constructive reviews and editorial comments. References ARVIDSSON, R, BEN-AVRAHAM, Z., EKSTROM, G. & WDOWINSKI, S. 1998. Plate tectonic framework for the October 9, 1996 Cyprus earthquake. Geophysical Research Letters, 25, 2241-2244. BEN-AVRAHAM, Z., KEMPLER, D. & GlNZBURG, A.

1988. Plate convergence in the Cyprean Arc. Tectonophysics, 146, 231-240. BORRADAILE, G. J. 1988. Magnetic susceptibility, petrofabrics and strain. Tectonophysics, 156, 1-20. BORRADAILE, G. J. 2001. Magnetic fabrics and petrofabrics: their orientation-distributions and anisotropies. Journal of Structural Geology, 23, 1581-1596. BORRADAILE, G. J. 2003. Statistics of Earth Science Data: Their Distribution in Space, Time, and Orientation. Springer-Verlag, New York, 293326. BORRADAILE, G. J. & HAMILTON, T. 2003. Limestones distinguished by magnetic hysteresis in threedimensional projections, Geophysical Research Letters, 30(18), 1973, 10.1029/2003GL017892. BORRADAILE, G. J. & HAMILTON, T. D. 2004. Magnetic fabrics may proxy as neotectonic stress trajectories, Polis Rift, Cyprus. Tectonics, 23, TCI001, doi:10.1029/2002TC001434.

539

BORRADAILE, G. J. & HENRY, B. 1997. Tectonic applications of magnetic susceptibility and its anisotropy. Earth Science Reviews, 42, 49-93. BORRADAILE, G. J. & KEHLENBECK, M. M. 1996. Possible cryptic tectono-magnetic fabrics in 'post-tectonic' granitoid plutons of the Canadian Shield, Earth and Planetary Science Letters, 137, 119-127. BORRADAILE, G. J. & LAGROIX, F. 2000. Magnetic characterization of limestones using a new hysteresis projection. Geophysics Journal International, 141, 213-226. BORRADAILE, G. J. & LAGROIX, F. 2001. Magnetic fabrics reveal Upper Mantle flow fabrics in the Troodos Ophiolite Complex, Cyprus. Journal of Structural Geology, 23, 1299-1317. BORRADAILE, G. J. & STUPAVSKY, M. 1995. Anisotropy of magnetic susceptibility: measurement schemes. Geophysical Research Letters, 15, 1957-1960. BORRADAILE, G. J., TELLA, S. & MCARTHUR, J. 1989. Magnetic fabric as a kinematic indicator of faults: a test case. Annals de Tectonicae, 3, 3-11. ClFELLI , F., ROSSETTI, F., MATTEI, M., HlRT, A. M.,

FUNICIELLO, R. & TORTORICI, L. 2004. An AMS, structural and paleomagnetic study of quaternary deformation in eastern Sicily. Journal of Structural Geology, 26, 29-46. DE WALL, H., BESTMANN, M. & ULLEMEYER, K. 2000. Anisotropy of susceptibility in Thassos marble: a comparison between measured and modelled data. Journal of Structural Geology, 22, 1761— 1771. FLINN, D. 1965. On the symmetry principle and the deformation ellipsoid. Geology Magazine, 102, 36-45. GIRDLER, R. W. 1961. The measurement and computation of anisotropy of magnetic susceptibility in rocks. Geophysical Journal of the Royal Astronomical Society, 5, 34-44. HENRY, B. 1983, Interpretation quantitative de 1'anisotropie de susceptibilite magnetique: Tectonophysics,9l, 165-177. HENRY, B. 1989. Magnetic fabric and orientation tensor of minerals in rocks: Tectonophysics, 165, 21-28. HENRY, B. 1990. Magnetic fabric implications for the relationships between deformation mode and grain growth in slates from the Borrowdale Volcanic Group in the English Lake District. Tectonophysics, 178, 225-230. HENRY, B. 1992. Modelling the relationship between magnetic fabric and strain in polymineralic rocks. Physics of Earth and Planetary Science International, 70, 214-218. HELLWEGE, K. H. & HELLWEGE, M. 1967. Magnetic properties, II. In: VON BORCHERS, H., HAUSEN, H., HELLWEGE, K. H., SCHAFER, K. & SCHMIDT, E. (eds), Landolt-Bornstein: Zahlenwerte und Funktionen aus Physik. Chemie, Astronomie, Geophysik und Technik, H/10, Springer, Berlin, 143. HUNT, C. P., MOSKOWITZ, B. M. & BANERJEE, S. K. 1995. Magnetic properties of rocks and minerals. In: AHRENS, T. J (ed.) A Handbook of Physical Constants. American Geophysical Union, Washington DC, 189-204.

