Slope Tectonics (Geological Society Special Publication 351)

Slope Tectonics

The Geological Society of London Books Editorial Committee Chief Editor

Bob Pankhurst (UK) Society Books Editors

John Gregory (UK) Jim Griffiths (UK) John Howe (UK) Rick Law (USA) Phil Leat (UK) Nick Robins (UK) Randell Stephenson (UK) Society Books Advisors

Mike Brown (USA) Eric Buffetaut (France) Jonathan Craig (Italy) Reto Giere´ (Germany) Tom McCann (Germany) Doug Stead (Canada) Gonzalo Veiga (Argentina) Maarten de Wit (South Africa)

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

It is recommended that reference to all or part of this book should be made in one of the following ways: Jaboyedoff, M. (ed.) 2011. Slope Tectonics. Geological Society, London, Special Publications, 351. Jarman, D., Agliardi, F. & Crosta, G. B. 2011. Megafans and outsize fans from catastrophic slope failures in Alpine glacial troughs: the Malser Haide and the Val Venosta cluster, Italy. In: Jaboyedoff, M. (ed.) Slope Tectonics. Geological Society, London, Special Publications, 351, 253–277.

GEOLOGICAL SOCIETY SPECIAL PUBLICATION NO. 351

Slope Tectonics

EDITED BY

M. JABOYEDOFF University of Lausanne, Switzerland

2011 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 over 10 000. 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 Indonesian Petroleum Association (IPA), the Geological Society of America (GSA), the Society for Sedimentary Geology (SEPM) and the Geologists’ Association (GA). Joint marketing agreements ensure that GSL Fellows may purchase these societies’ publications at a discount. The Society’s online bookshop (accessible from www.geolsoc.org.uk) offers secure book purchasing with your credit or debit card. To find out about joining the Society and benefiting from substantial discounts on publications of GSL and other societies worldwide, consult www.geolsoc.org.uk, or contact the Fellowship Department at: The Geological Society, Burlington House, Piccadilly, London W1J 0BG: Tel. þ 44 (0)20 7434 9944; Fax þ 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. þ 44 (0)1225 445046, Fax þ 44 (0)1225 442836) Online bookshop: www.geolsoc.org.uk/bookshop The publishers make no representation, express or implied, with regard to the accuracy of the information contained in this book and cannot accept any legal responsibility for any errors or omissions that may be made. # The Geological Society of London 2011. 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 Ltd, Saffron House, 6 –10 Kirby Street, London EC1N 8TS, UK. Users registered with the Copyright Clearance Center, 222 Rosewood Drive, Danvers, MA 01923, USA: the item-fee code for this publication is 0305-8719/11/$15.00. British Library Cataloguing in Publication Data A catalogue record for this book is available from the British Library. ISBN 978-1-86239-324-0 Typeset by Techset Composition Ltd, Salisbury, UK Printed by MPG Books Ltd, Bodmin, UK Distributors North America For trade and institutional orders: The Geological Society, c/o AIDC, 82 Winter Sport Lane, Williston, VT 05495, USA Orders: Tel. þ 1 800-972-9892 Fax þ 1 802-864-7626 E-mail: [email protected] For individual and corporate orders: AAPG Bookstore, PO Box 979, Tulsa, OK 74101-0979, USA Orders: Tel. þ 1 918-584-2555 Fax þ 1 918-560-2652 E-mail: [email protected] Website: http://bookstore.aapg.org India Affiliated East-West Press Private Ltd, Marketing Division, G-1/16 Ansari Road, Darya Ganj, New Delhi 110 002, India Orders: Tel. þ 91 11 2327-9113/2326-4180 Fax þ 91 11 2326-0538 E-mail: [email protected]

Contents Foreword JABOYEDOFF, M., CROSTA, G. B. & STEAD, D. Slope tectonics: a short introduction

vii 1

EL BEDOUI, S., BOIS, T., JOMARD, H., SANCHEZ, G., LEBOURG, T., TRICS, E., GUGLIELMI, Y., BOUISSOU, S., CHEMENDA, A., ROLLAND, Y., CORSINI, M. & PE´REZ, J. L. Paraglacial gravitational deformations in the SW Alps: a review of field investigations, 10Be cosmogenic dating and physical modelling

11

SAINTOT, A., HENDERSON, I. H. C. & DERRON, M.-H. Inheritance of ductile and brittle structures in the development of large rock slope instabilities: examples from western Norway

27

HENDERSON, I. H. C. & SAINTOT, A. Regional spatial variations in rockslide distribution from structural geology ranking: an example from Storfjorden, western Norway

79

BO¨HME, M., SAINTOT, A., HENDERSON, I. H. C., HENRIKSEN, H. & HERMANNS, R. L. Rock slope instabilities in Sogn and Fjordane County, Norway: a detailed structural and geomorphological analysis

97

MARTINOTTI, G., GIORDAN, D., GIARDINO, M. & RATTO, S. Controlling factors for deep-seated gravitational slope deformation (DSGSD) in the Aosta Valley (NW Alps, Italy)

113

BARON˘, I., KERNSTOCKOVA´, M., NOVOTNY´, R., BURIA´NEK, D., HRADECKY´, P., HAVLIC˘EK, P. & MELICHAR, R. Palaeostress analysis of a giant Holocene rockslide near Boaco and Santa Lucia (Nicaragua, Central America)

133

JABOYEDOFF, M., OPPIKOFER, T., DERRON, M.-H., BLIKRA, L. H., BO¨HME, M. & SAINTOT, A. Complex ˚ knes rockslide, landslide behaviour and structural control: a three-dimensional conceptual model of A Norway

147

PEDRAZZINI, A., JABOYEDOFF, M., FROESE, C. R., LANGENBERG, C. W. & MORENO, F. Structural analysis of Turtle Mountain: origin and influence of fractures in the development of rock slope failures

163

HENDERSON, I. H. C., LAUKNES, T. R., OSMUNDSEN, P. T., DEHLS, J., LARSEN, Y. & REDFIELD, T. F. A structural, geomorphological and InSAR study of an active rock slope failure development

185

BIANCHI FASANI, G., DI LUZIO, E., ESPOSITO, C., MARTINO, S. & SCARASCIA-MUGNOZZA, G. Numerical modelling of Plio-Quaternary slope evolution based on geological constraints: a case study from the Caramanico Valley (Central Apennines, Italy)

201

AMBROSI, C. & CROSTA, G. B. Valley shape influence on deformation mechanisms of rock slopes

215

GHIROTTI, M., MARTIN, S. & GENEVOIS, R. The Celentino deep-seated gravitational slope deformation (DSGSD): structural and geomechanical analyses (Peio Valley, NE Italy)

235

JARMAN, D., AGLIARDI, F. & CROSTA, G. B. Megafans and outsize fans from catastrophic slope failures in Alpine glacial troughs: the Malser Haide and the Val Venosta cluster, Italy

253

Index

279

Foreword Landslides represent serious problems for humans and infrastructure throughout the world. In order to cope with these problems, knowledge about the fundamental processes leading to landslides is of major importance. This Special Publication Slope Tectonics introduces a new term dealing with structures and processes formed mainly as a result of gravitation. Tectonics is a term that is normally linked to major structural or deformational features, and their relation, origin and evolution. Research on tectonic processes has, however, not been of main focus for people working on landslide analysis, although it can be also perfectly transformed to structures formed by gravitation. For example, structural geology and tectonics has not been the main focus in disciplines such as engineering geology and rock mechanics. Geological features and processes formed by slope tectonics are often the most important controlling factor for slope instabilities and failures. A better understanding of slope tectonics and its features are of vital importance for constructing reliable geological models for landslides, which again is necessary for stability analysis and for design and implementation of reliable mitigation measures. Slope Tectonics stresses the need and importance of a better understanding of slope tectonics, and the controlling geological structures for hazard mapping and analysis of large landslides. It presents a variety of aspects, including regional hazard analysis, structural geology, the evolution of slope failures, rock mass characteristics and strength, site-specific geological failure models, numerical modelling and the use of new remote-sensing technologies. The 13 papers and the introductory paper are based on presentations given at a research workshop

at the University of Lausanne, and are written by scientists from active landslide research groups in Italy, Switzerland, Norway, France, Czech Republic and Canada. They present a series of excellent case studies from Norway, Italy, France, Central America and Canada. The presented cases demonstrate the need for integrating different disciplines for a better understanding of landslide processes and their hazards. Slope Tectonics is edited by Professor and Director Michel Jaboyedoff from the Institute of Geomatics and Analysis of Risk at the University of Lausanne, Switzerland. Michel Jaboyedoff has a wide and long experience in landslide research, and is particularly well known for his interdisciplinary skills, creativity, ability to use new technology and to implement scientific results into practical use. In addition, he has, over the last few years, built up an excellent landslide group at the University of Lausanne. As a scientist responsible for the monitoring and early warning of large rockslides in inhabited areas in Norway, I hope that this book, Slope Tectonics, will generate enthusiasm and increased research activity of importance for hazard and risk handling in mountainous areas and fjord regions throughout the world.

Lars Harald Blikra Professor and Chief Geologist A˚knes/Tafjord Early Warning Centre and Sogn and Fjordane University College Norway April, 2010

Slope tectonics: a short introduction MICHEL JABOYEDOFF1*, GIOVANNI B. CROSTA2 & DOUG STEAD3 1

2

Institut de Ge´omatique et d’Analyse des Risques, University of Lausanne, Amphioˆle, 1015 Lausanne, Switzerland

Dipartimento di Scienze Geologiche e Geotecnologie, Universita` degli Studi di Milano Bicocca, Milano, Italy

3

Department of Earth Sciences, Simon Fraser University, Burnaby, B.C., Canada V5A 1S6 *Corresponding author (e-mail: [email protected])

Geomorphology, structural geology and engineering geology allow description of the main characteristics of a slope in distinct ways that can be combined to provide a complementary view of the operative slope processes. The subjects presented in this Special Publication include: slope morphology and evolution; mechanical behaviour of the material; modes of failure and collapse; influence of lithology and structural features; and the role played by controlling factors. This Slope Tectonics volume comprises a series of very different contributions that attempt to underline a multidisciplinary approach that should form the framework of slope instability studies. Slope Tectonics is adopted in this volume to mean deformation that is induced or fully controlled by the slope morphology and that generates features that can be compared to tectonic features. The stress field in a slope is the result of gravity, topography and the geological setting created by an ensemble of geodynamic processes. Active tectonics (also called neotectonics) generates a stress field that can control slope processes; a strong feedback existing between geological history, tectonics, lithology, geomorphological evolution and topography. As a consequence, a list of factors and their relative influence can be presented. (1)

Fabric induced by a local stress field within a slope: † discontinuities and local faults with cataclastic bands of variable thickness; † folds (Fig. 1), associated predominantly with brittle structures; † complex failure paths (stepped or multisurface); † local failures: rock bridge failures or extensional failures (graben-like or pseudograben-like); † subsidence due to weak or soluble materials causing complex sliding –toppling phenomena.

(2)

(3)

Reactivation of pre-existing faults, discontinuities, joints, foliations or rock anisotropies: † surfaces characterized by residual or lower than peak strength; † formation of composite failure surfaces. Regional tectonic movements inducing new slope morphologies: † uplift; † major fault movement; † pull-apart zones: † folding.

The boundary between classical tectonics and slope deformations, especially at a large scale, has always been indistinct as emphasized by Antoine (1988) (Schultz-Ela 2001). In his paper Antoine discussed mechanisms like ‘diverticulation’ defined by Lugeon (1943) and Badoux (1963), where part of Pre-Alpine nappes were reversed in geometry by huge landslides inverting the stratigraphy (Antoine 1988). A remarkable geometrical analogy exists between basin extensional tectonics (Wernicke 1981) and certain landslide spreading in clays (Voight 1973; Varnes 1978; Hutchinson 1988); this is despite the fact that Wernicke’s hypothesis demonstrated that Basin and Range regions were not produced by huge landslides, but by low-angle faulting induced by geodynamic processes. Regional extension has produced changes in topography and, as a consequence, significant gravity-induced deformations. Therefore, we suggest that the term ‘slope tectonics’ is justified and must be recognized as an important component in slope deformation. Slope instability implies movements driven by gravity that can produce irreversible deformations. In the past, slopes were viewed as privileged erosion zones (De la Noe & De Margerie 1888; Strahler 1950), whereas few erosion processes were attributed to landslides (Young 1972). Since the work of Simonett (1967) and Hovius et al. (1997), the link

From: Jaboyedoff, M. (ed.) Slope Tectonics. Geological Society, London, Special Publications, 351, 1 –10. DOI: 10.1144/SP351.1 0305-8719/11/$15.00 # The Geological Society of London 2011.

2

M. JABOYEDOFF ET AL.

Fig. 1. Kink-fold in the front of a translational landside in foliated amphibolites of northern Norway. This shows a similar structure to folds in thin skin tectonics style except that the lithostatic stress is not sufficient to close the voids. (From Braathen et al. 2004; reproduced with permission from the authors and Norwegian Journal of Geology.)

between landslides and erosion processes has been emphasized by various authors (Korup et al. 2010). This linkage has been suggested both for shallow and deep-seated landslides. In the first case, shallow landslides frequently evolve into debris flows along the slopes, allowing a faster and more complete transport to the fluvial network. For deepseated landslides the processes act in a slightly different way causing a slow removal of the material from the slope. In these cases the erosion and the rate of erosion induced by landslides apply over a different timescale and very often can be considered as a means for both morphological stabilization, through a decrease in slope gradient, and the introduction of marked changes in valley morphology and river sediment yield. Static and dynamic equilibrium can be important controls on the evolution of very large slope deformations. In fact, rivers are linear elements incising the topography and which do not erode the entire slope topographic surface. The slope surface is affected by erosive processes that, in fact, affect the entire topographic surface. As a consequence, diffused erosion or mass movements are required to displace material away from slopes into the fluvial system. Small and large slope instabilities clearly act differently in degradation of slopes and mountainous topography according to their volume, density and temporal frequency, and to the degree of disturbance that they cause to the slope materials. Gravity plays an important role in large deformations of the Earth’s crust (De Jong & Scholten 1973; Schultz-Ela 2001). Molnar & Lyon-Caen

(1988, 1989) showed that during mountain belt building, above a certain thickness of the Earth’s crust, the gravitational stress may be higher than the compressive strength. As a consequence, the mountain belts cease to elevate and begin to spread laterally (Delacou et al. 2004). Theoretically, this phenomenon can be considered to be similar to cases studied by Savage et al. (1985) and Savage & Varnes (1987) (Fig. 2). This hypothesis is also supported by the results of analogue modelling (Bachmann et al. 2009). Lugeon & Oulianoff (1922) have described the role of gravity-induced deformations on rock outcrops (Fig. 3) and began the documentation of a considerable number of large slope movements (Heim 1932). Terzaghi (1963) identified deepseated landsliding from both a mechanical and kinematic point of view. Subsequently, it has been shown by analytical solutions that gravitational stresses can induce slope deformations (Savage et al. 1985; Savage & Varnes 1987) (Fig. 4). More recent models support these findings for more complex geometries and material properties (Kinakin & Stead 2005; Ambrosi & Crosta 2006). Nevertheless, new techniques available for the observation of ground movements (e.g. InSAR (Interferometric Synthetic Aperture Radar), LiDAR (Light Detection And Ranging) and highresolution satellite imagery) provide the opportunity to show that the valley flanks within major mountain belts are moving both slowly and continuously over large areas (Singhroy 1995; Carnec et al. 1996; Dehls et al. 2002; Colesanti et al. 2003; Ambrosi

INTRODUCTION

3

Fig. 2. (a) Extension and compression zones deduced from the analytical model (modified after Savage & Varnes 1987). (b) Illustration of a DSGSD that presents similar features to those predicted by the model shown in (a) (modified after Agliardi et al. 2009).

& Crosta 2006; Colesanti & Wasowski 2006; Troisi 2007; Crosta et al. 2008b; Osmundsen et al. 2009) in a form of dynamic equilibrium that persists over thousands of years. The present landslides ‘paradigm’ did not predict this because most of the studies of mass movements were dedicated to areas displaying clearly present or past slope movements that were identified based on morphology. However, unfortunately, a lot of structures were either not identified or were confused with regional tectonic features. Again, the spatial and temporal scale typical of such phenomena suggest a strong link between gravity and tectonics, as well as between local morphological changes and the tectonic evolution of mountain belts (Crosta et al. 2009; see also Ambrosi & Crosta et al. 2011).

Structures associated with slope deformation Large slope movements have been encountered during dam projects (Lugeon 1933; Gignoux & Barbier 1955; Desio 1961) creating problems for dam stability and in meeting the impermeability requirements for foundation materials (Fig. 5). Drilling and excavation of hard rock masses at the slope toe has revealed that they can overlie recent alluvial and glacial deposits (Desio 1961; Terzaghi 1963), emphasizing that slow large rock mass movements can involve exceptional displacements (Figs 2b & 5). Lugeon & Oulianoff (1922) have shown the importance of toppling movements affecting entire

Fig. 3. Cross-sections showing three examples of ‘toppling’ in a slope near Martigny Switzerland or ‘Trois exemples de balancement superficiel des couches dans un versant’. (From Lugeon & Oulianoff 1922; reproduced with permission from the Socie´te´ Vaudoise des Sciences Naturelles.) The thickness of the affected zone can vary from a few metres to hundreds of metres.

4

M. JABOYEDOFF ET AL.

Fig. 4. Deep-seated gravitational slope deformation proposed by Terzaghi (1963).

slopes through the analysis of Alpine structures (Fig. 3). As previously noted, the observed deformational structures associated with slope movements are well documented, and have been variously described as ramps, folds thrusts, horst, graben and cataclastic bands (Choffat 1929; Terzaghi 1950; Voight 1973; Varnes 1978; Hutchinson 1988; Chigira 1992). Cruden (2000) made an attempt to classify slope-bedding attitude and potential failure mechanisms. Deformations at the level of large Alpine slopes were first observed during the 1940s (Ampferer 1939; Stini 1941; Terzaghi 1963), and the deep-seated gravitational slope deformations (DSGSD) were studied in detail by Zischinsky (1966, 1969), Nemcˇok et al. (1972) and Mahr & Nemcˇok (1977). Radbruch-Hall (1978) and Zaruba & Mencl (1982) summarized most of these features, providing an excellent illustration of uphill-facing scarps that was subsequently studied in detail by Bovis (1982). Recent work by Agliardi et al. (2001) and Ambrosi & Crosta (2006) shows the

importance of progressive slope failure mechanisms, and the close link between pre-existing tectonic structures and glacial unloading. They also showed that such slope deformations are the preferential site for the development of large landslides. A link between regional structural features and DSGSD is strongly suspected (Agliardi et al. 2001, 2009; Ambrosi & Crosta 2006) but remains enigmatic (Crosta & Zanchi 2000). It should be noted that in many slides the deformation is localized within a weak zone that is frequently coincident with bedding planes or foliation (Heim 1932; Cruden 1976; Braathen et al. 2004) or pre-existing lithological and tectonic features (Ambrosi & Crosta 2006). In addition, Agliardi et al. (2001) showed how active landslides can be found at the toe of slopes in extremely developed DSGSDs. These phenomena have been described as mass rock creep by Chigira (1992) and Chigira & Kiho (1994), who illustrated how many tectonic-like features (brittle faults and folds, cataclastic bands) can

Fig. 5. Cross-section showing the presence of an alluvial deposit below the sliding mass in the French Alps (modified after Gignoux & Barbier 1955).

INTRODUCTION

be produced at the slope scale by mass movement. These deformations lead to the degradation of the rock mass and are indicators of pre-failure and failure mechanisms. Chigira (1992) described these features as occurring in slopes of very different size, from a few tens of metres to thousands of metres in height. Depending on the material, the morphology and the presence of weakness zones or other external factors, the scale of gravitational deformations can vary from small to large sections of slope to entire mountain flanks and ridges (Beck 1968; Massironi et al. 2003) or multiple ridges and catchments (Giannini 1951; Argnani et al. 2003; Crosta et al. 2008a). Many ‘fossil mass movements’ can be recognized in sedimentary sequences in the form of olistostromes (Stow 1986), large slumps and chaotic deposits. The study of submarine landslides and of sedimentary formations show that such phenomena affect continental margins over distances of up to several hundreds of kilometres (Hampton et al. 1996; Locat & Meinert 2003), and that shear bands, folds and chaotic structures can exist at different scales. Finally, the limit between gravitational processes and purely tectonically induced movements can be quite difficult to identify (Fig. 6).

Mechanisms and processes The slope stability analyses performed for these types of phenomena require specific comment owing to the varied mechanical processes, and the variations in properties in space and time. Creep, progressive failure and degradation, spatial extent and continuity of the failure surface, and its changes with time (e.g. from a confined condition, as stated by Hutchinson 1988, to fully developed failure surfaces), together with the changes in type and intensity of perturbations with time, all need to be considered when dealing with DSGSD phenomena.

5

The role of geological structures and the importance of pore fluid pressure in the generation and formation of otherwise inexplicable structures was investigated in early work carried out by Hubbert & Rubey (1959) and Rubey & Hubbert (1959). Inherited structures like bedding, joints, faults, folds and weak rocks control many slope instabilities (Agliardi et al. 2001; Brideau et al. 2009), but an important issue that must be considered to further the understanding of gravity-driven deformations is time-dependency characteristics and the occurrence of pre-failure mechanisms (Petley & Allison 1997; Sjo¨berg 2000; Leroueil 2001). It is clear that the failure of most slopes is, in fact, time dependent (Agliardi et al. 2001; Eberhardt et al. 2004; Stead et al. 2006; Eberhardt 2008); but the true nature of this time-dependent failure is not yet well understood, sometimes requiring an assumed mode of progressive shear strength decrease in order to simulate such phenomena (Eberhardt et al. 2004). The failure of rock bridges (Scavia 1995) is not yet used routinely in slope stability studies owing to difficulties in determining in the field the percentage of intact rock bridges within a slope. Most work to date has considered two-dimensional rock ridges; yet, the three-dimensionality of rock bridges adds a further degree of complexity to rock slope analysis as noted by Elmo et al. (2007). Nevertheless, the concept of progressive failure is interesting (Bjerrum 1967) because the use of localized failure is suitable for explaining many observed phenomena (Eberhardt et al. 2004) both in soil and in rock. Recent discrete element modelling approaches have emphasized the importance of considering fracture propagation in the failure of large rock slopes and the effect of progressive rock mass damage (Stead et al. 2006; Alzo’ubi 2009; Brideau et al. 2009; Lorig et al. 2009; Vyazmensky et al. 2010). The geometry of observed deformations and structures both within and at the surface of the examined slopes is not always scale dependent. In fact, the rheology of a material makes it possible

Fig. 6. Sketch illustrating landslides affecting valley slopes and regions indicating the different orders of landslide deformation.

6

M. JABOYEDOFF ET AL.

for a similar structure to be present at the centimetre level or at the mountain chain level, as demonstrated for faults by Riedel (1929) and by Skempton (1966). The geometry of the failure surface is strongly dependent on the boundary condition of the system. For instance, the La Clapie`re landslide (France) developed a failure mechanism involving toppling failure during its first evolutionary phase (Giraud et al. 1990) followed by the development of a continuous sliding surface. The driving mechanism could have been the dissolution and deformation of the gypsum and anhydrite (Fig. 7) lying underneath the slope (Guglielmi et al. 2000). Rock slope weakening and the landslide activity can be linked to tectonic activity, such as tectonic uplift and uplift gradient, or to seismic activity (Keefer 1984; Burbank et al. 1996; Crosta et al. 2008a, 2009; Hovius et al. 2009), and to water-table changes and cycles (Zangerl et al. 2003; Jaboyedoff et al. 2009). The importance of considering three-dimensional (3D) influences on large-scale rock slopes and landslides, particularly when modelling slope deformations, includes not only kinematical release provided by structures but the influence of 3D slope geometry, water pressures and effects on seismicity. The importance of considering 3D effects in the modelling of the kinematics of large rock slopes is demonstrated by Brideau & Stead (2010), Brideau (2010) and Ambrosi & Crosta (2011).

In summary, the slope tectonics approach proposed herein consists of the location and identification in the field of features associated with pre-slope failure and failure deformation phases, and features that provide information on the transition from one phase to the other, as well as the influence of controlling factors.

Volume contents This volume includes a selection of contributions presented at the Slope Tectonics Conference, which was held at the University of Lausanne (Switzerland) on 15– 16 February 2008. These contributions provide an indication of the current state of knowledge and of the most recent research progress in the field we refer to as ‘Slope Tectonics’. The following provides a brief overview of the various manuscripts included. The first paper presents an overview of the problem of understanding time-dependent slope deformation. El Bedoui et al. propose to study the incipient deformation history of a slope after deglaciation. Using their multidisciplinary approach, starting from field investigation to modelling, they show that the slopes are progressively destabilized along pre-existing structures but also through progressive failures. They support the hypothesis of the development of secondary landsliding within DSGSDs.

Fig. 7. Illustration of different potential interaction mechanisms between weak zones in a rock mass with resulting slope instability and slope deformation. The slope can be stressed by (a) undercut, (b) weak zone or (c) material deformation located below the slope.

INTRODUCTION

The following papers are dedicated to reactivation of ancient structures. Saintot et al. demonstrate the importance of pre-existing structures, such as faults filled with gouges developed along the foliation planes. In a regional study in Norway, Henderson & Saintot show how pre-existing structures control the slope susceptibility to rockslides in the Storfjorden area (western Norway). Bo¨hme et al. show how the development of rock instabilities promoted by several different mechanisms in Sogn and Fjordane County (Norway) are often linked to the knick point along a slope profile. Martinotti et al. suggest that the dissolution and deformation of rocks can induce progressive slope failures (Aosta Valley, NW Alps). Several papers deal with the development of landslides promoted by pre-existing structures that control the formation of a series of induced structures. Baron et al. show how palaeostress at the regional scale can be controlled by landsliding and gravitational deformation in the volcanic province of Boaco and Santa Lucia (Nicaragua, Central ˚ knes America). A 3D conceptual model of the A (Norway) landslide is proposed by Jaboyedoff et al. This paper shows that several pre-existing structures control the stepped failure surface, the lateral release plane and the movement direction. The failure surfaces activated by the slope movement follow weaknesses zones within the rock mass. Pedrazzini et al. describe the influence of pre-existing discontinuities linked to regional folding in slope stability of the Frank Slide scar (Turtle Mountain, Canada). Henderson et al., who show that landslide activity can develop along active faults, support the important role played by active pre-existing structures. The InSAR technique is adopted for this purpose at a site in Norway. The role of tectonic activity is also illustrated for the case study of the Caramanico area of the Central Apennines, as discussed by Bianchi Fasani et al. The use of numerical models for the understanding of the role played by morphology and preexiting structures is presented in two papers. Ambrosi & Crosta, using 3D mechanical modelling, investigate the influence of the orientation of weak elements within the rock mass (e.g. jointing, foliation, layering) and of slope morphology on the development of unstable sectors. Their results are supported by observations made in the field at different sites in the European Alps. In a case study Ghirotti et al. demonstrate the agreement between mechanical modelling and field observations at the Celentino deep-seated gravitational slope deformation (Peio Valley, NE Italy). Finally, Jarman et al., by inspection of anomalous fans and of surrounding slope morphologies in Val Venosta (east-central Alps, Italy), suggest how

7

these could be reconstructed to demonstrate the occurrence of very large rock avalanches. The Slope Tectonics 2008 Congress received financial and logistical support from the Faculte´ des ge´osciences et de l’environnement of the University of Lausanne. The editor is grateful to the Advisory Committee, which supported this initiative: L.H. Blikra, Ch. Bonnard, A. Braathen, M. Chigira, J. Coe, R. Couture, M.-H. Derron, S.G. Evans, C.R. Froese, I. Henderson, V. Labiouse, J. Locat and A. Saintot, who gave their scientific support to the organization. We also thank the local group that helped us with the organization: J.-L. Epard, F. Baillifard, T. Oppikofer and A. Pedrazzini (IGAR). M. Jaboyedoff is also grateful to his colleagues at the Institute of Geomatics and Analysis of Risk at UNIL: P. Eraso, P. Horton, A. Loye, L. Perozzi, C. Schrocker and I. Spinello. We thank the reviewers without whom no publication would have been possible: F. Agliardi, A. Braathen, M.-A. Brideau, M. Chigira, J. Coe, R. Couture, E. Eberhardt, I. Evans, M. Geertsema, M. Giardino, J.S. Griffiths, A. Gu¨nther, R. Hermanns, M. Hurlimann, J. Hutchinson, O. Korup, C. Longchamp, A. Pedrazzini, G. Scarascia Mugnozza, A. Saintot, C. Squarzoni, and K. Sudmeier. Finally, we thank all the participants.

References Agliardi, F., Crosta, G. & Zanchi, A. 2001. Structural constraints on deep-seated slope deformation kinematics. Engineering Geology, 59, 83– 102. Agliardi, F., Zanchi, A. & Crosta, G. 2009. Tectonic v. gravitational morphostructures in the central Eastern Alps (Italy): Constraints on the recent evolution of the mountain range. Tectonophysics, 474, 250– 270. Alzo’ubi, A. M. 2009. The effect of tensile strength on the stability of rock slopes. Doctoral thesis, University of Alberta. Ambrosi, C. & Crosta, G. B. 2006. Large sackung along major tectonic features in the Central Italian Alps. Engineering Geology, 83, 183–200. Ambrosi, C. & Crosta, G. B. 2011. Valley shape influence on deformation mechanisms of rock slopes. In: Jaboyedoff, M. (ed.) Slope Tectonics. Geological Society, London, Special Publications, 351, 215–233. ¨ ber einige Formen der Ampferer, O. 1939. U Bergzerreißung. Sitzungsberichte der Akademie der Wissenschaften Wien. Mathematisch– Naturwissenschaftliche Klasse, 148, 1 –14. Antoine, P. 1988. Opening lecture: Some unexpected features of landslides. In: Bonnard, C. (ed.) Proceedings of the 5th International Symposium on Landslides, Lausanne, Switzerland, 10– 15 July 1988. AA Balkema, Rotterdam, 1389–1392. Argnani, A., Barbacini, G. et al. 2003. Gravity tectonics driven by Quaternary uplift in the Northern Apennines: insights from the La Spezia– Reggio Emilia geo-transect. Quaternary International, 101– 102, 13–26.

8

M. JABOYEDOFF ET AL.

Bachmann, D., Bouissou, S. & Chemenda, A. 2009. Analysis of massif fracturing during Deep-Seated Gravitational Slope Deformation by physical and numerical modeling. Geomorphology, 103, 130– 135. Badoux, H. 1963. Les unite´s ultrahelve´tiques de la Zone des Cols. Eclogae geologicae Helvetiae, 56, 1– 13. Beck, A. C. 1968. Gravity faulting as a mechanism of topographic adjustment. New Zealand Journal of Geology and Geophysics, 11, 191– 199. Bjerrum, L. 1967. Progressive failure in slopes of overconsolidated plastic clay and clay shales. Journal of the Soil Mechanics and Foundations Division ASCE, 93, 1– 49. Bovis, M. J. 1982. Uphill-facing (antislope) scarps in the Coast Mountains, southwest British Columbia. Geological Society of American Bulletin, 93, 804– 812. Braathen, A., Blikra, L. H., Berg, S. S. & Karlsen, F. 2004. Rock-slope failures in Norway; type, geometry and hazard. Norwegian Journal of Geology, 84, 67–88. Brideau, M.-A. 2010. Three dimensional kinematic controls on rock slope stability conditions. Doctoral thesis, Simon Fraser University, Vancouver, BC. Brideau, M.-A. & Stead, D. 2010. Controls on block toppling using a 3-dimensional distinct element approach. Rock Mechanics and Rock Engineering, 43, 241–260. Brideau, M.-A., Yan, M. & Stead, D. 2009. The role of tectonic damage and brittle rock fracture in the development of large rock slope failure. Geomorphology, 103, 30–49. Burbank, D. W., Leland, J., Fielding, E., Anderson, R. S., Brozovic, N., Reid, M. R. & Duncan, C. 1996. Bedrock incision, rock uplift and threshold hillslopes in the northwestern Himalayas. Nature, 379, 505– 510. Carnec, C., Massonet, D. & King, C. 1996. Two examples of the use of SAR interferometry on displacement fields of small spatial extent. Geophysical Research Letters, 23, 3579–3582. Chigira, M. 1992. Long-term gravitational deformation of rocks by mass rock creep. Engineering Geology, 32, 157– 184. Chigira, M. & Kiho, K. 1994. Deep-seated rockslideavalanche preceded by mass rock creep of sedimentary rocks in the Akaishi Mountains, central Japan. Engineering Geology, 38, 221– 230. Choffat, P. 1929. L’e´croulement d’Arvel (Villeneuve) de 1922. Bulletin de la Socie´te´ Vaudoise des Sciences Naturelles, 57, 5– 28. Colesanti, C., Ferretti, A., Prati, C. & Rocca, F. 2003. Monitoring landslides and tectonic motions with the Permanent Scatterers Technique. Engineering Geology, 68, 3– 14. Colesanti, C. & Wasowski, J. 2006. Investigating landslides with space-borne Synthetic Aperture Radar (SAR) interferometry. Engineering Geology, 88, 173– 199. Crosta, G. & Zanchi, A. 2000. Deep-seated slope deformations: huge, extraordinary, enigmatic phenomena. In: Bromhead, E., Dixon, N. & Isben, M.-L. (eds) Landslides in Research, Theory and Practice. Proceedings of the 8th International Symposium on

Landslides, Cardiff, UK, June 2000, Volume 1. Thomas Telford, London, 351 –358. Crosta, G. B., Agliardi, F., Frattini, P. & Zanchi, A. 2008a. Alpine inventory of Deep-Seated Gravitational Slope Deformations. Geophysical Research Abstracts, 10, EGU2008-A-02709. Crosta, G. B., Allievi, J., Frattini, P., Giannico, C. & Lari, S. 2008b. Monitoring Deep-Seated Gravitational Slope Deformations in the Alps by PSInSAR techniques. Geophysical Research Abstracts, 10, EGU2008-A-10662. Crosta, G. B., Agliardi, F., Frattini, P. & Malusa`, M. 2009. Deep-seated slope deformations, exhumation and relief in the Alps. Geophysical Research Abstracts, 12, EGU2010-13796. Cruden, D. M. 1976. Major rock slide in the Rockies. Canadian Geotechnical Journal, 13, 8 –20. Cruden, D. M. 2000. Some forms of mountain peaks in the Canadian Rockies controlled by their rock structure. Quaternary International, 68–71, 59–65. Dehls, J. F., Basilico, M. & Colesanti, C. 2002. Ground deformation monitoring in the Ranafjord area of Norway by means of the permanent scatterers technique. In: Proceedings of the IEEE International Geoscience and Remote Sensing Symposium (IGARSS 2002), Toronto, Canada, 1, 203 –207. doi: 10.1109/ IGARSS2002.1024988. De Jong, K. A. & Scholten, R. 1973. Gravity and Tectonics. Wiley, New York. Delacou, B., Sue, C., Champagnac, J.-D. & Burkhard, M. 2004. Present-day geodynamics in the bend of the western and central Apls as constrained by earthquake analysis. Geophysical Journal International, 158, 753–774. De La Noe, G. & De Margerie, E. 1888. Chap. III: Du fac¸onnement des versants. In: Les formes du Terrain. Imprimerie Nationale, Paris, 39– 47. Desio, A. 1961. Geological features of the reservoirs and of the dam foundation rocks. In: Association of Italian Electric Producer and Utility Companies (ed.) Dams for Hydroelectric Power in Italy. Unpublished. Eberhardt, E. 2008. Twenty-ninth Canadian Geotechnical Colloquium: the role of advanced numerical methods and geotechnical field measurements in understanding complex deep-seated rock slope failure mechanisms. Canadian Geotechnical Journal, 45, 484– 510. Eberhardt, E., Stead, D. & Coggan, J. S. 2004. Numerical analysis of initiation and progressive failure in natural rock slope – the 1991 Randa rockslide. International Journal of Rock Mechanics & Mining Sciences, 41, 69–87. Elmo, D., Yan, M., Stead, D. & Rogers, S. 2007. The importance of intact rock bridges in the stability of high rock slopes: A quantitative investigation using an integrated numerical modelling – discrete fracture network approach. In: International Symposium on Rock Slope Stability in Open Pit Mining and Civil Engineering, 12–14 September 2007, Perth, Australia. Australian Centre for Geomechanics, University of Western Australia, Perth, 253–266. Giannini, E. 1951. Alloctonia del macigno nella regione Bagni di Casciana-Chianni in Provincia di Pisa, Boll. Geological Society Italy, 70, 79–94.

INTRODUCTION Gignoux, M. & Barbier, R. 1955. Ge´ologie des barrages et des ame´nagements hydrauliques. Masson et Cie, Paris. Giraud, A., Rochet, L. & Antoine, P. 1990. Processes of slope failure in crystallophyllian formations. Engineering Geology, 29, 241– 253. Guglielmi, Y., Bertrand, C., Compagnon, F., Follacci, J. P. & Mudry, J. 2000. Acquisition of water chemistry in a mobile fissured basement massif: its role in the hydrogeological knowledge of the La Clapiere landslide (Mercantour massif, southern Alps, France). Journal of Hydrology. 229, 138–148. Hampton, M. A., Lee, H. J. & Locat, J. 1996. Submarine landslides. Reviews of Geophysics, 34, 33– 59. Heim, A. 1932. Bergsturz und Menschenleben. Fretz & Wasmuth, Zurich. Hovius, N., Meunier, P., Lin, C. W., Dadson, S. J., Chen, H. & Horng, M. J. 2009. Seismically induced erosion and the mass balance of a large earthquake. Geophysical Research Abstracts, 11, EGU2009-9183. Hovius, N., Stark, C. P. & Allen, P. A. 1997. Sediment flux from a mountain belt derived by landslide mapping. Geology, 25, 231–234. Hubbert, R. K. & Rubey, W. W. 1959. Role of fluid pressure in mechanics of overthrust faulting: I. Mechanics of fluid-filled porous solids and its application to overthrust faulting. Bulletin of Geological Society of America, 70, 115– 166. Hutchinson, J. N. 1988. General report: Morphological and geotechnical parameters of landslides in relation to geology and hydrogeology. In: Bonnard, C. (ed.) Proceedings of the 5th International Symposium on Landslides, Lausanne, Switzerland, 10–15 July 1988. AA Balkema, Rotterdam, 3–35. Jaboyedoff, M., Couture, R. & Locat, P. 2009. Structural analysis of the Turtle Mountain (Alberta) using digital elevation model: Toward a progressive failure. Geomorphology, 103, 5 –16. Keefer, D. K. 1984. Landslides caused by earthquake. Geological Society of America Bulletin, 95, 406–421. Kinakin, D. & Stead, D. 2005. Analysis of the distributions of stress in natural ridge forms: implications for the deformation mechanisms of rock slopes and the formation of sacking. Geomorphology, 65, 85– 100. Korup, O., Densmore, A. L. & Schlunegger, F. 2010. The role of landslides in mountain range evolution. Geomorphology, 120, 77–90. Leroueil, S. 2001. Natural slope and cuts: movement and failure mechanisms. Ge´otechnique, 51, 197– 243. Locat, J. & Meinert, J. 2003. Submarine mass movements and their consequences. In: First International Symposium on Advances in Natural and Technological Hazards Research, Volume 19. Kluwer Academic, Dordrecht. Lorig, L., Stacey, P. & Read, J. 2009. Chapter 10: Slope design methods. In: Read, J. & Stacey, P. (eds) Guidelines for Open Pit Slope Design. CSIRO Publishing, Collingwood, Victoria, Australia, 237– 264. Lugeon, M. 1933. Barrages et ge´ologie – Me´thodes de recherches – Terrassement et imperme´abilisation.

9

Librairies de l’Universite´, F. Rouge & Cie S.A., Lausanne. Lugeon, M. 1943. Une nouvelle hypothe`se tectonique: la diverticulation. Bulletin de la Socie´te´ Vaudoise des Sciences Naturelles, 62, 301–303. Lugeon, M. & Oulianoff, N. 1922. Sur le balancement superficiel des couches et sur les erreurs que ce phe´nome`ne peut faire commettre. Bulletin de la Socie´te´ Vaudoise des Sciences Naturelles, 54, 381–388. Mahr, T. & Nemcˇok, A. 1977. Deep-seated creep deformations in the crystalline cores of the Tatry Mts. Bulletin of the International Association of Engineering Geology, 16, 104–106. Massironi, M., Bistacchi, A. V., Dal Piaz, G., Monopoli, B. & Schiavo, A. 2003. Structural control on mass-movement evolution: A case study from the Vizze Valley, Italian Eastern Alps. Eclogae Geologicae Helvetiae, 96, 85– 98. Molnar, P. & Lyon-Caen, H. 1988. Somme simple physical aspects of the support, structure, and evolution of mountain belts. In: Clark, S. P., Jr (ed.) Processes in Continental and Lithospheric Deformation. Geological Society of America, Special Paper, 218, 179–207. Molnar, P. & Lyon-Caen, H. 1989. Fault plane solutions of earthquakes and active tectonics of the Tibetan Plateau and its margins. Geophysical Journal International, 99, 123 –153. Nemcˇok, A., Pasˇek, J. & Ryba´rˇ, J. 1972. Classification of landslides and other mass movements. Rock Mechanics, 4, 71–78. Osmundsen, P. T., Henderson, I., Lauknes, T. R., Larsen, Y., Redfield, T. F. & Dehls, J. 2009. Active normal fault control on landscape and rockslope failure in northern Norway. Geology, 37, 135– 138. Petley, D. N. & Allison, R. J. 1997. The mechanics of deep-seated landslides. Earth Surface Processes and Landforms, 22, 747– 758. Riedel, W. 1929. Zur Mechanik Geologischer Brucherscheinungen. Zentralblatt fur Mineralogie, Geologie und Paleontologie B, 1929, 354– 368. Radbruch-Hall, D. H. 1978. Gravitational creep of rock masses on slopes. In: Voight, B. (ed.) Rockslides and Avalanches, 1 – Natural Phenomena. Developments in Geotechnical Engineering, 14A, 607– 657. Rubey, W. W. & Hubbert, M. K. 1959. Role of fluid pressure in mechanics of overthrust faulting, 2. Geological Society of America Bulletin, 70, 167–206. Savage, W. Z. & Varnes, D. J. 1987. Mechanics of gravitational spreading of steep-sided ridges (‘sacking’). Bulletin of the International Association of Engineering Geology, 35, 31– 36. Savage, W. Z., Swolfs, H. S. & Powers, P. S. 1985. Gravitational stresses in long symmetric ridges and valleys. International Journal of Rock Mechanics and Mining Sciences & Geomechanics Abstracts, 22, 291– 302. Scavia, C. 1995. A method for the study of crack propagation in rock structures. Geotechnique, 45, 447–463. Schultz-Ela, D. D. 2001. Excursus on gravity gliding and gravity spreading. Journal of Structural Geology, 23, 725 –731.