540

T. D. HAMILTON ET AL.

IHMLE, P. F., HIRT, A. M., LOWRIE, W. & DIETRICH, D. 1989. Inverse fabric in deformed limestones of the Morales Nappe, Switzerland. Geophysical Research Letters, 16, 1383-1386. JACKSON, M. J., CRADDOCK, J. P., BALLARD, M., VAN DER Voo, R. & McCABE, C. 1989. Anhysteretic remanent magnetic anisotropy and calcite strains in Devonian carbonates from the Appalachian Plateau, New York. Tectonophysics, 161, 43-53. JELINEK, V. 1978. Statistical processing of anisotropy of magnetic susceptibility measured on groups of specimens. Studia Geophysica et Geodetica, 22, 50-62. JELINEK, V. 1981. Characterization of the magnetic fabric of rocks. Tectonophysics, 79, 63-67. KISSEL, C., BARRIER, E., LAJ, C. & LEE, T. Q. 1986. Magnetic fabric in 'undeformed' marine clays from compressional zones. Tectonics, 5, 769-781. KRISHNAN, K. S., GUHA, B. C. & BANERJEE, S. K. 1933. Investigation on magneto-crystallic action, Part I. Diamagnetics. Philosophical Transactions of the Royal Society London, Serial A, 231, 235-262. LAGROIX, F. & BORRADAILE, G. J. 2000. Tectonics of the Circum-Troodos Sedimentary Cover of Cyprus, from Rock Magnetic and Structural Observations. Journal of Structural Geology, 22, 453-469. LORD, A. R., PANAYIDES, I., URQUHART, E. & XENOPHONTOS, C. 2000. A biochronostratigraphical framework for the Late Cretaceous-Recent circum-Troodos sedimentary sequence, Cyprus. In: PANAYIDES, I., XENOPHONTOS, C. & MALPAS, J. (eds) Proceedings of the Third International Conference on the Geology of the Eastern Mediterranean. Cyprus Geological Survey Department, Nicosia, Cyprus, 289-297. MALPAS, J., MOORES, E., PANAYIOTOU, A. & XENOPHONTOS, C. (eds), Ophiolites: Oceanic Crustal Analogues, Proceedings of the Symposium 'Troodos 1987\ Cyprus Geological Survey Department, Nicosia, Cyprus. MATTEI, M., SPERANZA, F., ARGENTIERI, A., ROSETTI, F., SAGNOTTI, L. & FUNICIELLO, R. 1999. Extensional tectonics in the Amantea Basin (Calabria, Italy): a comparison between structural and magnetic anisotropy data. Tectonophysics, 307, 33-49. NAKAMURA, N. & BORRADAILE, G. J. 2001. Strain, anisotropy of anhysteretic remanence, and anisotropy of magnetic susceptibility in a slaty tuff. Physics of Earth and Planetary Science International, 125, 85-93. OWENS, W. H. & BAMFORD, D. 1976. Magnetic, seismic, and other anisotropic properties of rock fabrics. Philosophical Transactions of the Royal Society London, A283, 55-68. OWENS, W. H. & RUTTER, E. H. 1978. The development of magnetic susceptibility anisotropy through crystallographic preferred orientation in a calcite rock. Physics of the Earth and Planetary Interiors, 16,215-222. PAPAZACHOS, B. C. & PAPAIOANNOU, Ch. A. 1999. Lithospheric boundaries and plate motions in the Cyprus area. Tectonophysics, 308, 193-204.

PAYNE, A. S. & ROBERTSON, A. H. F. 2000. Structural evolution and regional significance of the Polis graben system, western Cyprus. In: PANAYIDES, I., XENOPHONTOS, C. & MALPAS, J. (eds) Proceedings of the Third International Conference on the Geology of the Eastern Mediterranean. Cyprus Geological Survey Department, Nicosia, Cyprus, 45-59. REILINGER, R. E., MCCLUSKY, S. C., ORAL, M. B., KING, R. W., TOKSOZ, M. N., BARKA, A. A., KINIK, I., LENK, O. & SANLI, I. 1997. Global position system measurements of present-day crustal movements in the Arabia-Africa-Eurasia plate collision zone. Journal of Geophysical Research, 102, 9983-9999. ROBERTSON, A. H. F. 1990. Tectonic evolution of Cyprus, In: MALPAS, J., MOORES, E., PANAYIOTOU, A. & XENOPHONTOS, C. (eds), Ophiolites: Oceanic Crustal Analogues, Proceedings of the Symposium 'Troodos 1987\ Cyprus Geological Survey Department, Nicosia, Cyprus, 235-250. ROBERTSON, A. H. F. 2000. Tectonic evolution of Cyprus in its Easternmost Mediterranean setting, In: PANAYIDES, I., XENOPHONTOS, C. & MALPAS, J. (eds.), Proceedings of the Third International Conference on the Geology of the Eastern Mediterranean. Cyprus Geological Survey Department, Nicosia, Cyprus, 11^4. ROBERTSON, A. H. F., KIDD, R. B., IVANOV, M. K., LIMONOV, A. F., WOODSIDE, J. M., GALINDOZALDICAR, J. & NIETO, L. 1995. Eratosthenes seamount: collisional proceses in the easternmost Mediterranean in relation to the Plio-Quaternary uplift of southern Cyprus. Terra Nova, 1, 254-264. ROCHETTE, P. 1988. Inverse magnetic fabric in carbonate-bearing rocks, Earth & Planetary Science Letters, 90, 229-237. ROCHETTE, P., JACKSON, M. J. & AUBOURG, C. 1992. Rock magnetism and the interpretation of anisotropy of magnetic susceptibility. Reviews of Geophysics, 30, 209-226. SAGNOTTI, L., FACCENNA, C., FUNICIELLO, R. & MATTEI, M. 1994.Magnetic fabric and structural setting of Plio-Pleistocene clayey units in an extensional regime: the Tyrrhenian margin of central Italy. Journal of Structural Geology, 16, 12431257. SAGNOTTI, L., SPERANZA, F., WINKLER, W., MATTEI, M. & FUNICIELLO, R. 1998. Magnetic fabric of clay sediments from the external northern Apennines (Italy). Physics of the Earth and Planetary Interiors, 105, 73-93. TARLING, D. H. & HROUDA, F. 1993. The Magnetic Anisotropy of Rocks. Chapman and Hall, London. VOIGHT, W. & KINOSHITA, S. 1907. Bestimmung absoluter Werte von Magnetisierungszahlen, insbesondere fur Kristalle. Annals of Physics, 24, 492-514. WERNER, T. & BORRADAILE, G. J. 1996. Paleoremanence dispersal across a transpressed Archean Terrain: Deflection by anisotropy or by late compression? Journal of Geophysical Research, 10, 5531-554.

Index Note: Page numbers in italics refer to Figures, those in bold refer to Tables. 3-D projection, magnetic susceptibility 64-7 AARM see anisotropy of anhysteretic remanent magnetization abnormal magnetic fabrics dykes 227-8, 237, 286 siderite-related 505 accretion 422, 426, 438 Achalian deformation 450, 470 aeolian deposits 145-73 AF demagnetization see alternating field demagnetization aggregate processes 345-7 AGRM see anisotropy of gyroremanent magnetization AIRM see anisotropy of isothermal remanent magnetization 'Alaskan' or 'wind-vigour' type loess 145-6, 170 albite-chlorite schists 486, 487 algorithms, micromagnetic 37-8 aligned rock, AMS mineral abundances 333-4 alignment mechanisms in crystals 306-8, 345 alternating field (AF) demagnetization 259, 260, 266-7 amphibolite-facies schist 62, 63, 64, 65 amphibolites 483-6 AMR see anisotropy of magnetic remanence AMS see anisotropy of magnetic susceptibility anatectic complexes 381-3, 382, 391 anhysteretic remanent magnetization (ARM) 318 see also anisotropy o f . . . anisotropy grain-scale 310-13 interactions in grain ensembles 313-15 magnetocrystalline 310-12 measurement 318-20 nontensorial 320-1 reliability 408, 409 anisotropy o f . . . anhysteretic remanent magnetization (AARM) cleavage/bedding angles 101 cleaved pelites 81-2, 83-4 deformed rocks 299 degree of anisotropy and shape parameter 95-6 metamorphism 347-8, 350-2 methods 25-7, 28-33 Ordovician shales 114, 119, 120, 123-4 paramagnetic limestones 535 principal axes orientation 91-5 rotation 26 gyroremanent magnetization (AGRM) 25, 27-31, 33-4, 321 isothermal remanent magnetization (AIRM) deformed rocks 302 low-grade metasediments 184-5 metamorphism 349 methods 24, 28, 31, 33, 38 Ordovician shales 114

magnetic remanence (AMR) distribution anisotropy 37-47 methods comparison 2, 21-35, 29, 32 results 28-31 summary of techniques 23 magnetic susceptibility (AMS) 28 see also ellipsoids Appalachian fold and thrust belt 120-4 batholiths 456-63 cleavage/bedding angle 77-107 cleavage and strain in mudrocks 193-5 crystal structure 311-12 deformed rocks 299-360 diamagnetic rocks 49-59 dykes 208-11, 212-13, 227-49, 291-4 field variation 71-3 flowing magma 227-49 high-field separation 519-21 instruments and field intensity 70 Klodzko Metamorphic Complex 479-90 lava flows 208-11 linear theory 69, 73 loess/palaeosol deposits 146, 151-2, 153-5, 153, 161-7 low-grade metasediments 77-107, 183-4 metamorphism 343, 347-50 methods 1-4,28,29,32 mineral orientation 9-20, 326-39 Ordovician shales 112-13, 115-20 orientation analysis 86, 87 paramagnetic limestones 533-5 phyllosilicates 366, 374, 376-7 principal axes orientation 117-19, 122 siderite-bearing pelitic rocks 495-505 Soultz granite 517-21 Stoddard Mountain laccolith 259-60, 263-4, 266-73, 275, 276-7 structural orientation comparison 185-7 sub-fabric identification 527-40 tectonic processes 339-52 temperature dependence 155, 157-60, 167, 268-9, 309, 363-5, 366 Toledo Complex, Spain 385-92 Trives and Veiga massifs 397-404 weakly deformed mudrocks 127-8, 130, 132-6, 191-203 partial anhysteretic remanent magnetization (pAARM) 348, 349, 350-2 rotational remanent magnetization (ARRM) 26, 27,28 saturation isothermal remanent magnetization (ASIRM) 38-46 thermoremanent magnetization (ATRM) 24, 28, 31-3 viscous remanent magnetization (AVRM) 24-5, 28

542

INDEX

annealing experiments 480, 487, 490 anticlinal structure, New England Orogen 422-4 anti-ferromagnetic minerals 310-11 aplitic dykes 467 Appalachian fold and thrust belt 109-26 Archaean granitoids 63, 66, 67 Archaean metagreywackes 61-8 Argentina, Devonian granitoids 447-74 arkose, AMS standardisation 330 ARRM see anisotropy of rotational remanent magnetization ASIRM see anisotropy of saturation isothermal remanent magnetization Asquempont Member 78, 79, 85-97, 100, 103 ATRM see anisotropy of thermoremanent magnetization Australia, oroclinal bending 423, 425-6 AVRM see anisotropy of viscous remanent magnetization axis-normal magnetic lineations 175-90 basalts AMS 211,218 pillow lavas 481 titanomagnetite 70, 72, 74 batholiths AMS 456-63 central Argentina 450-5 emplacement 466-9, 470 magnetic fabrics 463-6 structural interpretation 465-6 bedding 337-40 see also cleavage/bedding angle Belovo, Siberia 147 biotite 371, 464, 465, 521 blended sub-fabrics 334-7 Bohemian Massif 361-80, 475-91 boot-strap method 329 borehole, EPS-1 510, 511-14 Borrowdale volcanic slate 332, 348 Brabant Massif 78-81 Brazil, Rio-Ceara Mirim dyke swarm 285-98 Bude Formation, North Cornwall 175, 177, 179 bulk magnetic susceptibility Archaean greywackes 61-2 cleaved pelites 88-91 field variation 70-1 heat treatment of pelitic rocks 498-9 La Totora batholith 457-9 loess/palaeosol deposits 150 mafic dykes 286 mudrocks 197 orientation tensors 9 paramagnetic limestone 530, 531 rock composition and AMS ellipsoid-shape 324-6 second-rank tensor properties 315-18 Soultz granite 511 temperature 268-9

thermal demagnetization 270 Trives and Veiga massifs 406 calcite AMS orientations 305-6, 307 AMS parameters 52, 55, 56 petrofabrics 357, 528-30 Cambrian slates 78-9, 348 Carboniferous pelitic rocks 493-507 cataclastic granites 518-19, 522-3, 524 Central Siberia 154, 160-1, 167 characteristic remanence (ChRM) 136, 137-40, 140-2 chevron folds 177-9, 180, 187-8, 495 chilled zone, laccolith 256, 260, 264-6, 271-2, 277 'Chinese' or 'pedogenic' type loess 145-6, 170 chlorite interlayering with mica 368, 372 mineral fabric 113, 115-16, 118, 122 paramagnetic carrier 98, 99-102 pelitic rocks 493 tensors 371 ChRM see characteristic remanence cleavage AMS and strain in mudrocks 193-5 cogenetic 78-9 magnetic ellipsoids 193, 194-5 magnetic foliation relationship 434 mudrocks 140, 191, 199-200 oblate 493 slaty 110, 113, 124 stages of deformation 195-6 subfabrics 336 cleavage/bedding angle 77-107 degree of anisotropy relationship 90, 91 magnetic fabric orientation 100, 102-3 shape parameter relationship 89, 91 susceptibility axes 99-101 cleaved pelites 77-107 AARM81-2, 91-6 AMS 83-4, 86-91 ferromagnetic mineralogy 82-6 X-ray goniometry 96-9 climate, Siberian loess/palaeosol deposits 148-50 coercivity spectra 94 coercivity window 123 Coffs Harbour block 432-4 see also Texas-Coffs Harbour region cogenetic cleavage 78-9 Colorado Plateau, USA 252, 253 Columbia River Basalt 218 columnar basalts, AMS 211 combined study 127-43 computer simulations, AMS of flowing magma particles 229-37 cordierite 312 Cornwall, low-grade metasediments 175-90, 493-507 Crackington Formation, North Cornwall 177, 179 Crackington Haven, North Cornwall 501-2