10

M. JABOYEDOFF ET AL.

Simonett, D. S. 1967. Landslide distribution and earthquakes in the Bewani and Torricelli Mountains. In: Jennings, J. N. & Mabbutt, M. A. (eds) New Guinea in Landform Studies from Australia and New Guinea. Cambridge University Press, Cambridge, 64–84. Singhroy, V. 1995. SAR integrated techniques for geohazard assessment. Advances in Space Research, 15, 1167–1178. Sjo¨berg, J. 2000. Failure mechanism for high slopes in hard rock. In: Hustrulid, W. A., McCarter, M. K. & Van Zyl, D. J. A. (eds) Slope Stability in Surface Mining. Society for Mining, Metallurgy, and Exploration, Inc., Littleton, CO, 71–80. Skempton, A. W. 1966. Some observations on tectonic shear zones. In: Proceedings of the First International Congress of the International Society of Rock Mechanics, Lisbon, Volume 1. Laborato´rio Nacional de Engenharia Civil, Lisbon, 329 –335. Stead, D., Eberhardt, E. & Coggan, J. S. 2006. Developments in the characterization of complex rock slope deformation and failure using numerical modelling techniques. Engineering Geology, 83, 217– 235. Stini, J. 1941. Unsere Taler wachsen zu: Geologie und Bauwesen. 18, 71–79. Stow, D. A. V. 1986. Deep clastics seas. In: Reading, H. G. (ed.) Sedimentary Environments and Facies. Blackwell Scientific, Oxford, 399– 444. Strahler, A. N. 1950. Equilibrium theory of erosional slopes approached by frequency distribution analysis. American Journal of Science, 248, 673–696, 800– 814. Terzaghi, K. 1950. Mechanism of landslides. Engineering Geology, Berkley Volume, 83– 123. Terzaghi, K. 1963. Stability of steep slopes on hard unweathered rock. Geotechnique, 12, 251–270.

Troisi, C. 2007. Rilievi interferometrici satellitari con la tecnica PSinSARTM, le esperienze di Arpa Piemonte. Universita` europea d’estate 2007, Courmayoeur. Varnes, D. J. 1978. Slope movement types and processes. In: Schuster, R. L. & Krizek, R. J. (eds) Landslides, Analysis and Control – Special Report 176. National Academy of Sciences, Washington, DC, 11– 33. Voight, B. 1973. The mechanics of retrogressive block-gliding, with emphasis on the evolution of the Turnagain heights landslide, Anchorage, Alaska. In: De Jong, K. A. & Scholten, R. (eds) Gravity and Tectonics. Wiley, New York, 97– 121. Vyazmensky, A., Stead, D., Elmo, D. & Moss, A. 2010. Numerical analysis of block caving induced instability in large open pit slopes: A finite element/discrete element approach. Rock Mechanics & Rock Engineering, 3, 21– 39. Wernicke, B. 1981. Low-angle normal faults in the Basin and Range province: nappe tectonics in an extending orogen. Nature, 291, 645– 648. Young, A. 1972. Slopes. Oliver & Boyd, Edinburgh. Zangerl, C., Eberhardt, E. & Loew, S. 2003. Ground settlements above tunnels in fractured crystalline rock: numerical analysis of coupled hydromechanical mechanisms. Hydrogeology Journal, 11, 162–173. Zaruba, Q. & Mencl, V. 1982. Landslides and Their Control. Elsevier, Amsterdam. Zischinsky, U. 1966. On the deformation of high slopes. In: Proceedings of the First International Congress of the International Society of Rock Mechanics, Lisbon, Volume 2. Laborato´rio Nacional de Engenharia Civil, Lisbon, 179 –185. ¨ ber Sackungen. Rock Mechanics Zischinsky, U. 1969. U and Rock Engineering, 1, 30–52.

Paraglacial gravitational deformations in the SW Alps: a review of field investigations, 10Be cosmogenic dating and physical modelling S. EL BEDOUI1, T. BOIS2*, H. JOMARD3, G. SANCHEZ4, T. LEBOURG2, E. TRICS2, Y. GUGLIELMI5, S. BOUISSOU2, A. CHEMENDA2, Y. ROLLAND4, M. CORSINI4 & J. L. PE´REZ6 1

LRPC Nancy, 71 rue de la Grande Haie, 54510 Tomblaine, France

2

GEOAZUR CNRS-UNS-IRD-UPMC, UMR 6526, Nice Sophia-Antipolis University, 250 Avenue de Albert Einstein, 06560 Sophia-Antipolis, France

3

Institute of Radioprotection and Nuclear Safety (IRSN), Fontaine-aux-Roses, France

4

GEOAZUR CNRS-UNS-IRD-UPMC, UMR 6526, Nice Sophia-Antipolis University, Avenue de Valrose, 06000 Nice, France 5

Centre de Se´dimentologie – Pale´ontologie, Universite´ de Provence Aix-Marseille 1, 3 Place Victor Hugo, 13331 Marseille Cedex 03, France

6

Centre des Etudes Techniques de l’Equipement Me´diterrane´e, 52 Bd Stalingrad, 06520 Nice, France *Corresponding author (e-mail: [email protected]) Abstract: Catastrophic deep-seated landslides (DSL) are generally considered to be the result of large slope deformations also known as deep-seated gravitational slope deformation (DSGSD). This paper aims to build a synthesis of multiple studies made in the Tine´e Valley (southern French Alps) to assess the geometrical, kinematical, mechanical and chronological relationships between these two gravitational processes. At the scale of the valley, data issued from geological, geomorphological and 10Be dating indicate a clear geometrical link between DSGSD and DSL occurring at the base of the slope and suggest that gravitational slope evolution began after the glacial retreat (13 ka BP). This is supported by the example of the well-documented La Clapie`re slope. A continuous evolution process is characterized geometrically and temporally from geomorphic observations and analogue modelling. Coupling structural, geomorphological, physical and chronological studies allowed us to propose a four-dimensional (4D) deformation model mechanically correlated with progressive failure concept. The validity and variability of this reference site are discussed at the valley scale (taking Isola and Le Pra slope deformation as examples). It allows a rough estimation of the state of slope deformation at the valley scale to be constructed and the slope evolution with time to be considered. This 4D model could then be considered as a reference for other deep-seated gravitational slope deformations in comparable Alpine valleys.

Gravitational slope deformation plays an important role in relief evolution of mountain ranges (Jarman 2006); however, the interconnected processes leading from large spatial and timescale deep-seated gravitational slope deformation (DSGSD) (Dramis & Sorriso-Valvo 1994) to catastrophic rock slope failure remain poorly understood (Agliardi et al. 2001; Ballantyne 2002). Glacier retreat in Alpine valleys is often considered to be a major conditioning factor in slope destabilization (Evans & Clague 1994; Ballantyne 2002; Tibaldi et al. 2004; Bigot-Cormier et al.

2005; Hippolyte et al. 2006; Apuani et al. 2007). The main deglaciation effects on gravitational motion include topographic change of valleys (Savage & Varnes 1987; Augustinus 1995) and/or debuttressing of slopes leading to tensile stress state (Hutchinson 1988; Apuani et al. 2007). Both effects could strongly influence in situ stress conditions and rock strength parameters at the slope scale (Bachmann 2006). Structural heterogeneities, such as inherited tectonic faults and fractures, are also assumed to play a dominant role in gravitational slope failure

From: Jaboyedoff, M. (ed.) Slope Tectonics. Geological Society, London, Special Publications, 351, 11–25. DOI: 10.1144/SP351.2 0305-8719/11/$15.00 # The Geological Society of London 2011.

12

S. EL BEDOUI ET AL.

processes (Scavia 1995; Kaneko et al. 1997; Hermanns & Strecker 1999; Bachmann et al. 2004; Brideau et al. 2005; Jomard 2006; Bois et al. 2008). It is also commonly admitted that progressive failure within a rock slope can initiate and propagate through preferential weakened fault zones (Sartori et al. 2003; Bachmann et al. 2004; Willenberg 2004). This paper focuses on a post-glacial Alpine area: the Tine´e Valley (southern French Alps), which was affected by consecutive glaciations during the Quaternary period. Slopes located in this area are affected by recent DSGSD and deep-seated landslides (DSL) (Follacci 1987; Julian & Anthony 1996; Jomard 2006). This review paper is the synthesis of different approaches: field investigations, cosmonuclide dating of gravitational events and physical modelling that were performed at the valley scale and at the slope scale (the La Clapie`re slope) in order to: (i) establish chronological and physical links between DSGSD and DSL; and (ii) propose a model of slope deformations including some geomechanical considerations. To this end, the La Clapie`re slope is described in this paper as an observatory site used to understand the slope deformation evolution processes linking DSGSD and catastrophic rock collapses.

Regional-scale gravitational deformation in the upper Tine´e Valley General settings The Argentera –Mercantour massif is the southernmost external crystalline massif of the Western Alps. It is characterized by a polyphased deformation evolution from Variscan to Alpine orogenies (Corsini et al. 2004). The basement consists of highgrade metamorphic and intrusive rocks of late Carboniferous age (Ferrara & Malaroda 1969). It is unconformably covered by a marine sedimentary succession of Late Carboniferous –Cenozoic age, partly detached at the level of the Triassic evaporites (Faure-Muret 1955) and overthrusted by Penninic clastic units during the Late Eocene –Early Oligocene (Autapie nappe and Parpaillon nappe) (Tricart 1984). The upper Tine´e Valley represents the western boundary between the basement with its Permo-Triassic tegument and the detached Mesozoic sedimentary cover (Figs 1 & 2). The basement rocks are migmatitic paragneisses with meta-granodioritic intrusions (Faure-Muret 1955). Ductile fabrics present a global N1308 dipping to the NE foliation. The main slope directions are collinear to the major N1108 –1408 trending fault set of the massif and make a 708 angle towards a secondary N0008 – 0308 fault system (Fig. 1).

The Argentera –Mercantour massif and its foreland has been a high relief area since the Early Pliocene (Fauquette et al. 1999). Altitudes range between 400 and 3143 m a.s.l. (metres above sea level) (at Mt Ge´las) in the gneissic bedrock and 3051 m (at Mt Pelat) in the sedimentary cover (Fig. 1). The combined glacial and river network have deeply eroded and incised the massif, resulting in slope heights of 2000 m. The glaciations history of the area is complex and poorly documented but as in many other Alpine valleys, the morphology of the upper Tine´e area was strongly influenced by Rissian and Wu¨rmian glacial ages. Most of the current glacial landforms come from the Wu¨rmian glaciation. They are characterized by glacial polished surfaces of glacial deposit (Julian 1980). From 10Be dating of polished glacial surface and radiocarbon analysis, Bigot-Cormier et al. (2005) and Sanchez et al. (2009) reported that the Tine´e Valley glacier was totally deglaciated at 12 ka BP whereas higher slope parts were deglaciated around 8 ka BP (Julian 1980). Besides, successive positive pulses since the last main deglaciation have been evidenced in the massif, highlighting a discontinuous deglaciation since the Wu¨rm (Julian 1980). However, glaciers have not been as important as in the other northernmost massifs in the Alps with a southern extension of the ice to a 500 m a.s.l. altitude and a maximum thickness of 500 m in the valley. The summits were also not covered by any ice sheets. As a consequence, in the morphology of the Tine´e Valley the mean slope angles vary from 358 up to 1800 m to 258 above this altitude.

Large gravitational deformation inventory An inventory was established by Jomard (2006) (see also Bois et al. 2008). A distinction is made between deformations that are clearly linked to gravity (bounded by a failure surface and a large mobilized volume that correspond to DSL), and deformations that appear diffused in very large volumes of the massif. Such deformations display extensive graben-like features and correspond to large sagging zones of the slopes. Zischinsky (1966) first proposed the term sackung for those surface manifestations of deep-seated rock creeps of foliated bedrocks in the Alps. This kind of movement was then observed in almost all mountain ranges and most authors today use the generic term deep-seated gravitational slope deformation (DSGSD) introduced by Dramis & Sorriso-Valvo (1994) to name the landforms and geomorphic evidence, such as double crested ridges, troughs, antislope scarps and ridge depressions, associated with those deformations (Agliardi et al. 2001). The origin of these features (tectonic, gravitational or both) are still poorly understood even if recent

GRAVITATIONAL DEFORMATION IN THE SW ALPS

13

Fig. 1. Major tectonics features and gravitational deformations mapped in the Tine´e Valley (modified after Bois et al. 2008).

results strongly tend to demonstrate a gravitational origin (Bachmann 2006). These characteristic morphostuctural features were mapped at the Tine´e Valley scale (Jomard 2006). Field investigations within the Tine´e slopes showed that most of Wu¨rmian glacial morphologies

and deposits are affected by gravitational deformations, suggesting gravitational motions since the last deglaciation (Jomard 2006). Most of these observations have been made in the metamorphic basement, in the left bank of the valley. Eight DSGSD zones were recognized containing 20 DSL (Fig. 1).

14

S. EL BEDOUI ET AL.

Fig. 2. The La Clapie`re rockslide embedded in the Colle Longue DSGSD, showing the geometrical relationship between the inherited tectonic structures, DSL and DSGSD (modified after Jomard 2006).

A regional map representing DSGSD and DSL deformations (morphostructures) was compared to the structural map of the massif (Fig. 1). † The orientations of DSGSD’s morphostructures are strongly influenced by both the inherited tectonic framework and the mean N1308-trending foliation. In particular, the major N1108– 1408E fault set explains the localization of most of the internal deformations of DSGSD within the valley. † The other N0008–0308E fault set is also expressed in some cases through deformations secant to the main direction. † The most developed DSL occurred at the intersection of two fault set, as the La Clapie`re DSL between the N1108 –1408E and N0008–0308E fault set. The density of morphostructures within DSGSDs appears to be closely dependent on the angular relation between slope direction, foliation and faults orientations. Morphostructures are much more developed when this angle is close to zero. The most developed Colle Longue DSGSD could then be explained by the narrow angle existing between the N1408E main direction of the crests/ valley and the N1308E main faults and foliation direction. The Colle Longue DSGSD represents an area of about 45 km2. This zone, which presents extensive deformations spreading from the foot to the crests, has been chosen for a more accurate description. Counterscarps are recognized in the upper slope part down to 1800 m a.s.l., although troughs are observed in the lower slope part up to 2100 m a.s.l. Counterscarps can reach 20 m high and always are guided by the foliation planes whose main orientation is N1308–0408 NE. It also appears that counterscarps connect at depth to inherited tectonic

faults that guide the gravitational deformations. Troughs display 15 m wide apertures that are geometrically associated to downward-dipping (SW) normal and strike-slip fault zones. Both structures affect the slope down to the tributary valleys floor indicating that deformations should be deeper than these valley incisions that can reach 1000 m high. No failure surface enveloping the DSGSD was observed. Compressive features were only observed in some DSL feet encased in the DSGSD zone (like the large active La Clapie`re DSL). A long-term deformation of the slope is then characterized after the deglaciation. From the chronological point of view, two main observations are made: † In upper slope parts (.1800 m), gravitational deformations affect Wu¨rmian glacial morphologies of high-altitude tributary glaciers, rock glaciers and active screes. Glacial sediments filling cracks are also observed in a number of counterscarps. Deformations are mostly represented by counterscarps and scarps. † In lower slope parts, deformations affecting Wu¨rmian deposits and morphologies are represented by scarps. Troughs are filled with colluviums and no glacial sedimentary fillings are observed. On the La Clapie`re slope (Jomard 2006), a trough located on a down-slope normal fault was detailed (Fig. 3). The fault, characterized by its gouge and slickenslide planes, was recently toppled and created a large tension aperture (10 m) infilled by regular post-glacial colluviums sedimentation. The toppling of rock columns allowed an intense fracturing of the downward wall of the fault. Finally, this intense fracturing gives the possibility for sliding surfaces to develop and cross-cut the overall structure (Fig. 3). Chronological constraints based on 10Be dating of gravitational scarps, troughs and landslide surfaces in the Tine´e area show three successive periods of gravitational instabilities (Bigot-Cormier et al. 2005; Sanchez et al. 2009). The first closely post-dates the last deglaciation event (12–13 ka) with an age of 10–11 ka, a second destabilization occurs at 7 –9 ka BP and a third at 2.5– 5.5 ka BP. Thus, gravitational slope deformations have been effective at least since 12 ka BP, leading to the present active large deep-seated landslides such as La Clapie`re.

Slope-scale analysis of La Clapie`re Field analysis The La Clapie`re slope (Fig. 2) is one of the most active DSLs of the valley. This DSL affects 60  106 m3 of rock of the metamorphic basement.

GRAVITATIONAL DEFORMATION IN THE SW ALPS

15

Fig. 3. Trough cross-section (modified after Jomard 2006). Tilted normal fault gouge infilled with post-glacial colluvium affected by recent failure. (a) Photograph and (b) interpretation.

The first activity reported dates from the early 1930s, with a peak evolution between 1960 and 1990 leading to a 130 m-high scarp development in the middle of the slope. Thus, since the 1990s, structural investigations (Follacci 1987; Ivaldi 1991), geomechanicals triaxial experiments on rock samples, hydrogeological studies (Cappa et al. 2004; Guglielmi et al. 2005), numerical modelling and geophysical surveys (Lebourg et al. 2005; Jomard et al. 2007, 2010) have been performed underlining a complex post-failure behaviour. The La Clapie`re DSL is geometrically bound by N1108–1408E and N0108–0308E fault sets (Fig. 1). Other morphological signs of gravitational destabilization are guided by those tectonic orientations, and are represented inside and outside the DSL boundaries, mainly consisting of troughs aligned on tectonic fault scarps (Fig. 3). These troughs show, respectively, a trace linear far from the active DSL, curved close to the DSL and dislocated within the landslide body. Fifteen troughs were mapped with an average N1208 direction and a 100 –5000 m length (Fig. 4) that indicated a deep slope deformation in agreement with the observations made in the valleys bounding the slope (Fig. 2). Troughs were mainly observed from the toe to the middle of the slope (1500–2100 m in elevation) and their evolution is clearly linked to the initiation of the La Clapie`re DSL. Their average orientation is parallel to the slope, although troughs closed to the western boundary of the currently active DSL have an orientation close to N1308 (Fig. 4). Troughs 1– 4 that are the closest to the active DSL scarp are highly deformed (and even cut by the scarp) while torsion progressively vanishes from troughs 5 –15. Troughs 1, 6 and 14, and the west

lateral scarp propagation of La Clapie`re active DSL, were dated using an in situ produced 10Be cosmogenic approach (Bigot-Cormier et al. 2005; Sanchez et al. 2009). The result suggest that the troughs become progressively younger from trench 1 (10 ka BP), 6 (7.2 ka BP) to 14 (5.6 ka BP), meaning that a deformation of the slope propagated from the toe to the top of the slope at about 4.4 ka BP. The upper lateral scarp of the currently active DSL was dated at 3.6 ka BP, showing that after the up-slope deformation propagation a deep failure initiated in the middle part of the slope that ultimately bounds the currently active DSL. Two-dimensional (2D) physical modelling experiments were performed reproducing a NNE– SSW cross-section of the La Clapie`re slope in order to analyse the links between superficial and deep-seated deformations (Figs 5 & 6).

Two-dimensional physical modelling of the ‘La Clapie`re’ slope A complete description of the analogue material (called Slope1) and the loading device developed to perform our scaled physical models is available in Bachmann et al. (2004), Chemenda et al. (2005) and in Bois et al. (2008). A short description is available in the Appendix of this paper. For these models a scaling factor of 1/50 000 was chosen. The vertical faults have been numbered from F1 near the valley toe to F6 near the crest. Two distinct configurations of the slope were tested: † a homogenous model (Fig. 5) that must be considered as a massive homogeneously fractured massif without any major localized weak zone;

16 S. EL BEDOUI ET AL. Fig. 4. Evolution of the La Clapie`re slope deformation for the last 10 ka BP. (a) and (b) Troughs opening from 1600 to 2250 m elevation a.s.l.; (c) surface shearing close to the future rockslide area; (d) rockslide collapse and sliding (modified after El Bedoui et al. 2009).

GRAVITATIONAL DEFORMATION IN THE SW ALPS

17

Fig. 5. Homogeneous model of the La Clapie`re slope: (a) the non-deformed model; (b) the first deformation stage and its interpretative sketch; and (c) the last deformation stage and its interpretative sketch (modified after Bois et al. 2008).

† a model considering the N1408 fault zones previously presented (Fig. 6). The listric geometry of those faults was deduced from field investigations (Fig. 2). Homogeneous model. On the initial deformation stage (Fig. 5b), a deep sliding surface was formed inside the model.

Its maximum depth was equivalent to 1500 m. This sliding surface bounded a large unstable volume involving the entire massif. A 100 m- high scarp was formed behind the topographic crest. On the final deformation stage, the sliding surface becomes a more complex fracture network. Its width increased with the displacement of the

18

S. EL BEDOUI ET AL.

Fig. 6. A slope model of La Clapie`re cut by six inherited normal listric faults: (a) the non-deformed model; (b) the first deformation stage and its interpretative sketch; (c) the second deformation stage and its interpretative sketch; and (d) the last deformation stage and its interpretative sketch (modified after Bois et al. 2008).

sliding unit (Fig. 5c). This complex failure zone reached a maximum depth of 2200 m and was associated with two antithetic newly formed faults. The escarpment located behind the topographic crest kept growing to reach a maximum of 400 m. A slope cut by six inherited normal listric faults. During the initial deformation stage (Fig. 6a), most of the superficial non-elastic deformation

was localized on three inherited normal faults and on a newly formed gravitational fault. On the inherited fault close to the topographic crest (F6) a 100 m-high escarpment appeared (Fig. 6b). On the inherited faults close to the slope toe (F2 and F3) two troughs appeared (Fig. 6b). Finally, a newly formed gravitational fault formed at the slope toe that propagated inside the massif rather like a subhorizontal thrust fault

GRAVITATIONAL DEFORMATION IN THE SW ALPS

(Fig. 6b). Internal irreversible deformations took place along this newly formed thrust fault and also along an antithetic normal fault initiating from the inflection point of fault F6 towards the topographic surface (Fig. 6). During the second deformation stage, the deformation increased (Fig. 6c): another normal trench appeared at the top of the fault F5. The newly formed thrust fault propagated inside the model and a relatively small superficial DSL was finally triggered between the first normal fault and the slope toe. In the last deformation stage (Fig. 6d), the sliding surface propagated inside the model through the connection of the inflection points of faults F4– F6, delimiting a deep gravitational moving zone. The first antithetic fault propagated and reached the topographic surface. The hillside was affected by a second DSL, which was a retrogressive one. The fractured model deformation can then be summarized as follows. † The normal troughs are formed on the faults close to the valley immediately followed by the appearance of the first normal shifting on faults close to the topographic crest. † The first antithetic faults are formed, and a sliding surface (i.e. thrust fault) propagates from the toe. † New troughs are created higher on the slope due to a retrogressive deformation process. The sliding surface propagates and a small-scale DSL is triggered. † The sliding surface is connected to the F6 fault delimiting the total moving mass. The deformation of this part of the slope leads to the formation of successive landslides that affect the slope. Modelling major results. On one hand, the deformation pattern obtained in the homogeneous configuration showed that even if the kinematics of the rupture is coherent, the localization of the deformation is not evident. Indeed, it seems obvious that the ‘La Clapie`re’ slope cannot be considered to have fractured homogeneously. Some localized weak zones have to be taken into account, such as the N1408 faults zones. This is confirmed by the second model configuration for which it appears clearly that the deformation is mainly localized on pre-existing faults. However, on the other hand, those models are 2D models developed according to a NNE–SSW cross-section of the ‘La Clapie`re’ slope, and owing to this simplification some structures have not been considered (e.g. other inherited faults such as the N0308 ones). It is reasonable to say that those structures must also have had an influence on the

19

localization of the deformation and, at a larger scale, on the global deformation pattern of the massif.

From DSGSD to catastrophic failure at the slope scale Based on field investigations and absolute ages (10Be), El Bedoui et al. (2009) calibrated a model of the slope evolution for the last 10 ka. Physical modelling shows a very good agreement with this proposed model based on field and dating work underlining the link between those superficial deformations and the failure propagation at depth. From 10 to 5.6 ka BP (Fig. 4) extensional structures (troughs) spread from the toe to the top of the slope, showing a good correspondence with physical modelling (Fig. 6b, c) (phase I). The trench-like morphology is explained by the reactivation of inherited structures that could have been induced by the stress release effects in the slope related to glacial retreat of the last glaciation. This phase I could be related to the deep fracture retrogressive propagation showed by physical modelling (Fig. 6c, d). From 5 to 3.6 ka BP (Fig. 4) a shearing of troughs occurred in the western lower part of the slope that also displayed a high vertical displacement (phase II). This 3D twisting of surface troughs, which is related to shear deformations developing deep in the slope, could be induced by preferential tangential movements along a major N0308 reactivated vertical fault zone located in the east part of the slope. At 3.6 ka BP (Fig. 4) a failure initiates, wrapping this eastern lower part of the slope. In the last 50 years this failure evolved in a large-scale failure surface bounding the currently active La Clapie`re DSL (phase III). 10 Be ages indicate that phase I (troughs opening – DSGSD) propagates over a very long period (6 ka BP). Phase II (troughs shearing) is only constrained by two dated events: the trench opening (10 ka BP) and the first failure associated with the rockslide (3.6 ka BP). If physical modelling does not allow a kinematic model of the slope evolution, results underline the fact that the first rockslide event was synchronous with the upper and younger trench (dated at 5.2 ka BP). It could indicate that the trench shearing more probably occurred between 5.2 and 3.6 ka BP rather than immediately after the trench opening at 10 ka BP. The kinematics of slope deformation have been characterized in two dimensions and extended in three dimensions, and suggest a non-linear creep-like phenomenon. The significance of the deduced model needs to be discussed in a mechanical way and compared to other places in the valley to assess its validity.

20

S. EL BEDOUI ET AL.

Discussion Mechanical processes at the slope scale Rock slope failure in general results from many different processes and among these the influences of the pre-existing heterogeneities (such as bedding planes, foliation and fault zones) (Terzaghi 1962; Kato & Hada 1980; Chigira 1985; Agliardi et al. 2001) the mountain height and slope gradient appear to be predominant. Rock slope stability is thus highly dependent on the specific characteristics of the fault zones (e.g. density of the fault network, persistence at depth, the geometry of the faults, etc.) (Hermanns & Strecker 1999). Our study underlines the major role played by regional scale weak planes that correspond to inherited major fault zones (Bois et al. 2008). The temporal evolution of slope deformation at La Clapie`re is in good accordance with laboratory experiments that display highly non-linear creeplike phenomena even for hard rocks. In previous work conducted on the gravitational evolution of rock slopes (Bru¨ckl & Parotidis 2005; Petley et al. 2005; Apuani et al. 2007) the evolution from DSGSD to a localized catastrophic failure is physically and mechanically regarded as a creep-like phenomenon. Bru¨ckl & Parotidis (2005) numerically studied the slope instabilities resulting from deep-seated gravitational creep and the transition from a slow evolution phase to rapid sliding. The process of subcritical growth was considered to explain the primary phase of deep-seated gravitational creep because it allowed the progressive damage of the rock mass at a lower stress than the rock strength. It could correspond to the progressive failure growth at depth characterized by troughs opening and twisting at the surface. The total failure at depth then outlines the currently active rockslide, acceleration of the movements being related to the ‘smoothing of the basal surface’. Another first-order process that has an influence on strength reduction, and thus on damage process, is weathering controlled by climatic and fluids circulations (Hill & Rosenbaum 1998; Hall & Andre´ 2001; Pellegrino & Prestininzi 2007). It also has a strong control on the failure process, especially in the case of granitic rocks in the Alps (Girod 1999; Jaboyedoff et al. 2004). Indeed, the alteration of such rocks has two main influences: on the one hand, it leads to the formation of clays; thus reducing the fluid circulation and increasing the pressure (Girod 1999). This is particularly true in faulted zones where alteration is concentrated basically along localized weak zones: fractures and faults (Migon & Lidmar-Bergstro¨m 2002; Wyns 2002). On the other hand, weathering causes a progressive strength reduction of the rock material, which is

stress dependent. This softening is generally maximal at the surface and diminishes with depth (Chigira 2001). Even if this contributing factor has not been taken into account by the physical models, Chemenda et al. (2009) used a 2D finite-element numerical model to show that, in the case of the La Clapie`re slope, a progressive reduction in the mechanical resistance (due to a progressive reduction in the model cohesion), combined with the particular geometry of glacial Alpine valleys, can lead to the DSL. A time-dependent model (Fig. 7) calibrated on the La Clapie`re slope failure was proposed in order to fit the observed surface displacements as function of time (El Bedoui et al. 2009). This model presents three phases: (1) very slow displacements over a long time period (several mm year21); (2) an increase in surface displacements related to a deep slide plane (crack coalescence); and (3) a catastrophic evolution over a very short time period.

Extrapolation of the model at the valley scale The La Clapie`re slope is part of the Colle Longue DSGSD. Looking at the Tine´e Valley scale, different evolution stages were observed, from pre-failure stages characterized by troughs dislocation, to active and fossil DSL. Two representative localities were studied: the Isola slope and the Pra slope (Figs 1 & 8). The Isola slope is located downstream of the upper Tine´e Valley. This slope is also embedded in the Colle Longue DSGSD (Figs 1 & 8). The average slope is 358 in the basal part (from 850 to 1700 m elevation) and 258 in the upper part. A large series of N1208 troughs crop out in the basal part, extending laterally over distances of 500 m. A surface shearing of these troughs was located at elevations of between 1000 and 1500 m, which is similar to the La Clapie`re case and indicates a contrast of surface velocities between the upper stable part of the slope and the lower more intensely deformed area. The lower slope part shows signs of shearing that are as developed as those on the La Clapie`re slope (Fig. 8). It indicates that this slope can correspond to the second stage of the evolution model of slope deformation (Figs 6 & 7). This advanced stage of slope deformation is further exacerbated by very active rockfall activity and may indicate a future landsliding activity during the next century.

Conclusion Coupling the different approaches presented in this paper has allowed the elaboration of a 4D evolution

GRAVITATIONAL DEFORMATION IN THE SW ALPS

21

Fig. 7. Curve of time-dependent progressive failure calibrated on the La Clapie`re slope (modified after El Bedoui et al. 2009).

model of rock slope destabilization linking failure propagation at depth and subsequent deformations at the surface. The model seems to be well constrained for the La Clapie`re slope and seems also to be valid at the Tine´e Valley scale. Furthermore, as the spatial correlation between gravitational deformation and weakened zones of inherited faults suggest, the major structural framework has a key role in such processes. Thus, by comparison with other slopes, we suggest that kinematics of slope evolution are controlled by a critical spacing/ arrangement of major inherited faults. De facto, the deglaciation and the consecutive slope angle increase of the valley is supposed to have a strong influence on slope destabilization due to modifications of in situ stress condition and rock strength parameters. The main implication of these results is that the model proposed based on the La Clapie`re slope can be considered as a reference at the valley scale and, probably more widely, at the massif scale as proposed by Jomard (2006). Large-scale geomorphological studies (Jomard 2006) should be coupled with instrumental surveys (GPS) performed at the valley scale and mechanically scaled models; this

will, respectively, allow: (i) an estimation of volumes potentially mobilized during the destabilization (based on deformation of morphological signs; i.e. troughs); and (ii) a rough estimation of time prior to failure by comparison between instrumental velocities and the reference calibrated on the La Clapie`re slope. Authors sincerely express their gratitude to the GIS CURARE who financed this project and to the reviewers for their useful remarks.

Appendix Slope1 is a low frictional elasto-brittle– plastic analogue material with strain softening (Chemenda et al. 2005). This material represents a compositional system based on liquid and solid hydrocarbons. To create a model, the melted analogue material Slope1 is moulded into a rigid box at a temperature of 50 8C. In order to create the faults a series of openings cut in the two opposite lateral sides of the model box are used to position taut strings. After cooling to a temperature of 20 8C, at which the crystallized material is strong enough to be easily

22

S. EL BEDOUI ET AL.

Fig. 8. Morphostructures of the Isola slope.

handled without damage and easily cut with the strings without being damaged in areas other than the cuts, strings are translated along the slots to generate the faults and then removed. The model surface was then shaped to the desired topography (Figs 5a & 6a). The length of the model is thus equal to 14 cm and the width (third dimension) is equal to 30 cm. The third dimension has been chosen large enough to prevent any edge effects. Once the model is prepared it is loaded into a vertical accelerator table. The latter consists of a mobile platform that can be lifted up to 2 m and then released. During

its free fall the model reaches a maximum velocity of 6 m s21. The platform is then rapidly but smoothly decelerated to zero velocity when it comes into contact with a progressive shock absorber of 5 cm stroke. During this phase the model undergoes a strong vertical deceleration (up to 500 m s22). This deceleration acting in the same direction as gravity is repeated until failure develops, usually at approximately 100 cycles. Preliminary calibration tests are needed to determine which acceleration must be imposed onto a model for a given configuration (geometry, prefracturing state, etc.) in order to observe

GRAVITATIONAL DEFORMATION IN THE SW ALPS failure for a number of loading cycles ranging from 100 and 150. Model deformation can be observed accurately after each acceleration cycle. This discrete loading technique has proved to be equivalent to a continuous quasi-static loading (Chemenda et al. 2005). The main similarity criterion is:

sco scm ¼ ro go H o rm gm H m

(1)

where r is density and g is gravity acceleration, sc is the strength under uniaxial compression and H is the spatial scale of the phenomenon (the mountain height H, for example). The superscripts ‘o’ and ‘m’ mean original and model, respectively. To ensure that the deformation will be localized along brittle structures, as in nature, Slope1 has to exhibit a high degree of softening. Hence, the experiments were carried out at a fixed temperature of 20 8C. This mechanical behaviour is comparable to the strength degradation behaviour introduced into some numerical models. At this temperature the coefficient of friction measured on the pre-existing fractures is m ¼ 0.2. Cross-sections were made at the end of each experiment by cutting the model at various positions after cooling it to 10 8C in order to increase its strengths. Some experiments were stopped in the early stages of model deformation to analyse the corresponding evolution of internal slope deformation.

References Agliardi, F., Crosta, G. & Zanchi, A. 2001. Structural constraints on deep-seated slope deformation kinematics. Engineering Geology, 59, 83–102. Apuani, T., Masetti, M. & Rossi, M. 2007. Stress– strain– time numerical modelling of a deep-seated gravitational slope deformation: preliminary results. Quaternary International, 171–172, 80–89. Augustinus, P. C. 1995. Glacial valley cross-profile development: the influence of in situ rock stress and rock mass strength, with examples from the Southern Alps, New Zealand. Geomorphology, 14, 87–97. Bachmann, D. 2006. Mode´lisation physique tridimensionnelle des mouvements gravitaires de grande ampleur en milieu rocheux. PhD thesis, Universite´ de Nice Sophia Antipolis, Nice, France. Bachmann, D., Bouissou, S. & Chemanda, A. 2004. Influence of weathering and pre existing large scale fractures on gravitational slope failure: insights from 3-D physical modelling. Natural Hazards and Earth System Sciences, 4, 711 –717. Ballantyne, C.-K. 2002. Paraglacial geomorphology. Quaternary Science Reviews, 21, 1935–2017. Bigot-Cormier, F., Braucher, R., Bourle`s, D., Guglielmi, Y., Dubar, M. & Ste´phan, J. F. 2005. Chronological constraints on processes leading to large active landslides. Earth and Planetary Science Letters, 235, 141– 150.

23

Bois, T., Bouissou, S. & Guglielmi, Y. 2008. Influence of major inherited faults zones on gravitational slope deformations: A two-dimensional physical modelling of the La Clapie`re area (Southern French Alps). Earth and Planetary Science Letters, 272, 709– 719; doi: 10.1016/j-epsl_2008.06.006. Brideau, M.-A., Stead, D., Kinakin, D. & Fecova, K. 2005. Influence of tectonic structures on the Hope Slide, British Columbia, Canada, Enginereeing Geology, 80, 242–529. Bru¨ckl, E. & Parotidis, M. 2005. Prediction of slope instabilities due to deep-seated gravitational creep. Natural Hazards and Earth System Sciences, 5, 155– 172. Cappa, F., Guglielmi, Y., Soukatchoff, V. M., Mudry, J., Bertrand, C. & Charmoille, A. 2004. Hydromechanical modelling of a large moving rock slope inferred from slope levelling coupled to spring longterm hydrochemical monitoring: example of the La Clapie`re landslide (Southern Alps, France). Journal of Hydrology, 291, 67–90. Chemenda, A., Bois, T. & Bouissou, S. 2009. Numerical modeling of the gravity-induced destabilization of a slope: example of the Clapie`re landslide, southern France. Geomorphology, 109, 86– 93; doi: 10.1016/ j.geomorph.2009.02.025. Chemenda, A., Bouissou, S. & Bachmann, D. 2005. Three dimensional physical Modeling of Deep-Seated Landslides: new technique and first results. Journal of Geophysical Research, 110, F04004; doi: 10.1029/ 2004JF000264. Chigira, M. 1985. Mass rock creep of crystalline schist: minor structures formed by mass rock creep. Journal of Japanese Society of Engineering Geology, 26, 25–79. Chigira, M. 2001. Micro-sheeting of granite and its relationship with landsliding specially after the heavy rainstorm in June 1999, Hiroshima prefecture, Japan. Engineering Geology, 29, 219–231. Corsini, M., Ruffet, G. & Caby, R. 2004. Alpine and late Hercynian geomorphological constraints in the Argentera massif (Western Alps). Eclogae Geologicae Helvetiae, 97, 3 –15. Dramis, F. & Sorriso-Valvo, M. 1994. Deepseated gravitational slope deformations, related landslides and tectonics. Engineering Geology, 38, 231– 243. El Bedoui, S., Guglielmi, Y., Lebourg, T. & Pe´rez, J. L. 2009. Deep-seated failure propagation in a fractured rock slope over 10 000 years: the La Clapie`re slope, the south-eastern French Alps. Geomorphology, 105, 232–238; doi: 10.1016/j.geomorph. 2008.09.025. Evans, S. G. & Clague, J. J. 1994. Recent climatic change and catastrophic geomorphic processes in mountain environments. Geomorphology, 10, 107–128. Fauquette, S., Suc, J.-P. et al. 1999. Climate and biomes in the west Mediterranean during the Pliocene. Palaeogeography, Palaeoclimatology, Palaeoecology, 152, 15–36. Faure-Muret, A. 1955. Etudes ge´ologiques sur le massif de l’Argentera-Mercantour et ses enveloppes se´dimentaires. Me´moire de la Carte Ge´ologique de France.

24

S. EL BEDOUI ET AL.

Ferrara, G. & Malaroda, M. 1969. Radiometric age of granitic rocks from the Argentera massif (MAritime Alps). Bollettino della Societa` Geologica Italiana, 88, 311–320. Follacci, J. P. 1987. Les mouvements du versant de la Clapie`re a` Saint Etienne de Tine´e (Alpes Maritimes). Bulletin de Liaison des Ponts et Chausse´es, 150/151, 39–54. Girod, F. 1999. Alte´ration me´te´orique de roche granitique en milieu alpin: le cas de l’orthogneiss associe´ a` l’e´boulement de Randa (Mattertal, Valais, Suisse). PhD thesis, Universite´ de Lausanne. Guglielmi, Y., Cappa, F. & Binet, S. 2005. Coupling between hydrogeology and deformation of montainous rock slopes: insights from La Clapie`re area (southearn Alps, France). Comptes rendus Ge´osciences, 337, 1154–1163. Hall, K. & Andre´, M. F. 2001. New insights into rock weathering from high-frequency rock temperature data: an Antarctic study of weathering by thermal stress. Geomorphology, 41, 23– 35. Hermanns, R. L. & Strecker, M. R. 1999. Structural and lithological controls on large Quaternary rock avalanches (sturzstroms) in arid northwestern Argentina. Geological Society of America, 111, 934– 948. Hill, S. E. & Rosenbaum, M. S. 1998. Assessing the significant factors in a rock weathering system. Quarternary Journal of Engineering Geology and Hydrology, 31, 85– 94. Hippolyte, J.-C., Brocard, G. et al. 2006. The recent fault scarps of the Western Alps (France): tectonic surface ruptures of gravitational sackung scarps? A combined mapping, geomorphic, levelling, and 10Be dating approach. Tectonophysics, 418, 255– 276. Hutchinson, J. N. 1988. General Report: morphological and geotechnical parameters of landslides in relatio to geology and hydrogeology. Proceedings of the 5th International Symposium on Landslides, Lausanne, 3 –35. Ivaldi, J. P., Guardia, P., Follacci, J. P. & Terramorsi, S. 1991. Plis de couverture en e´chelons et failles de second ordre associe´s a` un de´crochement dextre de socle sur le bord nord-ouest de l’Argentera (AlpesMaritimes, France). Compte Rendus de L’Academie des Sciences, Paris, 313, 361–368. Jaboyedoff, M., Baillifard, F., Bardou, E. & Girod, F. 2004. Weathering, cycles of saturation-unsaturation, and strain effects as principal processes for rock mass destabilization. Quarterly Journal of Engineering Geology and Hydrogeology, 37, 95–103. Jarman, D. 2006. Large rock slope failures in the Highlands of Scotland: characterisation, causes and spatial distribution. Engineering Geology, 83, 161– 182. Jomard, H. 2006. Analyse multi-e´chelles des de´formations gravitaires du Massif de l’Argentera Mercantour. PhD thesis, Nice Sophia-Antipolis University. Jomard, H., Lebourg, T., Guglielmi, Y. & Tric, E. 2010. Electrical imaging of sliding geometry and fluids associated with a deep seated landslide (La Clapie`re, France). Earth Surface Processes and Landforms, 35, 588–599.