INDEX cryptofabric, mudrocks 196, 199-200 crystalline limestones 481-3, 484 crystallization of granitoids 469, 510-11 crystals alignment process 306-8, 345 PCOs 1,4 plasticity 344-5 siderite orientation 502-3 single-crystal deformation processes 341-5 structure and symmetry 303, 304, 307, 311-12 Culm Basin, SW England 175-90, 493-507 Cyprus calcite petrofabrics 351 neotectonic structures 527-40 Polis Rift Region 535-8 sub-areas and regional domains 532, 536, 537 tectonic features 529 Troodos southern slopes 538 Czech Republic, Bohemian Massif 361-80 dating, loess/palaeosol deposits 150 deformation see also faulting; folding; oroclinal bending; erogenic belts Achalian 450, 470 ductile 362-3 dykes 238-9, 240 fabric development stages 192-3, 195-6 lava flows 215-16, 219-20, 221 magmatic fabrics 231, 235 mechanisms 340-1, 344, 345-7 particulate flow 347 Permian sediments 440-1 polyphase 421-4, 425, 448-50, 479, 487-90 solid-state 455, 469 Variscan 177-8, 187-8, 362, 383 weakly deformed mudrocks 133, 136, 191-203 deformed rocks 229-360 crystal-aligning process 306-8 magnetic fabric characterization 308-21 mineral ODs 326-39 tectonic processes 339-52 tensor parameters 321-6 degree of anisotropy cleaved pelites 88-91, 95-6 diamagnetic rocks 49-50 granites 459-60 loess/palaeosol deposits 162-4 mudrocks 198-9 phyllosilicates 369, 372, 373 shape parameter relationship 92 Toledo Complex 387 demagnetization 85, 261-2, 457 deterministic approach, lava flows 216-17, 220 Devonian granitoids 447-74 limestones 486-7, 488, 490 metamorphic rocks 479

dextral transpression 178 diagenetic carbonate growth 503-4 diamagnetic rocks 49-59 AMS parameters 49-51 marble/limestone model 52-5, 56 quartzite/evaporite model 51-2, 56-7 diamagnetism 308 differentiation processes 343-4 diffusion 344 DIGICO see spinner magnetometer distribution anisotropy 37-47 ideal uniaxial SD grains 39-42 non-uniform magnetic structures 42-5 numerical model 37-9 ductile-brittle shear zones 391 ductile deformation 362-3 dykes see also mafic dyke swarm abnormal magnetic fabrics 227-8, 237, 286 AMS 208-11, 212-13, 239-44 aplite 467 flow direction 209-11, 213, 214, 216-19, 220-1, 227-9, 245, 285, 294-5 NE Brazil 285-98 rheological characteristics 208, 209 shear deformation 238-9, 240 shear of fluid elements 240-4 dynamic recrystallization 345 eigenvectors granites 399-400, 403-4 low-grade metasediments 181-2, 185 electron microscopy 111-12, 113-15, 120 ellipsoids bulk susceptibility and rock composition 324-6 Eocene marls, Spain 127-8, 130, 132-3, 136, 138-40 flowing magma 228, 229 granites 406-7 heating changes 349 low grade metasediments 183-4, 185, 187 mafic dyke swarm 295-6 magnetic susceptibility 103-4 movement 246-7 palaeosols 165-7 rotation 219-20, 221 shape parameters 321-4?? Soultz granite 517, 518 Trives and Veiga massifs 406-7 weakly deformed mudrocks 192, 193, 194-5, 197-8, 200 emplacement Argentinian batholiths 448, 466-9, 470 Iron Axis laccoliths 253-7 mafic dyke swarm 295, 296 mineralogy of magma 245-6 Soultz granite 524 Stoddard Mountain laccolith 253-7, 258, 277-9

543

544

INDEX

Eocene marine mudrocks 127-43 EPS-1 borehole 510, 511-14 evaporites 50, 51-2 fabrics see also magnetic fabrics; mineral fabrics; petrofabrics; subfabrics igneous 3-4, 387 magmatic295, 521-2, 524 non-magnetic 362 structural 301-2, 475-91 tectonic 3, 77, 300-2, 389 faulting North Cornwall 178-9, 180, 187-8 Sierra de San Luis 468, 469 FD see frequency-dependence of magnetic susceptibility ferrimagnetic fabric 504 ferrimagnetic mineralogy 113, 489 ferromagnetic granitoids 396, 448 ferromagnetic minerals 182 see also anti-ferromagnetic minerals carriers 99-102 clay-rich rocks 431 cleaved pelites 82-6 metamorphic rocks 489 ferromagnetism 309-10 field tests, palaeomagnetism 264-6 field variation in AMS 70-3 finite strain 339, 487-9 first-order folds 110 Flinn, D. 322-4, 322, 328-9, 328 flow direction see magma flow direction folding see also orogenic belts Appalachian fold and thrust belt 109-26 chevron, North Cornwall 177-9, 180, 187-8, 495, 507 orientation analysis 185-7 remanence 139 simultaneous 391-2 SW England 502-3 weakly deformed mudrocks 131, 132 foliations see also magnetic foliations New England Orogen 424 non-uniform magnetic structures 45 uniform SD grain assemblages 41-2 foreland basins 128-30, 495 frequency-dependence of magnetic susceptibility (FD) 145, 757-2, 153-4, 168-9 Fuller, M. 303 Gilgurry Mudstone 428-9, 437-8, 440 gneisses granulite fades 62, 63, 64, 65 Kapuskasing 332 pyrrhotite 77, 72, 73, 74

Gobelli granite 452, 453 goethite 182 Graham, J.W. 1, 303 grain axis orientations 11-19 grain ensemble interactions 313-15 grain scale anisotropy 310-13 deformation 340-1 grain shape anisotropy 312-13, 314 grain-strain 347 granites Argentina, Devonian 452, 453 degree of anisotropy 459-60 ferromagnetic 396 Iberian Hercynian belt 396-406 paramagnetic 395-420 Pyrenees 413-15 solid-state deformation 455 Soultz-sous-Forets 509-26 Spain, Montes de Toledo 382, 386-7, 388-9, 390-1 Trout Lake 332 granitoids Archaean 63, 66, 67 Devonian, central Argentina 447-74 ferromagnetic 448 paramagnetic 448 Spain 381-3 Stoddard Mountain laccolith 252 Variscan 510-11 granodiorites 405-6 granulite-facies gneiss 62, 63, 64, 65 greenschist-facies slate 62, 63, 64, 65 greenstones 486, 487 greywackes 61-8 greywacke-schists 325 gyroremanent magnetization see anisotropy of.. harzburgite 63, 66, 67 Hawaiian dykes 212 HDR see Hot Dry Rock project heat treatment of pelitic rocks 496-502 hematite clay-rich rocks 431 cleaved pelites 82-5 crystal structure 311 loess/palaeosol deposits 168 replacement 514, 522, 524 single crystal 71, 72 Hercynian Iberian Belt 381-94, 396-406 HFA see high-field magnetic anisotropy HFP, phyllosilicates 364, 374 high-field magnetic anisotropy (HFA) 365, 368 high-field magnetization/susceptibility 316-17 high-field separation 519-21 high-field torque 317 Hot Dry Rock (HDR) project 509 humidity, palaeoclimate 168-9, 170

INDEX hydrothermal fluids Soultz granite alteration 509-26 Upper Carboniferous sediments 493, 504, 505 hysteresis rocks 317 Iberian Hercynian belt 381-94, 396-406 ideal uniaxial SD grains 39-42 igneous fabrics 3-4, 387 see also magmatic fabrics igneous rocks see also batholiths; dykes; intrusions; lava flows AMS 205-25 models of fabric acquisition 207-20 Theological characteristics 208, 209 illite 522, 524 imbrication angle 212, 217 imperfect uniaxial grains 11-19 inner zone, Stoddard Mountain laccolith 260-6, 277 intrusions see also batholiths; dykes Colorado Plateau 253-6 Stoddart Mountain laccolith 251-83 inverse magnetic fabrics 311-12, 494, 523, 527 IRM see isothermal remanent magnetization Iron Axis laccoliths 252-7 iron sulphides 136, 141 isothermal remanent magnetization (IRM) 317 see also anisotropy of...; saturation isothermal... acquisition curves 267 anisotropy 24, 28, 31, 33 Soultz granite 514-16 Stoddard Mountain laccolith 259 thermal demagnetization 269, 270, 271, 430 Jaca-Pamplona basin, Pyrenees 127-43 Jeffery's equations ellipsoidal particles 209-10 fabric acquisition 236-7 particle orientation 219, 221, 227-8 Jelinek, V. 303, 322-4, 329-31, 397, 401 Kapuskasing gneiss 332 Khan's model 206-7, 209-10 kinematics 340 kinking, mineral fabric 113-15, 120, 124 Klodzko Metamorphic Complex 475-91 AMS 479-90 location 475, 476 mineral and magnetic fabric orientations 481 tectonic foliation and lineation 477-8 thrust sheets 475-6 KLY-2 Kappabridge 495, 519 KLY-3S Kappabridge 81, 150, 181, 268, 363, 432, 434, 435

KLY-4S Kappabridge 69-76, 259 Kuznetsk depression, Siberia 147-8, 153-4, 157-60, 161-7, 169

545

laccolith, Utah, USA 251-83 La Portena granite 452, 453 Las Chacras-Potrerillos batholith 454-5, 458-9, 461-2, 463, 465-6 Late Carboniferous oroclinal bending 426 Late Palaeozoic events 438 La Totora batholith 450-4 AMS data 456 AMS directional data 460-3, 466 AMS scalar parameters 459-60 magnetic mineralogy 457 lava flows 205-25 deterministic and systematic-variation approaches 216-19,220 flow direction 209-11, 213, 214-19, 220-1 internal deformation 215-16, 219-20, 221 magnetite 206-7 rheological characteristics 208, 209 titanomagnetites 206 unified model 219-20 layer-parallel shortening (LPS) 110 Layos granite, Toledo Complex 382, 388-9, 390, 391 Lehigh Gap, Pennsylvania 110, 113-20 limestones AMS 57, 52-5, 56, 307 crystalline 481-3, 484 paramagnetic 531 Upper Devonian 486-7, 488, 490 linear theory 69, 73 lineations see also magnetic lineations non-uniform magnetic structures 44-5 stretching 187, 389, 391, 421, 439 uniform SD grain assemblages 39-41 loess/palaeosol deposits 145-73 AMS 146, 151-2, 153-5, 153, 161-7 magnetic mineralogy 155-61, 167-8 palaeoclimatic reconstruction 168-9, 170 palaeowind directions 169-70 Lower Palaeozoic slate belt 78-81, 80, 348 low-field magnetic anisotropy 365, 367-8 low-grade metasediments 175-90, 493-507 see also weakly deformed mudrocks AIRM 184-5 AMS 183-4 cleaved pelites 77-107 magnetic fabrics 3 magnetic mineralogy 182-3 New England Orogen 421-45 petrofabrics 181-2 structural data 181-2 LPS see layer-parallel shortening L-S fabric scheme 328-9, 328, 528 mafic dyke swarm 285-98 AMS 291-4 ellipsoids 295-6 magma flow direction 294-5

546 mafic dyke swarm (cont.} magnetic mineralogy and oxide texture 287-8 plagioclase fabric 288-94 maghemite 167-8, 268-9 magma flow direction dykes 209-11, 213, 214, 216-19, 220-1, 227, 245, 285 lava flows 209-11, 275, 216-19, 220-1 mafic dyke swarm 294-5 magnetic fabrics 251-3, 277-9, 285 Mora granite 390-1 magmas batholiths 467 internal deformation 231, 235, 235 mineral content during emplacement 245-6 particle movement 227-49 shear deformation 238-9 magmatic fabrics mafic dyke swarm 295 Soultz granite 521-2, 524 magnetic carriers 77-107 AMS relationship 99-102 cleavage/bedding angle 102-3 ferromagnetic 99-102 paramagnetic 99-102 magnetic fabrics see also magnetic foliations; magnetic lineations abnormal 227-8, 237, 286, 505 acquisition mechanisms 137-40 Archaean greywackes 64-7 Argentinian batholiths 463-6 blocking processes 128 characterization methods 2-3 classification 308-10 clay-rich rocks 432-8, 439-40 Cyprus 536, 537 deformed rocks 299-360 development 109-26 dyke model 22-44 fresh granite 517-18, 521-2 heat treatment of pelitic rocks 496-502 hydrothermally altered granite 518, 521-4 igneous rock models 207-22 inverse 494 loess/palaeosol deposits 145-73 magma flow direction 251-3, 277-9, 285 marls 132-6, 137 mineral fabric relationship 370-2 mineral shapes influence 244 mudrocks 195-6 oblique 493-507 orientation, cleavage/bedding angle 100 paramagnetic granites 395-420 progress 303-8 siderite-related 505 structural fabrics relationship 475-91 terminology 300-2 timing of blocking 140-2