Jomard, H., Lebourg, T. & Tric, E. 2007. Identification of the gravitational boundary in weathered gneiss by geophysical survey: La Clapie`re landslide (France). Journal of Applied Geophysics, 62, 47–57. Julian, M. 1980. Les Alpes Maritimes Franco-Italiennes. Etude Ge´omorphologique. PhD thesis, Aix-Marseille II University, Marseille. Julian, M. & Anthony, E. 1996. Aspect of landslide activity in the Mercantour Massif and the French Riviera, southeastern France. Geomorphology, 15, 275–289. Kaneko, K., Otani, K. J., Noguchi, Y. & Togashiki, N. 1997. Rock fracture mechanics analysis of slope failure. Deformation and Progressive Failure. In: Asaoka, A., Adachi, T. & Oka, F. (eds) Geomechanics. Elsevier, New York, 671– 676. Kato, J. & Hada, S. 1980. Landslides of the Yoshino– Gawa water system and its geological aspects. Research Reports of the Kochi University. Natural Science, 28, 127– 140. Lebourg, T., Binet, S., Tric, E., Jomard, H. & El Bedoui, S. 2005. Geophysical survey to estimate the 3D sliding surface and the 4D evolution of the water pressure on part of a deep seated landslide. Terra Nova, 17, 399– 406. Migon, P. & Lidmar-Bergstro¨m, K. 2002. Europe: problems of dating and interpretation of geological records. Catena, 49, 25– 40. Pellegrino, A. & Prestininzi, A. 2007. Impact of weathering on the geomechanical properties of rocks along thermal-metamorphic contact belts and morphoevolutionary. Geomorphology, 87, 176 –195. Petley, D. N., Mantovani, F., Bulmer, M. H. & Zannoni, A. 2005. The use of surface monitoring data for the interpretation of landslide movement patterns. Geomorphology, 66, 133–147. Sanchez, G., Rolland, Y., Corsini, M., Braucher, R., Bourles, D., Arnold, M. & Aumaıˆtre, G. 2009. Relationships between tectonics, slope instability and climate change: cosmic ray exposure dating of active faults, landslides and glacial surfaces in the SW Alps. Geomorphology, 117, 1– 13; doi: 10.1016/ j.geomorph.2009.10.019. Sartori, M., Baillifard, F., Jaboyedoff, M. & Rouiller, J. D. 2003. Kinematics of the 1991 Randa rockslides (Valais, Switzerland). Natural Hazards and Earth System Sciences, 3, 423– 433. Savage, W. Z. & Varnes, D. J. 1987. Mechanics of gravitational spreading of steep-sided ridges (sackung). Bulletin of Engineering Geology and the Environment, 35, 31–36. Scavia, C. 1995. A method for the study of crack propagation in rock structures. Ge´otechnique, 45, 447–463. Terzaghi, K. 1962. Stability of steep slopes in hard unweathered rock. Ge´otechnique, 12, 251– 270. Tibaldi, A., Rovida, A. & Corazzato, C. 2004. A giant deep-seated slope deformaion in the Italian Alps studied by paleoseismological and morphometric techniques. Geomorphology, 58, 27– 47. Tricart, P. 1984. From passive margin to continental collision; a tectonic scenario for the western Alps. American Journal of Sciences, 284, 97–120.

GRAVITATIONAL DEFORMATION IN THE SW ALPS Willenberg, H. 2004. Geologic and kinematic model of a complex landslide in crystalline rock (Randa, Switzerland), PhD thesis, Swiss Federal Institute of Technology Zurich. Wyns, R. 2002. Climat, eustatisme, tectonique: quels controˆles pour l’alte´ration continentale? Exemple des se´quences d’alte´ration ce´nozoı¨ques en France.

25

Bulletin d’Information Ge´ologie des Bassins, Paris, 39, 5 –16. Zischinsky, U. 1966. On the deformation of high slopes. In: Proceedings of the First International Conference of the International Society of Rock Mechanics, Lisbon, Volume 2. International Society of Rock Mechanics, Lisbon, 179– 185.

Inheritance of ductile and brittle structures in the development of large rock slope instabilities: examples from western Norway A. SAINTOT1*, I. H. C. HENDERSON1 & M.-H. DERRON1,2 1

Geological Survey of Norway, Leif Eirikssons Vei 39, N-7491 Trondheim, Norway

2

Lausanne University, IGAR, Quartier UNIL-Sorge, CH-1015 Lausanne, Switzerland *Corresponding author (e-mail: [email protected]) Abstract: The high density of slope failures in western Norway is due to the steep relief and to the concentration of various structures that followed protracted ductile and brittle tectonics. On the 72 investigated rock slope instabilities, 13 were developed in soft weathered mafic and phyllitic allochthons. Only the intrinsic weakness of such rocks increases the susceptibility to gravitational deformation. In contrast, the gravitational structures in the hard gneisses reactivate prominent ductile or/and brittle fabrics. At 30 rockslides along cataclinal slopes, weak mafic layers of foliation are reactivated as basal planes. Slope-parallel steep foliation forms back-cracks of unstable columns. Folds are specifically present in the Storfjord area, together with a clustering of potential slope failures. Folding increases the probability of having favourably orientated planes with respect to the gravitational forces and the slope. High water pressure is believed to seasonally build up along the shallow-dipping Caledonian detachments and may contribute to destabilization of the rock slope upwards. Regional cataclastic faults localized the gravitational structures at 45 sites. The volume of the slope instabilities tends to increase with the amount of reactivated prominent structures and the spacing of the latter controls the size of instabilities.

In hard rock masses, the steepness of high slopes may not be the unique parameter to explain the development of large gravitational instabilities. Many studies document the role of pre-existing structural grain and tectonic weakening in the development of large instabilities in hard rock, and it is widely accepted that the failure surface(s) may be partly or fully structurally controlled (e.g. Nemcok et al. 1972; Cruden 1976; Varnes 1978; Giraud et al. 1990; Chigira 1992; Scavia 1995; Guzzetti et al. 1996; Julian & Anthony 1996; Sauchyn et al. 1998; Agliardi et al. 2001; Kellogg 2001; Krejcˇ´ı et al. 2002; Bachmann et al. 2004; Eberhardt et al. 2004; Brideau et al. 2005; Vilı´mek et al. 2007; El Bedoui et al. 2008; Brideau et al. 2009; Jaboyedoff et al. 2009). Worldwide, the structural development of many large rock slope instabilities may be controlled by regional faults (Kaneko et al. 1997; Sartori et al. 2003; Ambrosi & Crosta 2006; Bois et al. 2008). A large number of studies showed the link between high frequency sets of persistent joints and slope susceptibility to rockfalls (and especially discrete rockfalls as defined in Varnes 1978). Indeed, most of the pre-existing structures described in these studies belong to the domain of brittle deformation, and few publications demonstrate the influence of the pre-existing ductile fabrics and folds in the development of slope instabilities (Jones 1993; Badger 2002; Coe & Harp 2007). In Norway, the bedrock

geology reveals an extremely complex and protracted ductile and brittle tectonic history that may intuitively have a strong influence on the spatial distribution of instabilities along the steep slopes of the Norwegian coastal fjord areas. An intense collection of 5000 structural measurements allowed us not only to place 72 rock slope instabilities in a regional structural framework but also to precisely investigate the role of geology and specifically structural geology in the development of the rock slope instabilities. These were not only the gravitational structures, thus related to the slope instabilities themselves (open large fractures, basal sliding planes, transfer structures, etc.), but also the older structures (exfoliation, foliation, folds, joints, fault zones, etc.) that belong to the prolonged geological history of the area. With the field observations at these 72 sites, we established the clear relationships between the location of the failure surface(s) and pre-existing geological structures. We present in detail in this paper the sites that best exemplify the relationships between slope instabilities and inherited structures. We describe case studies that illustrate: (1) how the ductile structures (metamorphic foliation and folds) may focus the formation of gravitational structures; and (2) reactivation of brittle structures to produce gravitational failures. We also appraise the relationship between the estimated unstable volume, the number of involved preexisting structures and the type of the gravitational

From: Jaboyedoff, M. (ed.) Slope Tectonics. Geological Society, London, Special Publications, 351, 27–78. DOI: 10.1144/SP351.3 0305-8719/11/$15.00 # The Geological Society of London 2011.

28

A. SAINTOT ET AL.

slope deformation. Using field examples, we attempt to show how a combination of large and favourably orientated inherited structures (as fold hinges, brecciated/cataclastic tectonic faults) in a slope may specifically determine the final volume of the gravitationally deformed slope. We focus on identifying the geological factors that may render the steep slopes of western Norway more susceptible to instability development. Besides lithological factors (weak v. hard rocks) and the abundance of pre-existing planar structures (favourably orientated with regard to the slope), we have the opportunity in the Storfjord area to determine how high-grade metamorphism polyphase ductile folding may enhance slope susceptibility. We also examine the stability of some of the studied rockslides by a quantification of cohesion values and friction angles of shallow-dipping basal sliding planes.

An overview of the geology and rock slope instabilities in western Norway A large variety of mostly Lower Palaeozoic and Precambrian metamorphic rocks cover western Norway, and the area under investigation is known as the Western Gneiss Region (Fig. 1). The latter refers to the setting of the early Palaeozoic Caledonian Orogeny where the abundant gneissic rocks are found in both autochthonous and allochthonous positions. During the Caledonian Orogeny, the rocks were severely reworked by a general NW– SE-oriented bulk crustal shortening that resulted in a SE-directed nappe transport, in regimes ranging from the deepest ductile domain to the brittle– ductile transition (Roberts & Gee 1985; Hossack & Cooper 1986; Roberts 2003). Poorly documented Precambrian events (Tucker et al. 1990; Ska˚r & Pedersen 2003; Røhr et al. 2004) remain recorded in the protoliths of the Caledonian metamorphic rocks. In Phanerozoic times, after the Devonian semi-ductile syn- or post- orogenic Caledonian collapse (Hossack 1984; Fossen 2000), intense brittle tectonic events affected western Norway including the Permo-Triassic and Jurassic rifting phases and the opening of the North Atlantic Ocean (Torsvik et al. 1997; Valle et al. 2002; Mosar 2003). However, the stratigraphic records of these tectonic phases are lacking onshore western Norway and the major faults that pertain to the listed brittle tectonic events are far from being accurately identified. Started from a recent but still debated time, an ongoing high-rated uplift brings the intensively tectonized rocks to the surface. Norway also underwent significant glacial and post-glacial tectonic activity including large-magnitude earthquakes at the time of ice-cap melting with triggered rockfalls and possibly avalanches, isostatic rebound and further

neotectonic activity (Olesen et al. 2000, 2004). The deep incision of glaciers produced steep slopes along typical U-shaped valleys (Fig. 2). Etzelmu¨ller et al. (2007) summarized well the resulting relief of western Norway with these words: alpine, steep slopes, heavily over-deepened glacial valleys. The sudden unloading of the steep slopes due to a rapid ice-cap melting associated with a large discharge of water into the slope fractures generated a large amount of gravitational structures in Norway (Blikra et al. 2002, 2006b). Blikra et al. (2006b) invoked frost wedging (i.e. opening of preexisting fractures due to the ice expansion) as a primary factor of slope instability structural development. In the recent past, permafrost covered most of the slopes of western Norway and, during a Younger Drias dramatic cooling event, it even occurred down to sea level (Blikra & Longva 1995) causing ice wedging at the largest possible scale [the lower limit of permafrost today is at c. 1500 m a.s.l. (metres above sea level) in western Norway]. Most of the rock slope instabilities in Norway are believed to have been structured during those two periods and to be presently dormant or inactive (Table 1). The extreme Alpine relief of the coastal fjord areas of Norway combined with the high concentration of pre-existing geological structures, seasonal heavy precipitation, snow melting and frost periods makes these areas prone to catastrophic slope activity (Blikra et al. 2006b). As such, the two counties of Møre og Romsdal and Sogn og Fjordane of western Norway are characterized by a high density of past large rockfalls and avalanches (Anda & Blikra 1998; Blikra et al. 2000, 2002, 2006b; Braathen et al. 2004; Furseth 2006; Bo¨hme et al. 2011), with some 340 historical events recorded over an area of 90 000 km2 and by 72 potential large rock slope failures (Fig. 1 & Table 1). In Sogn og Fjordane County, some hundred historical events are identified and some 20 locations show potential for slope failures (Fig. 3) (see Bo¨hme et al. 2011). In Møre og Romsdal County, three areas were carefully studied. The first two investigated areas are the Romsdalen and Sunndalen valleys because of their concentration of past events. Romsdalen is a 30 km-long, typical U-shaped glacial valley of western Norway (Fig. 4a). Seventeen past large slope failures are recorded along the Romsdalen Valley. Five large –very large instabilities are identified along the southern slope of the valley (Fig. 4a). Along the 30 km-long Sunndalen Valley, 10 past events are identified and five instabilities are currently under examination (Fig. 4b). The third investigated area in Møre og Romsdal County is the 360 km2 Storfjord region (Henderson et al. 2006; Henderson & Saintot 2011) and its 42 potential large rock slope failures (Fig. 5). Field investigation

ROCK SLOPE INSTABILITIES IN NORWAY

29

Fig. 1. Lithological map of western Norway with the location of past and potential large rock slope failures (Geological Survey of Norway database: http://www.ngu.no/kart/skrednett/ and http://www.ngu.no/kart/bg250/; UTM co-ordinates, zone 32N). The frames limit the four areas shown in detail in Figures 3– 5.

30

A. SAINTOT ET AL.

Fig. 2. Maps showing (a) the high relief of western Norway with the present-day contours of uplift and (b) values of slope gradients along the fjords and valleys (derived from a 25-m digital elevation model; UTM co-ordinates, zone 32N).

Table 1. Summary of the field observations made at the 72 investigated sites of western Norway. A black line splits the table according to the lithology at the slope instability (soft v. hard rocks, in the top and bottom parts of the table, respectively) Studied sites (located/ described on)

Lithology

State*

Rock topple, rockfall Rockfall

Dormant?

Strandanipa (Fig. 3) Storfjord site 17b (Figs 5 & 6)

Micaschists

Steiggjeberg (Figs 3 & 28)

Pyroxene-granulites, Rockfall mafic gneiss, syenites, granodiorites and cataclasites Pyroxene-granulites, Rockfall mafic gneiss

Activity

Byttejuv (Fig. 3)

Pyroxene-granulites, Rock topple mafic gneiss

Dormant?

Nærøydalen (Fig. 3)

Mangerite, gabbro, gneiss, anorthosite, syenite Phyllite allochthon over gneiss autochthon Micaschists

Translational rockslide, rock topple, rockfall

Rockfall activity

Translational rockslide

?

Phyllites/gneiss

Several translational rockslides

Kloppuri (Fig. 3)

Vik (Midtfjellet and Hesten) (Fig. 3) Vollan (Figs 4, 18 & 19)

Hjellane (Figs 3 & 15)

Ultramafics

Exfoliation

Foliation

Activity

Activity

Regional fractures (faults, lineaments)

Reactivated east –west contact between gneiss and ultramafics Cataclasic fault zone

Unstable north– south vertical foliation NNE –SSW steep and opened Horizontal sole detachment crossing the slope (anorthosites over mafics) Sliding plane

? Translational (þrotational) rockslide, flexural toppling

?

Fold

Toppling along foliation

Large regional tectonic contacts between squeezed Caledonian allochthonous units and trending parallel to the slope Gneiss oversthrusting phyllites- thrust plane 308 dipping toward the fjord

Maximum No. of volume inherited (Mm3) prominent structures

Easting UTM 32

Northing UTM32

Reference

0

0.05

309472

6836201

Henderson et al. (2008)

1

0.001

398636

6905544

1

0.001

433412

6792621

Selmer-Olsen (1972), Henderson et al. (2008)

1

0.01

430926

6787851

1

0.02

431642

6795996

1

0.5

378839

6749543

Lied (1984, 1987), Henderson et al. (2008) Jørstad (1957), Henderson et al. (2008) Henderson et al. (2008)

1

1

370600

6774600

Henderson et al. (2008)

1

100

505863

6938057

Saintot et al. (2008)

1

100

416513

6811642

Henderson et al. (2008)

31

(Continued)

ROCK SLOPE INSTABILITIES IN NORWAY

Type*

Studied sites (located/ described on)

Lithology

32

Table 1. Continued Type*

State*

Phyllites

Rockslide, complex Active

Viddalen (Figs 3 & 14)

Allochthonous phyllites over autochthonous gneiss

Translational rockslide, rock topple

?

Tussen (Figs 3 & 15)

Phyllites

Translational rockslide

?

Storfjord site 6b (Figs 5 & 29)

Micaschists, garnet-rich micaschists, mica-rich gneiss

Translational rockslide

Dormant

Tirskardskreda (Fig. 3) Katlenova (Fig. 3)

Gneiss

Rockfall

Foliated quartz diorite Conglomerates

Migmatitic gneiss

Foliation

Fold

Regional fractures (faults, lineaments)

No. of Maximum inherited volume prominent (Mm3) structures

Easting UTM 32

Northing UTM32

Reference

Domaas et al. (2000, 2002), Blikra et al. (2000, 2002, 2006b), Braathen et al. (2004), Olesen et al. (2000, 2004), Henderson & Blikra (2008) Domaas et al. (2002), Henderson et al. (2008), Bo¨hme et al. (2011) Henderson et al. (2008), Bo¨hme et al. (2011)

1

200 (900 – 1500)

399394

6747916

NW –SE lineament þ horizontal sole detachment

2

1

405509

6743020

NNE– SSW lineaments as back-crack and structures associated and parallel to sole detachment NE– SW and north–south regional fractures as back-crack and NW – SE fractures as transfer

2

30

429274

6822214

5

15

389618

6887976

Henderson et al. (2006)

Active

0

0.001

329764

6789038

Rockslide

Activity

0

0.04

299673

6789681

Rock topple

Active?

0

0.1

295095

6790289

Rock topple

Active?

0

0.25

406596

6749619

W-Hyllestad, Conglomerates Lifjellet (Fig. 3)

Rockfall

Active?

0

0.25

294234

6790808

Henderson et al. (2008) Henderson et al. (2008) Henderson et al. (2008), Bo¨hme et al. (2011) Lied (1989), Henderson et al. (2008) Henderson et al. (2008), Bo¨hme et al. (2011)

Storfjord site 15 (Fig. 5)

Rockfall

?

1

0.001

398950

6894113

E-Hyllestad (Fig. 3) Terakamben (Fig. 3)

Gneiss

Gentle fold with axis azimuth parallel to slope dip direction.

Sliding plane

Fold axes parallel to sliding vector

Unstable fold hinge (crest)

A. SAINTOT ET AL.

Fla˚m (Fig. 3)

Exfoliation

0.001

398848

6900405

1

0.001

397444

6892294

1

0.001

406004

6898608

Sliding plane

1

0.001

416847

6898988

Sliding plane

1

0.001

418820

6900727

1

0.001

419260

6900300

1

0.005

382731

6785994

1

0.03

312548

6820914

1

0.05

343411

6790016

Sliding planes

1

0.5

422135

6913043

Sliding planes

1

0.5

423030

6913169

1

2

501134

6941198

Henderson & Saintot (2007), Saintot et al. (2008)

1

3

445567

6921062

Henderson & Saintot (2007)

1

7

439875

6920534

1

12.5

341573

6789323

1

40

437833

6924487

Henderson & Saintot (2007) Henriksen (2005), Eiken (2007), Henderson et al. (2008) Braathen et al. (2004), Blikra et al. (2006b)

Gneiss

Rockfall

?

Storfjord site 28 (Fig. 5) Storfjord site 31 (Fig. 5) Storfjord site 41b (Fig. 5) Storfjord site 42 (Fig. 5) Storfjord site 42b (Fig. 5) Hallandshammaren (Fig. 3) Gjøringbøfjellet (Fig. 3)

Gneiss

Rockfall

?

Gneiss

Rockfall

?

Gneiss

Translational rockslide Translational rockslide Rockfall

? ?

Translational rockslide Rock topple

Dormant

Gra˚berget (Fig. 3)

Gneiss

Rockslide, rock topple

Dormant

Storfjord site 39 (Fig. 5) Storfjord site 39b (Fig. 5) Ottem (Fig. 4)

Gneiss

Translational ? rockslide Translational ? rockslide Rockslide, complex ?

Flatmark (Figs 4 & 9)

Gneiss

Rockslide, rock topple

Dormant

Svartinden (Figs 4 & 8) Stopelen (Fig. 3)

Gneiss

Translational rockslide Rockslide

Dormant

Parallel to slope, steep and reopened as back-crack Sliding plane

Børa (Figs 4 & 9)

Gneiss

Gneiss Gneiss Gneiss Gneiss

Gneiss Gneiss

Foliated quartz monzonite

NNE– SSW regional fractures as open cracks Cracks along north–south lineaments Fault cutting the ridge

Active

Dislocation of fold hinges Sliding planes

Dormant?

Active?

Rockslide, complex Active

Folded gneiss

Sliding plane-opened steep slope-parallel exfoliation

Back-crack developed where small-scale vertical folds

Sliding planes

Parallel to slope, steep and reopened as back-crack

Reactivation of gneiss– granite contact parallel to cliff

Henderson et al. (2006)

Henderson et al. (2008) Moxnes (2002), Henderson et al. (2008) Henderson et al. (2008)

33

(Continued)

ROCK SLOPE INSTABILITIES IN NORWAY

1

Storfjord site 16 (Fig. 5)

Studied sites (located/ described on)

Lithology

34

Table 1. Continued Type*

State*

Gneiss

Rockfall

?

Foliated quartz monzonite

Translational rockslide

Dormant

Storfjord site 3 (Fig. 5) Storfjord site 38 (Figs 5 & 8) Storfjord site 1 (Fig. 5) Storfjord site 2 (Figs 5 & 22)

Gneiss

Rockslide

Gneiss Gneiss Gneiss

Easting UTM 32

Northing UTM32

Reference

396937

6884863

Sliding planes

North– south lineaments

2

0.1

383401

6843354

?

Sliding planes

2

0.5

396159

6898227

Translational rockslide Translational rockslide Translational rockslide

?

Sliding planes

2

0.5

419049

6914939

?

Sliding planes

Back-crack along NNW – SSE regional fractures North– south brecciated fault as transfer Back-crack NNE –SSW

2

1

396005

6901730

2

1

396090

6899135

Henderson et al. (2006)

Translational rockslide Translational rockslide Translational rockslide Translational rockslide Translational rockslide

?

Sliding planes

2

1.5

348366

6785839

?

Sliding plane

2

2

406765

6888339

Henderson et al. (2008) Henderson et al. (2006)

?

Sliding plane

North– south transfer

2

2

406224

6888422

?

Sliding plane

NNE– SSW transfer

2

2

400972

6889600

North– south transfer

2

2

403072

6890107

Crack along north–south regional fractures

2

2

409826

6906792

Large regional tectonic contacts between squeezed Caledonian allochthonous units and trending parallel to the slope Transfer zones along north– south brecciated faults

2

5

506485

6936915

Henderson & Saintot (2007), Saintot et al. (2008)

2

5

397952

6869539

Henderson et al. (2008)

NW –SE transfer

2

5

399662

6906560

Henderson et al. (2006)

?

Gneiss

Rockslide

?

Ivasnasen (Figs 4, 18 & 19)

Augen gneiss

Translational rockslide

Active?

Oppiga˚rdshyrna (Fig. 3)

Foliated quartz monzonite

Rockslide, rock topple

Dormant

Storfjord site 18 (Figs 5 & 9)

Gneiss

Translational? Rockslide

Dormant

Gneiss

Maximum No. of volume inherited (Mm3) prominent structures 0.001

Storfjord site 23 (Fig. 5)

Gneiss

Sliding plane

Regional fractures (faults, lineaments)

2

Gneiss

Gneiss

Fold

North– south lineaments

La˚nefjorden (Fig. 3) Storfjord site 11 (Figs 5 & 12) Storfjord site 11b (Figs 5 & 12) Storfjord site 12 (Figs 5 & 7) Storfjord site 12b (Figs 5 & 12)

Gneiss

Foliation

?

Main sliding plane N0408/0508 þ north– south fractures Slope-parallel trending lineaments Localization of back-crack

Localization of back-crack at the hinge Shalow fjord-dipping sliding planes Sliding planes

Localization of cracks along axial fold planes Steep opened plane

Lied (1975), Henderson et al. (2008)

Henderson et al. (2006)

Henderson et al. (2006) Henderson et al. (2006)

A. SAINTOT ET AL.

Storfjord site 8 (Figs 5 & 23) Rustøyane (Fig. 3)

Exfoliation

Gneiss

Rockslide, rock topple

Rockfall activity

Fold axes parallel to sliding vector

Storfjord site 14b (Fig. 5)

Gneiss

Dormant

Sliding plane

Storfjord site 2b (Fig. 5) Storfjord site 21 (Figs 5 & 10)

Gneiss

?

Sliding planes

Contact gneiss/ ultramafic rocks

Translational (þrotational?) Rockslide Translational rockslide Translational rockslide

Oppstadhornet (Oterøya) (Fig. 1)

Gneiss

Rockslide, complex Active

Hegguraksla (Fig. 5)

Banded mica-gneiss Rockslide, rock topple

Active

Sliding planes

Large-scale folds spatially constrain the gravitational deformation

Storfjord site 20 (Figs 5 & 11)

Gneiss

Translational rockslide

?

Vertical cracks

Sliding plane guided by fold geometry

Storfjord site 22 (Figs 5 & 26)

Gneiss

Translational rockslide

?

Opened vertical foliation perpendicular to slope

Storfjord site 26 (Figs 5 & 27)

Contact between two gneiss units

Translational rockslide

?

Sliding plane

Storfjord site 27 (Figs 5 & 27)

Contact between two gneiss units

Translational rockslide

?

Sliding plane

Storfjord site 10 (Figs 5 & 7)

Gneiss

Translational rockslide

?

Dormant

Sliding plane guided by fold geometry Sliding plane

Steep opened foliation as cracks Steep opened foliation as cracks Sliding planes

Large back-crack along north– south regional fractures Back-crack NNE –SSW

2

5

393283

6893897

Henderson et al. (2006)

2

7.5

395731

6891907

Henderson et al. (2006)

Main sliding plane N0308/458 North– south regional fractures as open cracks NW –SE regional fractures as transfer

2

10

395242

6900486

2

15

409275

6903646

Henderson et al. (2006) Henderson et al. (2006)

2

20

388951

6953563

2

10 to 60

415455

6907550

North– south cracks

3

0.1

405511

6904418

Back-crack along vertical north– south regional fault and sliding along west-dipping north – south faults North– south regional fractures as open cracks North– south regional fractures as open cracks North– south regional faults as transfer and NW –SE regional fractures as back-crack

3

1

409319

6906098

3

2

415515

6904360

3

2

416537

6902887

Henderson et al. (2006)

3

2

403639

6888157

Henderson et al. (2006)

Robinson et al. (1997), Blikra et al. (2001, 2002), Braathen et al. (2004), Derron et al. (2005) Furseth (1985, 2006), Harbitz et al. (2003), Braathen et al. (2004), Blikra et al. (2006a), Oppikofer et al. (2008a), Oppikofer & Jaboyedoff (2008) Henderson et al. (2006) Henderson et al. (2006)

35

(Continued)

ROCK SLOPE INSTABILITIES IN NORWAY

Storfjord site 4 (Figs 5 & 10)

36

Table 1. Continued Studied sites (located/ described on)

Lithology

Type*

State*

Exfoliation

Sliding plane

Foliation

Gneiss

Translational rockslide

?

Ha˚rstad (Fig. 4)

Gneiss

Translational rockslide

Dormant

Vertical, reopened, transfer Sliding planes

Storfjord site 10b (Figs 5 & 25)

Gneiss

Translational rockslide

?

Sliding plane

Storfjord site 13 (Fig. 5)

Gneiss

Translational rockslide

?

Sliding planes

Storfjord site 9 (Figs 5 & 24)

Gneiss

Translational rockslide

?

Sliding planes

Mannen (Figs 4, 13 & 21)

Gneiss

Translational rockslide

Active

Parallel to slope, steep and reopened as back-crack

Storfjord site 21b (Fig. 5)

Amphibolites and augen gneiss overthrusting gneiss

Translational rockslide

Dormant?

Sliding planes

Regional fractures (faults, lineaments)

No. of Maximum inherited volume prominent (Mm3) structures

Easting UTM 32

Northing UTM32

Reference

Back-crack along north– south regional faults

3

4

393396

6892440

Back-crack along 508 east-dipping brecciated fault and cracks along NW –SE regional fractures Back-crack along north– south lineamentsNW –SE regional fractures as transfer North– south regional faults as transfer and east –west lineaments as incipient back-crack North– south regional faults as transfer and east –west/NE– SW regional fractures as back-crack North– south faults as transfer zones

3

5

475507

6945271

Henderson & Saintot (2007), Saintot et al. (2008)

3

5

405423

6886351

Henderson et al. (2006)

3

5

394750

6889093

Henderson et al. (2006)

3

7.5

399845

6885400

Henderson et al. (2006)

3

10

436542

6925595

Henderson & Saintot (2007), Dahle et al. (2008)

3

20

409421

6903574

Localization of back-crack following an east– west recumbent fold Fold axial planes North– south þ NE– SW lineaments parallel to slope

A. SAINTOT ET AL.

Storfjord site 5 (Figs 5 & 20)

Fold

Gneiss

Translational rockslide

Dormant

Sliding planes

Storfjord site 7 (Figs 5 & 23)

Gneiss

Translational rockslide

Dormant

Sliding plane

˚ knes (Fig. 5) A

Gneiss

Translational rockslide

Active

Sliding plane

Gikling (Figs 4, 16 & 17)

Gneiss/ fault rocks

Active Translational rockslide (with a slight spreading of the allochthonous above the sole detachment)

Hinge of fold (bulge) limiting two blocks. Back-crack along tight fold hinges Fold axial planes localized part of the back-fracture

NE– SW þ north –south regional fractures as transfer – NW –SE fractures as back-crack NE– SW þ north –south regional fractures as transfer – NW –SE fractures as back-crack North– south regional fractures, large brecciated fault as eastern border /large fault as western border

4

5

393938

6886845

4

12

394063

6886297

5

80

395600

6895500

Sole detachment, large steep faults as limits of instability

5

100

492671

6945891

Henderson et al. (2006)

Blikra et al. (2006a), Venvik Ganerød et al. (2008), Oppikofer et al. (2008b), Jaboyedoff et al. (2011) Henderson & Saintot (2007), Saintot et al. (2008)

*Type of movement in bedrock according to Varnes (1978) and Cruden & Varnes (1996), and state of activity as in Cruden & Varnes (1996). Given measurement of planar structures: strike and dip angle (right-hand rule).

ROCK SLOPE INSTABILITIES IN NORWAY

Storfjord site 7b (Figs 5 & 23)

37

38 A. SAINTOT ET AL.

Fig. 3. The 22 studied slope instabilities in Sogn og Fjordane County with the distribution of historical slope failures reported on a geological map (Geological Survey of Norway database: http://www.ngu.no/kart/skrednett/ and http://www.ngu.no/kart/bg250/; legend of the lithologies as in Fig. 1; UTM co-ordinates, zone 32N).

ROCK SLOPE INSTABILITIES IN NORWAY

39

Fig. 4. (a) The past gravitational failures and the four studied slope instabilities in the Romsdalen Valley and a photograph illustrating the typical U-shape overdeepened glacial Romsdalen Valley (view to the east). (b) The past gravitational failures and the five studied slope instabilities in the Sunndalen Valley. (Background: scale 1/250 000 geological map; UTM co-ordinates, zone 32N: Nilsen & Wolff 1989; Tveten et al. 1998; Geological Survey of Norway Database: http://www.ngu.no/kart/skrednett/ and http://www.ngu.no/kart/bg250/).

40

A. SAINTOT ET AL.

Fig. 5. Map of the Storfjord area showing the historical slope failures and the 40 unstable rock slopes including the two ˚ knes and Hegguraksla (background: scale 1/250 000 geological map on a 25-m permanently monitored rockslides, A digital elevation model; UTM co-ordinates, zone 32N: Tveten et al. 1998; Geological Survey of Norway Database: http://www.ngu.no/kart/skrednett/ and http://www.ngu.no/kart/bg250/).

at the regional scale of Storfjord was decided not only because this area is characterized by the largest number of past events but also because two of the permanently monitored active large rock˚ knes and Hegguraksla rockslides of Norway, A slides, are located there (Blikra et al. 2006a)

(Fig. 5). Extensive fieldwork allowed a detailed examination of the known rock slope instabilities, for new ones to be identified and mapped (Henderson et al. 2006, 2008; Henderson & Saintot 2007; Saintot et al. 2008). It resulted in the total of 72 current slope instabilities. Most of the studied

ROCK SLOPE INSTABILITIES IN NORWAY

slopes show a high susceptibility to develop rockfalls of limited volumes (Derron et al. 2005; Crosta et al. 2007; Derron 2009) not only because of the steepness of the slopes but also because of the high density of interconnected tectonic joints in the rocks. The rockfalls that: (1) may threaten the community; (2) have a significant volume of 1000 m3 and more; and (3) show a high frequency of occurrence are included in the 72 analysed instabilities of this paper. Also, we encountered during our field studies 43 unstable volumes equal to or above 1 Mm3 (see Table 1) prone to evolve to rock avalanches (with a threshold value as in Evans et al. 2006). These are exclusively rockslides (deep-seated rockslides, and translational or rotational rockslides) or display complex slope movements combining rock topples and slides (according to Varnes 1978 and Cruden & Varnes 1996) (Table 1).

Influence of bedrock geology and the reactivation of pre-existing ductile and brittle structures in slope instability development The lithological factor in slope instability development The lithological map (Fig. 1) shows mainly the predominance of hard gneissic rocks in western Norway. They form the autochthonous basement but also most of the allochthons of the Caledonian tectonic units. However, the same map (Fig. 1) also displays some patches of other lithologies that are intrinsically weaker than the surrounding gneisses as amphibolites, schists and micaschists, and weathered mafic and ultramafics. Out of the 72 studied slope instabilities in western Norway, 13 developed in such rocks assumed to be softer than the gneisses. The ultramafics that we observed during the field campaign are strongly weathered and highly fractured compared to the host gneissic rocks. In detail and for example in Storfjord area, the frequency of fractures in the ultramafics is higher than 4 per m and three different orientations of interconnected joints are common (Fig. 6). Such brittle deformation leads to a mechanical weakening and enhances the chemical weathering. Consequently, the ultramafics can no longer be considered as hard rocks. When lying on steep slopes, the patches of ultramafics form very unstable packages that disaggregate and generate rockfalls. The failure surfaces generally coincide with the contact between the gneiss and the ultramafics. For the only Sogn og Fjordane County, 11 of the 20 slope instabilities developed in such soft rocks. Nine of them are specifically concentrated to the east

41

along deep narrow fjords that installed on a micarich phyllitic and mafic–ultra-mafic allochthonous belt (Fig. 3) (see Bo¨hme et al. 2011). This zone is assumed to be of high susceptibility to large gravitational slope deformation. The largest unstable slope of Norway developed along this belt at Fla˚m (location shown in Fig. 3) (Henderson & Blikra 2008). Its maximum volume may reach 200 Mm3. Therefore, the lithology appears to be an important factor controlling the development of unstable masses along the steep sides of the fjords and the failure plane(s) often reactivate(s) the contact between the different lithological units.

The role of pre-existing ductile structures The bedrock of Norway is of particular interest when investigating how the ductile fabrics influence slope stability. From Sogn og Fjordane to Møre og Romsdal counties and further north, to the Møre Trøndelag Fault Complex, the Caledonian ductile strain increases together with the regional metamorphism that ranges from low– middle to ultrahigh grades (with a recorded peak metamorphism of 800 8C at 100 km depth in the eclogites of the coastal area of Møre og Romsdal County: Tveten et al. 1998). The overall metamorphic foliation is rather flat or gently dipping along open folds in the whole Sogn og Fjordane County. Towards the north (i.e. towards the internal high–ultra-highgrade metamorphic core of the Caledonides) and in the Møre og Romsdal County, close –tight folds become more abundant and polyphase folds are common. In the Storfjord area, the ductile deformation of the basement is subsequent to at least three folding phases of the Caledonian Orogeny. At the Møre Trøndelag Fault Complex (Fig. 1), the ductile fabric tends to be parallel to the fault segments, that is, foliation and axial surfaces of close – tight folds are steep –vertical and ENE –WSW trending. It is assumed that the Møre Trøndelag Fault Complex was the locus for a very high shear strain that largely overprinted the older trends of ductile deformation. The following field case studies exemplify how the ductile grain may decrease the slope stability. The role of foliation in slope instability development. Studies of rock slope failures and associated discontinuities in intensively foliated rocks are documented; for example: in British Colombia, Canada, by Evans & Clague (1998, 1999); in the Central Italian Alps by Govi et al. (2002); in the Andes, Argentina, by Hermanns & Strecker (1999); as well as in Norway by Blikra et al. (2000, 2001, 2002, 2006b), Braathen et al. (2004), Henderson et al. (2006, 2008) and Venvik Ganerød et al. (2008) among others. In western Norway the

42

A. SAINTOT ET AL.

Fig. 6. (a) Aerial orthophotograph showing a lens of ultramafics (dunites) prone to rockfalls in the gneissic basement of Storfjord area (site 17b, located in Fig. 5). (b) Photograph of a NW–SE-orientated wall in the ultramafics with the NE– SW joints, gently NW-dipping joints and the steep foliation dividing the rock in blocks. (c) Stereonet of steep structures: joints in black and foliation in grey. The high frequency of joints in the ultramafics contributes to a mechanical weakening and enhances the chemical weathering.

ROCK SLOPE INSTABILITIES IN NORWAY

foliation is a secondary metamorphic foliation with, for gneissic rocks, a differentiated compositional banding between felsic and mafic minerals, and, for phyllitic rocks, cleavage and schistosity from platy mineral planar alignments (e.g. Twiss & Moore 1992; Passchier & Trouw 2005). It is actually the main penetrative structural grain, and the first and inherent factor of rock strength anisotropy to consider when studying rock slope instabilities (Braathen et al. 2004; Blikra et al. 2006b). Several field examples given below focus on the reactivation of foliation planes as gravitational structures on overdip –underdip cataclinal slopes (with anaclinal and cataclinal slopes as defined by Powell 1875: see in Cruden & Hu 1996; Cruden 2000). Foliation reactivated as basal shear planes of rockslides on overdip cataclinal slopes (i.e. with respect to foliation). Thirty of the 72 studied slope instabilities correspond to rockslides with a basal shear plane parallel to foliation on overdip cataclinal slopes (Table 1). Most of them are located within the dioritic –granitic gneiss of the Caledonian autochthonous basement of western Norway. Compositional variation in this host rock is such that some layers are extremely mafic and consist of almost 100% biotite and/or hornblende. We observed that the shear planes at the base of the rockslides are commonly and preferentially developed along the hornblende –biotite-rich layers and not along feldspar – quartz-rich layers. The inherently weak properties of mica and amphiboles (their low shear strength compared to quartz and feldspar) and their platy nature focus strain at the base of the sliding block. Subsequently, gouges and breccias developed along these weak planar structures attesting to shearing along the plane (see also in Henderson & Saintot 2011). Figures 7 and 8 illustrate four localities where rockslides developed with the reactivation of the foliation as basal shear surfaces. In these four cases rockslides occurred in the past and the current slope instabilities are propagating from the previous failure surfaces. Site 10 of the Storfjord area (Fig. 7a–c; location in Fig. 5) displays foliation that dips 208 –358 towards the fjord. A back escarpment produced by the opening of a steep fracture delimits the potential unstable block with an estimated volume of 2 Mm3. Site 10 is one of the potential rockslides lying on the approximately 558 dipping southern face of Geiranger Fjord, along which the foliation dips towards the fjord and is systematically reactivated as basal shear planes (sites 7, 7b, 9, 10 and 10b located in Fig. 5). The same setting is observed at site 12 in Storfjord (Fig. 7d, e; located in Fig. 5) (Henderson et al. 2006) with, in addition, an exfoliation that occurred along the foliation planes and that contributed to

43

their weakening by opening. Such a process is also observed all along the slope in the vicinity of site 38 in Storfjord (Fig. 8a) (Henderson et al. 2006). Exfoliation developed parallel to the 308 dipping foliation and may have facilitated the development of shear planes at the base of the large unstable slices at sites 38, 39 and 39b (located in Fig. 5). Figure 8b, c displays a 7 Mm3 unstable block, called Svartinden, in the Romsdalen Valley (located in Fig. 4a) that slid on a foliation-parallel plane which moderately dips towards the valley (Henderson et al. 2004; Henderson & Saintot 2007). The friction created a 20 cm-thick microbreccia along the sliding plane. This foliationparallel basal sliding plane is also the westwards prolongation of a previous failure surface (Fig. 8b). Propagation of failure surfaces along steep – vertical foliation. Eleven investigated instabilities of large volumes (sites 5, 18, 20, 22, 26 and 27 in Storfjord, see Fig. 5; Mannen, Flatmark, Børa in the Romsdalen Valley, see Fig. 4; Kloppuri and Byttejuv in Sogn og Fjordane, see Fig. 3 & Table 1) are located on slopes where the foliation is steep – vertical, and 10 of them are found on underdip– dip cataclinal slopes (with respect to foliation). Site 18 of Storfjord (Fig. 9a –c) accurately illustrates how the steep –vertical foliation along the cliffs can promote large columnar failures. This site has foliation that steeply dips towards the fjord. Large extensional structures developed parallel to the foliation. A down-dip displacement along these opened fractures resulted in the formation of several metre-high steps on the top of the unstable block. There is, therefore, evidence of a combination of shear and tension on the foliation planes. However, it is not clear what is controlling the sliding of the block at its base (Fig. 9b, c). The southern face of the Romsdalen Valley is characterized by a steep –vertical cliff parallel to foliation. The two large unstable fields of Flatmark and Børa (Fig. 9d, e; located in Fig. 4a) show the opening of foliation under gravitational forces to form large cracks at the back of downwardsdisplaced blocks. At Børa, on the plateau, a palaeoglacial valley trends parallel to the cliff (and therefore to the foliation; Fig. 9e). The ice-river is assumed to have partly contributed to the destabilization of the subglacial bedrock (at the back of the present-day cliff; Fig. 9e) by a partial opening of the foliation (probably by hydraulic fracturing, cf. Boulton & Caban 1995). The final destabilization at the edge of the plateau was coeval with debuttressing subsequent to ice retreat in the Romsdalen Valley. Folds as a factor for slope destabilization. Field observations at 17 localities allow identification of

44

A. SAINTOT ET AL.

Fig. 7. (a) Aerial photograph of a potential rockslide of 2 Mm3 at site 10 in Storfjord (location in Fig. 5) showing the presence of back escarpments (dashed line). A previous slide failed from the same structure (solid line). Additional scarps are marked in dotted lines. (b) Down extension of the back-crack structure (white dashed line) to a foliation parallel sliding plane (black line) (view to east; site 10 in Storfjord). (c) Close view of opened relay-stepped structures subsequent to the shear along the foliation-parallel sliding plane (site 10 in Storfjord). (d) Photograph of 2 Mm3 remaining instability west of a past failure at site 12, 900 m a.s.l., in Storfjord (located in Fig. 5) and (e) stereonets of structures (Schmidt’s projection, lower hemisphere). Exfoliation occurred along the foliation planes, and probably contributed to their loss of strength and to their reactivation as basal sliding planes. Note the open vertical joints acting as ‘back-crack’ type fractures and a conjugate system of joints with an intersection normal to the shear plane. These sets of joints are limited to the block and particularly consistent with newly developed structures related to downwards movement of the unstable block.

the role of fold geometries in the slope destabilization. Besides the intrinsic weakening of the rock volume during folding under ductile conditions (due to the flattening and stretching of minerals, in particular along the limbs), the fold geometries certainly guided the development of new failure surfaces. As mentioned earlier, the interferences of several phases of Caledonian ductile folding is particularly well preserved in inner Møre og Romsdal County where three phases can be distinguished. The Caledonian ductile folding tends to attenuate southwards in Sogn og Fjordane County to lead to gentle open folds. To the north it was totally overprinted by a late high shear strain along the Møre Trøndelag Fault Complex that developed a parallel-to-fault vertical foliation. Fourteen of the 17 sites where we observed the gravitational structures guided by pre-existing folds are therefore in

the inner parts of Møre og Romsdal County and in the Storfjord region, and exemplify the loss of slope stability at a regional scale owing to interferences of several ductile folding phases. Chosen field examples show the influence of centimetre- to kilometre-scale amplitude folds in the development of gravitational structures. At sites 4, 11, 11b, 12b, 20, 21 and 21b in the Storfjord area (located in Fig. 5), basal shear planes developed according to the geometrical constrains of the preexisting folds. The attitude of folds is described according to McClay (1991, table 3.2, p. 49). At site 4 (Fig. 10) the instability consists of at least seven column-shaped blocks with a potential total volume of 2–5 Mm3. These are currently detached from the north –south vertical cliff by a set of north–south deeply eroded regional faults from which extensional cracks propagate. A survey