INDEX Toledo Complex 383-5, 391-2 weakly deformed mudrocks 127-43 magnetic foliations Argentinian batholiths 460, 461, 463, 465, 469 diamagnetic rocks 49-50 regional cleavage relationship 434 reliability 415, 416 sedimentary rocks 3 Soultz granite 517-18, 520,521-2 Stoddard Mountain laccolith 275 tectonics related 389, 438 Texas-Coffs Harbour region 437-8, 438 magnetic lineations Argentinian batholiths 460, 462, 463, 466 axis-normal 175-90 diamagnetic rocks 49-50 distribution 412 mafic dyke swarm 287, 290-1 Mora granite 387, 388 paramagnetic granites 396 reliability 407-15, 410 rotation 128 sedimentary rocks 3 Soultz granite 518, 520, 522 Stoddard Mountain laccolith 275 stretching lineation relationship 389 Texas beds 434 Texas-CofTs Harbour region 437, 439 Toledo Complex 387-8 magnetic mineralogy clay-rich rocks 438 cleaved pelites 82-6 La Totora batholith 457 loess/palaeosol deposits 155-61, 167-8 low-grade metasediments 182-3 mafic dyke swarm 287-8 Ordovician shales 113 preferred orientation 9-20 magnetic petrology 3, 63 magnetic remanence see anisotropy of magnetic remanence magnetic susceptibility see also anisotropy of magnetic susceptibility 3-D projection 64-7 ellipsoids 103-4 EPS-1 borehole 511-14 Klodzko Metamorphic Complex 480 metamorphic control 61-8 temperature dependence 268, 363-5, 366 magnetic susceptibility axes AARM91-5 cleavage/bedding angle 99-101 cleaved pelites 78, 83-4, 87 crystal symmetry 303, 304, 307 loess/palaeosol deposits 162-4, 166 orientation distribution 327-31 peak density 330 principal axes orientation 117-19, 122

INDEX rotation 137-40 Stoddard Mountain laccolith 272, 276 stress trajectories 350-2 magnetite AMS 306 bulk magnetic susceptibility 62-3 clay-rich rocks 431 cleaved pelites 85 crystal structure 310-11 decomposition 514 dykes 246 granite 521 lava flows 206-7 loess/palaeosol deposits 167-8 low-grade metasediments 182 metamorphic rocks 489 mudrocks 197 oxidation 522, 524 Stoddard Mountain laccolith 268-9 magnetite-free granites 395 magnetocrystalline anisotropy 310-12 map, deformation mechanisms 344, 345-7 marble 52-5, 56 marls 128-30, 132-6 Martinsburg Formation, Pennsylvania 109-26 MD see multidomain particles MDFs see median destructive fields mean magnetic suceptibility 428 mean tensors, limestone 533-4 median destructive fields (MDFs) 155-6, 160-1, 267 meta-andesites 483-6 metabreccias 483, 484 metagabbros 483-6 metagranites 483-6 metagreywackes 61-8 metamorphic complexes central Argentina 449, 450 Klodzko, Poland 475-91 metamorphic rocks see gneisses; low-grade metasediments; schists; slates metamorphism AARM 347-8, 350-2 AMS 343, 347-50 control of magnetic susceptibility 61-8 deformation mechanisms 345-7 overprinting 489, 490 metarhyolites 486, 487 meta-sandstones 481-3 metasediments 3, 63, 486, 487 see also low-grade metasediments mica interlayering with chlorite 368, 372 mineral fabric 113, 115-16, 118, 122 paramagnetic carrier 99-102 micromagnetic algorithm 37-8 microstructures, granites 455 mineral abundances AMS for aligned rock 333-4

547

AMS-magnitude relations 338 susceptibility magnitudes 337 mineral fabrics development 109-26 Klodzko Metamorphic Complex 481 mafic dyke swarm 294 magnetic fabric relationship 370-2 Soultz granite 513 mineralogy see also magnetic mineralogy magma emplacement 245-6 mudrocks 192, 196-7 orientation distribution 326-39 mineral shapes 244 Minusa and Rybinsk depressions, Siberia 148, 169 modelling AMS development in weakly deformed mudrocks 198-9 AMS of a dyke 237-44 AMS in flowing magma particles 229-37 AMS of phyllosilicates 376-7 igneous rock fabrics 207-22 mafic dyke swarm emplacement 295, 296 magma flow direction 228, 277-9 modified Lowrie-Fuller test 267-8 monocline, Pyrenees 128, 130-2 Montes de Toldo area, Spain 381-94 monzogranite 452-3, 455 monzonite magma 467 Mora granite 382, 386-7, 388, 390-1, 391 mudrocks see also weakly deformed mudrocks classification 191-2 composition and AMS 196-8 magnetic fabric 127-43, 195-6 mudstones79-81,3tf£ multidomain (MD) particles 21-2 multiple regression parameters 64-7 mylonites 383, 389, 391 mylonitization 479 nappe pile 476-7 natural remanent magnetization (NRM) 21, 150, 155-6, 160-1 Near-Ob' crest plain, Siberia 147, 153, 155-7, 161, 168-9 neotectonics 350-2, 527-40 neutron diffraction 362 neutron texture goniometry 362, 365-6, 368-70 New England Orogen, eastern Australia 421-45 see also Texas-Coffs Harbour region nonlinear magnetizations 320-1 non-magnetic fabric analysis 362 nontensorial anisotropy 320-1 non-uniform magnetic structures 42-5 NRM see natural remanent magnetization