ROCK SLOPE INSTABILITIES IN NORWAY

45

Fig. 8. (a) Structures observed from the helicopter at site 38, 1000 m a.s.l., in Storfjord (located in Fig. 5). The foliation dips moderately towards the valley along which sliding planes may have developed (dashed line). A weakening of the foliation surfaces occurred by opening when exfoliation superimposed. Along the potential shearing plane a blocky talus can be observed and can be the result of rock disintegration subsequent to movement along the shear plane. There is some evidence for steep, extensional fractures (dotted line) and the presence of an old sliding plane (from Henderson et al. 2006). The volume of the unstable slice is estimated to 0.5 Mm3. (b) Photograph of the remaining 7 Mm3 unstable volume called Svartinden, Romsdalen Valley, at 1300 m a.s.l. (located in Fig. 4a). The previous slide occurred along a foliation-parallel surface that propagated westwards and defined the brecciated basal shear plane of the remaining volume (Henderson et al. 2004; Henderson & Saintot 2007). (c) Same unstable block, Svartinden in Romsdalen Valley, viewed to the east.

by helicopter identified shallow fjord-dipping basal planes at the foot of the columns. A possible contributing factor to the basal shear plane development here is the presence of fjord-plunging inclined small-scale folds in the gneiss (Fig. 10b). These have a trend of approximately 1008N, exactly parallel to the extensional direction of the column failure (i.e. perpendicular to the main set of opened joints: Fig. 10c), and also a plunge of 308 –358ESE (Fig. 10d), which is parallel to the supposed dip of the low-angle sliding planes. The favourably orientated pre-existing ductile fabric in this area therefore promotes a slippage of material at a moderate – shallow angle eastwards towards the fjord. At site 21 in Storfjord (located in Fig. 5: Henderson et al. 2006) the slope displays an impressive rockfall (Fig. 10e), which threatened the small community

of Norddal village some years ago. Examination of the mountainside shows that instabilities still remain present. Two basal surfaces limit an estimated 15 Mm3 unstable volume and define a SW-facing wedge (Fig. 10e, f ). The 458 SE-dipping basal shear plane is subparallel to a lithological contact between mafic rocks at the base and a grey dioritic gneiss above. The gneiss unit displays steep open cracks (Fig. 10f). In addition, the two units display upright 458 plunging microfolds, and the shear plane is perpendicular to the axial planes of the microfolds and contains the fold axes (Fig. 10f, g). The second basal shear plane was not accessible and could only be observed from the helicopter. The rock slope instability at site 20 (Storfjord; located in Fig. 5: Henderson et al. 2006) lies on

46

A. SAINTOT ET AL.

Fig. 9. (a) and (b) Structures observed at site 18 of Storfjord (located in Fig. 5): steep foliation parallel to slope and fractures opened along the foliation and (c) sketch of hypothetical rock failure process (in Henderson et al. 2006). (d) and (e) Photographs showing the opening of the foliation at the edge of the plateau, southern side of the Romsdalen Valley (Fig. 4a). (d) The unstable block collapsing towards the valley at Flatmark along the nearly vertical foliation (Henderson & Saintot 2007). (e) View to the WNW of Børa slope instability (from www.norgei3d.no). Cracks developed parallel to steep foliation planes trending parallel to the cliff of the Romsdalen Valley. The palaeoglacial valley trends parallel to the cliff. The ice loading is assumed to have partly opened the steep foliation and to have therefore contributed to the destabilization of the edge of the plateau. The total destabilization occurred when the ice melted in the Romsdalen Valley.

an approximately 508 west-facing slope and is restricted to the Caledonian allochthonous augen gneiss units above the thrust surface (Fig. 11a). The southern side of the slope instability is observed in a deep gully that is a free boundary with respect to the unstable zone, frequently affected by rockfalls (Fig. 11a). At approximately 250 m altitude along the gully, the steeply SSE dipping foliation is clearly reactivated and promotes large open fractures (Fig. 11b, c). The large NNE –SSW joints are also reopened (Fig. 11b). Further up the gully the gneiss is largely folded and, at the base of what we

define to be the uppermost unstable block, we observed a large 408 dipping shear plane parallel to the axial planes of the small-scale folds (Fig. 11b, d). The instabilities at this site are quite small (0.1 Mm3 maximum) and consist of detached slabs from the southern free border. The limits of the slabs are the steep foliation, or the axial fold planes, and the NNE–SSW fracture sets (Fig. 11b). Sites 11, 11b and 12 in Storfjord (located in Fig. 5) lie on the approximately 608 northern side of Geiranger Fjord, where listric structures dipping towards the fjord are observed (Fig. 12). Some

ROCK SLOPE INSTABILITIES IN NORWAY

47

Fig. 10. Folds controlling the geometry of the basal sliding planes at sites 4 and 21 (Storfjord, located in Fig. 5: see Henderson et al. 2006). (a) Large detached column-shaped blocks at site 4 on a 1-m digital elevation model (UTM co-ordinates, zone 32N). North– south regional faults are deeply eroded and the resulting canyons are free boundaries at the back of the blocks. Cracks opened parallel to the fault zones. (b) Photograph from the helicopter of a basal shear plane (dashed line) at the foot of a block and traces of metre-scale folds (underlined by dotted lines). The inferred dip direction along the shallow fjord dipping basal shear planes is parallel to the fold axes at site 4. (c) Photograph from the helicopter of north– south cracks that detached columns from the north–south cliff and traces of metre-scale folds (dotted lines). The trend of fold axes (arrows) is parallel to the cracks. Stereonet of field data with north– south cracks as black dashed lines, foliation as grey lines and fold axes as dots. (d) Photograph of one of the inclined folds outcropping on the detached columns. (e) Orthophotograph of site 21 (UTM co-ordinates, zone 32N) showing the rockfall deposits and the unstable wedge (limited by the dashed line). (f ) Photograph of the unstable wedge of site 21 (from the helicopter). Stereonet of field data: the NW basal shear plane of the wedge of site 21 (dotted lines) contains the fold axes (dots) and developed perpendicular to the fold surfaces (full lines). (g) Sketch of the geometrical relationships between basal shear planes and inclined or upright gently plunging folds as observed at sites 4 and 21, respectively.

48

A. SAINTOT ET AL.

Fig. 11. Unstable slope at site 20 in Storfjord (UTM co-ordinates, zone 32N; located in Fig. 5: see Henderson et al. 2006). (a) The gravitational deformation is restricted to the overlying Caledonian allochthonous augen gneiss unit. The present-day activity is along the free boundary on the northern side of a gully. Traces of foliation underlined with dotted lines: foliation is undulating and steep in the lower part of the slope and displays gently inclined folds in the upper part of the slope (b) Stereonets of structures measured on the field along the northern active edge of the gully. (c) Photograph of the structures that detach large slabs from the wall: NNE–SSW open fractures and east–west steep foliation surfaces. The east–west steep surface also displays the horizontal hinges of the undulating steep foliation. (d) Photograph of a sliding plane (dashed line) localized parallel to the inclined axial surfaces of the small-scale horizontal folds.

rock slope failure activity occurred in the past along such listric planes (Fig. 12a–c). An important feature at these three sites is the role of the preexisting horizontal fjord-inclined fold that guided the formation of the single listric faults. These

faults are parallel to the inclined axial fold surface and act as sliding planes at the base of the instabilities. They correspond to back-cracks where they steepen up towards the surface (Fig. 12a–c: Henderson et al. 2006). At site 12b the listric fault has an

ROCK SLOPE INSTABILITIES IN NORWAY

49

Fig. 12. (a) – (c) Observations made at sites 11, 11b and 12b in Storfjord (located in Fig. 5; see Henderson et al. 2006). The listric failure planes on this slope are guided by the geometry of horizontal inclined folds. (a) Site 11b is at the position of a previous slide (picture from the helicopter). Traces of foliation as dotted white lines displaying the hinge of an inclined horizontal fold. (b) The low-angle basal shear plane is parallel to the inclined axial surface and steepens up into an extensional back-crack structure (picture of site 11 from the helicopter). (c) Structures observed at site 12b. The basal shear plane has a listric geometry dipping 308 –358 towards the fjord and parallel to the foliation and steepening up into an extensional fracture along the axial plane of an antiform. A block already detached from this structure and was laterally guided by a north–south-trending old fault zone that acted as a transfer fault. (d) Horizontal inclined folds controlling the geometry of the basal sliding planes. At five sites in the Storfjord area (sites 11, 11b, 12b, 20 and 21b located in Fig. 5) the sliding plane at the base of the rock slope instability is developed parallel to the moderately inclined axial plane of horizontal folds. The dip-slip direction of blocks is inferred to be perpendicular to the fold hinges.

angle of 308 to 358 towards the fjord, parallel to the axial plane of an antiform, and steepens up as it reaches the surface to form an extensional fracture truly cutting the hinge of the antiform. The eastern limit of the failed block at site 12b is a north– south epidote-rich fault plane that acted as a transfer fault (i.e. the lateral boundary between the unstable and stable parts of the slope: Fig. 12c). The basal shear planes of rock slope instabilities at

sites 11, 11b, 12b, 20 and 21b in Storfjord (located in Fig. 5: Henderson et al. 2006) developed parallel to axial surfaces of horizontal folds moderately inclined towards the fjord or valley and the inferred slip direction of the instabilities is perpendicular to the fold axes (Fig. 12d). At the active rockslide of Mannen (Romsdalen Valley, located in Fig. 4: Henderson & Saintot 2007; Dahle et al. 2008), the back-crack of the

50

A. SAINTOT ET AL.

collapsing block is localized on a folded zone and opens perpendicularly to the axial surfaces of recumbent horizontal folds (Fig. 13). This site is a good example of the weakening of the rock mass due to folding. The folds promote anisotropy in terms of strength that will be the locus for the development of gravitational failure surfaces. The field examples above demonstrate that the presence of centimetre- to hectometre-scale close – tight ductile folds influence the stability of the slopes. Folded zones can focus the gravitational deformation (as localized folded narrow zones). The contact between folded and unfolded units can be the locus of gravitational structures (due to a contrast of rheology between the units as a result of folding). Development of sliding planes at the base of rockslides can be guided by the geometry of centimetre-scale folds (where favourably orientated with respect to the slope), either parallel to the fold axes or parallel to the axial surfaces. Where decametre- or hectometre-scale fjordinclined fold limbs intersect the slope, they are reactivated as basal sliding planes and the upwards development of the failure planes (back-cracks)

is often deviated from the fold hinges. The latter confirms that the ductile folding decreases the strength of the rocks in the limbs and not in the hinges of folds (see the discussion in the last section of this paper). Influence of the sole detachments of the Caledonian allochthons on the slope stability. Western Norway is characterized by the remnants of the allochthonous sheets of the Caledonian Orogeny (Roberts & Gee 1985; Hossack & Cooper 1986; Roberts 2003). Most of the contacts between the different units (displayed in the map in Fig. 1) are tectonic in origin. The sole detachments of the allochthons developed in the ductile domain of deformation and they are commonly rather flat or shallowdipping. However, in a few places some of these contacts are found to be vertical. Field studies allowed the identification of eight rock slope instabilities that are located in the vicinity of the sole detachments. Such a setting provides an opportunity to evaluate how the presence of sole detachments may contribute to the gravitational destabilization of the slopes.

Fig. 13. The back-crack of the collapsing block opens perpendicularly to the axial surfaces of recumbent horizontal folds (white dashed line) at Mannen (Romsdalen Valley, located in Fig. 4a; see Henderson & Saintot 2007).

ROCK SLOPE INSTABILITIES IN NORWAY

At the sites called Vik, Viddalen, Tussen and Hjellane in Sogn og Fjordane County (located in Fig. 3), Gikling in the Sunndalen Valley (located in Fig. 4b) and site 20 in Storfjord (located in Fig. 5; also illustrated in Fig. 11) the gravitational deformation is limited to the Caledonian allochthonous units above the flat-lying–shallow-dipping sole detachments that intersect the steep slopes. Vik, Viddalen, Tussen and Hjellane slope instabilities (in Sogn og Fjordane County, located in Fig. 3) developed in the phyllitic nappes. As mentioned earlier, steep phyllitic slopes are highly susceptible to rupture only because of the low strength of the rocks. However, and in addition, at the four sites of Vik, Viddalen, Tussen and Hjellane there is the sole detachment underneath the gravitational structures. At Viddalen the instability developed on an anaclinal slope with foliation of the phyllites at about 308. Despite such geometry that may enhance the stability, a collapsing zone of approximately 1 Mm3 developed (Fig. 14: see also in Bo¨hme et al. 2011). The lowermost limit of the instable volume coincides with the 58 south-dipping sole detachment cropping out along the cliff (Fig. 14a, b). The qualitative assessment to explain the development of Viddalen instability against the unfavourable geometry of the foliation is to consider: (1) the intrinsic weakness of the rocks; (2) its pre-existing vertical joint sets (see in Bo¨hme et al. 2011); and (3) the presence of the sole detachment and its hydraulic properties. Because the sole detachment is underlined by water springs, it may act as a conduit for groundwater; the subjacent gneiss being, in this case, quasi-impermeable. Considering the heavy rainfall and snow melting periods, large water pressures, consequent to the rising of the hydraulic head, may build up along the sole detachment, contributing to local destabilization and additional opening of pre-existing joints (Terzaghi 1962; Giraud et al. 1990; Crosta & Agliardi 2003; Cruden 2003; Sartori et al. 2003; Grimstad 2005 among others). Also during wintertime, when the outflow of water at the springs gets blocked by frost, the hydraulic head may lift up yielding large water pressures at the sole detachment. The fact that the water springs never dry from spring to autumn may indicate that the water flows from far inwards in the high mountain plateau (Fig. 14c) and an approximately 400 m-high hydrostatic column may then be considered to cause pressures of some units of MPa at the sole detachment. Such pressures would exceed the low tensile strength of the jointed phyllites (the anisotropic tensile strength of intact phyllites ranging from 1.25 to 15 MPa: Hanssen et al. 1990; Ramamurthy et al. 1993; Saroglou & Tsiambaos 2008) and may subsequently generate (1) further opening of the tectonic joints

51

and (2), by decreasing the normal stress acting on the sole detachment, a spreading-like displacement of the overlying unit specifically directed towards the free cliff surface (Fig. 14c). Tussen is a site in Sogn og Fjordane (located in Fig. 3) that displays two rock slope instabilities (Fig. 15a). The southernmost instability was the focus of the field study and consists of two blocks (namely blocks A and B; Fig. 15b: Henderson et al. 2008; Bo¨hme et al. 2011). Both of the instabilities are located in the allochthonous phyllitic sheet thrusting over the autochthonous gneiss along a shallow south-dipping sole detachment (Fig. 15c, d). The cleavage in the phyllites is parallel to the sole detachment (see in Bo¨hme et al. 2011). It is likely that some shears occurred along the foliation coeval with the large shear along the nearby sole structure, rendering the south-dipping foliation more susceptible to being sheared again under gravitational forces. It would imply a southwardsdirected sliding of the blocks, oblique relatively to the NNE– SSW-striking cliff. The opening of large back-bounding crevasses also clearly took place along slope-parallel segments of kilometre-scale NNE –SSW lineaments (Fig. 15a, b). Six different potential rockslides are reported at Hjellane in Sogn og Fjordane (Fig. 15e, f; located in Fig. 3). Five of them lie on the thin gneissic allochthonous cover, overlying a phyllitic allochthon. The sole detachment of the allochthonous gneiss is dipping approximately 258 towards the fjord. Such a position favours the gravitational sliding of the gneiss along the reactivated sole detachment (Fig. 15e, f ). One potential rockslide developed in the weak phyllitic rocks at the knick-point of the slope (see Fig. 8c in Bo¨hme et al. 2011). The site called Gikling in the Sunndalen Valley (located in Fig. 4b) (Henderson & Saintot 2007; Saintot et al. 2008) is a very large active and complex unstable zone that covers an area of approximately 1 km2. It lies on a 258 slope, the uppermost limit of the deformed area being at 1400 m a.s.l. and the inferred toe zone below 900 m (Figs 16 & 17). Our field study at this site emphasized the important role of pre-existing structures in the development of slope instabilities. Not only were the pre-existing structures clearly reactivated by gravitational forces but they also defined the extent of the deformed volume. The foliation is, in general, dipping to the south (Fig. 16). The uppermost limit of the unstable zone corresponds to a spatial arrangement of pre-existing structures (Fig. 17). As such, segments A and B in Figure 16 are reactivated old brecciated faults (Faults A and B in Fig. 17a–d). Segment C (Fig. 16) comprises a north– south crack several metres wide and deep, that displays stepped walls. The steps are made up of the planes of the south-dipping foliation and of

52

A. SAINTOT ET AL.

Fig. 14. (a) Photograph of the Viddalen slope instability (located in Fig. 3) that developed at the edge of the cliff in the phyllitic Caledonian allochthon (see also Bo¨hme et al. 2011). The lowermost limit of the instability coincides with the sole detachment between the allochthon and the autochthonous gneiss. The tectonic contact is underlined by water springs. (b) Topographic map of the surrounding of the slope instability with hydrographical system and tectonic units (UTM co-ordinates, zone 32N). (c) Topographic profile (located in b) with the possible path of groundwater (grey arrows) from subsurface to the sole detachment (dotted black lines represent the foliation). The instability may have partly developed by high water pressures along the flat-lying sole detachment, when outflow of water gets blocked by the frost in wintertime or when high rates of precipitation occur (see text for further explanation).

ROCK SLOPE INSTABILITIES IN NORWAY

53

Fig. 15. (a) Three-dimensional view (www.norgei3d.no) of the two instabilities at Tussen (Sogn og Fjordane County; located in Fig. 3: see Henderson et al. 2008; Bo¨hme et al. 2011). (b) Three-dimensional view (www.norgei3d.no) of the southern instability. Large back-bounding crevasses follow NNE– SSW regional lineaments. (c) Aerial photograph with tectonic units and (d) geological section across the unstable slope. The instability developed in the phyllitic Caledonian allochthonous nappe. (e) Locations of the six rock slope instabilities at Hjellane (Sogn og Fjordane County; located in Fig. 3: see Henderson et al. 2008) on the geological map (UTM co-ordinates, zone 32N) and (f ) profile. Five of the six potential rockslides developed in the thin Caledonian allochthonous gneissic sheet and the main basal shear plane is probably the reactivated 258 fjord-dipping sole detachment. One of the potential rockslides developed in the underlying allochthonous phyllites at the knick-point of the slope.

54

A. SAINTOT ET AL.

Fig. 16. Detailed structural mapping on the geological map (UTM co-ordinates, zone 32N: Tveten et al. 1998) and cross-section showing the peculiar tectonic setting of the large complex slope instability of Gikling (Sunndalen Valley, located in Fig. 4b: see Saintot et al. 2008). Letters A –D refer to specific segments that limit the gravitational deformation, and A–C to pre-existing structures displayed in photographs in Figure 17.

ROCK SLOPE INSTABILITIES IN NORWAY

55

Fig. 17. Photographs illustrating the pre-existing structures that limit the complex slope instability at Gikling (Sunndalen Valley, located in Fig. 4b: Saintot et al. 2008). (a) Picture to the NW from the helicopter. Faults A and B, respectively, refer to the segments A and B in Figure 16 that limit the unstable slope. (b) Reactivated pre-existing brecciated NE-dipping fault (fault A in picture a, and segment A in Fig. 16) that marks a sharp border between the unstable and stable slopes. (c) Reactivated pre-existing steep SE-dipping fault with gouge and breccias (fault B in picture (a) and segment B in Fig. 16). (d) Close-up of the fault B with fine-grained breccia and gouge (white frame in c; view to SW). (e) Localization of tensional deformation to form segment D (as referred to in Fig. 16) along a zone of small amplitude gently plunging inclined folds, the newly formed gravitational fractures at this place trend parallel to the azimuth of fold axes (view to the west). (f) Picture to the west from the helicopter of the gravitational graben-like structures on the uppermost unstable sub-block (see the location in a).

56

A. SAINTOT ET AL.

the north –south-trending steep fractures. Segment D (on Fig. 16) is the uppermost limit that we observed. It is a wide zone of newly formed tensional fractures that are localized along a zone of small-amplitude gently plunging inclined folds and whose planes are oriented parallel to the fold axes (Fig. 17e). The lowest limit of the instability is not reachable and therefore not well constrained by direct field observation. However, a semi-horizontal line of springs is observed at about 800 m a.s.l. (Fig. 16). It coincides with the shallow-dipping sole of the gneissic allochthonous unit, in which are confined the gravitational structures (Fig. 16). The sole of the allochthon is composed of metasandstones and schists that are weaker that the overlying gneiss, and that would intuitively not support such a material where cutting across the slope (considered as a free border that will allow the propagation of the gravitational deformation). This level is a good candidate to be the basal limit of the slope instability and, therefore, the volume of the entire slope instability

would be greater than 100 Mm3. In addition, a large vertical fault zone lies at the eastern foot of the unstable zone. It separates the two autochthonous units of the valley and also limits to the east the allochthonous sheet (Fig. 16). Gikling is, therefore, in a very peculiar structural setting that makes the slope vulnerable to gravitational forces. At the sites called Vollan and Ivasnasen, in the Sunndalen Valley (location given in Fig. 4b), the Caledonian contacts between the five squeezed tectonic units are steep to vertical and NE–SW trending (Fig. 18). The potential rockslide identified at Ivasnasen (located in Fig. 4b) is underlain by allochthonous augen gneiss (Fig. 18). The site displays an old slide with a maximum failed volume of 5 Mm3 (Figs 18 & 19a, b). The top of the failed volume was at approximately 500 m a.s.l. and the bottom of the valley is at 200 m a.s.l.. Extensive fractures appear to propagate from the previous failure plane towards the SW and towards the NE, isolating

Fig. 18. Geological map (UTM co-ordinates, zone 32N: Nilsen & Wolff 1989) of Vollan and Ivasnasen surroundings (Sunndalen Valley, located in Fig. 4b: Saintot et al. 2008). Note that sector 1 is a topographic promontory called Ivasnasen, with three free borders that may favour the destabilization of the whole volume.

ROCK SLOPE INSTABILITIES IN NORWAY

57

Fig. 19. (a) Three-dimensional view (www.norgei3d.no) of the slope instability at Ivasnasen (Sunndalen Valley, located in Fig. 4b) and locations of the different sectors of the slope instability (Saintot et al. 2008). (b) Northeastern part of the back-fracture of the previous failed volume and slope instability that propagates towards the NE (sector 2). (c) Southwestern slope instability (sector 1) and open NE–SW fractures that propagated from the failed back-crack. (d) Aerial photograph (www.norgeibilder.no) of Vollan site in the Sunndalen Valley (located in Fig. 4b) with limits of the slope instability (dashed line: Saintot et al. 2008). The arcuate shape of the cracks that developed within the instability is a typical deformation pattern in weak rocks. (e) The back-scarp of the rock slope instability is a reactivated segment of a regional NE–SW vertical tectonic contact between the phyllites (in which developed the instability) and the quartzites (with a foliation parallel to the contact).

the two new unstable zones (sectors 1 and 2; Figs 18 & 19). The southwestern instability is coincident with a topographic promontory (sector 1; Figs 18 & 19a, c). The three free borders of the topographic promontory may favour the destabilization of the whole body. The limiting back-scarp and the parallel fractures downwards (Fig. 19) strike NE–SW parallel to the steep thrust planes displayed on the geological map (Fig. 18). The estimated volume of sector 1 may be up to 5 Mm3. The large slope instability of Vollan is located in phyllites and micaschists (Fig. 18). Figure 19a shows an aerial photograph with the 1 km2 estimated extent of this large slope instability. Given the arcuate shape of the cracks that developed within the instability, it appears that most of the displacement was accommodated by creeping as a

‘semi-ductile downwards flow’, that is a typical deformation pattern in weak rocks. The opening of the back-scarp (at 1050 m a.s.l.) occurred by reactivation of a segment of the NE– SW-trending steep regional tectonic contact between quartzites and micaschists (Fig. 19b). The lowermost level of the deformed area may also correspond to the tectonic contact between the allochthonous micaschists and the quartzites (Fig. 18) at 400 m a.s.l. These field examples enhance the need to systematically check the slope stability where the allochthonous sole detachments crop out. When dipping towards the valley or fjord, they are reactivated as sliding surfaces. They can be also composed of a de´collement layer of weak rocks (e.g. sandstones, micaschists) that may not bear the overlying allochthon rocks, even if their dip angle and

58

A. SAINTOT ET AL.

direction are not optimum (i.e. flat-lying). They are also often underlined by water springs and act as water conduits. Seasonally, when (1) excessive precipitation occurs and snow melts and when (2) in wintertime the outflow of water is stopped by frost at the springs, the hydraulic head may significantly rise, and hydrostatic pressures of some units of MPa can build up along the plane. Such pressures may be large enough to promote: (i) further opening of the pre-existing fractures in the overlying rocks; and (ii) the free slope surface-directed gliding of the overlying units by decreasing the resistance to shear along the plane.

Reactivation of brittle structures in rockslide development It is understood that the susceptibility of a given slope to rockfalls or rockslides increased with its density of brittle structures. Of note herein is our definition of ‘density of structures’ as a qualitative estimation of the concentration of structures within an area in relation to its size, which includes both the frequency and persistence of structures. As mentioned before, after the ductile Caledonian Orogeny the bedrock of Norway was deformed by a significant amount of tectonic events and a high density of brittle structures (showing, therefore, a high frequency and persistence) are observable at all scales. In the following, we demonstrate through selected field examples how the presence of these brittle structures affects the slope stability. Reactivation of exfoliation. The most recent development of brittle structures in Norway is a widespread exfoliation process caused by the sudden ice-cap removal. However, the unstable volumes along exfoliated slopes are generally too small to be classified with the 72 studied instabilities. The shape and the size of the instabilities are, indeed, constrained by the small spacing of exfoliation joints and, therefore, the instabilities mostly correspond to relatively thin slices of rocks. The most prominent exfoliation we observed during our field campaigns is found in Tafjord and cuts across the metamorphic foliation (for example, between sites 26 and 27 located in Fig. 5). The exfoliation process also occurred along the parallel-to-slope, or nearly, metamorphic foliation and contributed to its weakening through opening (as along the western side of La˚nefjorden in Sogn og Fjordane County; Fig. 3). We also observed that at the scale of western Norway there are few remnants of slopes that display exfoliation joints. This may be due to the weak downwards penetration of the exfoliation and the subsequent superficial rockslides having already failed. At site 5 of Storfjord (location shown in Fig. 5) there is an example of a 5 Mm3

unstable block that slid along an exfoliated plane. The block is also limited by one of the north – south regional faults and by an east –west crack. The latter follows the vertical foliation (Fig. 20). The role of regional faults in slope instability development. The regional faults of western Norway cover a large spectrum in strike and dip (Fig. 1). However, some prominent strikes are identified: north– south; ENE– WSW to NE –SW; and WNW –ESE to NW –SE (cf. in Gabrielsen et al. 2002). These large structures largely controlled the hydrographic pattern and fjord system. As such, on the fjord-sides, we can generally observe an important set of mesoscale brittle structures (and among them, striated faults) trending parallel to the fjord axes and believed to form the damage zone of these regional faults. The ENE –WSW strike is the prominent brittle grain towards, and of, the Møre Trøndelag Fault Complex. The north –south steep faults design the coast westwards of Sogn og Fjordane County. North– south regional faults were also observed in the Storfjord area and the Romsdalen Valley, and played a role in the development of slope instabilities. When cutting the gneissic rocks, the faults are large topographic linear depressions. This differential erosion suggests that they are weaker than the host gneisses. They, indeed, often present a metre-scale core filled by a juxtaposition of cohesive and non-cohesive fault rocks; as, for example, epidote-rich cataclasites, and hematiteand chlorite-rich breccias and gouges. Some field examples described herein illustrate the reactivation or the control of such structures in slope instability development. The role of regional-scale lineament orientations and their angular intersection and subsequent interaction with the fjord geometries is further discussed by Henderson & Saintot (2011). At the site of Mannen (in the Romsdalen Valley, location shown in Fig. 4a) we observed that the western border of the instability is a nearly north – south-trending vertical epidote-rich brecciated fault, largely opened under gravitational forces. It also marks a clear limit between the unstable eastern zone, where numerous secondary parallel pre-existing fractures are opened, and the more stable western zone, where such structural grain is less developed (Fig. 21a). The active collapsing block of Mannen is highly disrupted by the same system of opened fractures (Fig. 21b). The perpendicular fractures opened along the east –west steep foliation (Fig. 21c) and the main back-crack is localized at a folded zone (see Fig. 13). Several sets of moderately- to shallow-dipping joints were also measured and some of them are good candidates to define basal sliding planes (Fig. 21c). The regional system of north–south-trending vertical faults is very well developed in the SW

ROCK SLOPE INSTABILITIES IN NORWAY

59

Fig. 20. (a) Photograph showing the intense network of north–south regional faults with one of them reactivated as a back-crack of the potential rockslide at site 5 of Storfjord (located in Fig. 5). (b), Photograph showing the unstable block and its limits as reactivated structures. (c) Stereonet of structures measured at the foot of the block: foliation planes as black lines, exfoliation surfaces as grey lines, large regional lineaments as dashed lines. (d) Close view at the foot of the unstable zone, there is a recent rockfall activity at the front of the unstable block, that is, precisely above the basal shear plane (exfoliation and foliation planes as dashed white and dotted black lines, respectively).

part of the studied Storfjord area (Fig. 5). Along the north–south-trending fjords either the faults are largely reopened as back-cracks where they are vertical, as observed at sites 1, 4, 5 (see in Fig. 20a) and 6b (location shown in Fig. 5) (Henderson et al. 2006), or reactivated as basal shear planes where

they are favourably dipping towards the fjord, as observed at sites 2 and 2b (location shown in Fig. 5) (Henderson et al. 2006). Sites 2 and 2b in Storfjord are the locations of two potential rockslides that developed on an approximately 658 dipping east-facing slope with

60

A. SAINTOT ET AL.

Fig. 21. (a) Photograph from the helicopter of the opened roughly north–south-trending epidote-rich faults and parallel or nearly parallel fractures (black dashed lines) on the unstable edge of the plateau at Mannen (Romsdalen Valley, located in Fig. 4a: Henderson & Saintot 2007; Dahle et al. 2008). The opening of the system occurred under gravitational forces. A north– south main fault marks the limit between the unstable and stable areas. The foliation is steep and marked with a white dotted line (b) Photograph from the helicopter of the collapsing block of Mannen with an important set of north–south trending opened fractures. The dotted line is the unstable block limit in this plane view. (c) Stereonet of field data.

ROCK SLOPE INSTABILITIES IN NORWAY

reactivation as a basal sliding plane of the same system of regional faults dipping 408 towards the fjord. These faults are subparallel to, or slightly steeper than, foliation (their development is assumed to have been partly guided by the preexisting metamorphic foliation). Site 2 corresponds to an approximately 80 m-high block. The unstable block is also detached from the slope by the roughly north–south steep regional fractures that are largely reopened and act as back-crack structures. The basal sliding plane is identified along a reactivated 50 cm-thick epidote-rich cataclastic fault zone, with the main sliding surface exactly developed at the interface between the cohesive fault rocks and the host gneiss (Fig. 22). The instability at site 2b

61

is located approximately 900 m above sea level on a parallel epidote-rich cataclastic fault that reaches the surface inwards the plateau. This is the whole prominent ridge with an estimated volume of 10 Mm3 that is involved in the gravitational sliding. Examination of the basal shear plane showed that a fine-grained breccia was created by the sliding of the block and superimposed on the pre-existing fault core. Along the east –west-orientated fjord system the north –south regional steep faults behave as transfer fractures that allow the sliding of blocks from their sides (typically at sites 7, 7b and 9, see their locations in Fig. 5; see also Henderson & Saintot 2011). A detailed mapping from west of sites 7 and 7b to

Fig. 22. (a) Basal sliding plane at the foot of an unstable block, site 2 in Storfjord (located in Fig. 5). (b) The basal sliding plane is well developed and reactivated an epidote- rich cataclastic fault. (c) Closer view of a lens of epidote-rich cataclasites in the fault core. The planes of weakness are at the interface between the cohesive cataclasites and the host gneiss.

62

A. SAINTOT ET AL.

site 8 of Storfjord (Fig. 23) showed the importance of these north–south large cataclastic fault zones in assessing the slope stability. The offset of bedrocks along the faults juxtaposed gneiss with foliation dipping towards the fjord and gneiss with foliation dipping into the mountain; in other words, the faults delimit unstable cataclinal and stable anaclinal parts of the slope. Sites 7 and 7b are on the cataclinal part of the slope and are surrounded by large areas where failures occurred along the fjorddipping foliation reactivated as sliding planes. Large WNW –ESE vertical cracks opened in these two remaining instabilities (Fig. 23), they are subparallel to the second prominent set of regional lineaments of the large Storfjord area that partly controlled trends of the fjord (Fig. 1). A fine-grained breccia formed because of the block sliding along a foliation-parallel basal shear plane at site 7b (Henderson et al. 2006; Henderson & Saintot 2011). This susceptible part of the slope is limited westwards and eastwards by two north– south cataclastic large faults, west and east of which the slopes are anaclinal (Fig. 23). This study along such a slope showed again the high relevance of the foliation as a parameter to define the slope susceptibility. It showed also, however, that other structures should be taken into account when carrying out a slope stability analysis. As such, and near site 8, one of these faults juxtaposed a western anaclinal slope and an eastern cataclinal slope (Fig. 23). Whereas it is thus predicted that the western zone would be more stable, unsteady slabs formed because of the occurrence of approximately 558 dipping slope-parallel exfoliation surfaces reactivated under gravity. At site 9 on the 558 dipping southern side of Geiranger Fjord (located in Fig. 5) the foliation dips towards the fjord and ENE– WSW cracks are widely opened in the potential sliding mass (Fig. 24a). Fieldwork demonstrated that these ENE– WSW structures consist of large (several kilometres long and several metres wide) preexisting fault zones that were reactivated by recent extensional and down-dip movement towards the fjord (Fig. 24b –d). No low-angle shearing plane at the base of the unstable block was evidenced in the field owing to the inaccessibility of the area and the severity of the topography. However, it is postulated that the extensional displacement on the steep structures is taken up on shearing along some basal structures, parallel to foliation, and that the down-dip displacement may be coupled with a rotation of the blocks. The sliding mass is guided by north–south pre-existing faults parallel to the motion of the block (Fig. 24e). Site 9 is a good example of how a combination of two sets of regional faults and of a fjord-dipping foliation increases the susceptibility of a steep slope to rockslides (Fig. 24e).

Some field examples in Storfjord also clearly show that these faults, where oblique to the slope, guided the structures that developed under gravitational forces. However, such instabilities formed because of the presence of other pre-existing zones of weakness favourably orientated relatively to the fjord or valley. This is the case at sites 10b, 22, 26 and 27 in Storfjord (location shown in Fig. 5) presented below, where a combination of pre-existing structures favoured rockslide formation. The unstable block at site 10b may be of 2 Mm3 maximum and lies on a 508 dipping SW slope of Geiranger Fjord (Fig. 25). Low-angle basal shear planes are developed subparallel to the foliation and recent movement along the shear structures is evidenced by the development of a fault-rock gouge. The block is fully delimited by: (1) an extensional vertical crack, oblique to the slope, which is a reactivated long-lived regional north–south lineament; and (2) a NE–SW steep large structure that transfers the downwards motion of the block (Fig. 25). The NE –SW steep structures strikes parallel to a system of small-scale inherited epidoteand chlorite-coated vertical faults that we observed at the toe zone of the block (Fig. 25d). One of the best examples of a combination of structural sets was observed at site 22 in Storfjord on a 708 west-dipping slope (Fig. 26; location shown in Fig. 5) (Henderson et al. 2006). Open NE– SW vertical fractures, together with the opening of the NW–SE vertical foliation, form a columnar fabric into the unstable part of the slope. The southern part of the main back-crack follows a NE –SW fracture. Two different types of faults, both characterized by epidote-rich coated surfaces also contribute to the development of an unstable rock mass. Gently west-dipping faults intersect the columns and may form the basal sliding plane required to detach the rock mass at the foot of the columns. The pre-existing epidote-rich vertical north– south faults guided the development of the northern part of the main crack at the back of the unstable area. The main mechanism here is likely to be columnar failure, thereby reducing the likelihood of a large volume of unstable material. The slope instabilities of sites 26 and 27 in Storfjord (location shown in Fig. 5) are found where two gneiss units are juxtaposed along a NW–SE steep contact cutting the 558 –608 dipping parallel slope (Fig. 27). In addition, several steep north–southtrending regional-scale faults obliquely cut the NW– SE-striking slope (Fig. 27a). Prominent fractures trend parallel to these large main structures and are largely opened. The east –west-trending vertical foliation planes guided the development of the second trend of open fractures (Fig. 27c). These two nearly perpendicular vertical trends contributed to detaching large high blocks from the slope. Field

ROCK SLOPE INSTABILITIES IN NORWAY 63

Fig. 23. Detailed topographic map (UTM co-ordinates, zone 32N) from sites 7 and 7b to site 8 in Storfjord (located in Fig. 5): north–south regional cataclastic fault zones acting as transfer zones and limiting unstable and stable parts of the slope with regard to the attitude of foliation from one side to another side of the faults. At site 8 the foliation dips into the mountain slope but the exfoliation dips towards the fjord, and the latter was the locus for the development of gravitational structures. The data of stereonets correspond to sites 7 and 7b: the observed basal shear planes of the potential rockslides are parallel to the foliation planes; the roughly north–south set of steep fractures corresponds to the trend of transfer faults and the WNW– ESE set to the trend of back-cracks.

64

A. SAINTOT ET AL.

Fig. 24. (a) Structures of the rockslides at site 9 in Storfjord (located in Fig. 5): ENE– WSW extensional fractures steeply dipping towards the fjord and postulated shearing along the undulating shallow– moderate fjord-dipping foliation to accommodate the extensional movement in the steep structures. A down-dip component of displacement is also inferred along the same steep ENE –WSW fractures and may be accommodated by a rotational component of motion of the blocks. (b) Steep pre-existing 3 m-wide ENE –WSW-trending fault zone that consists of a ‘damage zone’ of many smaller fractures. These are partly cemented by quartz. This weak zone focused recent movement, producing three opened fractures. (c) and (d) Secondary structures that evidence both extensional and down-dip (fjord-directed) movement along the steep ENE–WSW fractures. (e) Stereonets of the three pre-existing structures that are reactivated under gravity and led to a completely structured rockslide.

observations at the toe zones of the instabilities documented sliding planes that trend parallel to well-developed 308 –408 fjord-dipping exfoliation joints. Steiggjeberg in Sogn og Fjordane County (location shown in Fig. 3) displays a vertical face with frequent rockfall activity (Fig. 28a). The site is located in the footwall and, more specifically, in the damage zone of a moderately fjord-dipping Devonian regional normal detachment (Fig. 28b). The erosional process deeply incised the slope

down to the thick cataclasites of the fault core and formed steep walls within the overlying granulites. A dense network of faults (Fig. 28c) that characterize the wide damage zone of the detachment dramatically increases the instability of the wall. At Steiggjeberg, the pre-existing tectonic architecture, the subsequent rheological contrast and relief favour the destabilization of the rock under gravitational forces. Our observations at 45 sites lead to the conclusion that whatever their orientation relative to

ROCK SLOPE INSTABILITIES IN NORWAY

65

Fig. 25. (a) Photograph from the helicopter to the south of site 10b in Storfjord (located in Fig. 5) and of the large north– south regional fault zone. (b) Orthophotograph of the site showing the triangular shape of the site limited by the large north–south and NE– SW pre-existing faults (UTM co-ordinates, zone 32N). The white and black arrows highlight, respectively, the north– south and the NE–SW reactivated faults. (c) Closer view of the site, to the SW. The dashed line is the basal shear plane that developed parallel to foliation. (d) Stereonets of geological structures.

the slope the regional faults may contribute to slope destabilization and rockslide formations. This is particularly true in the gneissic areas (that covers more than the three-quarters of the western Norway area) where the fault cores appear weaker

than the host rocks and focus the gravitational structures (from back-cracks to transfer fractures). Also, and as expected, these regional faults are characterized by wide damage zones that are a dense network of mainly parallel faults and

66

A. SAINTOT ET AL.

Fig. 26. (a) Picture from the helicopter towards the NE showing the 1 Mm3 volume of rock (limited by a dotted line) detached from the slope by a NE–SW vertical crack, site 22, Storfjord (located in Fig. 5). (b) Picture to the SE from the helicopter of the SW– NE crack. The steep NW– SE foliation is seen. (c) Mapping of structures (UTM co-ordinates, zone 32N). (d) Stereonets of field measurements.

ROCK SLOPE INSTABILITIES IN NORWAY 67

Fig. 27. Geological structures observed on the SW side of Tafjord (sites 26 and 27 of Storfjord region; located in Fig. 5). The slope instabilities are largely developed subsequent to the reactivation of the pre-existing structural pattern with: (1) opening of cracks along north–south steep regional faults and the east– west vertical foliation; and (2) development of basal sliding planes along the moderately fjord-dipping exfoliation joints. (a) Orthophotograph (UTM co-ordinates, zone 32N) with the two unstable sites, the foliation attitude, the north– south steep regional faults (solid white lines) and the geological contact between two gneissic tectonic units (solid black lines). (b) Stereonets of structures with joints as thin black dashed lines, pre-existing faults as thin black filled lines, foliation as thick black lines, exfoliation as thick grey lines and sliding planes as thick black dashed lines. (c) Photograph of the main crack at site 27 that developed along the east–west vertical foliation.

68

A. SAINTOT ET AL.

Fig. 28. (a) Photograph from the helicopter of the unstable face at Steiggjeberg in Sogn og Fjordane (located in Fig. 3). (b) Extract of the geological map with the peculiar tectonic setting of the unstable scarp (star) that lies in the damage zone of a moderately fjord-dipping Devonian regional normal detachment (UTM co-ordinates, zone 32N). (c) Close view of one (supposed normal fault) of the numerous faults that characterizes the damage zone of the Devonian detachment and that contributes to weaken the scarp.

fractures. The observations fit with the consensus that links many gravitational slope deformations worldwide with the presence of regional faults (see the references cited in introductory section of this paper). In that sense, Bo¨hme et al. (2011) show the spatial relationships between the rock slope instabilities in western Norway and the largest fault zones. Also, because large fault zones structurally control the hydrographical network in Norway, the side slopes of the valleys and fjords may contain a non-negligible amount of associated fractures. The latter mainly striking parallel to the slope (to the main fault segment) serve to localize the back-crack of instabilities. In addition, our study reinforces the concept that the high susceptibility to rockslides of the steep slopes in western Norway is directly linked to the high spatial density of faults at various scales and with various orientations. This high fault density is related to the age of the Norwegian bedrocks that cumulated a long history of successive brittle

tectonics from the time of the Caledonian Orogeny. Subsequently, it is not rare to observe two cross-cutting orientations of faults and associated fractures at one site that promote the slope deformation (Table 1).

Combination of favourably orientated structures with respect to the slope in rockslide development Intuitively, and subsequently to the field analyses, it appears that a combination of several geological factors, which taken one by one decrease the slope stability, will dramatically further enhance the destabilization. Site 6b in Storfjord (location shown in Fig. 5) is one example where a combination of pre-existing geological factors allowed the development of a large unstable volume of 15 Mm3 on an approximately 458 SE-dipping slope (Fig. 29). The site is

ROCK SLOPE INSTABILITIES IN NORWAY

69

Fig. 29. (a) Orthophotograph, (b) structural mapping (UTM co-ordinates, zone 32N) with open fractures in thin lines, wide back-bounding cracks in thick lines, basal sliding plane in festooned line and (c) stereonets of the potential rockslide at site 6b in Storfjord area (located in Fig. 5). An unfortunate combination of six pre-existing geological factors allowed the development of a large unstable volume of 15 Mm3: (1) the instability developed in the weak mica-rich rocks; (2) foliation dips towards the fjord and guided the creation of a basal shear plane, underlined by a newly formed 20 cm thick gouge; (3) the basal shear plane may be located at the interface between the unfolded upper and the folded lower parts of the slope; (4) and (5) NE–SW and NW– SE reactivated pre-existing regional faults (see stereonet) are open fractures and transfer fractures, respectively, into the unstable block and the back-crack on its northern part also developed along a NE–SW vertical fault (thick white line in b); and (6) the southern part of the back-crack (thick white line in b) follows a north–south regional cataclastic fault.

located on the mica-rich units, and these rocks are prone to destabilization because of their intrinsic weakness. The foliation dips approximately 308 SE towards the fjord and a shear plane with newly

formed 20 cm-thick gouge developed subparallel to it at the base of the unstable volume (see also Henderson & Saintot 2011). An important field observation is that below the base of the unstable

70

A. SAINTOT ET AL.

volume the rocks are folded. Because of the contrast in rheology between folded and unfolded rocks, we propose that the basal shear plane formed at the interface between the two units. Open joints in the unstable block trend roughly NE–SW. NW– SE-trending joints act as transfer zones. An analysis in the surrounding area of the unstable part of the slope revealed that both the NW –SE and NE–SW vertical fractures belong to the pre-existing structural brittle grain and are commonly epidote-rich faults. On the northern side, the main back-crack follows a NE –SW structure. Towards the south, the back-crack with a width of 3 m is an extensional fracture that trends north–south, similar to the inherited regional cataclastic faults. Three regional trends of pre-existing vertical faults, the fjorddipping foliation, the different rheology between the folded lower and the unfolded upper parts of the slope (with, at their interface, the localization of the basal sliding plane), and the intrinsic weakness of the mica-rich rocks represent six factors that led to the development of the rockslide.

Discussion, concluding remarks and future work Table 1 summarizes the field observations at the 72 examined sites of western Norway. The table is divided in two parts according the host rocks of the instabilities as we assume that one of the primary

factors to consider is the lithology of the rocks, that is, weak v. hard rocks and, respectively, amphibolites, schists and micaschists, weathered mafic and ultramafics v. gneisses (see the subsection ‘The lithological factor in slope instability development’). Out of the 72 studied instabilities 13 sites are grouped as being hosted by weak rocks and are set in the top part of Table 1. A second criterion of site classification in Table 1 is the number of structures that hold or guide the gravitational deformation (as documented by detailed field examples in previous sections) and a third criterion is the estimated maximal unstable volume, with the criterion values increasing from top to bottom in each of the two lithological groups. Table 1 also provides by sites qualitative assessments on the type and nature of the pre-existing structures involved in the gravitational deformation. Table 1 and the diagram in Figure 30a show that 10 out of the 13 instabilities that developed in the weak rocks do not necessarily comprise reactivation of more than one prominent (i.e. large and persistent at the scale of the instability) pre-existing structure. Only two out of the 13 instabilities hosted by the weak rocks developed with the reactivation of two pre-existing structures. Site 6b in Storfjord is the major exception of these 13 sites, with five preexisting geological structures involved in its development. However, this analysis tends to confirm that such rocks cannot hold on the steep slopes of western Norway simply because they are intrinsically weak (i.e. pre-existing structures would not

Fig. 30. (a) Diagram of the type of gravitational slope deformation and of the estimated maximum volume of the 72 studied instabilities of western Norway as a function of the number of pre-existing structures that guide or hold the gravitational deformation. There is a slight correlation between the numbers of pre-existing structures and the volume of the instability. (Colour code: white and black for sites in weak and hard rocks, respectively; shape code: circle, triangle and star for rockfall, rock topple and rockslide, respectively.) (b) Hypothetical scheme (perspective view) of pre-existing structures that control in a way the volume of the unstable zone (grey part): the old large fault zone focuses a back-bounding crack, a transfer fracture is guided by a axial fold plane and the basal shear plane forms along foliation (with h being a sufficient height to generate weight/force that can destabilize a mica-rich layer and allow the sliding).

ROCK SLOPE INSTABILITIES IN NORWAY

be necessary to the development of a gravitational instability). That is not the case if we analyse the 59 unstable sites developed in the harder gneissic rocks of Norway. We observed that at such 54 localities pre-existing prominent structures are required to develop the instabilities (in agreement with worldwide structural studies on rockslides; see the references in the introductory section to this paper) and, to a greater extent, that at 37 localities more than one inherited prominent structures played a significant role in the formation of the instability (Table 1 & Fig. 30a). This does not preclude that the other necessary gravitational structures (assuming that three orientations of gravitational structures lead to fully detach a rockslide at the back, at the side and at the bottom) are not reactivated structures. They may also be reactivated but would be smaller and more local pre-existing structures (like typical sets of joints) than the prominent structures that we presently count (like regional brecciated/cataclastic faults or decametre- and hectometre-scale folds). The gravitational deformation in the hard rock like the gneisses of western Norway is only possible because of the anisotropy of strength subsequent to the tectonics; that is, the metamorphic foliation, the presence of folds and brittle structures at all scale. The strength of such hard rock slopes will dramatically decrease if structural features are favourably orientated with regard to the slope. For instance, in the metamorphic rocks of western Norway, one obvious structural factor that controls development of slope instabilities is the foliation and, more specifically in the dioritic gneiss, its mafic weak layers. As observed at 11 localities, the foliation is vertical, roughly slope-parallel and contributes to columnar shape-like failures. The foliation can also be fjord- or valley-dipping and reactivated to form basal sliding planes as observed at 30 localities. However, many field examples showed that foliation dipping towards the fjord or valley is far from being the only structural factor that led to the full development of slope instabilities. The associated large failure surfaces (back-crack and transfer fractures) responsible for the detachment of a block from the slope should be obviously present and these are rarely newly formed structures. When such failure planes are newly formed they are, indeed, guided by the ductile grain of the rock (as a particular fold geometry), but very commonly gravitational forces simply reactivate preexisting brittle structures. It follows that at most of the sites there is a combination of several favourably orientated structures with respect to the steep slopes that will dramatically decrease the slope stability and promote gravitational failures. Also, Henderson & Saintot (2011) present a new qualitative approach to the spatial variation in susceptibility

71

at the scale of Storfjord based on structural criteria and demonstrate that the sites with the highest susceptibility record the greatest number of identifiable gravitational structures. An important loss of rock-mass strength is linked to the high density of pre-existing ductile and brittle structures of various orientations that characterized the bedrock of western Norway. The direct consequence is that there are few slopes exempt of gravitational deformation in western Norway. The rock mass of western Norway suffered at a large scale an important tectonic weakening. Any further weathering (as meteoric) will enhance the destabilization of the slope by the weakening of rocks and preferentially of its geological structures (Bachmann et al. 2004). This last point will also provide the assumption that the volume of the slope instability may be in a certain limit, controlled by the presence and the reactivation of the weakest parts of the slope, that is, the large geological structures (Fig. 30b). Indeed, we observed at 24 sites that borders of the gravitationally deformed slope coincide with large inherited structures. This may explain the slight correlation between the numbers of pre-existing prominent structures and the volume of the slope instability (Fig. 30). Where the gravitational deformation occurred with a destabilization involving two or more pre-existing prominent structures the volume of the instability tends to exceed 1 Mm3 (with the exception of one rockfall of 0.001 Mm3; Fig. 30a). The dimension of the unstable zone may be constrained by the hundreds-of-metre-scale spacing and persistence of large structures (as well exemplified by the distribution of the slope instabilities in the vicinity of sites 7, 7b and 8 in Storfjord; see Fig. 23). Henderson & Saintot (2011) established this relationship in more detail in the Storfjord area where the largest unstable volumes are at sites with the largest number of well-developed gravitational structures. These gravitational structures reactivated preexisting structures and also clearly correspond to the borders of the instabilities. Therefore, the preexisting large structures not only locate the gravitational deformation, but may limit and impose the unstable volume. The two best examples to illustrate ˚ knes (described in Jaboyedoff such setting are the A et al. 2011) and Gikling rockslides (Saintot et al. 2008) (see Figs 16 & 17). The analysis of the spatial distribution of the number of pre-existing structures observed by instabilities coupled with the unstable volume over western Norway tends to confirm this point. To the north of western Norway and in Møre og Romsdal County, where large ductile and brittle structures are more abundant, there is a clear increase in the volume of instabilities together with an increase in the number of structures involved (Fig. 31).

72

A. SAINTOT ET AL.

Fig. 31. Map of the 72 slope instabilities of western Norway with an indication of the type of gravitational slope deformation, the estimated volume and the number of pre-existing structures that guide or hold the gravitational deformation (background is Fig. 2b).

In addition, the spatial distribution of potential rock slope failures over western Norway shows that there exists a specific clustering in Storfjord (Fig. 31). The assumption to explain such clustering

is, in addition to the steepness of the slopes and the gradient of the ongoing high rate uplift (see Fig. 2 and discussion in Henderson & Saintot 2011), the occurrence of close –tight metre-scale polyphase

ROCK SLOPE INSTABILITIES IN NORWAY

folds in the gneiss (which is not so pronounced towards the Møre Trøndelag Fault Complex and towards Sogn og Fjordane County). We propose that there is a strong correlation between the occurrence of ductile folds and slope stability decrease, and that fold interferences will provide further loss in slope stability. This may be because ductile folding is expected to: (1) decrease the intrinsic strength of minerals (and thus rocks) by stretching and flattening, specifically along the limbs; and (2) increase along a given slope the probability to have favourably orientated grains prone to be reactivated by gravitational forces (or to guide newly formed gravitational structures). Interference of several phases of folding would increase these factors. The typical gravitational slope deformation of western Norway mostly corresponds to rockslide or complex slide (also involving toppling and flexural toppling) with, out of the 72 studied sites, 56 instabilities that developed along a basal sliding planes (i.e. all classified as rockslides in Figs 30a & 31). We previously mentioned that 30 rockslides have a basal sliding plane that formed along the metamorphic foliation on cataclinal slopes. According to Figures 30a and 31, the type of slope deformation is exclusively slide with the increase of both: (1) the number of pre-existing structures involved in the instability formation; and (2) the volume of the instability. Five rock topples and 11

73

rockfalls are observed along slopes where none or only one geological structure is clearly reactivated in the process of destabilization. Although most of the studied slope instabilities display well-developed gravitational structures (see above and Henderson & Saintot 2011), thus testifying to quite significant down-slope cumulative displacements, many of the basal sliding planes of the rockslides are rather shallow dipping. Such geometry implies high values of normal stresses that prevent any gliding of the blocks (Table 2). In order to explore this point, we calculated for 20 sites the potential values of cohesion, C, as a function of various friction angles, w, of the sliding basal planes (Fig. 32). Using Selby’s equation (1993, p. 328):

C ¼ r  g  H

sin(b  a)  (sin a  cos a  tan w) 2sin b

where r is the density taken equal to 2700 kg m23, g is the acceleration of gravity (9.8 m s22), H is the the height of the unstable volume, b is the slope gradient and a is the angle of the sliding plane (Table 2). At most of the sites the basal sliding plane is not accessible due to the steepness of the slope. Hence, the analysis is carried out on a relatively limited number of 20 sites for which the geometry of the basal sliding plane is constrained.

Table 2. Parameters used for the calculation of the possible cohesion and friction angle of 20 sliding planes (see Fig. 32) Sites Site 1, Storfjord Site 2, Storfjord Site 2b, Storfjord Site 5, Storfjord Site 6b, Storfjord Site 7, Storfjord Site 7b, Storfjord Site 9, Storfjord Site 10, Storfjord Site 10b, Storfjord Site 12, Storfjord Site 14b, Storfjord Site 22, Storfjord ˚ knes, Storfjord A Oppstadhornet Ha˚rstad Rustøyane Viddalen Tussen Hjellane (down-slope part)

H (m)

b (8)

a (8)

t (MPa)

sn (MPa)

200 100 300 300 100 200 60 200 250 150 100 250 130 200 100 200 80 150 200 100

60 65 65 55 46 60 45 55 60 50 60 65 70 45 32 45 40 45 60 40

40 50 45 25 28 25 35 33 30 38 30 30 25 30 30 35 25 10 20 30

3.4 2.0 5.6 3.4 1.2 2.2 0.9 2.9 3.3 2.4 1.3 3.3. 1.5 2.6 1.3 3.0 0.9 0.7 1.8 1.3

4.1 1.7 5.6 7.2 2.3 4.8 1.3 4.4 5.7 3.1 2.3 5.7 3.1 4.6 2.3 4.3 1.9 3.9 5.0 2.3

a dip angle of the sliding plane; b, slope gradient; H, height of the unstable block; in addition, first-order magnitudes of the shear and normal stresses, t and sn, respectively, acting on the plane due to the weight of the lithostatic column of height H.

74

A. SAINTOT ET AL.

Fig. 32. Calculation of the possible cohesion and friction angle of 20 sliding planes (from Selby 1993; see Table 2 for the values of parameters). The shaded area represents the physically meaningless part of the graph with, towards the black side, unrealistic friction angle values decreasing to 08.

The analysis shows very low cohesion values (from 0 to approximately 1000 kPa) when varying the friction angle values reasonably taken above 58 (Fig. 32). Of note herein is the precise determination of the friction angle close to 308 at the ˚ knes rockslide (the cohesion being negliactive A gible as the mass is moving). At the other sites that are considered inactive or dormant, cohesion values are clearly one –two orders lower than shear stress magnitudes (some units of MPa; Table 2). Hence, the analysis predicts that these instabilities would fail but actually they do not. A first explanation is that the planes still display asperities, typically bridges, giving higher cohesion values than the one predicted by Selby’s equation. This roughness of the planes will be smoothed out by chemical and/or mechanical weathering. Such process is part of the long-term fatigue of the medium and contributes to the required maturation of the basal sliding planes. An additional reason is probably the high values of the normal stresses (Table 2). However, the calculated normal stresses

(Table 2) are first-order values, for example without taking into account fluctuations due to the water pressure. The normal stress may dramatically drop off with large seasonal water inputs, thus triggering motion (Terzaghi 1962; Giraud et al. 1990; Crosta & Agliardi 2003; Cruden 2003; Sartori et al. 2003; Grimstad 2005 among others). For example, the shear and normal stress magnitudes due to the weight of the lithostatic column on the extreme flat-lying plane at the base of the rock slope instability at Viddalen (see the earlier subsection ‘Influence of the sole detachments of the Caledonian allochthons on the slope stability’; Fig. 14) would be approximately 0.7 and 3.9 MPa, respectively (Table 2). In this case, an increase in the isotropic water pressure to a value of only 1.6 MPa would be enough to provoke the reversing of the shear and normal stress relationships triggering the gliding along the plane. Water pressures of some few units of MPa are predictable along most of the basal planes of the potential investigated rockslides when considering the important relief

ROCK SLOPE INSTABILITIES IN NORWAY

at the sites and, thus, the height of the hydrostatic columns. Both roughness of the plane and normal stress variation interplay for the stability of the rockslide. Finally, this geological study provides a basis to construct a map of susceptibility to large gravitational deformation along selected slope gradients taking into account lithology (weak v. strong rocks), anisotropy due to structures (i.e. considering their orientation and their types) and the density of structures. The following list summarizes some crucial parameters that we observed during the fieldwork and may be used to index the loss of stability at a regional scale along high slopes with gradients above 358 in western Norway: (1) weak rocks; (2) foliation towards the fjords or the valley or steep foliation striking roughly slope-parallel; (3) folds and interference folds; (4) Caledonian thrusts cutting the slope; and (5) regional brecciated/ cataclastic faults close to the slope. If all of these features are present, gravitational deformation is very likely to occur. The susceptibility map based on geological/structural criteria, combined with an analysis of past events, will help to complete hazard mapping of western Norway. We are grateful to M. Bo¨hme, E. Anda, H. Dahle, S. Sætre, Tim Redfield and all our companions during the fieldwork in western Norway; and L. Blikra and R. Hermanns for discussions on slope stability. We thank two anonymous reviewers for their help in improving the manuscript. The work described in this contribution is partially supported by the Research Council of Norway through the International Centre for Geohazards (ICG). Their support is gratefully acknowledged. This is ICG contribution No. 279.

References Agliardi, F., Crosta, G. B. & Zanchi, A. 2001. Structural constraints on deep-seated slope deformation kinematics. Engineering Geology, 59, 83–102. Ambrosi, C. & Crosta, G. B. 2006. Large sackung along major tectonic features in the central Italian Alps. Engineering Geology, 83, 183– 200. Anda, E. & Blikra, L. H. 1998. Rock-avalanche Hazard in Møre & Romsdal, Western Norway. Norwegian Geotechnical Institute Publication, 203, 53– 57. Bachmann, D., Bouissou, S. & Chemenda, A. 2004. Influence of weathering and preexisting large scale fractures on gravitational slope failure: insights from 3-D physical modelling. Natural Hazards and Earth System Sciences, 4, 711 –717. Badger, T. C. 2002. Fracturing within anticlines and its kinematic control on slope stability. Environmental and Engineering Geoscience, 8, 19– 33. Blikra, L. H. & Longva, O. 1995. Frost-shattered debris facies of Younger Drias age in the coastal sedimentary

75

successions in western Norway: palaeoenvironmental implications. Palaeogeography, Palaeoclimatology, Palaeoecology, 118, 89–110. Blikra, L. H., Anda, E., Høst, J. & Longva, O. 2006a. A˚knes/Tafjord-prosjektet. Geological Survey of Norway, Report 2006.039 [in Norwegian]. Blikra, L. H., Braathen, A., Anda, E., Stalsberg, K. & Longva, O. 2002. Rock avalanches, gravitational bedrock fractures and neotectonic faults onshore northern West Norway: Examples, regional distribution and triggering mechanisms. Geological Survey of Norway, Report 2002.016. Blikra, L. H., Braathen, A. & Skurtveit, E. 2001. Hazard evaluation of rock avalanches; The Baraldsnes– Oterøya area. Geological Survey of Norway, Report 2001.108. Blikra, L. H., Dehls, J. F. & Olesen, O. 2000. Gravitational-slope features from Odda in Hardanger to Aurland in Sogn, Western Norway. In: Dehls, J. & Olesen, O. (eds) Neotectonics in Norway. Geological Survey of Norway, Annual Technical Report 2000.01, 26–30. Blikra, L. H., Longva, O., Braathen, A., Anda, E., Dehls, J. & Stalsberg, K. 2006b. Rock slope failures in Norwegian fjord areas: examples, spatial distribution and temporal pattern. NATO Science Series. Series IV, Earth and Environmental Sciences, 49, 475– 496. Bois, T., Bouissou, S. & Guglielmi, Y. 2008. Influence of major inherited faults zones on gravitational slope deformation: A two-dimensional physical modelling of the La Clapiere area (Southern French Alps). Earth and Planetary Science Letters, 272, 709– 719. Boulton, G. S. & Caban, P. 1995. Groundwater flow beneath ice sheets: Part II – Its impact on glacier tectonic structures and moraine formation. Quaternary Science Reviews, 14, 563–587. Braathen, A., Blikra, L. H., Berg, S. S. & Karlsen, F. 2004. Rock-slope failures in Norway; type, geometry, deformation mechanisms and stability. Norwegian Journal of Geology, 84, 67–88. Brideau, M.-A., Stead, D., Kinakin, D. & Fecova, K. 2005. Influence of tectonic structures on the Hope Slide, British Columbia, Canada. Engineering Geology, 80, 342–359. Brideau, M.-A., Yan, M. & Stead, D. 2009. The role of tectonic damage and brittle rock fracture in the development of large rock slope failures. Geomorphology, 103, 30–49. Bo¨hme, M., Saintot, A., Henderson, I. H. C., Henriksen, H. & Hermanns, R. L. 2011. Rock slope instabilities in Sogn & Fjordane County, Norway: a detailed structural and geomorphological analysis. In: Jaboyedoff, M. (ed.) Slope Tectonics. Geological Society, London, Special Publications, 351, 97–111. Chigira, M. 1992. Long-term gravitational deformation of rock by mass rock creep. Engineering Geology, 32, 157– 184. Coe, J. A. & Harp, E. L. 2007. Influence of tectonic folding on rockfall susceptibility, American Fork Canyon, Utah, USA. Natural Hazards and Earth System Sciences, 7, 1– 14.

76

A. SAINTOT ET AL.

Crosta, G. B. & Agliardi, F. 2003. Failure forecast for large rock slides by surface displacement measurements. Canadian Geotechnical Journal, 40, 176–191. Crosta, G. B., Frattini, P., Agliardi, F., Andreolli, M. & Blikra, L. H. 2007. Modelling rockfall hazard in the Storfjorden area, Western Norway. Geophysical Research Abstracts, 9, 06437. Cruden, D. M. 1976. Major slides in the Rockies. Canadian Geotechnical Journal, 13, 8 –20. Cruden, D. M. 2000. Some forms of mountain peaks in the Canadian Rockies controlled by their rock structure. Quaternary International, 68–71, 59– 65. Cruden, D. M. 2003. The shapes of cold, high mountains in sedimentary rocks. Geomorphology, 55, 249– 261. Cruden, D. M. & Hu, X.-Q. 1996. Hazardous modes of rock slope movement in the Canadian Rockies. Environment and Engineering Geoscience, 2, 507– 516. Cruden, D. M. & Varnes, D. J. 1996. Chapter 3: Landslide types and processes. In: Turner, A. K. & Shuster, R. L. (eds) Landslides: Investigation and Mitigation. Special Report, 247, Transportation Research Board, National Research Council, Washington, DC, 36–75. Dahle, H., Anda, E., Saintot, A. & Sætre, S. 2008. Faren for fjellskred fra fjellet Mannen i Romsdalen. Geological Survey of Norway, Report 2008.087 [in Norwegian]. Derron, M.-H. 2009. Rock Fall Susceptibility Map of Norway. www.skrednett.no Derron, M.-H., Jaboyedoff, M. & Blikra, L. H. 2005. Preliminary assessment of rockslide and rockfall hazards using a DEM (Oppstadhornet, Norway). Natural Hazards and Earth System Sciences, 5, 285– 292. Domaas, U., Blikra, L. H., Rosenvold, B. S., Lægreid, O., Gunleiksrud, O. & Grimstad, E. 2000. Studie av fjellskred og dalsidestabilitet i fyllittomra˚der. Norwegian Geotechnical Institute, Report 20001132-1 [in Norwegian]. Domaas, U., Rosenvold, B. S. et al. 2002. Studie av fjellskred og dalsidestabilitet i fyllittomra˚der. Norwegian Geotechnical Institute, Report 20001132–2 [in Norwegian]. Eberhardt, E., Stead, D. & Coggan, J. S. 2004. Numerical analysis of initiation and progressive failure in natural rock slopes – the 1991 Randa Rockslide. International Journal of Rock Mechanics and Mining Sciences, 41, 69–87. Eiken, T. 2007. Rapport om deformasjonsma˚linger i potensielle fjellskred Sogn og Fjordane. University of Oslo, Institute of Geosciences, Report 2005– 07 [in Norwegian]. El Bedoui, S., Guglielmi, Y., Lebourg, T. & Pe´rez, J.-L. 2008. Deep-seated failure propagation in a fractured rock slope over 10,000 years: the La Clapie`re slope, the south-eastern French Alps. Geomorphology, 105, 232– 238; doi: 10.1016/j.geomorph.2008.09.025. Etzelmu¨ller, B., Romstad, B. & Fjellanger, J. 2007. Automatic regional classification of topography in Norway. Norwegian Journal of Geology, 87, 167–180. Evans, S. G. & Clague, J. J. 1998. Rock avalanche from Mount Munday, Waddington Range, British Columbia, Canada. Landslide News, 11, 23– 25.

Evans, S. G. & Clague, J. J. 1999. Rock avalanches on glaciers in the Coast and St. Elias Mountain, British Columbia. In: Proceedings of the 13th Symposium on Slope Stability and Landslides, 28 May 1999, Vancouver Geotechnical Society, Vancouver, B.C., 115– 123. Evans, S. G., Scarascia Mugnozza, G., Strom, A. L., Hermanns, R. L., Ischuk, A. & Vinnichencko, S. 2006. Landslides from massive rock slope failure and associated phenomena. NATO Science Series. Series IV, Earth and Environmental Sciences, 49, 3– 52. Fossen, H. 2000. Extensional tectonics in the Caledonides: synorogenic or postorogenic? Tectonics, 19, 213– 224. Furseth, A. 1985. Dommedagsfjellet – Tafjord 1934. Gyldendal Norsk Forlag A/S [in Norwegian]. Furseth, A. 2006. Skredulykker i Norge. Tun Forlag, Oslo [in Norwegian]. Gabrielsen, R. H., Braathen, A., Dehls, J. & Roberts, D. 2002. Tectonic lineaments of Norway. Norwegian Journal of Geology, 82, 153–174. Giraud, A., Rochet, R. & Antoine, P. 1990. Processes of slope failure in crystallophyllian formations. Engineering Geology, 29, 241 –253. Govi, M., Gulla, G. & Nicoletti, P. G. 2002. Val Pola rock avalanche of July 28, 1987, in Valtellina (Central Italian Alps). Review of Engineering Geology, XV, 71–89. Grimstad, E. 2005. The Loen rock slide – an analysis of the stability. In: Senneset, K., Flate, K. & Larsen, J. O. (eds) Landslides and Rock Avalanches ICFL 2005 Norway. Taylor & Francis, London. Guzzetti, F., Cardinali, M. & Reichenbach, P. 1996. The influence of tectonic setting and lithology on the landslide type and pattern. Environmental and Engineering Geoscience, 2, 531– 555. Hanssen, T. H., Stjern, G. & Sørløkk, T. 1990. Bergartens mekaniske egenskaper. SINTEF Report STF36 F90048 (internal) [in Norwegian]. Harbitz, C., Løvholt, F., Domaas, U. & Blikra, L. H. 2003. Tsunamis generated by rockslides in Geiranger- and Tafjorden – scenarios and model comparison. Norwegian Geotechnical Institute, Report 20001622–1. Henderson, I. H. C. & Blikra, L. H. 2008. Ustabile fjellpartier i fyllittomra˚dene i Fla˚m – Aurland. Geological Survey of Norway, Report 2008.033. Henderson, I. H. C. & Saintot, A. 2007. Fjellskredundersøkelser i Møre og Romsdal. Geological Survey of Norway, Report 2007.043 [in Norwegian]. Henderson, I. H. C. & Saintot, A. 2011. Regional spatial variations in rockslide distribution from structural geology ranking: an example from Storfjorden, western Norway. In: Jaboyedoff, M. (ed.) Slope Tectonics. Geological Society, London, Special Publications, 351, 79– 95. Henderson, I H.C., Derron, M. H. & Jaboyedoff, M. 2004. A Combined Structural Geology and GIS Approach to Rockslides: An Example from Western Norway. American Geophysical Union, Fall Meeting 2004, abstract H41B. American Geophysical Union, Washington, DC. Henderson, I. H. C., Saintot, A., Bo¨hme, M. & Henriksen, H. 2008. Kartlegging av mulig ustabile fjellpartier, Sogn og Fjordane. Geological Survey of Norway, Report 2008.026 [in Norwegian].

ROCK SLOPE INSTABILITIES IN NORWAY Henderson, I. H. C., Saintot, A. & Derron, M. H. 2006. Structural mapping of potential rockslide sites in the Storfjorden area, Western Norway: the influence of bedrock geology on hazard analysis. Geological Survey of Norway, Report 2006.052. Henriksen, H. 2005. Fjellskredprosjektet Sogn og Fjordane. Geologiske undersøkelser pa˚ Stopelen, Høyanger. Report Høyanger kommune/Sogn og Fjordane Fylkeskommune [in Norwegian]. Hermanns, R. L. & Strecker, M. R. 1999. Structural and lithological controls on large Quaternary rock avalanches, sturzstroms, in arid northwestern Argentina. Geological Society of America Bulletin, 111, 934– 948. Hossack, J. R. 1984. The geometry of listric growth faults in the Devonian basins of Sunnfjord, west Norway. Journal of the Geological Society, London, 135, 705–711. Hossack, J. R. & Cooper, M. A. 1986. Collision tectonics in the Scandinavian Caledonides. In: Coward, M. P. & Ries, A. C. (eds) Collision Tectonics. Geological Society, London, Special Publications, 142, 863– 873. Jaboyedoff, M., Couture, R. & Locat, P. 2009. Structural analysis of Turtle Mountain (Alberta) using digital elevation model: Towards a progressive failure. Geomorphology, 103, 5 –16. Jaboyedoff, M., Oppikofer, T., Derron, M.-H., Blikra, L. H., Bo¨hme, M. & Saintot, A. 2011. Complex landslide behaviour and structural control: a three˚ knes rockslide, dimensional conceptual model of the A Norway. In: Jaboyedoff, M. (ed.) Slope Tectonics. Geological Society, London, Special Publications, 351, 147– 162. Jones, P. B. 1993. Structural geology of the modern Frank Slide and ancient Bluff Mountain Slide, Crownest, Alberta. Bulletin of Canadian Petroleum Geology, 41, 232 –243. Julian, M. & Anthony, E. 1996. Aspects of landslide activity in the Mercantour Massif and the French Riviera, southeastern France. Geomorphology, 15, 275–289. Jørstad, F. 1957. Steinsskredfare i A˚rdal, Sogn og Fjordane. Norwegian Geotechnical Institute, Report S.56.S.141 [in Norwegian]. Kaneko, K., Otani, K. J., Noguchi, Y. & Togashiki, N. 1997. Rock fracture mechanics analysis of slope failure. In: Asaoka, A., Adachi, T. & Oka, F. (eds) Deformation and Progressive Failure in Geomechanics. Elsevier, New York, 671– 676. Kellogg, K. S. 2001. Tectonic controls on a large landslide complex: Williams Fork Mountains near Dillon, Colorado. Geomorphology, 41, 355– 368. Krejcˇ´ı, O., Baronˇ, I., Bı´l, M., Hubatka, F., Jurova´, Z. & Kirchner, K. 2002. Slope movements in the Flysch Carpathians of Eastern Czech Republic triggered by extreme rainfalls in 1997: a case study. Physics and Chemistry of the Earth, 27, 1567–1576. Lied, K. 1975. Vurdering av steinskredfare for byggetomter pa˚ Rustøyen, Stryn. Norwegian Geotechnical Institute, Report 75422–1 [in Norwegian]. Lied, K. 1984. Vurdering av fare for flodbølge ved skred i Kleppura. Norwegian Geotechnical Institute, Report 83483– 1 [in Norwegian]. Lied, K. 1987. Vurdering av skredfare ved Kleppura. Norwegian Geotechnical Institute, Report 86495–1 [in Norwegian].

77

Lied, K. 1989. Kyrkjedøri, Terakamben; Aurland kommune, Sogn og Fjordane;Vurdering av faren for fjellskred. Norwegian Geotechnical Institute, Report 85465-1 [in Norwegian]. Mcclay, K. R. 1991. The Mapping of Geological Structures. Geological Society, London, Handbook Series. Mosar, J. 2003. Scandinavia’s North Atlantic passive margin. Journal of Geophysical Research, 108, 2360; doi: 10.1029/2002JB002134. Moxnes, A. B. 2002. Fjellskred i Sogn og Fjordane: Registrering og kartlegging. Master’s thesis, Høgskolen i Sogn og Fjordane [in Norwegian]. Nemcok, A. J., Pasek, J. & Rybar, J. 1972. Classification of landslides and other mass movements. Rock Mechanics, 4, 71–78. Nilsen, O. & Wolff, F. C. 1989. Geologisk kart over Norge, berggrunnskart Røros og Sveg, 1: 250,000. Geological Survey of Norway, Trondheim. Olesen, O., Dehls, J. et al. 2000. Neotectonics in Norway, Final Report. Geological Survey of Norway, Report 2000.002. Olesen, O., Blikra, L. H. et al. 2004. Neotectonic deformation in Norway and its implications: a review. Norwegian Journal of Geology, 84, 3– 34. Oppikofer, T. & Jaboyedoff, M. 2008. A˚knes/Tafjord Project Analysis of ancient rockslide scars and potential instabilities in the Tafjord area & Laser scanner monitoring of instabilities at Hegguraksla. University of Lausanne, Technical Report IGARTO-009. Oppikofer, T., Blikra, L. H., Derron, M.-H., Jaboyedoff, M. & Pedrazzini, A. 2008a. Geological, structural and geomorphic analysis of potential instabilities at Hegguraksla (Tafjord, Norway). In: 33rd International Geological Congress, (33IGC), Oslo, Norway, 6 –14 August, 2008, CD-ROM. Oppikofer, T., Bo¨hme, M., Blikra, L. H., Jaboyedoff, M. & Saintot, A. 2008b. Geological and structural ˚ knes landslide (Norway). In: 33rd model of A International Geological Congress, (33IGC), Oslo, Norway, 6 –14 August, 2008, CD-ROM. Passchier, C. W. & Trouw, R. A J. 2005. Microtectonics, 2nd edn. Springer, Berlin. Powell, J. W. 1875. Exploration of the Colorado River of the West and its Tributaries. Government Printing Office, Washington, DC. Ramamurthy, T., Venkatappa Rao, G. & Singh, J. 1993. Engineering behaviour of phyllites. Engineering Geology, 33, 209–225. Roberts, D. 2003. The Scandinavian Caledonides: event chronology, palaeogeographic settings and likely modern analogues. Tectonophysics, 365, 283 –299. Roberts, D. & Gee, D. G. 1985. An introduction to the structure of the Scandinavian Caledonides. In: Gee, D. G. & Sturt, B. A. (eds) The Caledonide Orogen—Scandinavia and Related Areas. Wiley, Chichester, 55– 68. Robinson, P., Tveten, E. & Blikra, L. H. 1997. A postglacial bedrock failure at Oppstadhornet, Oterøya, Møre og Romsdal; a potential major rock avalanche. Bulletin of the Geological Survey of Norway, 433, 46–47. Røhr, T. S., Corfu, F., Austrheim, H. & Andersen, T. B. 2004. Sveconorwegian U– Pb zircon and

78

A. SAINTOT ET AL.

monazite ages of granulite-facies rocks, Hisarøya, Gulen, Western Gneiss Region, Norway. Norwegian Journal of Geology, 84, 251– 256. Saintot, A., Bo¨hme, M., Redfield, T. & Dahle, H. 2008. Field studies of unstable slopes in Sunndalen Valley. Geological Survey of Norway, Report 2008.049. Saroglou, H. & Tsiambaos, G. 2008. A modified Hoek– Brown failure criterion for anisotropic intact rock. International Journal of Rock Mechanics & Mining Sciences, 45, 223–234. Sartori, M., Baillifard, F., Jaboyedoff, M. & Rouiller, J. D. 2003. Kinematics of the 1991 Randa rockslides (Valais, Switzerland). Natural Hazard and Earth System Sciences, 3, 423–433. Sauchyn, D. J., Cruden, D. M. & Hu, H. Q. 1998. Structural control of the morphometry of open rock basins, Kananaskis region, southwestern Alberta. Geomorphology, 22, 313–324. Scavia, C. 1995. A method for the study of crack propagation in rock structures. Geotechnique, 45, 447–463. Selby, M. J. 1993. Hillslope Materials and Processes. Oxford University Press, Oxford. Selmer-Olsen, R. 1972. Steggjaberget riksveg 53, ˚ rdal. Oppdragsrapport, 46-S30, Veglaboratoriet [in A Norwegian]. Ska˚r, Ø. & Pedersen, R. B. 2003. Relations between granitoid magmatism and migmatization: U–Pb geochronological evidence from the Western Gneiss Complex, Norway. Journal of the Geological Society, London, 160, 935 –946. Terzaghi, K. 1962. Stability of steep slopes on hard unweathered rock. Geotechnique, 12, 251–270. Torsvik, T. H., Andersen, T. B., Eide, E. A. & Walderhaug, H. J. 1997. The age and tectonic

significance of dolerite dykes in Western Norway. Journal of the Geological Society, London, 154, 961–973. Tucker, R. D., Krogh, T. E. & Ra˚heim, A. 1990. Proterozoic evolution and age – province boundaries in the central part of the Western Gneiss Region, Norway: results of U –Pb dating of accessory minerals from Trondheimsfjord to Geiranger. In: Gower, C. F., Rivers, T. & Ryan, B. (eds) Mid-Proterozoic Laurentia-Baltica. Geological Association of Canada, Special Paper, 38, 149– 173. Tveten, E., Lutro, O. & Thorsnes, T. 1998. Geologisk kart over Norge, berggrunnskart A˚lesund, 1: 250,000. Geological Survey of Norway, Trondheim. Twiss, R. J. & Moore, E. M. 1992. Structural Geology. W.H. Freeman & Co., New York. Valle, P., Faerseth, R. B. & Fossen, H. 2002. Devonian– Triassic brittle deformation based on dyke geometry and fault kinematics in the Sunnhordland region, SW Norway. Norwegian Journal of Geology, 82, 3 –17. Varnes, D. J. 1978. Slope movement types and processes. In: Schuster, R. L. & Krizek, R. J. (eds) Landslides, Analysis and Control. Transportation Research Board Special Report, 176. National Research Council, Washington, DC, 11– 33. Venvik Ganerød, G., Grøneng, G. et al. 2008. Geo˚ knes Rockslide, Western logical Model of the A Norway. Engineering Geology, doi: 10.1016/ j.enggeo.2008.01.018 Vilı´mek, V., Zvelebil, J., Klimesˇ, J., Patzelt, Z., Astete, F., Kachlı´k, V. & Hartvich, F. 2007. Geomorphological research of large-scale slope instability at Machu Picchu, Peru. Geomorphology, 89, 241– 257.

Regional spatial variations in rockslide distribution from structural geology ranking: an example from Storfjorden, western Norway IAIN H. C. HENDERSON* & ALINE SAINTOT Geological Survey of Norway, Leiv Eirikssons Veg 39, 7491 Trondheim, Norway *Corresponding author (e-mail: [email protected]) Abstract: Structural geology has recently become a key topic in landslide research. However, the link between regional structures, their cumulative contribution to rockslide development and their significance in a spatial framework is uncertain. We examine the influence of structures on rockslide susceptibility in the Storfjorden area, a 900 km2 fjord complex in western Norway that ˚ knes and Heggursaksla. We have newly identified 52 includes the monitored rockslide sites of A potential rockslide sites from aerial photographs. The structural features critical for the development of large rockslides (fjord-dipping foliation, basal shear plane and breccia, extensional fracture and transfer fault) have a spatial bias in orientation. Rockslides are more likely to develop in specific fjord orientations that have favourably oriented structures. Therefore, the development of rockslides has a marked spatial distribution that we describe qualitatively with an inventory of structural features. Sites with the full plethora of features display the most movement, the largest volumes and are already under the closest scrutiny with regard to monitoring. These sites are also spatially biased, the largest clustering occurring in west Sunnylyvsfjorden. The utilization of structural criteria can show trends in spatial distribution of rockslide potential and on a regional scale can be an important tool in susceptibility analysis.

The study of rockslides has increased rapidly in recent years owing to the increasing needs of society and as a result of the expanding use of remotesensing techniques (e.g. Van Westen et al. 1997; Dai & Lee 2001; Jaboyedoff et al. 2005). Despite this, the development of multi-scalar, regional- to outcrop-scale field studies (e.g. Chigira 1992; Hungr et al. 1999; Agliardi et al. 2001) are still the exception. Recent studies examine the spatial distribution of rockslides where tectonic, structural and topographical data are combined with spatial analysis (DeGraff 1978; Nichol et al. 2002; Ambrosi & Crosta 2006, 2011; Di Crescenzo & Santo 2007). Other studies shed new light on the underlying geometrical, structural and tectonic causes of rockslides (e.g. Cruden 1976; Varnes 1978; Braathen et al. 2004; Redfield & Osmundsen 2009; Agliardi et al. 2009b; Osmundsen et al. 2009). Our understanding of the underlying geological causes of, and controls on, the development of rockslides and their regional spatial distribution is often based on intuitive or subjective geological observation. However it is only seldom that a more objective and empirical control on rockslide causes and distribution is forthcoming (Sauchyn et al. 1998; Kawabata et al. 2004). Studies focusing on the combination of individual structural controls on rockslides (e.g. Sauchyn et al. 1998; Amin 1999; Bo¨hme et al. 2011), particularly in relation to the interplay between lineament orientation and valley orientation is sometimes overlooked. For example, the control of fall-line sloping structures parallel to planar slopes is a

direct and intuitive control on rockslide development and it is one which will be clearly shown to be an important factor in our field study. However, studies of the interaction of a varying large-scale fjord (valley) architecture relative to different structural trends are rather few and are mostly confined to the field of geomorphology (e.g. Schmidt & Montgomery 1995). Very few have focused on structural aspects of susceptibility (e.g. Irigaray et al. 2003; Romana et al. 2003). Some studies focus on the controlling factors for the spatial distribution of rockslides (Osmundsen et al. 2009; Redfield & Osmundsen 2009) without invoking susceptibility considerations, whereas others have focused on the susceptibility from geographical information systems (GIS), slope aspect and numerical modelling (Van Westen et al. 1997; Dai & Lee 2001; Dorren et al. 2004; Guthrie & Evans 2004; Mansor et al. 2004) without detailed structural analysis. Our aim is therefore not to contribute to further development of geotechnical models on a site scale (e.g. Markland 1972; Hocking 1976; Hoek & Bray 1981) or on a regional scale (e.g. Ward et al. 1981; Fall & Azzam 2001; Fall et al. 2006) but rather to integrate more conventional structural geology techniques with a spatial study of rockslide susceptibility. Such studies of slope tectonics have been relatively few until recently (Gu¨nther 2003; Braathen et al. 2004; Roering et al. 2005; Agliardi et al. 2009a; Saintot et al. 2011). The Western Gneiss Region (WGR) of Norway (Fig. 1a) is historically a place with many

From: Jaboyedoff, M. (ed.) Slope Tectonics. Geological Society, London, Special Publications, 351, 79–95. DOI: 10.1144/SP351.4 0305-8719/11/$15.00 # The Geological Society of London 2011.