548

INDEX

numerical modelling see also modelling distribution anisotropy 37-9 mineral orientation 10-11 oblate cleavage fabric 493 oblate ellipsoids 521-2 oblate magnetic fabric 496 oblique magnetic fabric 493-507 OD see orientation distribution Oisquercq Formation 79, 91, 93, 101 see also Asquempont Member; Ripain Member Omanian dykes 213 optical microscopy 111-12 Ordovician, Martinsburg Formation 109-26 orientation analysis see also magnetic fabrics; preferred orientations cleaved pelites 86, 87 magnetic minerals 9-20 principal axes 91-5 structural comparison 185-7 orientation distribution (OD) 302-3, 324 AMS determination 326-7 AMS specimen tensors 331-3 anomalies 327 blended sub-fabrics 334-7 deformed rocks 299 isolating subfabrics 337-9 linear elements 342 L-S fabric scheme 328-9, 328 mineral abundances 333-4 specimen AMS axes 327-31 orientation model 206, 207 orientation tensors axial distribution 14-19, 17 estimation 72-73, 75-76 marble/limestone 53-6 numerical modelling 10-19 oroclinal bending 423, 425-6, 438, 440-1 erogenic belts Argentina 450 eastern Australia 421-45 Variscan 494-5 oxide texture 287-8, 288 pAARM see anisotropy of partial anhysteretic remanent magnetization palaeoclimatic reconstruction 168-9, 170 palaeo-fluid flow direction 502 palaeomagnetism Stoddard Mountain laccolith 257-9, 260-6, 273-6 weakly deformed mudrocks 134-5, 136-7 palaeosols see loess/palaeosol deposits palaeowind directions 169-70 paramagnetic granitoids 395-420, 448 paramagnetic limestones 531, 535 paramagnetic minerals AMS 305, 307

carriers 99-102 inverse anisotropy 311-12 phyllosilicates 182, 361 paramagnetism 308-9 particle movement, flowing magma 227-49 particulate flow deformation 347 PCOs see preferred crystallographic orientations PDOs see preferred dimensional orientations 'pedogenic' (Chinese) type loess 145-6, 170 pelitic rocks see also cleaved pelites heat treatment changes 496-502 magnetic susceptibility 77-107 regional variations in AMS fabrics 496 Upper Carboniferous siderite-bearing 493-507 pencil structures Martinsburg Formation 111, 113, 124 weakly deformed mudrocks 737, 132, 133 Pennsylvania, Martinsburg Formation 109-26 peraluminous granite 388-9 petrofabrics see also magnetic fabrics blended sub-fabrics 334-7 calcite 357, 528-30 coaxial ellipsoids 421 deformed rocks 299-360 development stages 136 low-grade metasediments 3, 181-2 Mora granite 387 sampling 303-6 terminology 301-2 petrography of EPS-1 borehole 511-14 petrology, Archaean metasediments 63 petrostructure, Toledo Complex 390-2 phyllites 486, 487 phyllosilicates 182 see also chlorite; mica AMS model 376-7 HFP 364, 374 mudrocks 198 mudstones 368 preferred orientation 361-80 qualitative correlation 375-6 quantitative correlation 376 shape parameters 375 tensors 377 pillow lavas 481 plagioclase fabric 288-94, 296 Pleistocene loess/palaeosol deposits 145-73 plutons see also batholiths; granites; granitoids central Argentina 447-74 fabrics 3 Poland, Klodzko Metamorphic Complex 475-91 Polis Rift Region, Cyprus 535-8 polyphase deformation Australia 421-4, 425

INDEX central Argentina 448-50 Variscan basement 479, 487-90 porphyritic granite 454 post-transpressional structures 178 preferred crystallographic orientations (PCOs) 1, 4 preferred dimensional orientations (PDOs) 1, 4 preferred nucleation 343 preferred orientations determination from AMS 9-20 hydrothermally altered granites 518 phyllosilicates 361-80 plagioclase in dykes 288-94 shape preferred orientation 288-94, 455 siderite crystals 502-3 Soultz granite 521 Pyrenees granite plutons 413-15 Jaca-Pamplona basin mudrocks 127-43 pyrrhotite clay-rich rocks 431, 438-9 cleaved pelites 85-6 gneiss 71, 72, 73, 74 schists 70, 72 quartz 305-6 quartzite 51-2, 56-7 quartz monzonite porphyry 256 Rayleigh law 75, 320-1 recrystallization, dynamic 345 regional scale 538 reliability anisotropy 408, 409 magnetic foliation 415, 416 magnetic lineation 407-15, 410 shape 406-7, 408, 409 Renca batholith 455, 458-9, 461-2, 463, 465 Rhenohercynian sediments 361-80, 494-5 rheology of igneous rocks 208, 209 Rhine Rift 509-26 rhyolites 486, 487, 490 rigid body rotation 341-3 Rio-Ceara Mirim dyke swarm, Brazil 285-98, 286 Ripain Member 78, 79, 82, 86-97, 99, 100 rock composition and susceptibility 324-6 rock magnetism clay-rich rocks 427-31 Soultz granite 514-17 Stoddard Mountain laccolith 259, 266-7 weakly deformed mudrocks 136-7 rock salt 52 Ronquieres Formation 79, 81, 86-102 rotation 26, 27, 28 characteristic remanence 137-40 magnetic fabrics influence 244 magnetic lineation 128 magnetic susceptibility axes 137-40

549

particles in flowing magma 231 rigid body 341-3 rotational remanent magnetization see anisotropy of... sampling, petrofabrics 303-6 saturation isothermal remanent magnetization (SIRM)317 anisotropy 38-46 Texas-Coffs Harbour region 427-31 thermal demagnetization 150, 155, 156-7, 160, 167 scanning electron microscopy 365, 366-7 schists 22 albite-chlorite 486, 487 amphibolite-facies 62, 63, 64, 65 greywacke 325 pyrrhotite 70, 72 SD see single-domain particles second-rank tensors 315-18, 383 sedimentary fabrics 3 sediments see also low-grade metasediments; mudrocks mudstones 79-81,368 Ordovician shales 109-26 Permian 440-1 Rhenohercynian 361-80, 494-5 Upper Carboniferous 493, 504, 505 self-demagnetization 312-13, 314 serpentinite 63, 66, 67 shales, Martinsburg Formation 109-26 shape anisotropy 312-13, 314 shape parameters cleaved pelites 88-91, 95-6, 97 degree of anisotropy relationship 92 diamagnetic rocks 49-50 ellipsoids 321-6 granites 459-60 loess/palaeosol deposits 162-4 low-grade metasediments 186, 188 mudrocks 198-9 phyllosilicates 375 shape preferred orientation (spo) 288-94, 455 shape reliability 406-7, 408, 409 Shawangunk Formation, Pennsylvania 110-11 shear deformation, dykes 238-9, 240 shear zones, Sierra de San Luis 468, 469 Siberian loess/palaeosol deposits 145-73 methodology and sampling 150-3 study area and climate 147-50 siderite abnormal magnetic fabrics 505 pelitic rocks 493-507 substrate-controlled growth 502 Sierra de San Luis, central Argentina 448-54, 468 Silurian slates 79-81 simple shear, North Cornwall 187-8 simultaneous folding 391-2 single-crystal processes 341-5