80 I. H. C. HENDERSON & A. SAINTOT Fig. 1. (a) Regional location map showing the study area (black box) relative to the MTFC and the relatively high density of rockslides Precambrian basement is in pink, Caledonian thrust sheets are in greens and yellows. (b) Profile X–Y through A showing uplift data (Olesen et al. 2000) relative to the clustering of both historic rockslides and potential rockslides. (c) Storfjorden field area showing bedrock geology and simplified foliation form lines (dashed red) in the Precambrian basement. Blue lines are open folds in the Precambrian basement. Green dots are identified potential rockslides.

STRUCTURAL GEOLOGY RANKING

documented rockslides. Enhanced clustering of rockslides in this area may be owing to the presence of a strong gradient in regional uplift (Fig. 1b). Other studies suggest that areas with strong uplift gradients are more prone to rockslide clustering (Jaboyedoff et al. 2003). Our study area, in the Storfjorden Fjord Complex (Fig. 1c), is approximately 900 km2, and comprises Sunnylyvsfjorden, Geirangerfjorden, Norddalsfjorden, Tafjorden and the southern part of Storfjorden, where rockslide occurrence is exceptionally enhanced. We have identified 52 new rockslide sites from initial aerial photograph analysis, airborne LiDAR-DEM (Light Detection and Ranging digital elevation model) and digital orthophotographs. Our study incorporates an inventory of structure types present on these sites (the reader is referred to Saintot et al. 2011 for a detailed description of the sites), regional lineament analysis from the remote sensing, and orientation analysis of the different structures present at the sites relative to the many different fjord and slope orientations. We determined which orientations and types of structures were likely to increase the spatial susceptibility depending on variable fjord orientation. From a qualitative numerical ranking of the different structural types we present a novel method of demonstrating the regional spatial susceptibility of rockslide occurrence.

81

gneisses and the overlying Caledonian thrust sheets (Fig. 1c). Over 80% of the surface area consists of granodioritic –dioritic gneiss of the Precambrian basement. Small amounts of quartz mica gneiss and coarse-grained granitic gneiss are present. In the eastern part of the field area a Caledonian thrust sheet consisting of augen gneiss and minor mica schist is present. The majority of potential rockslide sites, up to 70%, are located within the basement granodioritic –dioritic gneiss and appear to display no localization along lithological or tectonic contacts, with the exception of sites 20–23 and Heggursaksla in Tafjord, which are located on the base of the Caledonian thrust sheet.

Methods Stereographical aerial photograph analysis initially identified the sites, followed by helicopter reconnaissance and detailed field structural analysis. Volume calculations and regional lineament mapping were undertaken using regional airborne LiDAR-DEM and associated high-resolution digital orthophotographs. Regional structural data were collected, as well as site-specific structural data.

Results Regional setting Tectonic Our study area is located in an area of enhanced rockslide occurrence associated with enhanced present-day uplift gradients (Fig. 1a, b) that are, in part, glacially-controlled but can also be related to neotectonic activity (e.g. Blikra 1990; Dehls et al. 2000; Blikra et al. 2002; Braathen et al. 2004; Olesen et al. 2004; Redfield et al. 2004) on the Møre– Trøndelag Fault Complex (MTFC). Redfield & Osmundsen (2009) postulated a long-lived topographical high related to lithospheric flexure and normal faulting, creating a steep footwall margin (‘Great Escarpment’). Our study area lies directly on the western edge of this footwall high. This is controversially purported to be still active today. In addition, the only definitive description of a neotectonically active fault in southern Norway appears to directly segment the study area (Anda et al. 2002; Blikra et al. 2002), and can be an influencing factor in the development and spatial clustering of these rockslides.

Lithological The field area transects the contact between the Western Gneiss Region crystalline basement

Site-specific structural controls on rockslide development Our regional mapping has identified 52 rockslide sites. Site-specific structural analysis has generated over 3000 measurements of different types: foliation, exfoliation, joints, fractures, sliding planes, breccias and pre-existing faults, which often form mappable regional lineaments (Fig. 2). On all of the rockslides where the block is fully detached, a well-defined basal sliding plane (BSP) lies subparallel to a fjord-dipping foliation and sites that demonstrate fully developed extensional fractures are localized on fjord-dipping foliation with 208 –358 dipping sliding planes (Fig. 2a). The BSP is most often parallel/subparallel to foliation or is localized in a mica-rich sliver or along a pre-existing breccia structure that is very often subparallel to foliation. In rare cases the BSP is parallel to exfoliation, which may be unrelated to foliation orientation. Block detachment varies from several centimetres up to 10 m. Where the extensional fractures are not fully opened a topographical depression is observed at the rear of the block. Examples of evidence for down-slope gravitational motion are gravitational duplexes at the base of the block (Fig. 2b), partial disintegration of the hanging-wall block by extensional fracture development above the BSP

82

I. H. C. HENDERSON & A. SAINTOT

Fig. 2. Examples of the structures observed in the field common to most of the examined sites. (a) Site 6B, which shows a detached block with a several-metre-wide back-crack and several secondary subvertical fractures in the block, fjord-dipping foliation and a sliding plane at the base of the block subparallel to foliation. (b) Brittle extensional duplex at the base of site 2B showing down-slope movement. (c) Extensional fractures in the hanging wall to the basal shear zone at site 2B. (d) Non-cohesive fault rock breccia on the sliding plane at the base of site 6B. The inset photographs shows a SEM (scanning electron microscope) image of the breccia texture.

(Fig. 2c) and the development of a non-cohesive breccia (Henderson et al. 2009) where sufficient movement along the BSP has occurred (Fig. 2d). The main regional lineament trend is NNE– SSW or NE –SW (Fig. 3), with minor trends of WNW–ESE and east– west. These consist of steeply dipping, cemented breccia fault zones (Fig. 2e). Translational rockslides are by far the most predominant geometrical type of rockslide and often have large potential volumes, occurring where the pre-existing structural geometries are favourable: the presence of fjord-dipping foliation, a BSP, a fault breccia on the BSP, the presence of exfoliation (to weaken the foliation planes), a steep extensional fracture at the back of the block and a through-going transfer fault to detach the block. A conceptual model of the development of rockslides within the Storfjorden area is shown in Figure 2g and is similar to that developed by others (Hoek & Bray 1981; Chigira 1992; Braathen

et al. 2004). Where few or none of these structural criteria are present, there is generally no evidence for activity in the hanging-wall block, such as preferential hanging-wall disintegration by fracturing (Fig. 2c), kinematic evidence (Fig. 2b) and breccia development (Fig. 2d). Extension fractures are localized on, and reactivate many, pre-existing geological lineaments. Small-fold development within the foliation also appears to have a large effect on the development and movement of rockslides. More details of the relationships between pre-existing ductile structures and the geometrical development of potential rockslides are discussed by Saintot et al. (2011).

Regional structural controls on rockslide development Regional mapping demonstrates that the foliation forms a series of complex recumbent isoclinal

STRUCTURAL GEOLOGY RANKING

83

Fig. 2. (Continued) (e) A 1 m-wide pre-existing cataclastic zone forming the eastern margin of site 7B acting as a transfer structure. (f) Basal shear zone at site 2B intersecting topography showing a 15 cm displacement away from underlying talus. This suggests recent movement. (g) Schematic model of the assemblage of different structural types necessary to cause failure, based on the 52 sites studied here.

folds in the host granodioritic –dioritic gneisses, which is then refolded on open folds with ENE – WSW fold axes where the dip of the foliation is between 208 and 408 (Fig. 1c). The fjord system in Storfjorden consists of several east –west- and several north –south-trending fjords. We contend that the combination of the different structural

types, their regional variation in orientation, the variation in regional lineament orientation and how these different structural aspects interact with a varying fjord orientation can produce a spatial variation in rockslide susceptibility. For example, we identify eight areas where the regional lineament orientation is different. Lineament mapping (Fig. 3,

84 I. H. C. HENDERSON & A. SAINTOT Fig. 3. Digital aerial photographs acquired simultaneously with airborne LiDAR-DEM data over the whole study area. Both the digital orthophotographs and the airborne LiDAR-DEM data have been used to create a lineament map. We divided the area into eight different domains and demonstrated that different lineament sets are present in different ˚ knes site). areas of the study area (the inset shows an example of the combined airborne LiDAR-DEM and orthophotography from the A

STRUCTURAL GEOLOGY RANKING 85

Fig. 4. The orientation of different key structures relative to the fjord geometry (‘quarter’ stereonets are structure dip). The background map is the detailed digital orthophotography. Structures have a distinct preferred orientation, suggesting that specific geographical areas will have an enhanced potential for rockslides if several structures have favourable orientations and that there will therefore be a spatial variation in rockslide susceptibility.

86

I. H. C. HENDERSON & A. SAINTOT

Fig. 5. Matrix constructed from the structural mapping of the different sites and the ranking of these based on the number of structural criteria present that we contend are important for the development of a rockslide.

STRUCTURAL GEOLOGY RANKING

areas A –H) from the digital orthophotographs and airborne LiDAR-DEM shows a wide variation in lineament orientation in different fjords. (The inset in Fig. 3 shows an example from one site with the combined digital orthophotographs, airborne LiDAR-DEM and simplified mapping.) Area A (Fig. 3a) displays predominantly WNW– ENE lineaments that may, for example, act as transfer structures, whereas a limited number of north– south lineaments may act as extensional backfractures. For a rockslide in area B (Fig. 3b) both east –west (predominant) and NNE– SSW (minor) structures may act as transfer structures and extensional back-fractures, respectively. Area C (Fig. 3c) displays a major NE– SW trend (and a minor north–set set) that may act as transfer structures, whereas a limited number of NW–SE structures may act as extensional back-fractures. In area D (Fig. 3d) a predominant set of WNW –ESE structures may act as extensional back-fractures.

87

There are several minor lineament trends but not in any abundance to function as north –south transfer structures. Very few potential rockslides are observed here (except for sites 11 and 12) relative to the other areas. In area E (Fig. 3e) where there are also very few lineaments developed (with many oblique to the slope). This area is also limited in rockslide development (cf. Fig. 1c). Area F (Fig. 3f) has two well-developed lineament trends, ENE –WSW and north–south. The NW –SE orientation of the fjord means that these structures are not ideally oriented for potential rockslides, except in the NW part where several instabilities are observed (see Fig. 1c). Area G (Fig. 3g) has a major NE–SW and a minor NW–SE trend. These orientations are not ideal for rockslide development as they are both highly oblique to the east –west fjord orientation and, indeed, this is a geographical domain with limited rockslide potential. Area H (Fig. 3h) has few north–south lineaments and also has poor

Fig. 6. The spatial distribution of different localities plotted by volume (as calculated using the airborne LiDAR-DEM data). The localities with the largest volumes are concentrated on the western side of Sunnylyvsfjorden.

88

I. H. C. HENDERSON & A. SAINTOT

rockslide potential. In area H there is extremely limited evidence for back fractures. Insufficient lineaments are present to localize fracture and those that are present have a high obliquity to the slope. The different structural types have a strong orientation bias (Fig. 4), which may mean that their influence in rockslide development is biased towards particular orientations of fjords. For example, foliation dips, on average, 308 but shows a strong bias to SE-dipping (Fig. 4a). Therefore, foliation is likely to play a major role in developing potential rockslides (Jaboyedoff et al. 2011) preferentially on SE-facing fjords. This is indeed the case ˚ knes, Heggursakin reality where the largest sites (A sla, 2B, 4, 5, and 6B) are all located on SE-facing slopes. Extensional fractures have a wide variation in orientation (Fig. 4d), but have a predominant NNE–SSW trend, suggesting that north –south

fjords (in this case predominantly Sunnylyvsfjorden) are more susceptible to rockslides. On a regional scale, structures potentially acting as transfers have a predominantly north –south trend (Fig. 4e). Therefore, fjord systems that are north – south trending may develop rockslides more easily because structures able to function as transfer structures are already present. Basal shear planes and associated breccias have two main trends. These are NW- and SE-dipping and NNE-dipping (Fig. 4f), suggesting that SE-facing parts of fjords (the southern part of Sunnylyvsfjorden) and the outer northern part of Tafjord, the southern parts of Tafjord and Norddalsfjorden, and south Geirangerfjorden would be more susceptible to rockslides. Joint orientation shows an extremely variable orientation and does not represent a strong bias for any fjord orientation on a regional basis and is isotropic on a regional scale (Fig. 4b). However, small-fold

Fig. 7. The spatial distribution of different localities plotted by the amount of structural criteria present (based on the structural mapping). This, like the volume data presented in Figure 6, demonstrates that the sites with the greatest number of structural criteria are located in the same areas.

STRUCTURAL GEOLOGY RANKING

orientations show a very distinct SE-plunging orientation, suggesting that SE-facing slopes would be more susceptible (Fig. 4c). Exfoliation (Fig. 4g) has two main trends; either NNE- or SSW-dipping or SE-dipping, many exfoliation planes are parallel to foliation and are therefore controlled by foliation orientation. All three of these orientations would affect susceptibility if the exfoliation enhanced weakening of foliation planes, therefore allowing easier development of BSPs.

Structural ranking of rockslide sites We constructed a simple cumulative structural inventory that categorized the 52 rockslides in terms of structures favourable for sliding (Fig. 5). We also included geomorphological evidence for recent activity and GPS (global positioning system) measurements. These last two factors

89

were weighted relative to the structural criteria because they constitute direct evidence for movement (these have a value of 2; Fig. 5). The results represent a relative susceptibility to failure of the various sites on a regional scale. First, a very wide spectrum of cumulative structural features was observed per rockslide site, based on a scale from 0 to 9. Secondly, approximately 80 –85% of sites show three criteria or less. It is notable that none of these sites display a fault breccia on the sliding plane, suggesting that the presence of a fault rock is necessary to facilitate movement. Seventy per cent of the sites with four structural criteria or more in the structural inventory have a gouge. In addition, it is clear that both geomorphological evidence for recent activity and GPS movement occur only where a breccia is present on the sliding plane. Of the 52 sites only 21 (c. 40%) have a fully developed sliding plane. Of this subset of localities with a

Fig. 8. The spatial distribution of different localities plotted by evidence for recent activity (based on geomorphological observations in the field, such as those shown in Fig. 3f) and those displaying a gouge in the BSP. This shows remarkable spatial agreement with both Figure 6 and Figure 7 (volume and amount of structural criteria, respectively).

90

I. H. C. HENDERSON & A. SAINTOT

sliding plane, only eight (15% of the total sites) have a fault rock developed. Figure 5 also shows that 80% of the localities with four or more structures have small-scale structures that are favourable for slide development, such as fjord-plunging fold axes. This point is discussed in more detail in Saintot et al. (2011).

Spatial variation in susceptibility There is a distinct spatial distribution of the different rockslide sites relative to their volumes (Fig. 6), with the largest potential slides concentrated in west Sunnylyvsfjorden or in Tafjord. The regional variation in structural criteria presented in Figure 5 is also notable (Fig. 7). Sites with the most structural criteria (eight or nine) are found exclusively on the western side of Sunnylyvsfjorden or in Tafjord. Site 21 is the only exception and is probably the result of its localization on the

Caledonian thrust zone. Figure 8 shows that most of the sites that contain a gouge have also had recent activity. Comparison of these features with the structural criteria presented in Figure 7 demonstrates that there is also a relationship between structural criteria and the presence of a gouge. Figure 9a shows the relationship between fjorddipping foliation orientation and potential rockslide location: the shaded areas show where foliation dips between 208 and 358 down towards the fjord. There is a remarkable correlation between the areas with favourable foliation for sliding and those areas where the biggest potential slides are observed (notably including both of the key localities presently ˚ knes and Tafjord). Only two being monitored: A minor rockslides volumes occur outside of these foliation demarcated areas. Figure 9b shows a susceptibility map constructed from the site information on volume alone and this shows remarkable correlation with areas of fjord-dipping foliation.

Fig. 9. (a) Susceptibility map based on fjord-dipping foliation relative to mapped areas with potential failures. There is good agreement between the inferred susceptible areas and the area where rockslides are actually mapped. These delineated higher susceptibility areas agree remarkably well with the largest mapped potential rockslides, most of which lie within these areas. Only two small volume rockslides are located outside of these areas.

STRUCTURAL GEOLOGY RANKING

West Sunnylyvsfjorden and Tafjord have the highest average volumes of potential rockslides and also areas with the most consistently fjord-dipping foliation. It is also notable that these are the areas with the two highest structural criteria inventory. Figure 10 demonstrates that the vast majority (over 75%) of the potential rockslides are less than 1 Mm3. There is also a rapid drop-off in potential rockslide volume over approximately 5 Mm3. ˚ knes and Heggursaksla (Fig. 1c) are the only two A sites that show active movement and these happen to be amongst the largest sites in terms of volume. Movement here is documented from several millimetres up to 8 cm per year on average. Therefore the fact that these sites are high-priority monitored objects is noticeable from both our structural inventory ranking and from the spatial distribution data that we present.

91

Discussion We have newly identified 52 potential rockslides clustering in an area of approximately 900 km2 in western Norway. Braathen et al. (2004) provided a comprehensive review of the structural geometries of the most prominent rockslide sites in Norway; we identify the same structures here and demonstrate that these are cumulatively important for the development of rockslides. Only the most active and/or the largest potential rockslides have a full inventory of these structures (Henderson et al. 2006; Saintot et al. 2011). We demonstrate that a spatial variation of these structural elements leads to a spatial variation in the susceptibility to rockslide development. Ganerød et al. (2008) have identified the critical geological structures on the ˚ knes slide that are identical to the highly active A

Fig. 9. (Continued) (b) Susceptibility map based on the average slide volume for individual geographical areas (fjords). The average structural criteria inventory is shown as numbers. The areas of west Sunnylyvsfjorden and both south and north Tafjord have the largest average volumes. West Sunnylyvsfjorden and north Tafjord also have the greatest number of structural criteria.

92

I. H. C. HENDERSON & A. SAINTOT

Fig. 10. (a) Frequency histogram of the occurrence of potential rockslide sites with respect to their volume. This demonstrates that most sites have volumes of less than 1 Mm3 and only nine sites have volumes above 8 Mm3. However, ˚ knes and Heggursaksla). (b) Log-normal two of the largest sites are also those that are documented as being active (A plot of structural criteria with rockslide volume. This demonstrates that it is only the largest volumes that display the greatest possible number of structural features.

structures we observe regionally and play an obvious role in the development of the gravitational structures over the whole fjord system. Ambrosi & Crosta (2006) presented a similar study concerning the spatial distribution of rockslides or ‘sackung’ within the Alps related to tectonic features. They demonstrate the importance of similar lithological and structural constrains on rockslide development, and their spatial distribution. Bovis & Evans (1996) also described an identical suite of structures to those that we determined as critical to develop large potential rockslide volumes. Such regional analytical approaches to rockslide susceptibility based on geological information have also been carried out. Hungr et al. (1999) provided an extensive study (up to 3500 sites) of rockslides and rockfalls on a regional scale in Canada, and outlined a risk analysis based on magnitude–cumulate frequency considerations. However, we believe that our approach, where the relative susceptibility of different rockslides on a regional scale is based on a spatial variability of structural elements, is a valuable concept that could be applied elsewhere. The regional clustering of rockslides may be related to the proximity of the long-lived and still active MTFC to the NW. This is also the location of the highest gradient of uplift recorded in western Norway (Hicks et al. 2000; Bungum et al. 2005). We propose that rockslide development in this area is, in part, related to recognized, but hitherto unassessed, neotectonic factors (Blikra et al. 2002; Olesen et al. 2004). We document an

interplay of structural controls, their spatial variability relative to fjord geometry, and what role they may play in potential rockslide development and distribution. The make-up of the structural inventory (fjord-dipping foliation, sliding plane and breccia development, extensional detachment, and transfer structures) and how these structures interact with a variable fjord geometry is critical to the geographical distribution of rockslides. The spatial pattern of rockslides is also determined by the preexisting lineaments, which we have mapped as having different orientations in the different fjords. The underlying geology and, particularly, the lineament pattern often plays a critical role in rockslide distribution (Bachmann et al. 2004). We are therefore able to determine areas of spatially enhanced rockslide susceptibility based on the structural inventory. We observe that fjorddipping foliation is generally the main precursor for the development of fjord-dipping sliding planes and the development of subsequent gouges that weaken the base of the rockslide. Suitably oriented foliation is therefore an important factor in determining spatial variation in rockslide susceptibility on a regional scale but cannot solely account for the distribution of rockslides. Ranking of the 52 different sites based on their structural inventory to assess the structural influence ˚ knes and Hegon rockslides development places A gursaksla, which have the full plethora of structural criteria, as the most susceptible. This confirms that this novel approach is a viable method of identifying

STRUCTURAL GEOLOGY RANKING

the key structures, structural processes and the most important sites within a rockslide susceptible area on a regional scale. We also observe a distinct spatial clustering of rockslides with the largest volumes and with the most favourable structures for rockslide development. Our data suggest that the west side of Sunnylyvsfjorden and Tafjorden are the two areas most prone to rockslide development (where both ˚ knes & Heggursaksla are located). A Notably, the clustering of rockslides that we observe has little to do with the pattern of regional bedrock geology. The area with the clustering of most rockslides, displaying the highest average volumes and the most structural criteria in west Sunnylyvsfjorden, lies in the middle of the Precambrian Basement granodioritic gneiss, far from the contacts with the overlying Caledonian cover. The only exception to this is the northern side of Tafjord, where the Heggursaksla slide is located close to the autochthon–allochthon interface. Application of this methodology to similar areas containing a high concentration of potential rockslides, where no monitoring is yet present, is a powerful and inexpensive tool in prioritizing sites for monitoring. We believe that the analysis of regional structural data in this way in active rockslide systems can: (1) help to determine the spatial distribution of potential rockslide activity; (2) help to understand the contributing factors promoting failure; and (3) provide a better geological foundation for susceptibility analysis.

Conclusions † The 900 km2 Storfjorden area is coincident with a high regional uplift gradient, adjacent to the still active Møre –Trøndelag Fault Complex (MTFC). Fifty-two new potential rockslide sites have been discovered during this study, their clustering showing a correlation with uplift and historical rockslides pattern. † Regional lineaments and mapped structures (fjord-dipping foliation, sliding planes, basal breccias, extension fractures and transfer structures) that delineate potential rockslides have preferred orientations, meaning that rockslides are more likely to develop in specific fjord orientations that have favourably oriented structures. † Cumulative ranking of all sites by structural features demonstrates that those with the highest structural ranking are those that display movement, form amongst the largest volumes and are already being monitored. † The largest volumes and those with the most structural criteria are spatially biased, the largest clustering occurring in west Sunnylyvsfjorden.

93

† Comparison of the structural ranking method shows that areas with suitably oriented fjorddipping foliation (between 208 and 408) are coincident with sites displaying the bestdeveloped fault gouges and also sites that have the largest volumes. † The spatial distribution of potential rockslides appears to have only a limited relationship to the regional-scale bedrock pattern, as demonstrated by the clustering of sites within the large, crystalline, hard granodioritic mass on the west of Sunnylyvsfjorden and the few located on the minor lithological units and/or close to the major tectonic contacts. L. H. Blikra is thanked for providing interesting discussions in the field. Reviews by G. Crosta and an anonymous reviewer greatly improved the manuscript. This project was funded by the Norwegian Government through the intermunicipality project ‘Aaknes-Tafjord’. K. Jogerud and T. Bergeng from Stranda Kommune are thanked for logistical help in the field. Partial funding was also received from the International Centre for Geohazards. This work is ICG publication No. 278.

References Agliardi, F., Crosta, G. & Zanchi, A. 2001. Structural constraints on deep-seated slope deformation kinematics. Engineering Geology, 59, 83–102. Agliardi, F., Crosta, G. B., Zanchi, A. & Ravazzi, C. 2009a. Onset and timing of deep-seated gravitational slope deformations in the eastern Alps, Italy. Geomorphology, 103, 113–129. Agliardi, F., Zanchi, A. & Crosta, G. B. 2009b. Tectonic vs. gravitational morphostructures in the central Eastern Alps (Italy): constraints on the recent evolution of the mountain range. Tectonophysics, 474, 250–270; doi:10.1016/j.tecto.2009.02.019. Ambrosi, C. & Crosta, G. B. 2006. Large Sackung along major tectonic features in the central Italian Alps. Engineering Geology, 83, 183–200. Ambrosi, C. & Crosta, G. B. 2011. Valley shape influence on deformation mechanisms of rock slopes. In: Jaboyedoff, M. (ed.) Slope Tectonics. Geological Society, London, Special Publications, 351, 215–233. Amin, A. A. 1999. Geologic hazards along part of Al-Sayl– Alkabeir Al-Jammum road in Saudi Arabia. Environmental Geology, 39, 20– 24. Anda, E., Blikra, L. H. & Braathen, A. 2002. The Berill Fault – First evidence of neotectonic faulting in southern Norway. Norwegian Journal of Geology, 82, 175 –182. Bachmann, D., Bouissou, S. & Chemenda, A. 2004. Influence of weathering and pre-existing large scale fractures on gravitational slope failure: insights from 3-D physical modelling. Natural Hazards and Earth System Sciences, 4, 711–717. Blikra, L. H. 1990. Geological mapping of rapid mass movement deposits as an aid to land-use planning. Engineering Geology, 29, 365–376.

94

I. H. C. HENDERSON & A. SAINTOT

Blikra, L. H., Braathen, A., Anda, E., Stalsberg, K. & Longva, O. 2002. Rock avalanches, gravitational bedrock fractures and neotectonic faults onshore northern west Norway. Geological Survey of Norway Report 2002.016. Bovis, M. J. & Evans, S. G. 1996. Extensive deformations of rock slopes in southern Coast Mountains, southwest British Columbia, Canada. Engineering Geology, 44, 163– 182. Bo¨hme, M., Saintot, A., Henderson, I. H. C., Henriksen, H. & Hermanns, R. L. 2011. Rock slope instabilities in Sogn & Fjordane County, Norway: a detailed structural and geomorphological analysis. In: Jaboyedoff, M. (ed.) Slope Tectonics. Geological Society, London, Special Publications, 351, 97–111. Braathen, A., Blikra, L. H., Berg, S. S. & Karlsen, F. 2004. Rock-slope failures in Norway, type, geometry, deformation mechanisms and stability. Norwegian Journal of Geology, 84, 67–88. Bungum, H., Lindholm, C. & Faleide, J. I. 2005. Postglacial seismicity offshore mid-Norway with emphasis on spatio-temporal –magnitudal variations. Marine and Petroleum Geology, 22, 137–148. Chigira, M. 1992. Long term gravitational deformation of rocks by mass rock creep. Engineering Geology, 32, 157– 184. Cruden, D. M. 1976. Major slides in the Rockies. Canadian Geotechnical Journal, 13, 8 –20. Dai, F. C. & Lee, C. F. 2001. Terrain-based mapping of landslide susceptibility using a geographical information system: a case study. Canadian Geotechnical Journal, 38, 911 –923. Degraff, J. V. 1978. Regional landslide evaluation: two Utah examples. Environmental Geology, 2, 203– 214. Dehls, J. F., Olesen, O., Olsen, L. & Blikra, L H. 2000. Neotectonic faulting in northern Norway; the Stuoragurra and Nordmannvikdalen postglacial faults. Quaternary Science Reviews, 19, 1447–1460. Di Crescenzo, G. & Santo, A. 2007. High-resolution mapping of rock fall instability through the integration of photogrammetric, geomorphological and engineering– geological surveys. Quaternary International, 171, 118– 130. Dorren, L. K. A., Maier, B., Putters, U. S. & Seijmonsbergen, A. C. 2004. Combining field and modelling techniques to assess rockfall dynamics on a protection forest hillslope in the European Alps. Geomorphology, 57, 151–167. Evans, S. G., Mugnozza, G. S., Strom, A. & Hermanns, R. L. 2006. Single-event landslides resulting from massive rock slope failure: characterising their frequency and impact on society. NATO Science Series. Series IV, Earth and Environmental Sciences, 49, 53– 73. Fall, M. & Azzam, R. 2001. An example of multidisciplinary approach to landslide assessment in coastal area. International conference on landslide. In: Proceedings of the International Conference on Landslides: Causes Impacts and Countermeasures. Glu¨ckauf, Davos, Switzerland, 45–54. Fall, M., Azzam, R. & Noubactep, C. 2006. A multimethod approach to study the stability of natural

slopes and landslide susceptibility mapping. Engineering Geology, 82, 241– 263. Ganerød, G. V., Grøneng, G. et al. 2008. A geological ˚ knes rockslide, western Norway. Enginmodel of the A eering Geology, 102, 1– 18. Gu¨nther, A. 2003. SLOPEMAP: programs for automated mapping of geometrical and kinematical properties of hard rock hill slopes. Computers & Geosciences, 29, 865–875. Guthrie, R. H. & Evans, S. G. 2004. Analysis of landslide frequencies and characteristics in a natural system, coastal British Columbia. Earth Surface Processes and Landforms, 29, 1321–1339. Henderson, I. H. C., Saintot, A. & Derron, M. H. 2006. Structural mapping of potential rockslide sites in the Storfjorden area, western Norway: the influence of bedrock geology on hazard analysis. Geological Survey of Norway Report. 2006.052. 86. Henderson, I. H. C., Venvik Ganerød, G. & Braathen, A. 2009. The relationship between particle characteristics and frictional strength in basal fault breccias: Implications for fault-rock evolution and rockslide susceptibility. Tectonophysics, 486, 132–149. Hicks, E. C., Bungum, H. & Lindholm, C. D. 2000. Seismic activity, inferred crustal stresses and seismotectonics in the Rana region, northern Norway. Quaternary Science Reviews, 19, 1423– 1436. Hocking, G. 1976. A method for distinguishing between single and double plane sliding of tetrahedral wedges. International Journal of Rock Mechanics and Mining Science, 13, 225 –226. Hoek, E. & Bray, J. W. 1981. Rock Slope Engineering. Institution of Mining and Metallurgy, London. Hungr, O., Evans, S. G. & Hazzard, J. 1999. Magnitude and frequency of rock falls and rock slides along the main transportation corridors of southwestern British Columbia. Canadian Geotechnical Journal, 36, 224–238. Irigaray, C., Ferna´ndez, T. & Chaco´n, J. 2003. Preliminary rock-slope-susceptibility assessment using GIS and the SMR classification. Natural Hazards, 30, 309– 324. Jaboyedoff, M., Baillifard, F. & Derron, M. H. 2003. Preliminary note on uplift rates gradient, seismic activity and possible implications for brittle tectonics and rockslide prone areas: the example of western Switzerland. Bulletin de la Socie´te´ Vaudoise des Sciences Naturelles, 88, 393– 412. Jaboyedoff, M., Baillifard, F., Derron, M. H., Coutoure, R., Locat, J. & Locat, P. 2005. Modular and evolving rockslope hazard assessment methods. In: Flaate, K. & Larsen, J. O. (eds) Landslides and Avalanches. ICFL Norway. Senneset. Jaboyedoff, M., Oppikofer, T., Derron, M.-H., Blikra, L. H., Bo¨hme, M. & Saintot, A. 2011. Complex landslide behaviour and structural control: a three˚ knes rockslide, dimensional conceptual model of A Norway. In: Jaboyedoff, M. (ed.) Slope Tectonics. Geological Society, London, Special Publications, 351, 147 –162. Kawabata, D., Bandibas, J. & Urai, M. 2004. Landslide susceptibility index determination using artificial neural networks. Eos, Transactions of the American

STRUCTURAL GEOLOGY RANKING Geophysical Union, 85, (Fall Meeting Supplement), Abstract H41B-0296. Mansor, S., Shariah, M. A., Billa, L., Setiawan, I. & Jabar, F. 2004. Spatial technology for natural risk management. Disaster Prevention and Management, 13, 364–373. Markland, J. T. 1972. A useful technique for estimating the stability of rock slopes when the rigid wedge sliding type of failure is expected. Imperial College Rock Mechanics Research Report 19. Nichol, S. L., Hungr, O. & Evans, S. G. 2002. Large-scale brittle and ductile toppling of rock slopes. Canadian Geotechnical Journal, 16, 773– 788. Olesen, O., Blikra, L. H. et al. 2004. Neotectonic deformation in Norway and its implications: a review. Norwegian Journal of Geology, 84, 3– 34. Olesen, O., Dehls, J. et al. 2000. Neotectonics in Norway, Final Report. Geological Survey of Norway Report 2000.002, 135. Osmundsen, P. T., Henderson, I. H. C., Lauknes, T. R., Larsen, Y., Redfield, T. F. & Dehls, J. 2009. Active normal fault control on landscape and rock-slope failure in northern Norway. Geology, 37, 135– 138. Redfield, T. F. & Osmundsen, P. T. 2009. The Tjellefonna fault system of Western Norway: Linking late-Caledonian extension, post-Caledonian normal faulting, and Tertiary rock column uplift with the landslide-generated tsunami event of 1756. Tectonophysics, 474, 106–123. Redfield, T. F., Torsvik, T. H., Andriessen, P. A. M. & Gabrielsen, R. H. 2004. Mesozoic and Cenozoic tectonics of the Møre Trøndelag Fault Complex, central Norway: constraints from new apatite fission track data. Physics and Chemistry of the Earth, 29, 673–682. Roering, J. J., Kirchner, J. W. & Dietrich, W. E. 2005. Characterizing structural and lithologic controls on

95

deep-seated landsliding: implications for topographic relief and landscape evolution in the Oregon Coast Range, USA. Geological Society of America Bulletin, 117, 654–668. Romana, M., Sero´n, J. B. & Montalar, E. 2003. SMR Geomechanics classification: Application, experience and validation. In: ISRM 2003 – Technology Roadmap for Rock Mechanics. South African Institute of Mining and Metallurgy, Johannesburg, 1– 4. Saintot, A., Henderson, I. H. C. & Derron, M. H. 2011. Inheritance of ductile and brittle structures in the development of large rock slope instabilities: examples from western Norway. In: Jaboyedoff, M. (ed.) Slope Tectonics. Geological Society, London, Special Publications, 351, 27– 78. Sauchyn, D. J., Cruden, D. M. & Hu, X. Q. 1998. Structural control of the morphometry of open rock basins, Kananaskis region, Canadian Rocky Mountains. Geomorphology, 22, 313–324. Schmidt, K. M. & Montgomery, D. R. 1995. Limits to relief. Science, 270, 617– 620. Varnes, D. J. 1978. In: Schuster, R. L. & Krizek, R. J. (eds) Landslides, Analysis and Control. Transportation Research Board Special Report 176. National Research Council, Washington, DC, 11–33. Ward, T. J., Li, R. M. & Simons, D. B. 1981. Use of a mathematical model for estimating potential landslide sites in steep forested drainage basins. In: Timothy, R. H. & Pearce, A. J. (eds) Erosion and Sediment Transport in Pacific Rim Steeplands Symposium. International Association of Hydrological Sciences Publications, 132, 21–41. Van Westen, C. J., Rengers, N. & Terlien, M. T. J. 1997. Prediction of the occurrence of slope instability phenomena through GIS-based hazard zonation. Geologische Rundschau, 86, 4004–4414.

Rock slope instabilities in Sogn and Fjordane County, Norway: a detailed structural and geomorphological analysis ¨ HME1,2*, ALINE SAINTOT1,2, IAIN H. C. HENDERSON1,2, MARTINA BO HELGE HENRIKSEN3 & REGINALD L. HERMANNS1,2 1

Geological Survey of Norway, 7491 Trondheim, Norway

2

International Centre for Geohazards (ICG), P.O. Box 3930, 0806 Oslo, Norway

3

Høgskulen i Sogn og Fjordane, Avdeling for Ingeniør og naturfag, P.O. Box 133, 6851 Sogndal, Norway *Corresponding author (e-mail: [email protected]) Abstract: More than 250 rock slope failures have occurred in Sogn and Fjordane County in historical times. So far, 28 sites are known where open cracks indicate that rock slope failures may occur in the future. Detailed structural and geomorphological analyses of these sites have been conducted, and form the basis for an evaluation and comparison of the unstable rock slopes. Four of these sites are described in detail herein. The main characteristics for rock slope instabilities in Sogn and Fjordane are: (1) a preferred location within relatively weak rock units, such as phyllites and weathered mafic gneisses; and (2) the development of most instabilities at convex slope breaks, which are evident as knick-points in the slope profile. Sogn and Fjordane is compared with other Norwegian regions, particularly Møre and Romsdal County, with respect to the spatial distribution of past and current rock slope instabilities. Sogn and Fjordane shows the greatest number of historical slope failures, whereas in Møre and Romsdal a larger amount of potential instabilities is observed. We propose that the larger amount of unstable rock slopes in Møre and Romsdal may be controlled by a locally high gradient of ongoing postglacial uplift and a higher rate of neotectonic activity.

During the last century Norway has suffered several natural disasters with large losses of life caused by rockslides and related tsunamis (Furseth 2006). Owing to the present situation of global climatic change, which leads to a warmer and wetter climate in Norway (Flatøy et al. 2008), an increase in the frequency of rockslides is likely. Therefore a better understanding of present-day rock slope instabilities will optimize monitoring and further research on the susceptibility of future rock slope failures in that region. Historical data and geological studies show that a high concentration of post-glacial gravitational slope failures as well as current instabilities can be found in two Norwegian counties: Møre and Romsdal, and Sogn and Fjordane, both situated in western Norway (Blikra et al. 2006). Our summary shows that Sogn and Fjordane is the county that is historically most affected by rock slope failures, and has the largest loss of life resulting from rock slope failures and related tsunamis (Fig. 1). Seven large historical events in Loen Lake are documented, all of them triggering displacement waves. Two of these lake tsunamis killed a total of 135 people in the years 1905 and 1936 (Bjerrum & Jørstad 1968; Grimstad & Nesdal 1990). In spite of this, a systematic registration

and detailed study of all ancient and current rock slope instabilities has not yet been carried out in Sogn and Fjordane. However, to predict rock slope failures and, therefore, prevent such extensive humanitarian disasters in the future, it is important to investigate the potential for rock slope failures in this county and to systematically study current rock slope instabilities. We present an overview of unstable rock slope areas detected so far in the county of Sogn and Fjordane, and summarize historical events (Fig. 2). To highlight that local geology is one of the major conditioning factors for slope instabilities, the individual characteristics of four selected sites are presented. The different geological features are discussed in order to define the controlling factors for rock slope instabilities in Sogn and Fjordane. Finally, these controlling factors are compared with those of rock slope instabilities in other Norwegian regions and discussed within the framework of international studies. We define rock slope failures as gravitational mass movements of rock masses down a mountain slope and rock slope instabilities as the sources of these failures. Both are classified according to Cruden & Varnes (1996), based on the type of observed or expected slope movement.

From: Jaboyedoff, M. (ed.) Slope Tectonics. Geological Society, London, Special Publications, 351, 97–111. DOI: 10.1144/SP351.5 0305-8719/11/$15.00 # The Geological Society of London 2011.

¨ HME ET AL. M. BO

98

1000

Number of casualties

Sogn & Fjordane Møre & Romsdal

100

Hordaland Oppland Rogaland Troms Nordland Sør-Trøndelag Vest Agder Buskerud Telemark Aust Agder Finnmark Vestfold Nord-Trøndelag

10

Østfold

1 1

10

100

1000

Number of rock-slope failures Fig. 1. Number of rock slope failures and related casualties for each affected Norwegian county, (Geological Survey of Norway (NGU) – Landslide database). Sogn and Fjordane County is historically most affected by rock slope failures and has the greatest loss of life caused by rock slope failures and related tsunamis.

Fig. 2. Overview of the study area including historic rock slope failures in Sogn and Fjordane (NGU – Landslide database), identified rock slope instabilities from the NGU’s database and localities described within this paper. For the numbers see Table 1. The inset shows the location of the study area within Norway and indicates the two discussed counties.

ROCK SLOPE INSTABILITIES IN NORWAY

Methodology All of the 28 localities were surveyed from a helicopter for evidence of active movement and for critical factors such as the presence of open fractures at the back of potentially moving blocks (Table 1). In a second step we carried out a structural analysis on the sites following the method suggested by Henderson & Saintot (2011). This included the investigation of: (1) dip and azimuth of the foliation with respect to the slope; (2) the presence of a weakened plane at the bottom of the block that could form a basal sliding plane; and (3) the presence of lateral boundaries of the block, such as transfer structures that could possibly act as structures detaching the block from the mountainside. Volumes that are given in this paper are approximations of the expected maximum volume. They have been calculated with an assumption about the shape of the rockslides based on bordering structures visible in the field.

Overview of rock slope instabilities in Sogn and Fjordane County The foundation of this study is a database with approximately 30 localities that have been identified from previous studies mainly based on the interpretation of aerial photographs or from field observations (Blikra et al. 2002; Bøe et al. 2002) (Fig. 2 and Table 1). These localities could represent a prominent risk for inhabitants and infrastructure, and have therefore been the subject for field analysis (Henderson et al. 2008). Most of the steep cliffs that were indicated as possible rock slope instabilities are indeed potential rockfall areas with small unstable volumes, but do not show all of the structures for large rock slope instability development (Table 1). Henderson & Saintot (2011) demonstrate that rockslides with the largest volumes, that is at least 3 Mm3, only develop when the full set of critical structures are present, that is, a valley-dipping foliation and a weakened plane at the base of the potentially unstable block, as well as existing lateral boundaries of the unstable block. Localities where several of these structures are developed form a high potential for future rock slope failure and, therefore, were the subject for subsequent fieldwork. This is the case for the four localities presented in detail in this study, namely Tussen, Viddalen, Strandanipa and Hyllestad (Fig. 2).

Tussen This potential rockslide area is located at the NE end of the inner fjord system of Sognefjorden (Fig. 2). It is situated 1100 m above the valley

99

floor on a NNE–SSW-trending slope that has an average gradient of 358, but exceeds 608 in the uppermost part where the instability is located at the edge of the plateau (Fig. 3a). In this region a Caledonian thrust boundary forms the contact between the phyllitic nappe, in which the instability is located, and the underlying Precambrian gneiss. However, a direct relationship between the presence of this sole detachment and the development of the rock slope instability is unlikely because the detachment is located relatively deep (.500 m) below the unstable area, as well as shallowly dipping (c. 108) and trending oblique to the valley. The bedrock of the unstable area consists of mica schists and phyllites, as well as some regions with mylonitic quartzrich gneisses and quartzites. The foliation dips approximately 308 to SSW and is, therefore, oblique to the valley (Fig. 3b). The whole area is characterized by two sets of regional structures. NNE –SSW-striking lineaments with a persistence of several hundreds of metres contribute to the major structural influence of the unstable area. They are most probably reactivated structures that have now developed large graben-like structures behind the unstable area (Fig. 3b, c). A set of less-developed SE– NW-striking joints act as transfer structures (Fig. 3b). The currently most apparent active block comprises only a part at the front of the whole instability and the volume of this block is estimated to be approximately 7 Mm3 (Fig. 3a, c). It is bounded by a 300 m-long graben-like structure to the NW, while the other three sides are free (Fig. 3c, d). Northeastwards and southwestwards are scarps of older slope failures with widespread talus at their base. The instability is thus a remaining part of the former slope. Several smaller fractures behind the main unstable block have been observed until approximately 250 m inside the plateau (Fig. 3a). Those fractures are up to 1.5 m open and up to several tens of metres deep, but die out laterally. Using these fractures to delimit a maximum volume and to estimate a worst-case scenario gives a volume of 10 Mm3. This unstable rock slope area is therefore one of the largest instabilities in Norway. Measured opening vectors, determined by matching definite pairs of edges on the opposite walls of a fracture, indicate movement directions perpendicular to the NNE–SSW-striking grabenlike structures (Fig. 3b). A major basal sliding plane could not be observed in the field, but thick scree deposits may cover its trace. However, a limited number of field data indicate a third set of joints, dipping approximately 308 to the ESE and hence towards the valley (Fig. 3b). These structures are clearly visible on both side-cliffs and may contribute to the formation of a basal sliding plane. A sliding plane at the base of the unstable block with

100

Table 1. Overview of the 28 investigated sites No.

Studied sites

Lithology

Type of expected movement*

Open fractures at the back of potentially moving blocks

Tussen

Phyllite

Rockslide

Yes

2

Viddalen

Phyllite

Yes

3

Hyllestad

Conglomerate

4

Strandanipa

Phyllite

5

Almenipa

6

Byttejuvet

Quartz monzonite (metamorphic) Gneiss (mafic)

Rockslide, rock topple Rockfall, rock topple Rock topple, rockfall Rockslide

7

Aurland-Fla˚m

8 9

Yes Yes Yes

Rock topple

Yes

Phyllite

Rockslide

Yes

Fosse Gjøringbøfjellet

Phyllite Gneiss (felsic)

No Yes

10

Gra˚berget

Gneiss (felsic)

Rockfall Rockfall, rock topple Rock topple

11

Hallandsberget

Gneiss (felsic)

Rockslide

Yes

12

Hjellane

Phyllite / Gneiss (felsic)

Rock topple, rockslide

Yes

Yes

Geomorph. model (Fig. 8)

Yes

b

Yes

c

Yes

a

Yes

a

Yes

b

One side defined Diffuse lateral boundaries

Yes

a

Yes

c

Not defined One side defined One side defined Not defined

Yes Yes

a a

Yes

a

Yes

c

Diffuse lateral boundaries

Yes

c

Weakened plane at the bottom of the block

Lateral boundaries of the unstable block

Oblique to valley, moderate dip Towards mountain, moderate dip No foliation

Not observed

Towards mountain, moderate dip Towards valley, steep dip

Not observed

All sides defined All sides defined All sides defined All sides defined All sides defined

Towards mountain, steep dip Gentle fold with axis azimuth perpendicular to slope strike and shallow dip towards valley (southern part) Unknown Towards mountain, moderate dip Towards mountain, moderate dip Varying, lower part towards valley Towards fjord, moderate dip

Not observed

Yes Not observed

Yes

Not observed

Not observed Not observed Not observed Yes Not observed

¨ HME ET AL. M. BO

1

Knickpoint

Dip and azimuth of the foliation with respect to the slope

Jølstravatnet

Gneiss (felsic)

Rockfall

No

Not observed

Not defined

No

–

Not observed

No

–

Not observed

One side defined Not defined

No

–

Yes

Not defined

Yes

b

Not observed

All sides defined Not defined

No

–

Yes

Towards valley, moderate dip Towards mountain, moderate dip Oblique to valley, vertical Towards valley, moderate dip Oblique to valley, moderate dip Unknown

14

Katlenova

Rockslide

Yes

15

Kloppuri

Quartz diorite (metamorphic) Gneiss (mafic)

Rockfall

No

16

La˚nefjorden

Gneiss (felsic)

Rockslide

Yes

17

Nepangerhaugane

Gneiss (felsic)

Yes

18

Nærøydalen

Gneiss (mafic)

19

Oppigardshyrna

20

Osmundsneset

Quartz monzonite (metamorphic) Gneiss (felsic)

Rockfall, rockslide Rockfall, rock topple Rock topple, rockslide

No

–

Yes

Towards mountain

Yes

One side defined

Yes

c

Rockslide

Yes

Not observed

b

Rockslide

Yes

Yes

b

Rockslide

Yes

Yes

c

Rockfall Rockslide

No Yes

Unknown Oblique to valley, steep dip

Not observed Not observed

Not defined Not defined

Yes Yes

a c

Terakamben Tirskardskreda

Quartz monzonite (metamorphic) Gneiss (mafic) Quartz monzonite (metamorphic) Gneiss (felsic) Gneiss (felsic)

All sides defined All sides defined One side defined

Yes

Phyllite

Oblique to valley, moderate dip Towards valley, moderate dip Towards valley, shallow dip

21

Ovrisdalen

22

Rustøyane

23 24

Steiggjeberget Stopelen

25 26

Rock topple Rockfall

Yes No

Not observed Not observed

Not defined Not defined

Yes No

a –

27

Vidme

Phyllite

Rockslide

Yes

Not observed

c

Vik (Midtfjellet and Hesten)

Phyllite

Rockfall, rockslide

No

All sides defined Not defined

Yes

28

Unknown Vertical, slope parallel Towards valley, moderate dip Subhorizontal

Yes

c

Not observed

Not observed Yes

Yes

ROCK SLOPE INSTABILITIES IN NORWAY

13

*According to Cruden & Varnes (1996).

101

102

¨ HME ET AL. M. BO

Fig. 3. Tussen rockslide area. (a) Sketch cross-section of the unstable area. Structures that dip 308 towards the valley display weakened planes that could develop into a basal sliding plane. (b) Structural field data (equal-area projection, lower hemisphere; in black: foliation poles and mean plane; in grey: fracture poles and mean planes, grey arrows: opening vectors, black, dashed: general slope trend). Grey crosses represent the joint set that may form a basal sliding plane. (c) Three-dimensional view with aerial photograph overlay (www.norgei3d.no) of the study area showing the spatial arrangement of structures. View to the SW. The NW border of the unstable block is a back-bounding graben-like structure and the other three sides are free. Dashed line A –B marks the location of the cross-section. (d) Unstable block, view to the NE.

a dip of 308 coincides with the furthest open fractures going inwards the plateau (Fig. 3a). Recent rockfall activity is detected along the cliff at both ends of the back-bounding graben-like structure, as well as at the front of the block where the proposed sliding plane may crop out. Both ends of the graben-like structure exhibit several smaller blocks or columns with average volumes of 90 m3 that are almost completely detached from the cliff.

Viddalen The unstable rock slope of Viddalen is located in the Aurland region at the SE end of the inner fjord

system of Sognefjorden (Fig. 2). It is situated 260 m above a water reservoir on a north–southtrending slope with an average gradient of 408 (Fig. 4a). The bedrocks are mica-rich schists and phyllites with a foliation that dips moderately SE towards the mountains (Fig. 4b). A flat-lying thrust plane, which forms the contact with subjacent Precambrian gneiss, is cropping out approximately 200 m below the unstable area (Fig. 4a). This site is especially important for detailed investigation because of the possible formation of a post-failure displacement wave in the reservoir and subsequent overtopping of the dam, similarly to the Vaiont slide in 1963 (Hendron & Patton 1987). The

ROCK SLOPE INSTABILITIES IN NORWAY

103

Fig. 4. Viddalen unstable area. (a) Sketch cross-section of the unstable area. (b) Structural field data (keys for stereonet as for Fig. 3). (c) Three-dimensional view of a shaded digital elevation model (DEM) of the unstable area with a resolution of 1 m, based on airborne LiDAR (Light Detection and Ranging) data. View to the SE. The dashed line indicates the extent of the unstable block on the surface and the dotted line A–B marks the location of the cross-section. (d) Overview of the area displaying a spring level that coincides with the sole detachment between the phyllitic and the gneissic unit. The dashed line marks the limits of the unstable block. View to the east.

104

¨ HME ET AL. M. BO

consequences for the local community in such a scenario would be catastrophic. A 150 m-long NNE –SSW-striking backbounding crevasse and a large NNW–SSE-striking transfer fracture limit the unstable block to the east and to the north, respectively; whereas only a ground depression indicates the southern border (Fig. 4c, d). In the field, the back-bounding crevasse shows a maximum down-throw of 13 m and an opening of up to 15 m in the northern part of the instability. The volume of the unstable block is uncertain because the depth of the unstable part is not known. On the surface it covers an area of 13 600 m2, but it is separated into several smaller blocks by three main fracture sets (Fig. 4b). One set, which includes the back-bounding crevasse, strikes NNE –SSW and is thus slightly oblique to the general slope trend. The second set strikes NW–SE, while the last set strikes west–east (Fig. 4b). Several fractures are open with a width of up to 1 m and a visible depth of maximum 10 m, but there are also some fractures that are only expressed as ground depressions on the surface. The unstable area is highly fractured at its front, as highlighted by the occurrence of many smaller detached blocks (Fig. 4d). Present-day rockfall activity is observed at the front and the toe of the unstable area. A spring is situated along the outcropping thrust plane and, consequently, indicates a layer with high permeability that drains the surface (Fig. 4a, d). This sole detachment may form the lower limit of the unstable area and may also have contributed to destabilize the slope. However, it can probably not act as a basal sliding plane because it is horizontal. Two theories exist about how the instability might

have developed. The first theory is based on the toppling of big rock columns that are separated by steep discontinuities. Flat lying discontinuities, which might be fractures parallel to the main thrust plane, constrain the column height. Another possible failure mode is a planar failure along a complex basal sliding plane that develops by connecting various pre-existing planes with unfavourable orientations by failure of rock bridges. Field observations suggest that the gravitational deformation is propagating towards the south. The back-bounding crevasse shows the largest opening, as well as the largest down-throw, in the northern part of the instability, whereas only a depression is visible as a continuation of it at the southern end of the unstable block (Fig. 4c). Furthermore, a higher rockfall activity at the front of the northern part indicates that the deformation rate seems to be higher in this area (Fig. 4d). However, this needs to be confirmed by monitoring data.

Hyllestad There are two unstable rock slope areas located on the northern coast of the peninsula of Hyllestad in the SW coastal area of Sogn and Fjordane (Fig. 2). Both lie only 1 km from each other and are situated directly above the sea on the edge of a subvertical, 440 m-high cliff that consists of massive Devonian conglomerates (Figs 5 & 6). There were two larger registered rockfalls in 1992 and 1998 with volumes of approximately 30 000 m3, one from the eastern unstable area that is described below, and another one further east along the same fjord at Katlenova, location 14 (Fig. 2 & Table 1). Each

Fig. 5. Hyllestad unstable area, western part. The overhanging unstable block is detached from the cliff by a subvertical, 120 m-long fracture that has an opening of up to 1 m. (a) Sketch cross-section of the unstable area. (b) View to the east. Dashed line A–B marks the location of the cross-section. (c) View to the west.

ROCK SLOPE INSTABILITIES IN NORWAY

105

Fig. 6. Hyllestad unstable area, eastern part. (a) Sketch cross-section of the unstable area. (b) Overview. Solid white line marks the border of the whole unstable area. Dashed lines indicate open fractures and depressions that separate the area into several blocks. Short black lines indicate the location of the two bolts. Dotted line A–B marks the location of the cross-section. View to the SE. (c) High rockfall activity is present at the front of the cliff. View to the south.

created tsunamis that reached a maximum run-up height of 6 m on the opposite side of the fjord and caused severe damage (Harbitz et al. 2001). The westernmost unstable area forms a massive block of approximately 0.17 Mm3 that is detached from the cliff by a 120 m-long and about 100 m-deep back-bounding fracture (Fig. 5). This subvertical extensional fracture has an opening of up to 1 m. The block is overhanging and shows almost no internal fracturing. Failure would occur when all rock bridges of this major crack have broken, resulting in a free fall. As the volume of this block is much larger than the previously observed events in this area, run-up heights of the triggered tsunami would significantly exceed previous ones. The eastern unstable area is divided into several blocks by a set of cliff-parallel fractures and several fractures perpendicular to the cliff (Fig. 6a, b). Fractures are subvertical with an opening of up to 1 m and visible depths of up to approximately 15 m or are only expressed as depressions in the topography. The lower limit of the unstable area cannot be accurately defined, but fractures on the cliff and a lowermost level of rockfall sources along the cliff suggests a depth of 60–100 m (Fig. 6c). All the blocks together have an approximate maximum volume of 90 000 m3, whereas the largest single block presents a maximum volume of 10 000 m3. The single blocks become smaller towards the rear of the instability. Intense rockfall activity is observed at the front of the cliff (Fig. 6c) and there are some smaller completely detached columns with volumes of up to 200 m3. Two bolts at one of the major cliff-parallel fractures indicate a movement of 2 cm within the last 4 years (Fig. 6b) (Henderson et al. 2008). It is likely that the failure of one of the frontal blocks triggers the

failures of the rear blocks by pressure release and induces a rapid succession of several rockfall events.

Strandanipa This unstable rock slope area is located in the central coastal area of Sogn and Fjordane (Fig. 2). It is situated on top of a 100 m-high cliff 620 m above the fjord on a WNW –ESE trending-slope that has an average gradient of 358 (Fig. 7a). The bedrock consists of mica schist and phyllite, and the foliation is in general dipping NE towards the mountains, but folds are locally observed. Widespread blocky rockfall deposits are present at the base of the cliff, some reaching the fjord. The unstable part of Strandanipa is characterized by an irregular pattern of steep fractures that separate the unstable volume into many small blocks (Fig. 7b). Several fractures are open with widths of up to 0.5 m and visible depths of 5 m, while others are only expressed as topographic depressions on the surface (Fig. 7b). Fractures that are visible on the top surface are not clearly noticeable on the cliff. A well-defined boundary at the bottom of the unstable block that could act as a basal sliding plane could not be observed in the field. Instead the bedrock shows irregular layering and fracturing increasing with depth (Fig. 7c). Below an approximately 10 m-thick upper massive layer the bedrock is highly fractured. Hence, the strength of the highly fractured bedrock is lower and, therefore, probably quite unstable. This suggests, similar to observation of sub-recent rockfall events, that this unstable area might only be affected by successive smaller rockfall or toppling events, and not by large slope failures that could reach the fjord and cause a tsunami.

106

¨ HME ET AL. M. BO

Fig. 7. Strandanipa unstable area. (a) Sketch cross-section of the unstable area. (b) Overview. The solid line marks the border of the unstable area, and the dashed lines indicate open fractures and depressions that separate the block into several parts. Dotted line A– B marks the location of the cross-section. View to the south. (c) Exposed cliff on which the unstable part is located. The bedrock is highly fractured and is therefore probably quite unstable below the uppermost approximately 10 m-thick massive layer. View to the NE.

Summary and discussion of controlling factors of rock slope instabilities in Sogn and Fjordane County Rock slope instabilities in Sogn and Fjordane County occur basically in six different rock types, which are: Devonian conglomerate, phyllites, mafic and felsic gneisses, metamorphic quartz monzonite as well as metamorphic quartz diorite. All these rock types have undergone a certain degree of metamorphism and most of them show a distinct foliation. Foliation is one of the most important discontinuity sets regarding the development of rock slope instabilities in western Norway (Saintot et al. 2011). The majority of the possible unstable rock slopes in Sogn and Fjordane (about 46%, Table 1) are situated within relatively weak rocks, mainly intrinsically weak rocks like phyllites, but also in rocks that were prone to an intense weathering like mafic gneisses and thus became weak over time. Among them, we can find the largest unstable volumes like Tussen or the Aurland-Fla˚m area

(previously described by Braathen et al. 2004; Blikra et al. 2006). Also, a large number of rock slope instabilities, about 36%, are within felsic gneisses. Only four localities, accounting for 14% of all instabilities observed so far, are situated within metamorphic quartz monzonite. The localities at Hyllestad and Katlenova form exceptions, as they developed within massive Devonian conglomerate and metamorphic quartz diorite, respectively. The presence of the four types of critical structures, including open fractures at the back of the potentially moving block, a valley-dipping foliation, a weakened plane at the base of the potentially unstable block, as well as existing lateral boundaries of the unstable block, was systematically checked for each of the investigated sites in Sogn and Fjordane County (Table 1). As a result it can be shown that in none of the localities all the critical structures are present. Most sites (79%) show a distinct open fracture at the back of the potentially moving blocks, and those that do not comprise potential rockfall areas with minor involved volumes (,5000 m3). Lateral boundaries, such as

ROCK SLOPE INSTABILITIES IN NORWAY

transfer structures, are only completely defined for 32% of the localities, whereas no lateral borders at all could be found for 39% of the localities. The latter comprise mainly rockfall areas. Structural studies of many high-priority sites, like Tussen or Viddalen, show that the foliation is dipping towards the mountain or oblique to the valley so that its activation as a basal sliding plane is not favoured. From the rock slope instabilities where the orientation of the foliation is known, 13 sites show a foliation that is favourable to the observed initial failing mechanisms. Of these sites eight have a foliation that dips towards the valley and thus might be reactivated as a basal sliding plane. Back-bounding fractures or transfer structures may develop along a vertical or steep foliation, which has been observed at five sites. The limit to the base of the instability and thus the vertical thickness remains undetermined for most sites (75%). It is assumed that this is mainly due to the limited accessibility of the very steep lower parts of the studied rock slope instabilities. A possible basal sliding plane is observed only at six sites, accounting for 25% of all observed instabilities. These are actually sites where the lower parts of the slope are accessible or where the structures are distinct enough to be visible from the helicopter. As movement rates are probably small, it is likely that erosion at the front masks deformations due to the slope instability and that the basal sliding planes are therefore not cropping out. Also, no distinct basal sliding plane might be visible on the face if a complex sliding surface develops by destroying rock bridges and stepping down along foliation or other discontinuities. In any case, for defining the shape of the lower limits of the unstable area, further geophysical investigations or drilling would be required. Lineaments on aerial photographs indicate a connection of some rock slope instabilities with pre-existing faults or regional structures. Agliardi et al. (2001) showed that the structural control of pre-existing brittle structures is one of the most important factor in the development of large rock slope instabilities. Pre-existing fractures may form the lateral limits of an unstable area and backbounding crevasses or transfer structures preferentially develop along them (Mahr 1977; Bachmann et al. 2004). For example at the Tussen rockslide area, the back-bounding graben-like structures developed along pre-existing regional lineaments. In Saintot et al. (2011) the influence of reactivated brittle structures on the slope stabilities in western Norway are described in more detail. A slope profile was drawn for each of the rock slope instability sites using elevation data with a resolution of 25 m. An analysis of the location of the instabilities on the corresponding profile shows

107

that 79% of the rock slope instabilities in Sogn and Fjordane County have developed along convex slope breaks, which appear as knick-points in the slope profile or at the top of a slope, that is, at the edges of plateaus (Table 1). Figure 8 illustrates all 22 slope profiles where the slope instability developed at a knick-point aligned at this knickpoint. It is so argued that these instabilities are part of the general processes of denudational slope development that try to even knick-points towards an equilibrium state of the slope (Ahnert 1987; Selby 1993, p. 369). Likewise, Holm et al. (2004) observed a concentration of rockfall along convex slope breaks that are interpreted to be of glacial origin.

Discussion and comparison with other areas in Norway As mentioned in the previous section, the slope instabilities in Sogn and Fjordane County are situated at knick-points in the slope profiles, as well as at the edges of high plateaus (Fig. 9a–c). This is similar to several unstable sites along the Romsdalen or Sunndalen valleys (Braathen et al. 2004; Blikra et al. 2006; Saintot et al. 2008). Counterexamples can be found in other highly susceptible ˚ knes rockslide areas of Norway, for example at the A (Braathen et al. 2004; Blikra et al. 2005; Ganerød et al. 2008; Jaboyedoff et al. 2011), where the slope instabilities are developed directly on steep slopes (Fig. 9d). Braathen et al. (2004) defined three geometrical models for rock slope instabilities in Norway based on pre-avalanche deformation patterns, which are: (1) rockfall areas: (2) rockslide areas: and (3) complex fields. Rockfall areas are located on subvertical slopes, and steep cliffparallel or -oblique extension fractures limit the unstable blocks. Rockslide areas are found on slopes with moderate dips (,468) and are characterized by slope-parallel basal sliding planes that develop along zones of weakness. Complex fields show an intricate deformation pattern with the development of a wide range of different structures. However, all complex fields have a low-angle basal sliding plane in common. These geometrical groups also show the typical location of the instability on the slope as we have discussed above. Rockfall areas and complex fields defined by Braathen et al. (2004) are located at the edge of a plateau and thus agree with our three geomorphological models (a), (b) and (c) in Figure 9, respectively, whereas rockslide areas are located directly on the slope and consequently agree with model (d) in Figure 9. Field investigations have been carried out in several areas of Norway and therefore allow a

¨ HME ET AL. M. BO Elevation

108

Location of knick-point and rock-slope instability

Distance

500 m

Fig. 8. All 22 slope profiles of rock slope instabilities in Sogn and Fjordane that developed at a knick-point. Profiles are based on a 25 m DEM. The location of current rock slope instabilities coincides with the location of knick-points. However, the intensity of the knick-points is varying. While most sites are situated at very clear knick-points, others show only small breaks in the slope. This results in a relatively broad distribution of slopes below and above the knick-point. Nevertheless, all slope profiles are located below the straight dashed reference line close to the knick-point.

back-bounding fracture

knick-point in slope profile

back-bounding fracture

knick-point in slope profile

basal sliding plane basal plane

(c)

(a) back-bounding fracture

knick-point in slope profile

back-bounding fracture

basal sliding plane

basal sliding plane

(b)

(d)

Fig. 9. General sketch of types of rock slope instabilities in Norway based on pre-failure geomorphology. Rock slope instabilities in Sogn and Fjordane are located at unstable edges of plateau-like surfaces (a and b), as well as at knick-points of slopes (c) and are not situated directly on a steep slope with constant slope angle (d). The expected types of movement are fall or topple for (a) and slide for (b) –(d).

ROCK SLOPE INSTABILITIES IN NORWAY

comparison of the spatial distribution of current rock slope instabilities in different regions. In the Storfjorden region, Møre and Romsdal County, for example, more than 40 localities display structures, which are critical for the development of large rock slope instabilities (Henderson & Saintot 2011). In several localities in the Storfjorden area that have been identified from aerial photographs all critical factors are present, such as: (1) distinct open fractures at the back of possible moving blocks; (2) foliation dipping towards the valley; (3) a potential basal sliding plane; and (4) lateral free boundaries of the block, such as well-developed transfer structures (Henderson et al. 2006; Henderson & Saintot 2011). In contrast, none of the localities in Sogn and Fjordane displays all the critical structures that are named above. Many of the localities that have been identified from aerial photographs eventually only showed some rockfall activity but no signs of large unstable rock slopes that may develop into rockslides. Furthermore, the majority of the recognized unstable rock slopes in Sogn and Fjordane County are situated within highly weathered and, hence, weak phyllites as well as mafic gneisses, whereas the rock slope instabilities in the Storfjorden region are mainly located within Precambrian gneiss. Henderson & Saintot (2011) argue that a fjord- or valley-dipping foliation orientation is one of the main influencing factors for the development of large rock slope instabilities (.3 Mm3) in the Storfjord region of Møre and Romsdal County. This cannot be demonstrated in the same order of magnitude for Sogn and Fjordane County, as discussed in the previous section. It is shown that there is a discrepancy in the amount of historical and current rock slope instabilities in Møre and Romsdal and in Sogn and Fjordane. Whereas the historical record indicates a much larger extent of rock slope failures in Sogn and Fjordane County, the number of potential current instabilities is greater in Møre and Romsdal. This could be due to the incompleteness of the historical and present database. However, the higher density of large current rock slope instabilities in Møre and Romsdal County may result from a locally high gradient of present-day post-glacial rebound combined with a higher tectonic activity in post-glacial times that has strongly disrupted the slopes (Olesen et al. 2000; Henderson & Saintot 2011). For example, the post-glacial Berill fault, up to now the only described neotectonically active fault in southern Norway, lies in this county (Anda et al. 2002). A large historic rockslide occurred in its vicinity and a complex rock slope instability currently develops in its hanging wall (Anda et al. 2002; Blikra et al. 2006). Finally, many branches of the large Møre– Trondelag Fault

109

Complex (MTFC) dissect the northern part of the Møre and Romsdal County and are still active. Redfield & Osmundsen (2009) argued that an earthquake that originated at one of the fault segments of the MTFC triggered the Tjelle rockslide in the eighteenth century. It has been discussed in many studies that active faults strongly influence the development and spatial distribution of gravitational slope instabilities (e.g. Hasegawa 1992; Hermanns et al. 2001; Sauro & Zampieri 2001). Ambrosi & Crosta (2006) showed for a study in the Central Italian Alps that major tectonic features control the strength of the surrounding rock mass and thus support the development of failure zones. However, regional uplift can be the cause for an increased relief and, therefore, may negatively affect the stability of slopes (Martino et al. 2004; Galadini 2006). In Sogn and Fjordane County, the gradient of present-day post-glacial uplift is lower and no active faults have been reported up to now (Olesen et al. 2000; Henderson & Saintot 2011). The differing amounts of current rock slope instabilities in the two counties may be controlled by differences in the effect of permafrost or climatic conditions. However, there is no significant difference in the permafrost-affected areas in both counties. Furthermore, climatic variations within each county are much larger than potential differences between the two counties and thus cannot explain the observed distribution of current rock slope instabilities.

Conclusions From these first results, the most important factor for the development of instabilities in Sogn and Fjordane County appears to be the rock type, as the majority of the current rock slope instabilities are situated within relatively weak rock units, such as phyllites and weathered mafic gneisses. Furthermore, a very important controlling factor for the location of rock slope instabilities in Sogn and Fjordane County is a particular slope profile with knick-points. Instabilities develop preferentially at convex slope breaks and are thus part of the general processes of denudational slope development that try to even out knick-points of a slope profile towards an equilibrium state of the slope. The larger number of potential current rock slope instabilities in Møre and Romsdal may be explained by a locally high gradient of present-day postglacial uplift and the greater tectonic activity that occurred in post-glacial times in this county. Therefore, the location of current rock slope instabilities in Sogn and Fjordane seems to have controlling factors other than those found in the Storfjorden region of Møre and Romsdal County.

110

¨ HME ET AL. M. BO

We are grateful to Sogn and Fjordane County for support, as well as to L.H. Blikra and M.-H. Derron for their helpful discussions during our work on this manuscript. Also, thanks go to the editor of this Special Publication, M. Jaboyedoff, for a constructive review and to the two anonymous persons for reviewing this paper. The work described in this paper is partially supported by the Research Council of Norway through the International Centre for Geohazards (ICG). Their support is gratefully acknowledged. This is ICG contribution No. 266.

References Agliardi, F., Crosta, G. B. & Zanchi, A. 2001. Structural constraints on deep-seated slope deformation kinematics. Engineering Geology, 59, 83–102. Ahnert, F. 1987. Process-response models of denudation at different spatial scales. Catena Supplement, 10, Cremlingen, 31– 50. Ambrosi, C. & Crosta, G. B. 2006. Large sackung along major tectonic features in the Central Italian Alps. Engineering Geology, 86, 183– 200. Anda, E., Blikra, L. H. & Braathen, A. 2002. The Berill fault – first evidence of neotectonic faulting in southern Norway. Norsk Geologisk Tidsskrift, 82, 175– 182. Bachmann, D., Bouissou, S. & Chemenda, A. 2004. Influence of weathering and pre-existing large scale fractures on gravitational slope failure: insights from 3-D physical modeling. Natural Hazards and Earth System Sciences, 4, 711– 717. Bjerrum, L. & Jørstad, F. A. 1968. Stability of Rock Slopes in Norway. Norges geotekniske institutt, Oslo, Publications, 79, 1– 11. Blikra, L. H., Braathen, A., Anda, E., Stalsberg, K. & Longva, O. 2002. Rock avalanches, gravitational bedrock fractures and neotectonic faults onshore northern west Norway: Examples, regional distribution and triggering mechanisms. Norges geologiske undersøkelse Report 2002.016. Blikra, L. H., Braathen, A., Derron, M.-H., Kveldsvik, V., Grøneng, G., Dalsegg, E. & Elvebakk, H. 2005. Rockslope failure at A˚kernesremna, Stranda, western Norway. Norges geologiske undersøkelse, Report (ISSN 0800–3416). Blikra, L. H., Longva, O., Braathen, A., Anda, E., Dehls, J. F. & Stalsberg, K. 2006. Rock slope failures in Norwegian fjord areas; examples, spatial distribution and temporal pattern. NATO Science Series, Series IV, Earth and Environmental Sciences, 49, 475– 496. Bøe, R., Lepland, A., Blikra, L. H., Longva, O. & Sønstegaard, E. 2002. Postglacial mass movements in western Norway with special emphasis on the 2000–2200 BP and 2800– 3200 BP periods – final report. Norges geologiske undersøkelse Report 2002.020. Braathen, A., Blikra, L. H., Berg, S. S. & Karlsen, F. 2004. Rock-slope failures in Norway; type, geometry, deformation mechanisms and stability. Norsk Geologisk Tidsskrift, 84, 67–88. Cruden, D. M. & Varnes, D. J. 1996. Landslide types and processes. In: Turner, A. K. & Shuster, R. L. (eds)

Landslides: Investigation and Mitigation. Special Report 247, Transportation Research Board, National Research Council, Washington, DC, 36– 75. Flatøy, F., Barstad, I. & Sorteberg, A. 2008. The GeoExtreme project, Module B: Description of the current climate and investigation of future climate scenarios. In: 33rd International Geological Congress, (33IGC), Oslo, Norway, 6– 14 August, 2008. Furseth, A. 2006. Skredulykker i Norge. Tun Forlag, Oslo [in Norwegian]. Galadini, F. 2006. Quaternary tectonics and large-scale gravitational deformations with evidence of rock-slide displacements in the Central Apennines (central Italy). Geomorphology, 82, 201– 228. Ganerød, G. V., Grøneng, G. et al. 2008. Geological ˚ knes rockslide, western Norway. Enginmodel of the A eering Geology, 102, 1– 18. Grimstad, E. & Nesdal, S. 1990. The Loen rockslides – A historical review. In: Barton, N. & Stephansson, O. (eds) Rock Joints, Proceedings of the International Symposium on Rock Joints, Loen, Norway. A.A. Balkema, Rotterdam, 3– 8. Harbitz, C. B., Domaas, U. & Varlid, E. 2001. Rock slide generated tsunamis – probability and hazard ˚ fjorden, Western Norway. In: Sharma, zoning in A V. M. & Saxena, K. R. (eds) Fjellsprengningsteknikk/Bergmekanikk/Geoteknikk, Oslo 2001. A.A. Balkema, Rotterdam, 27.1– 27.13. Hasegawa, S. 1992. Large-scale rock mass slides along the fault scarp of the Median Tectonic Line in northeastern Shikoku, southwest Japan. In: Bell, D. H. (ed.) Landslides: Proceedings of the 6th International Symposium on Landslides, Christchurch, New Zealand, Volume 1. A.A. Balkema, Rotterdam, 119–125. Henderson, I. H. C. & Saintot, A. 2011. Regional spatial variations in rockslide distribution from structural geology ranking: an example from Storfjorden, western Norway. In: Jaboyedoff, M. (ed.) Slope Tectonics. Geological Society, London, Special Publications, 351, 79– 95. Henderson, I. H. C., Saintot, A., Bo¨hme, M. & Henriksen, H. 2008. Kartlegging av mulig ustabile fjellpartier, Sogn og Fjordane. Norges geologiske undersøkelse Report 2008.026. Henderson, I. H. C., Saintot, A. & Derron, M.-H. 2006. Structural mapping of potential rockslide sites in the Storfjorden area, western Norway: the influence of bedrock geology on hazard analysis. Norges geologiske undersøkelse Report 2006.052. Hendron, A. J. Jr. & Patton, F. D. 1987. The Vaiont slide – a geotechnical analysis based on new geologic observations of the failure surface. In: Leonards, G. A. (ed.) Dam Failures. Engineering Geology, 24, 475–491. Hermanns, R., Niedermann, S., Villanueva Garcia, A., Gomez, J. S. & Strecker, M. R. 2001. Neotectonics and catastrophic failure of mountain fronts in the southern intra-Andean Puna Plateau, Argentina. Geology, 29, 619 –623. Holm, K., Bovis, M. & Jakob, M. 2004. The landslide response of alpine basins to post-Little Ice Age glacial thinning and retreat in southwestern British Columbia. Geomorphology, 57, 201–216.

ROCK SLOPE INSTABILITIES IN NORWAY Jaboyedoff, M., Oppikofer, T., Derron, M.-H., Blikra, L. H., Bo¨hme, M. & Saintot, A. 2011. Complex landslide behaviour and structural control: a three˚ knes rockslide, dimensional conceptual model of A Norway. In: Jaboyedoff, M. (ed.) Slope Tectonics. Geological Society, London, Special Publications, 351, 147– 162. Mahr, T. 1977. Deep-reaching gravitational deformations of high mountain slopes. Bulletin of the International Association of Engineering Geology, 16, 121–127. Martino, S., Moscatelli, M. & Mugnozza, G. S. 2004. Quaternary mass movements controlled by a structurally complex setting in the central Apennines (Italy). Engineering Geology, 72, 33–55. Olesen, O., Dehls, J. et al. 2000. Neotectonics in Norway – final report. Geological Survey of Norway – Report 2000.002. Redfield, T. F. & Osmundsen, P. T. 2009. The Tjellefonna fault system of Western Norway: Linking

111

late-Caledonian extension, post-Caledonian normal faulting, and Tertiary rock column uplift with the landslide-generated tsunami event of 1756. Tectonophysics, 474, 106– 123. Saintot, A., Bo¨hme, M., Redfield, T. & Dahle, H. 2008. Field studies of unstable slopes in Sunndalen Valley. Norges geologiske undersøkelse Report 2008.049. Saintot, A., Henderson, I. H. C. & Derron, M. H. 2011. Inheritance of ductile and brittle structures in the development of large rock slope instabilities: examples from western Norway. In: Jaboyedoff, M. (ed.) Slope Tectonics. Geological Society, London, Special Publications, 351, 27–78. Sauro, U. & Zampieri, D. 2001. Evidence of recent surface faulting and surface rupture in the fore-Alps of Veneto and Trentino (NE Italy). Geomorphology. 40, 169 –184. Selby, M. J. 1993. Hillslope Materials and Processes. Oxford University Press, New York.

Controlling factors for deep-seated gravitational slope deformation (DSGSD) in the Aosta Valley (NW Alps, Italy) G. MARTINOTTI1, D. GIORDAN2, M. GIARDINO1* & S. RATTO3 1

Dipartimento di Scienze della Terra, Universita` degli Studi di Torino, Torino, Italy

2

Consiglio Nazionale delle Ricerche, Istituto di Ricerca per la Protezione Idrogeologica, Torino, Italy

3

Regione Autonoma Valle d’Aosta, Dipartimento difesa del suolo e risorse idriche, Aosta, Italy *Corresponding author (e-mail: [email protected]) Abstract: Deep-seated gravitational slope deformation (DSGSD) is a common and widespread type of large slope instability in the Alps. In the Aosta Valley region in NW Italy, DSGSDs occupy at least 13.5% of the regional territory. In this study, regional distribution analyses have been coupled with local detailed geological and geomorphological surveys of individual phenomena to detect the controlling factors, deformation processes and evolution stages of DSGSD. Data and maps from field and remote-sensing investigations have been supported by drill data and geomechanical and hydrogeochemical analyses from project studies for hydroelectric plants and tunnels. Several phenomena related to DSGSD have been studied thoroughly: gravity-induced stresses, tectonic– metamorphic setting, morphostructural relations, glacial and periglacial morphodynamics, recent tectonic evolution, hydrogeological conditions and karst phenomena have been generically indicated as controlling factors. In the studied area three of the controlling factors were crucial in differentiating the form and evolution of DSGSDs: deep dissolution, surface tectonics, and tectonostructural setting. They are presented as possible end members of a classification scheme for DSGSDs.

Large portions of mountain relief along the Alpine chain are affected by deep-seated gravitational slope deformation (DSGSD) (Mortara & Sorzana 1987; Ambrosi & Crosta 2006). These huge gravitational phenomena are characterized by their large areal extension (multiple km2), and by distinctive complex geometrical, geomorphological and geomechanical settings (Crosta 1996). Over the past four decades DSGSD have been studied throughout the world (Jahn 1964; Ter-Stepanian 1966; Beck 1968; Nemcock 1972; Mahr 1977; Savage & Varnes 1987; Varnes et al. 1989; Chighira 1992; Sorriso Valvo 1995; Crosta 1996), but their original recognition and first detailed studies (Zischinsky 1966, 1969) were conducted in the European Alps where they are widespread. Historically, there have been two different kinds of approaches to the studies of DSGSD: one based on geomorphological and another based on structural data. By integrating these two approaches it is possible to analyse DSGSD geometry and form, interpret the controlling factors, and model their evolution. Recent papers on DSGSD by French researchers accomplished this comprehensive objective by considering both geomorphological (relief forms, weathering of slope material) and structural– tectonic (large fracture patterns, inherited faults) factors as influential in large slope

instabilities (Bachmann et al. 2004, 2006); they made a significant contribution to both two- and three-dimensional (2D and 3D) modelling of DSGSD (Chemenda et al. 2005; Bois et al. 2008; Bachmann et al. 2009). As stated above, activity of DSGSD implies different elements performing interactively as a complex system. To analyse and interpret a DSGSD, it is necessary to conduct a multidisciplinary research project (e.g. Agliardi et al. 2001) by collecting, organizing and processing a considerable amount of data on different features: † geomorphological data on landslide-related elements and other landforms; † stratigraphical and sedimentological data on surficial deposits; † lithological and geostructural data on bedrock units; † hydrological data on surface waters; † hydrogeological and geochemical data on groundwater; † geotechnical and other monitoring data on deformed materials and related deformational features; † pedological and vegetation data pertaining to land use; † historical data on past instability events.

From: Jaboyedoff, M. (ed.) Slope Tectonics. Geological Society, London, Special Publications, 351, 113– 131. DOI: 10.1144/SP351.6 0305-8719/11/$15.00 # The Geological Society of London 2011.

114

G. MARTINOTTI ET AL.

Different data sources mean different acquisition methods and, possibly, different data formulations. In remote-sensing studies, geometrical data are required for thematic mapping, and numerical data are required for geotechnical in situ analysis. From field surveys, geometrical data are obtained from geological and geomorphological sketch maps. From the microscopic analysis of rock-forming minerals, geometrical, but not mappable, data are derived. Such a complex multidisciplinary research project has to comply with the interpretation needs of different variables and overcome the difficulties of their different formulations. For this reason, all of these aspects have been investigated in the DSGSDs presented in this work, representing a selection of some of the most interesting Aosta Valley large-slope instability phenomena. The Aosta Valley has a great number of types and dimension DSGSD that makes it a great study field in the Western Alps. The Regione Autonoma Valle d’Aosta (‘Aosta Valley Autonomous Region’) is the smallest Italian region, with an area of 3262 km2; it is one of the main Alpine valley systems surrounded by the greatest peaks of the Pennine and Graie Alps. It covers the whole mountain zone of the Dora Baltea hydrographical basin. The elevations range from 400 m a.s.l. (metres above sea level) of the valley mouth into the Po Plain up to 4810 m a.s.l. of Mont Blanc, the highest peak of the European Alps. From a geological point of view, the Aosta Valley belongs to the Western Alps, the axial zone of the Alpine chain where it is recognizable as an imbricated stack of continental crust and oceanic units (Dal Piaz 1992). They are the result of the convergence phenomena between the European and Insubric (‘Adria’) palaeocontinent (Polino et al. 1990). The Alpine tectonometamorphic units of the Aosta Valley (Fig. 1) also show signs of postcollisional tectonic activity and neotectonic fault systems (e.g. Aosta-Ranzola fault: Balle`vre et al. 1986; Bistacchi et al. 2001). Similar to other Alpine regions, all of these geostructural and tectonic characters influence not only the Alpine relief evolution and uplift (Hunziker et al. 1992) but also the mountain slope dynamics (Giardino & Polino 1997). Of all of the morphodynamic agents that have influenced the regional geomorphology, the glaciers are the most important, especially by means of the Pleistocene glacial advance phases. The valley configuration still preserves, at various observation scales, the direct traces of glacial modelling (Carraro & Giardino 2004) both as erosional and as depositional landforms. Glaciation also influences indirectly the actual slope dynamics because of the pressure released by decaying glaciers (Panizza 1974).

The erosive activity of watercourses affected all of the Aosta Valley territory through the progressive deepening of glaciated side valleys, and the erosion of glacial landforms and deposits. Important fluvial deposits and landforms are present on the valley bottoms, as are imposing mixed alluvial and debrisflow fans, products of recurrent meteorological events (e.g. the last major one occurring on 13–16 October 2000: Ratto et al. 2003). Gravity has played a fundamental role in the geomorphic evolution of the Aosta Valley and has operated in tandem with a variety of other exogenic processes in shaping the landscape. Gravitational phenomena on mountain slopes can be differentiated by size and type: from simple shallow landslides (mainly soil slips) to large slope instability phenomena (mainly DSGSDs) (Fig. 2) (Giardino et al. 2004). The latter are widespread in the Aosta Valley, as indicated by the Inventory of Landslides in Italy (IFFI Project: Amanti et al. 2001; Colombo et al. 2005): they cover at least 13.5% (441 km2) of the valley (Ratto et al. 2007). Among the many DSGSD phenomena in the Aosta Valley, four representative case studies (Hoˆne, Villeneuve, Breuil– Cervinia and Quart) have been chosen to highlight possible end members of the different factors that control the structure, form and dynamic evolution of large slope instabilities. Figure 1 shows the locations of studied DSGSD in the Aosta Valley with respect to the main geological units in the region. The selection has been made based on geological and geomorphological characteristics considered useful to determine the timedependent and time-independent factors that controlled their formation and continuing movement. For each DSGSD we collected data and produced geotemathic maps by means of field and remote-sensing investigations. Interpretations have been supported by drill data, and by results from geomechanical and hydrogeochemical analysis carried out during the project of hydroelectric plants and tunnels in the Aosta Valley. The result of our studies have been used in this paper to develop a generic classification scheme for DSGSDs in the Aosta Valley: by condensing a large amount of information from the four selected case studies and by comparing to the results of the regional inventory of landslides (IFFI Project), we propose to use the controlling factors as an approach to DSGSD studies in general, regardless of their geographical location.