550

INDEX

single-domain (SD) particles 21-2, 31 foliation 41-2 lava flows 207, 221 lineation 39-41 SIRM see saturation isothermal remanent magnetization slates AMS 325 Appalachian fold belt 109-26 Borrowdale volcanic 332, 348 greenschist-facies 62, 63, 64, 65 Lower Palaeozoic 78-81, 80, 348 slaty cleavage 110, 113, 124 solid-state deformation 455, 469 Soultz-sous-Forets granite 509-26 fresh rock 512-14, 516 hydro thermally altered 514, 516-17, 523 magnetic fabrics 517-24 microfabrics 513 rock magnetic properties 514-17 Spain Montes de Toldo area 381-94 Trives and Veiga massifs 396-406 spinner magnetometer (DIGICO) 432, 434 spo see shape preferred orientation standard deviatoric susceptibility 372, 374 statistical data, Trives and Veiga massifs 397-404, 415-18 statistical significance, magnetic fabric data 395-420 Stoddard Mountain laccolith 251-83, 254-5 AMS 259-60, 263-4, 266-73, 275, 276-7 chilled zone 256, 260, 264-6, 271-2, 277 emplacement and field relations 253-7, 258, 277-9 inner zone 260-6, 277 magma flow path models 278-9 palaeomagnetism 257-9, 260-6, 273-6 strain AMS relationship 78, 104 cleavage and AMS in mudrocks 193-5 finite 339, 487-9 grain-strain 347 history 340 stretching lineation 187, 389, 391, 421, 439 strong-field remanence 321 structural data low-grade metasediments 181-2 Texas block 436 Texas-CofTs Harbour region 427, 435 Toledo Complex 384 weakly deformed mudrocks 130-2, 134-5 structural fabrics 301-2, 475-91 structure Argentinian batholiths 465-6 La Totora batholith 450-4 orientation comparison 185-7 Sierra de San Luis 468 subduction, eastern Australia 426, 438 subfabrics

bedding and cleavage 336 identification 527-40 orientation 341 substrate-controlled carbonate growth 502-5 superimposed stretching 187-8 susceptibility tensors, marble/limestone 54, 56-7 SW England, Culm Basin 175-90, 493-507 synclinorium, Pyrenees 128, 130 syn-kinematic features 469, 487 syn-orogenic basin 177 systematic-variation approach, lava flows 217-19, 220 TDS see temperature dependence of magnetic susceptibility tectonic fabrics 3, 300-2 foliation 77 lineation 389 tectonic processes aggregate processes 345-7 evolution of central Argentina 450 grain-scale processes 340-1 metamorphism 347-50 neotectonics 350-2, 527-40 single-crystal processes 341-5 temperature 309, 311 temperature dependence of magnetic susceptibility (TDS) 155, 157-60, 167, 268-9, 309, 363-5, 366 tensors see also orientation tensors bulk susceptibility properties 315-18 limestone 533-4 orientation distribution 331-3 parameters and representation 318-20, 321-6 phyllosilicates 371 susceptibility tensors 54, 56-7 terminology anisotropy and magnetism 300-1 structural and petrofabrics 301-2 Terrica beds 426-7, 437-8, 440 Texas beds 434-7 Texas block 434-7 Texas-Coffs Harbour region, Australia 421-45, 423, 424 AMS mean tensorial data 433 magnetic fabrics 432-8, 439-40 metasediment AMS 440 rock magnetism 427-31 textural reorientation 109-10 thermal demagnetization (TH) IRM 269, 270, 271, 430 SIRM 150, 155, 156-8, 160-1, 167 Soultz granite 514 thermal remanence (TRM) 313, 318 thermomagnetic heating curves 367 thermoremanent magnetization see anisotropy o f . . . thrusting Klodzko Metamorphic Complex 475-6 North Cornwall 178

INDEX titanomagnetite lava flows 206 mafic dyke swarm 287-8 volcanics 70, 72, 74 Toledo Complex, Spain 382 AMS and structural data 384 directional AMS data 386-8, 389-90 footwall AMS data 388-90, 391-2 hanging wall AMS data 385-8, 390-1 petrostructural interpretation 390-2 shear zone 391 tourmaline 311-12 Trives and Veiga massifs 396-406 magnetic foliation reliability 415, 416 magnetic lineation reliability 407-15, 410 shape reliability 406-7 TRM see thermal remanence Troodos microplate, Cyprus 528, 529 Trout Lake granite 332 turbidites, New England Orogen 426 uniform SD grain assemblages foliation 41-2 lineation 39-41 Universal stage 362 Upper Carboniferous low-grade metasediments 175-90 pelitic rocks 493-507 Upper Rhine Graben 509-26 USA igneous rocks 218, 252, 253 Martinsburg Formation 109-26 Stoddard Mountain laccolith 251-83

Variscan basement rocks 475-81 deformation 177-8, 187-8, 362, 383 granitoids 510-11 orogenic belt 494-5 Vichenet Formation 79-81, 85-95, 100-3 viscous remanent magnetization see anisotropy of volcanic arc-forearc basin-accretionary wedge complex 422, 426 weak cleavage, mudstones 140 weak-field susceptibility 315-16 weakly deformed mudrocks 127-43, 191-203 AMS data 127-8, 130, 132-6, 134-5, 195-8 characteristic remanence 136-7, 140-2 cleavage and strain 193-5 models of AMS development 198-9 palaeomagnetic data 134-5, 136-7 structural data 130-2, 134-5 weak magnetic fields 69-76 west-central Siberian transect 145-73 Widemouth Bay, North Cornwall 175, 176, 178-9, 187 wind strength, Siberian deposits 168 'wind-vigour' (Alaskan) type loess 145-6, 170 X-ray diffraction 495, 499 X-ray pole figure goniometry cleaved pelites 82, 96-9 Ordovician shales 112, 113-15, 116, 120, 121 phyllosilicates 362

551