Hoˆne DSGSD: the role of pre-existing shear zones in the stability of large slopes The Hoˆne DSGSD is located on the right slope of Champorcher Valley (lower Aosta Valley; Fig. 3),

CONTROLLING FACTORS FOR DSGSD Fig. 1. Geological sketch map of the Aosta Valley (modified from Bonetto & Gianotti 1998) and the locations of the studied DSGSDs (1, Hone; 2, Villeneuve; 3, Cervinia; 4, Quart).

115

116

G. MARTINOTTI ET AL.

Fig. 2. DSGSD distribution in the Aosta Valley showing the location of the studied phenomena (1, Hone; 2, Villeneuve; 3, Cervinia; 4, Quart): (a) shows the percentages of landslide types in the Aosta Valley in regard to the total number of landslides; and (b) shows the percentage of landslide type area in regard to total landslide area (modified from Ratto et al. 2007). Only the 8% of phenomena mapped in the IFFI Project (Giardino et al. 2004) are classified as DSGSD, but they occupy more than 74% of the total landslide area (Ratto et al. 2007). The IFFI Project (Amanti et al. 2001) recognized 5218 landslide phenomena (Ratto et al. 2007), corresponding to 580 km2 (17.8% of the Aosta Valley territory).

at elevations of between 500 and 2200 m a.s.l. The area is characterized by erosional features of both glacial and fluvial origin. The middle and lower slopes are crossed by the deeply incised Ayasse Torrent. In the higher part of the area, landforms are mainly the result of shallow-slope instabilities related to gravitational and cryogenic processes. Micaschists and gneisses in the lower part of the valley have been assigned to the Sesia Lanzo complex (Compagnoni et al. 1977; Venturini 1995). Bedrock is fairly competent from a geomechanical point of view, the instability of the Hoˆne slope is conditioned by a moderately dipping, planar shear zone that is many kilometres long.

The Hoˆne DSGSD was documented first in 1996 during studies conducted for a hydroelectric plant. Detailed geological and geomorphological surveys and local geostructural analyses carried out by the authors outlined a large (23 km2) area of gravitational sliding characterized by fragmented rock overlying more stable bedrock (Giordan 2006; Martinotti & Giordan 2008). Structural geological surveys and drilling holes verified the characteristics and role of the planar shear zone in the control of the DSGSD dynamics and evolution. A geological cross-section of the south slope of the Champorcher Valley is shown in Figure 4. The major controlling shear plane dips 258NW: it separates a relatively unjointed gneiss and micaschists

CONTROLLING FACTORS FOR DSGSD

117

Fig. 3. Vertical view of the Hoˆne DSGSD (dashed pink line). In the upper-right part of the image the Ayasse Torrent confluences into the Dora Baltea River, close to Hoˆne village. The red line represents the Hoˆne hydroelectric power plant. Line A– A0 is the cross-section profile of Figure 4.

lower unit from a heavily fractured gneiss and micaschists upper unit, locally evolving to rock block slide deposits. The thickness of the shear zone ranges from a few metres to tens of metres. Microscale geostructural analyses conducted on oriented samples from rocks cropping out along the shear zone (Fig. 5) suggest that displacement along the shear zone evolved from a deeper crustal position, with ductile deformation, to a shallow one, involving brittle reactivations. Interpretation of petrographical data from thin sections showed the presence of both cataclasites and ultracataclasites. Cohesive breccias with silicified, epitodized and sericitized bedrock clasts represent the product of a hydrothermal phase in lower greenschist condition. They are associated with foliated cataclasites, overprinting older mylonite kinematic indicators. Discordant

pseudotachilites, associated with non-foliated cataclasites, represent a further shallower brittle deformational phase. Clay-rich fault gouges are the latest product of superficial deformation. The microscale observations have been used together with field evidences of the outcropping deformed bedrock to realize a 3D model of the shear plane (Giordan 2006) whose large extent and lateral continuity has been proved in the Hoˆne slope. Interactions between the complex distribution of geomorphological processes and landforms (first: valley glacial modelling; later: deep slope incision by running water) and the relatively homogeneous structural setting (constant dip of shear plane) indicate that the DSGSD involved local, but coherent, sliding in different areas of the Champorcher Valley (Fig. 5). The main shear plane emerges along adjacent tributary valleys separated by

118

G. MARTINOTTI ET AL.

Fig. 4. Schematic cross-section of the Hoˆne DSGSD showing the position of the reactivated shear plane.

slightly deformed ridges. The hanging wall is intensely fractured with open cracks near the surface and local water seepage on the slope. When conditions were favourable for initial slope instability, the superficial rock mass moved

down-slope along the shear plane. Later, on the higher part of the slope, the DSGSD evolved as a collapse: very coarse gravitational deposits formed, with a large number of huge towering blocks lying over a structural slope subparallel to

Fig. 5. Three-dimensional reconstruction of the relationships between the Hoˆne DSGSD shear plane and the topography; arrows indicate the outcrops of the shear plane mapped during the study; red shaded areas show the projection of the shear plane. A and B are outcrops where the samples for microstructural analysis were taken.

CONTROLLING FACTORS FOR DSGSD

the shear plane. Some drill holes penetrated these collapsed deposits, reaching loose and disjointed rocks at depth of tens of metres, below which are more stable rock masses (Martinotti & Giordan 2008).

Villeneuve DSGSD: the role of deep dissolution in slope deformation The DSGSD at Villeneuve affects a north–southelongated ridge on the south slope of the middle Aosta Valley, between the confluence of the

119

Cogne and Valsavaranche tributaries valleys. The DSGSD was identified in the late 1980s after geological surveys were conducted for the underground power station of a planned hydroelectric plant below the Mt Poignon ridge (Fig. 6). As shown in the geological map in Figure 7, the ridge is composed of micaschists and gneisses of the crystalline basement (Grand St Bernard unit), and calcshists, prasinites, marbles and carnioles of the Mesozoic Piemonte nappe system. The rocks were complexly folded during two isoclinal phases and an open folding event. The folds are crossed by younger shear planes (Fig. 8).

Fig. 6. Hillshade view of the Villeneuve ridge elongated north to the main Aosta Valley bottom, and separating the Cogne Valley (left ¼ east) and Valsavaranche (right ¼ west). The dotted red shape represents the borders of a collapse zone affecting the slope. The dashed yellow line is the projected hydroelectric penstock down to the Dora Baltea River; the solid red line is the alternative penstock, relocated and realized after DSGSD studies.

120

G. MARTINOTTI ET AL.

Fig. 7. Geological sketch map of the Villeneuve area showing the cross-sections of Figure 8.

CONTROLLING FACTORS FOR DSGSD

121

Fig. 8. Sequence of geological cross-sections of the Villeneuve area with the limits of the collapsed area.

The Villeneuve area has a variety of glacial landforms and deposits at different elevations. Tributary valleys have steep rock walls and have been deeply incised by post-glacial fluvial erosion. The slope affected by the DSGSD heads at Mt Poignon (1500 m a.s.l.) and extends down the valley side to 800 m a.s.l. Detailed geostructural and geomorphological field analysis and mapping identified evidence of an anomalous gravitational instability localized in the central part of the slope. The origin of this feature was established through deep drilling (500 m), which enabled the integration of surficial and underground data. The surficial data indicated the presence along the slope of a zone of disjointed and loose rocks, bordered by diffused rockslides and rockfalls. A depression (area c. 200 000 m2) is located in the central part of the ridge, into which glacial landforms and deposits have subsided (Fig. 9a). The zone is characterized by both active and inactive surficial deformation: open fractures, trenches, antislope scarps and closed depressions. The depression, covered by a dense conifer forest, is peculiar because of the absence of streams or springs. The surrounding slopes are much steeper and entirely stable because the rocks are not jointed (Fig. 9a). Surface data suggest the presence of a sinkhole, a phenomenon normally caused by the dissolution of underlying soluble rocks, such as evaporites. Anhydrite and gypsum, although not present at the surface of the depression, may be present at depth,

near the St Bernard–Piemonte nappe contact (Martinotti 1989). Because the underground power plant was to be built in a possible zone of dissolution, drilling was undertaken to investigate the phenomenon. Of particular importance were the results from the deepest (500 m) drill hole (Martinotti 1989). It penetrated, in sequence: landslide deposits, undisturbed– moderately jointed calcschists with many open fractures, folded marbles and calcschists with extremely bad geomechanical characteristics [rock quality designation (RQD) ¼ 0: Fig. 9c], anhydrite breccias, and joint-free Great St Bernard micaschists (RQD nearly 100). As shown in the cross-section in Figure 9b, deep solution disarticulated the rock mass above the soluble rocks, propagating deformation to the surface. Rock collapse has occurred below the surface depression. This interpretation has been confirmed by geological data provided by two highway tunnels in the lower part of the slope, and by nearby surveys and structural analysis. In summary, the Villeneuve DSGSD is mainly caused by dissolution of deep anhydrite. This phenomenon, first documented in the Western Alps at the end the nineteenth century, was first described as an active process by a Swiss geologist (Schardt 1905, 1906) while studying the Sempione railway tunnel. In the Villeneuve case, most of the dissolution probably occurred at the Last Glacial Maximum when the main valley glacier introduced abundant cold water under pressure to the valley bottom. The poor ion content of glacial water is an

122

G. MARTINOTTI ET AL.

Fig. 9. (a) Geological and geomorphological map of the central-lower part of the Villeneuve DSGSD; (b) schematic representation of the collapsed area and location of the 500 m-long drill hole; and (c) drill cores of very poor RQD, highly fractured marbles at the contact with anhydrite.

enhancing factor of water ‘aggressiveness’ toward soluble rocks (Alberto et al. 2008). As a first stage, gypsum and anhydrite rocks experienced massive dissolution and consequent volume reduction. Subsequently, by means of sulphate-ion-rich fluids, induced dissolution extended to marbles and calcschists (Alberto et al. 2007). This process is still active, although at a lower rate, as shown by the fresh appearance of surficial deformation and by the high sulphate content of the deep waters sampled in the highway tunnel mentioned earlier.

Breuil – Cervinia: litho-structural conditions on DSGSD in high mountain relief The Breuil –Cervinia DSGSD is located on the west side of the upper part of the Valtournanche, near the local ski resort. It is a large slope instability (8 km2) that has been subject to recent intensive urbanization and is crossed by a pipeline connecting Goillet Lake and the Perreres hydroelectric station (Fig. 10). In comparison to the other case studies presented in this paper, the Breuil –Cervinia DSGSD has the

best-preserved and freshest evidence of gravityinduced slope deformation. The deformation extends from high elevations, near Motte de Plete`Cime Bianche (2900 m a.s.l.), which were deglaciated at the end of the Little Ice Age, to Breuil– Cervinia (1900 m a.s.l.), deglaciated during Late Glacial time (Fig. 10). The deformed area therefore has been affected by different geomorphic processes, with respect to those described in the preceding case studies: here fluvial activity is of lesser importance than glacial activity. The high elevation, hummocky topography and abundance of rock outcrops allow reliable geological mapping of the surface geology. The evolution of the Breuil– Cervinia DSGSD is complex. Deformation is concentrated and more active in the lower and marginal portions of the slope than in higher areas. A 10-year record of deformation of a penstock for conveying water down to the Perreres hydroelectric plant on the SW portion of the DSGSD (Fig. 10 section B) shows mean movement rates of 2–3 cm year21, with displacement vectors parallel to the slope. The bedrock comprises units of the Piemonte Nappe System (Vannay & Alleman 1990): the

CONTROLLING FACTORS FOR DSGSD

123

Fig. 10. Panoramic view and geological model of the Cervinia DSGSD. Mont Fort unit is shown in black. Dashed line A is the hydroelectric tunnel and solid line B is the penstock; the circle is the DSGSD–tunnel intersection, where the most damaged sections of the tunnel are located.

Zermatt – Saas Unit (serpentinites, serpentineschists and metabasites, with minor metasediments) and the Combin Zone Unit (sensu stricto) (Fig. 1). The latter consists of two superposed sub-units (Marthaler 1984; Sartori 1987): the Tsate´ Unit (calcshists and prasinites) and the Mont Fort Unit (quartzites, dolomitic marbles, calcareous marbles, pseudocarnioles and, possibly, evaporites). The main schistosity in the high Valtournanche dips by about 108 to the west, which favours the activation of deep-seated slope deformation. The geomorphological and geological analysis indicates that only the Combin Zone units have been involved in the gravitational phenomena; the underlying Zermatt –Saas Unit has not been affected. The lateral variations in thickness, the presence nearby of pseudocarnioles (Alberto et al. 2005, 2008) and the poor strength of the Mont Fort Unit, which is bounded by two large shear zones, show that the deep-seated gravitational phenomena result from the progressive deformation of the Mont Fort Unit (Fig. 10). Either the pseudocarnioles or, possibly, minor evaporite masses along the main basal plane of the DSGSD probably acted as weak and soluble layers, playing an active role in the activation of DSGSD. The overlying rocks of the Tsate` Unit were consequently deformed and displaced by the reactivation of the pre-existing joint systems. Confirmation of this hypothesis came from a geological inspection of the damaged tunnel connecting the Goillet Dam to the Perreres hydroelectric pipe. The tunnel passes through the Zermatt– Saas unit, which is only slightly fractured, and into the Mont Fort Unit, which is highly fractured. Tunnel remedial work was concentrated at the tectonic contact between the two units, where

water seepage and ground deformation were encountered and the tunnel lining was damaged. During tunnel remedial work, clay-rich cataclasite and fault gouge were found along this contact, but not in outcrops outside the tunnel. These observations confirm the role that DSGSD played in deformation of the tunnel. Other indications of large-scale slope instability are visible at the foot of the Breuil– Cervinia DSGSD, where highly fractured and altered rock masses alternate with thick coarse debris units. Here walls of buildings and other structures have been progressively fractured and tilted in a manner consistent with DSGSD. This DSGSD is characterized, in addition to the typical slow and progressive movements, by several rapid episodes of movement localized at the base of the DSGSD. An example is the Motta di Plete` rockslide, with a surface area of about 0.3 km2.

The Quart DSGSD: the dynamic role of fault zones in the stability of large valley slopes The Quart DSGSD extends over several kilometres of the slope on the north side of the broad middle section of the Aosta Valley. The section of the valley owes its morphology, in part, to the Aosta – Ranzola fault zone, one of the more important neotectonic lineaments in the NW Alps that has been active from the Oligocene to the present (Giardino 1995; Bistacchi & Massironi 1999). The geomorphology of the Quart DSGSD (Fig. 11) is a product of several agents: glaciers, responsible for erosional and depositional landforms (roches moutonne´es, spillways, erratics,

124

G. MARTINOTTI ET AL.

Fig. 11. Evidence of the deep-seated deformations related to the Quart DSGSD are visible at the base of the slope in a hydroelectric tunnel and plant. Three-dimensional view of the Quart DSGSD and principal geomorphological and deformational elements: 1, glacial deposits, distributed at different elevation as a response to the deepening of the Aosta Valley by the Balteo Glacier; 2, main DSGSD down-slope scarp, the upper one being a reactivated tectonic discontinuity of the Aosta –Ranzola system; 3, main DSGSD counterslope scarps; 4, outcrops of bedrock thrusts over glacial deposits; 5, DSGSD displacement vectors; 6, the Vollein area; 7, hydroelectric penstock; 8, hydroelectric power plant. (a) Subvertical attitude of the tectonic discontinuity along the main DSGSD upper scarp. (b) Schematic representation of the kinematic relations between the master fault of the Aosta–Ranzola system (in the lower part of the image) and the Quart DSGSD.

tills), gravity, tectonic processes, and deep dissolution. These agents are responsible for nearsurface deformational features, including trenches, counterslope scarps and closed basins, many of which displace glacial landforms. Pre-Quaternary bedrock comprises rocks of the Piemonte Domain (calcshists with intercalations of mica marbles and metabasites) and the Austroalpine System (orthogneisses, fine-grained gneisses and subordinate amphibolite lenses) (Canepa et al. 1990). The rocks have undergone multiphase isoclinal ductile deformation (phases 1 and 2), overprinted by post-foliation folding (phase 3) with east– west horizontal fold axes. Ductile low-angle shear planes cut the calcschist units; some of these shears have been reactivated and locally involve Quaternary deposits. The reactivated shear planes are possibly phase 3 structures. Rock outcrops along the slope are cut by several high-angle fractures, commonly showing relative

displacements. The two principal groups of discontinuities – both subvertical, striking N58E – N258E and N808E–N1108E – show striations, furrows, rock and calcite steps, and calcite, quartz and other fibrous mineral growths. Kinematic analysis of these high-angle planes suggests the presence of fault surfaces with carbonate or silty in-fillings, showing later striae from mechanical effects. In the eastern area there is field evidence that the last displacement was left-lateral strike–slip. The highangle faults acted as headscarps for gravitational sliding along low-angle shear planes. The evolution of the Quart DSGSD has been controlled by several factors, whose importance and mutual interaction have changed through time. The principal factors are: † neotectonic activity of the Aosta –Ranzola shear zone (Giardino 1995); † deep dissolution of evaporites, confirmed by the high sulphate content of the deep waters sampled

CONTROLLING FACTORS FOR DSGSD

in the hydroelectric tunnel (Paolina et al. 1986; Alberto et al. 2008); † glacial and fluvial–glacial erosion, which overdeepened the valley and produced steep valley walls, leading to later failure during and after deglaciation; † gravity, which induced both deep-seated and shallow-slope movements. Faults associated with the left-lateral Aosta– Ranzola shear zone control the geometrical pattern and the kinematics of the Quart DSGSD. Although primarily a strike– slip structure, the Aosta – Ranzola master fault has a significant normal component responsible for recent uplift of the south slope of the Aosta Valley. The Quart DSGSD is actively deforming within this conjugate extensional system (Fig. 11b). Ancient valley floors at different elevation testify to the progressive deepening of the Aosta Valley resulting from the Balteo Glacier and the Dora Baltea watercourse during the Quaternary. The study of Quaternary landforms and deposits show that the DSGD was active at least from the Late Pleistocene. The Vollein necropolis, dated to between 3000 and 1550 years BC and located on the lower part of the DSGSD, was itself involved in the gravitational deformation (Giardino et al. 2000). But, from a geological point of view, this gravitational activity represents only a single episode of the long history of DSGSD of the Quart slope (Fig. 12). Deformation of modern buildings and other recent engineered structures, in addition to levelling and GPS measurements, confirm that the DSGDG is still active. Deformation of one important underground structure allowed an estimate to be made of present-day activity. Damage to the concrete lining of the Quart hydroelectric station at the base of the slope (600 m a.s.l.) has been monitored since the start of the operation in 1950. Remedial work has been supported by studies and investigations that have provided a detailed description of the underground phenomena and a quantification of the movements (Paolina et al. 1986; Fornero et al. 1988; Chiesa et al. 1991). Deformation has also been observed in the underground hydroelectric power plant located below the present-day valley floor and in the large feeder tunnel located at the base of the slope. Movement has also been recorded in a drill hole equipped with an inclinometer and extensometer that is located above the power station. The offsets in the concrete have extensional and lateral components, which is in contrast to the main movements parallel to the slope along tectonic features and also parallel to the deep-seated collapse or progressive settling caused by deep dissolution phenomena.

125

Precise topographical measurements were made between 1994 and 1998 at four sites located on the west side of the Vollein necropolis, between 650 and 940 m a.s.l. Although the periods of observation are brief, meaningful and coherent movements of 0.5 –0.8 cm have been found on the slope. Figure 12 summarizes the complex evolution of the Quart DSGSD. The present-day slope is the result of a complex interaction between glacial and fluvial processes and gravitational movements, operating over a long period of time. The DSGSD is characterized by an upper extensional zone and a lower compressional one. As the valley was deepened by erosion, displacements on the slope propagated to progressively lower levels, creating complex interference patterns with the local geological structures; former extensional sectors were later subject to compressional forces, and vice versa. At present, extensive portions of the lower slope are subject to gravitational activation along pre-existing tectonic or gravitational discontinuities. Continuing dissolution of gypsum and anhydrite beneath the slope may contribute to this instability.

Controlling factors and a classification of DSGSD In the Alpine geological literature, DSGSD is commonly defined on the basis of geomorphological features and surficial deformation (Mortara & Sorzana 1987; Varnes et al. 1989). Important DSGSD criteria include: † movement over a large area, commonly an entire slope (Agliardi et al. 2001), in some cases on both sides of a drainage divide; † small displacements in comparison to size of the mass, but occurring over thousands of years (Cruden & Hu 1993; Ballantyne 2002); † toe bulging (Crosta 1996); † the presence of large masses of fractured rock; † the lack of a drainage network or, in one is present, it is anomalous with strong structural control; † the presence of trenches, closed and linear depressions, scarps and counterscarps, and double ridges (Crosta 1996; Radbruch-Hall et al. 1997); † small discrete landslides within the area of the DSGSD (Agliardi et al. 2001; Tibaldi et al. 2004). In the past 10 –15 years great progress has been made in defining these geomorphological features and their evolution. These features provide a good starting point for DSGSD classification. Our studies in the Aosta Valley indicate that, in addition to classic geomorphological investigations,

126

G. MARTINOTTI ET AL.

Fig. 12. Idealized cross-sections of the Quart slope showing the long-term DSGSD evolution related to the river and glacial erosional activity. Glacial deposits of different age involved in the low-angle shear planes at different elevations confirm different stages of activity stages of the DSGSD. The ‘zip’ lines show the changing limit between the extensional and compressional zones of the unstable slope (diagrams modified from Giardino 1995).

a detailed geological and structural study of the associated basement rocks involved in the gravitational deformation is an essential element in understanding the phenomenon. Moreover, classifications

based solely on landforms are inadequate because different sectors of the same DSGSD can have different kinematic behaviour, spanning more than one geomorphically based classification field.

CONTROLLING FACTORS FOR DSGSD

Another limitation of the geomorphological approach is that DSGSDs are long-lasting phenomena and thus can be affected by a variety of changing phenomena; for example, large-scale climate change, variable tectonic activity, earthquakes and changes in slope morphology. The size of DSGSDs and their persistence in time force the researcher to thoroughly study bedrock, geological structure, slope geomorphology, surficial deposits, hydrogeology and changing land use. By using this approach, one can identify the main factors that control DSGSD evolution. Some attempts at classification have been based on predisposing and activating factors. This exercise has proven to be complicated and unfruitful; it thus has been set aside in preference for a different approach based on ‘controlling factors’, which condition and differentiate the evolution of DSGSD. † Lithotectonic aspects: for example, the involvement of foliated rocks such as phyllites or calcschists that can behave plastically (Zischinsky 1969); and the presence of shear planes, even those unrelated to recent tectonic evolution, that can be reactivated by gravity under suitable geomorphological conditions (Crosta & Zanchi 2000; Ambrosi & Crosta 2006). † Neotectonic history: recent tectonic activity can serve as an effective driver of large slope instability (Carraro et al. 1979; Hippolyte et al. 2006; Agliardi et al. 2009). † Relations between lithotectonic and geomorphological features and slope evolution (Hu¨rlimann et al. 2006). † Deep-seated dissolution, which is widespread and underappreciated in the Alps because outcrops of evaporites are rare (Schardt 1906). In all cases of DSGSD, where deep drilling data are available, evaporitic units have been intercepted at depth even though they do not crop out at the surface. Examples include Crodo and the Simplon Tunnel (Ossola Valley), Villeneuve (Aosta Valley region) and the Vizze Valley (Bolzano Province). Figure 13 shows a possible classification of the DSGSD in the Aosta Valley and their distinctive features based on the above-mentioned factors. Case (a) is a DSGSD that evolves based mainly on movement along pre-existing, mainly slopeparallel, shear planes or other planar tectonic elements. Case (b) involves plastic flow, as described by Zischinsky (1969), where the deformation of the slope is controlled by a fairly homogeneous rock mass under the influence of gravity. Specific lithotypes, such as phyllites, and the absence of a single shear plane are the distinctive features of this class of DSGSD. Case (c) is a

127

DSGSD involving deep dissolution and related surface collapse. Case (d) is similar to case (a), but is partly driven by neotectonic activity. Case (e) involves a combination of two or more of the above mechanisms. It is common in the Alpine environment owing to the complex interplay of structural and metamorphic evolution, neotectonic activity and geomorphological processes. It should be emphasized that DSGSD are longlasting phenomena, in which many controlling factors come into play at different times and with different effects on slope evolution. A single factor can, at different times, activate movement or, alternatively, predispose the slope to movement. Recognition that relations between different factors can change over time has led to the possibility of discriminating time-independent factors from time-dependent factors. The main timeindependent factors are: † lithology; † Alpine tectonometamorphic structure, including inactive deformational elements of ductile and brittle style; † gravity; The main time-dependent factors are as follows. † Relations between the lithotectonic framework, geomorphology and the slope evolution. Slope evolution can condition the potential energy of the slope. Valley forms can change, and it is possible that specific lithotectonic factors can switch from passive to active causing the slope to transition from stable to unstable (Tibaldi et al. 2004). For example, erosional deepening of the Ayasse stream has probably played an important active role in activating the Hoˆne’s tectonic shear zone. † On a temporal scale of thousands of years, as is the case with our DSGSD, recent tectonics can be considered a time-dependent variable. There are instances where changes in the local stress field have been caused by the creation of new structures or the reactivation of preexisting ones (Koons 1986; Bistacchi & Massironi 2000). † Glacial and fluvial processes can change with climate, affecting both surface drainage and the groundwater circulation, inducing or accelerating dissolution of evaporites or limestones at depth. The Quart DSGSD is a good example of how complex the evolution of a DSGSD can be. The Aosta – Ranzola shear zone plays an important role in slope deformation at Quart because the basement rocks along the shear zone are highly fractured. Secondary permeability induced by shearing allows surface and near-surface waters to reach evaporites

128

G. MARTINOTTI ET AL.

Fig. 13. Possible classification scheme for the DSGSD in the Aosta Valley based on their controlling factors: (a) DSGSD characterized by gravity-induced reactivation of pre-existing shear planes; (b) DSGSD with plastic flow, mainly in schists and calcschists bedrocks; (c) DSGSD with collapse and gravity-induced movements related to deep-seated dissolution phenomena; (d) DSGSD with gravity-induced movements mainly caused by the activity of neotectonic structures; and (e) DSGSDs controlled by mixed factors.

layers at depth. These waters have dissolved gypsum and anhydrite, activating upwards collapse that has extended to the surface. Deformation has also propagated downwards because of progressive erosional down-cutting of the Aosta Valley. The

evolution of the Quart DSGSD is therefore dependent on a series of factors: † the regional tectonic structure (Aosta –Ranzola shear zone) and its recent evolution;

CONTROLLING FACTORS FOR DSGSD

† the relation of the slope to exogenous factors such glacial erosion, deglaciation and fluvial deepening of the valley. All of these factors, combined in space and time, and with differing levels of importance, produced the present the Quart slope. DSGSDs commonly have a complex history that makes it difficult to evaluate separately each controlling factor. However, the data we have collected allow us to formulate an interpretative model for each of these factors. The model must take into account that relations among controlling factors may vary through time. The implication is that the representation of DSGSD can not be confined to a simple time, but is better represented as an evolutionary trend, during which the relative importance of the various controlling factors can vary.

Conclusion The paper presents results of studies on DSGSDs in the Aosta Valley, based on detailed geological and geomorphological surveys of four large slope instabilities and on comparison of data at a regional scale. Over 13.5% of the area of the Aosta Valley is affected by many types and sizes of DSGSDs, making the valley an ideal environment for studying this phenomenon in the Western Alps. Four case studies have been selected from the regional database on the basis of their representativeness in terms of size, geological and geomorphological features, state of activity, and related hazards and risks. Data on geomorphology and structural geology are presented and compared for the four selected DSGSDs (Hoˆne, Villeneuve, Breuil – Cervinia and Quart). Special attention was paid to tectonic, hydrogeological, geoengineeering and archeological features where relevant to the interpretation of individual DSGSDs. Data and maps from field and remote-sensing investigations are supported by drill hole data and geomechanical and hydrogeochemical analyses obtained from studies and remedial work carried out the for hydroelectric plants and tunnels. By analysing and comparing the entire dataset of each case study, we identified and interpreted controlling factors on DSGSD activity and evolution. We propose a fivefold classification scheme based on the controlling factors of the observed DSGSD: ‘Tectonic’, ‘Plastic flow’, ‘Dissolution’, ‘Neotectonic’ and ‘Mixed’. The five types of DSGSD effectively integrate the geomorphological research approach with the geological and structural ones. The widespread occurrence of similar features in other areas of the Alps suggests that the ‘controlling factor’ concept is useful for classifying DSGSD outside the Aosta Valley. This concept may help

129

in interpreting long-term, complex, large slope instability phenomena by enabling analysis of regional, time-independent factors controlled by structural geology and local time-dependent factors controlled by geomorphological and seismotectonic processes. Moreover, the approach using the controlling factors could also enhance applied studies by facilitating recognition of hazardous processes, interpretation of their causes and the design of targeted remedial measures.

References Agliardi, F., Crosta, G. & Zanchi, A. 2001. Structural constraints on deep-seated slope deformation kinematics. Engineering Geology, 59, 83– 102. Agliardi, F., Crosta, G., Zanchi, A. & Ravazzi, C. 2009. Onset and timing of deep-seated gravitational slope deformations in the eastern Alps, Italy. Geomorphology, 103, 119–129. Alberto, W., Carraro, F., Giardino, M. & Tiranti, D. 2005. Proposta di classificazione delle ‘pseudocarniole’ dell’alta Valle di Susa (Alpi Occidentali). Il Quaternario, 18, 187 –200. Alberto, W., Carraro, F., Giardino, M. & Tiranti, D. 2007. Genesis and evolution of ‘pseudocarniole’: preliminary data from the Susa Valley (Western Alps). In: Schreiber, B. C., Lugli, S. & Ba˛bel, M. (eds) Evaporites Throughout Time and Space. Geological Society, London, Special Pubblications, 285, 129– 142. Alberto, W., Giardino, M., Martinotti, G. & Tiranti, D. 2008. Geomorphological hazards related to deep dissolution phenomena in the Western Italian Alps: Distribution, assessment and interaction with human activities. Engineering Geology 99, 147– 159. Amanti, M., Bertolini, G. & Ramasco, M. 2001. The Italian landslides inventory – IFFI Project. In: Castaneda, M., Jorge, E. & Olarte Montero, J. (eds) Proceedings of III Panamerican Symposium on Landslides, Cartagena, Colombia, 29 July– 3 August 2001, Volume 2. Societad Colombiana de Geotecnica, Bogota`, 841– 846. Ambrosi, C. & Crosta, G. B. 2006. Large sacking along major tectonic features in the Central Italian Alps. Engineering Geology, 83, 183–200. Bachmann, D., Bouissou, S. & Chemenda, A. 2004. Influence of weathering and preexisting large scale fractures on gravitational slope failure: insights from 3-D physical modelling. NHESS, 4, 711– 717. Bachmann, D., Bouissou, S. & Chemenda, A. 2006. Influence of large scale topography on gravitational rock mass movements: New insights from physical modelling. Geophysical Research Letters, 33, L21406. Bachmann, D., Bouissou, S. & Chemenda, A. 2009. Analysis of massif fracturing during Deep-Seated Gravitational Slope Deformation by physical and numerical modelling. Geomorphology, 103, 130–135. Ballantyne, C. K. 2002. Paraglacial geomorphology. Quarterly Science Review, 21, 1935–2017.

130

G. MARTINOTTI ET AL.

Balle`vre, M., Kienast, J. R. & Vuichard, J. P. 1986. La nappe de la Dent Blanche (Alpes Occidentales): deux unit’es austroalpines independantes. Eclogae Geologicae Helvetiae, 79, 57– 74. Beck, A. C. 1968. Gravity faulting as a mechanism of topographic adjustment. New Zealand Journal of Geology & Geophysics, 11, 191– 199. Bistacchi, A. & Massironi, M. 1999. Brittle tectonics in the Pennine-Graian Alps. Memorie di Scienze Geologiche, 51–52, 488–492. Bistacchi, A., Dal Piaz, G. V., Massironi, M., Zattin, M. & Balestrieri, M. L. 2001. The Aosta–Ranzola extensional fault system and Oligocene Present evolution of the Austroalpine–Penninic wedge in the north-western Alps. International Journal of Earth Sciences (Geologisches Rundschau), 90, 654– 667. Bistacchi, A. & Massironi, M. 2000. Post-nappe brittle tectonics and kinematic evolution of the north-western Alps: an integrated approach. Tectonophysics, 327, 267– 292. Bois, T., Bouissou, S. & Guglielmi, Y. 2008. Influence of major inherited fault zones on gravitational slope deformation: a two-dimensional physical modeling of the La Clapire area (Southern French Alps). Earth and Planetary Science Letters, 272, 709 –719. Bonetto, F. & Gianotti, F. 1998. Guida al Giardino delle Rocce. Comune di Pollein, Regione Autonoma Valle d’Aosta. Canepa, M., Castelletto, M., Cesare, B., Martin, S. & Zaggia, L. 1990. The Australpine Mont Mary nappe (Italian Western Alps). Memorie di Scienze Geologiche, Universita` di Padova, 42, 1 –17. Carraro, F., Dramis, F. & Pieruccini, U. 1979. Large-scale landslides connected with neotectonic activity in the Alpine and Apennine ranges. In: Panizza, M., Carton, A. & Piacente, S. (eds) Proceedings of the 15th Plenary Meeting of the IGU Commission on Geomorphological Survey and Mapping, Modena, Catania, 7–15 September. Universita degli Studi, Istituto di Geologia, Modena, 213 –230. Carraro, F. & Giardino, M. 2004. Quaternary glaciations in the Western Italian Alps – a review. In: Ehlers, J. & Gibbard, P. L. (eds) Quaternary Glaciations – Extent and Chronology. Part 1. Developments in Quaternary Science, 2, 201–208 (CD-ROM included). Chemenda, A., Bouissou, S. & Bachmann, D. 2005. 3-D physical modeling of deep-seated landslides: new technique and first results. Journal of Geophysical Research, 110, F04004. Chiesa, S., Fornero, I., Frassoni, A., Mazza, G., Zanchi, A. & Zaninetti, A. 1991. Gravitational instability phenomena concernine a Hydrelettric plant in Italy. In: Proceeding of the 7th International Congress of International Society for Rock Mechanics. A.A. Balkema, Rotterdam, 6. Chighira, M. 1992. Long-term gravitational deformation at affiction Creek, southern Coast Mountains, southwest British Columbia. Geological Society of American Bulletin, 93, 804–812. Colombo, A., Lanteri, L., Ramasco, M. & Troisi, C. 2005. Systematic GIS-based landslide inventory as the for effective landslide-hazard management. Landslides, 2, 291–301.

Compagnoni, R., Dal Piaz, G. V., Hunziker, J. C., Gosso, G., Lombardo, B. & Williams, P. F. 1977. The Sesia-Lanzo Zone: a slice of continental crust with alpine high pressure–low temperature assemblages in the Western Italian Alps. Rendiconti dellaSocieta` Italiana di Mineralogia e Petrologia, 33, 281–334. Crosta, G. 1996. Landslide, spreading, deep seated gravitational deformation: analysis, examples, problems and proposal. Geografia Fisica e Dinamica Quaternaria, 19, 297–313. Crosta, G. B. & Zanchi, A. 2000. Deep-seated slope deformations. Huge, extraordinary, enigmatic phenomena. In: Bromhead, E., Dixon, N. & Ibsen, M. (eds) Proceedings of the 8th International Symposium Landslides, Cardiff: Landslides in Research, Theory and Practice. Thomas Telford, London, 351–358. Cruden, D. M. & Hu, X. Q. 1993. Exhaustion and steady-state models for predicting landslide hazards in the Canadian Rocky Mountains. Geomorphology, 8, 279–285. Dal Piaz, G. V. 1992. Le Alpi dal M. Bianco al Lago Maggiore. Societa` Geologica Italiana, Guide Geologiche Regionali, 3. Fornero, I., Frassoni, A., Giuseppetti, G. & Mazza`, G. 1988. The restoration of a Concrete Tunnel Lining: the role of field measurements in the safety and control of the structural stability. In: Proceedings of 1st International Conference on Computer Methods and Water Resources, Rabat, Morocco. Computer Aided Engineering in Water Resources, 4. Computational Mechanics Publications, Boston, 411–421. Giardino, M. 1995. Analisi di deformazioni superficiali: metodologie di ricerca ed esempi di studio nella media Valle d’Aosta. PhD thesis, Turin University. Giardino, M., Giordan, D. & Ambrogio, S. 2004. G.I.S. technologies for data collection, management and visualization of large slope instabilities: two applications in the Western Italian Alps. Natural Hazards and Earth System Sciences 4, 197 –211. Giardino, M., Martinotti, G. & Mezzena, F. 2000. Dinamica ambientale ed evoluzione del sito archeologico di Vollein, Media Valle d’Aosta. In: Lollino, G. (ed.) Proceedings of GeoBen 2000: ‘Condizionamenti Geologici e Geotecnica nella Conservazione del Patrimonio Storico Culturale’. Gruppo Nazionale Difesa Catastrofi Idrogeologiche, 2133, 527–537. Giardino, M. & Polino, R. 1997. Le deformazioni di versante dell’alta Valle di Susa: risposta pellicolare dell’evoluzione tettonica recente. Il Quaternario, 10, 293–298. Giordan, D. 2006. Lo studio dei fenomeni franosi: dai sistemi informativi territoriali al monitoraggio in tempo reale. PhD thesis, Turin University. Hippolyte, J.-C., Tardy, M. & Nicoud, G. 2006. Les failles re´centes des Grands-Moulins (Savoie): un Sackung (tassement gravitaire) majeur dans les Alpes franc¸aises. Comptes Rendus Geoscience, 338, 734– 741. Hunziker, J. C., Desmons, J. & Hurford, A. J. 1992. Thirty-two years of geocronological work in the Central and Western Alps: a review on seven maps. Me´mories de Ge´ologie (Lausanne), 13, 1– 52.

CONTROLLING FACTORS FOR DSGSD Hu¨rlimann, M., Ledesma, A., Corominas, J. & Prat, P. C. 2006. The deep-seated slope deformation at Encampadana, Andorra: Representation of morphologic features by numerical modelling. Engineering Geology, 83, 343– 357. Jahn, A. 1964. Slopes morphological features resulting from gravitation. Zeitschrift fu¨r Geomorphologie, Supplement 5, 59–72. Koons, P. 1986. Relative geobarometry from highpressure rocks of quartzo-feldspathic compositions from the Sesia zone: western Alps. Contributions to Mineralogy and Petrology, 93, 322– 344. Mahr, T. 1977. Deep-reaching gravitational deformation of high mountain slopes. Bulletin of Engineering Geology and the Environment, 16, 121–127. Marthaler, M. 1984. Geologie des unite´s penniques entre le val d’Annivers et le val de Tourtemagne (Valais, Suisse). Eclogae Geologicae Helvetiae, 77, 395–448. Martinotti, G. 1989. Studio geologico per l’impianto idroelettrico di Villeneuve. Unpublished report for ENEL DCO. Martinotti, G. & Giordan, D. 2008. Studio geologico correlato alla progettazione per il rifacimento dell’impianto Idroelettrico di HONE II – sito nella Valle di Champorcher (AO). Unpublished report for CVA (Compagnia Valdostana Acque). Mortara, G. & Sorzana, P. F. 1987. Fenomeni di deformazione gravitativa profonda nell’arco alpino occidentale italiano. Considerazioni litostrutturali e morfologiche. Bollettino de Societa` Geologica Italiana, 106, 303–314. Nemcock, A. 1972. Gravitational slope deformation in high mountains. In: Gill, J. E. (ed.) Proceedings of the 24th International Geological Congress, Montreal, Harpell’s Press Co-operative, Quebec, 13, 132–141. Panizza, M. 1974. Fenomeni franosi connessi ad azioni di glaciopressione. Bollettino del Comitato Glaciologico Italiano, Serie 2, 22, 45–48. Paolina, R., Fauro, R., Fornero, I., Fanelli, M., Giuseppetti, G. & Mazza, G. 1986. La Galleria di Quart. Memorie del dipartimento di Ingegneria Strutturale, Politecnico di Torino, MIR86, 16, 1 –12. Polino, R., Dal Piaz, G. V. & Gosso, G. 1990. Tectonic erosion at the Adria margin and accretionary processes for the Cretaceous orogeny in the Alps. Memorie della Societa` Geologica France, 156, 345–367. Radbruch-Hall, D. H., Varnes, D. J. & Colton, R. B. 1977. Gravitational spreading of steep-sided ridges (‘Sackung’) in Colorado. Journal of the Research of the US Geological Survey, 5, 359– 363.

131

Ratto, S., Bonetto, F. & Comoglio, C. 2003. The October 2000 flood in Valle d’Aosta (Italy): event description and land planning measures for the risk Mitigation. JRBM (International Journal of River Basin Management), 2, 105–116. Ratto, S., Giardino, M., Giordan, D., Alberto, W. & Armand, M. 2007. Carta dei fenomeni franosi della valle d’Aosta, 1:100 000 in scale. Tipografia valdostana, Aosta. Sartori, M. 1987. Structure de la Zone du Combin entre le Diablons et Zermatt. Eclogae Geologicae Helvetiae 80, 789 –814. Savage, W. Z. & Varnes, D. J. 1987. Mechanics of gravitational spreading of steep-sides ridges (sackung). Bulletin of the International Association of Engineering Geology, 35, 31– 36. Schardt, H. 1905. les re`sultats scientifiques du percement du tunnel du Simplon. Bulletin technique de la Suisse romande, 10, 35–41. Schardt, H. 1906. Note sur la valeur souterraine de l’e´rosion par l’action de sources. Bulletin de la Socie´te´ Belge de ge´ologie, 20, 86– 94. Sorriso Valvo, M. 1995. Considerazioni sul limite tra deformazione gravitativa profonda di versante e frana. Memorie della Societa` Geologica Italiana, 50, 179– 185. Ter-Stepanian, G. 1966. Type of depth creep of slopes in rock masses. Problems in Geomechanics, 3, 49–69. Tibaldi, A., Rovida, A. & Corazzato, C. 2004. A giant deep-seated slope deformation in the Italian Alps studied by paleoseismological and morphometric techniques. Geomorphology, 58, 27–47. Vannay, J. C. & Alleman, R. 1990. La zone pie´montaise dans le Haut–Valtournanche (Val d’Aoste, Italie). Eclogae Geologicae Helvetiae, 83, 21–39. Varnes, D. J., Radbruch-Hall, D. & Savage, W. Z. 1989. Topographic and Structural Conditions in Areas of Gravitational Spreading of Ridges in the Western United States. US Geological Survey, Professional Papers, 1496. Venturini, G. 1995. Geology, Geochemistry and Geochronology of the Inner Central Sesia Zone (Western Alps – Italy). Me´moires de Ge´ologie (Lausanne), 25. Zischinsky, U. 1966. On the deformation of high slopes. In: Proceedings of the First International Conference of the International Society of Rock Mechanics, Lisbon, Volume 2. International Society of Rock Mechanics, Lisbon, 179–185. Zischinsky, U. 1969. Uber Sackungen. Rock Mechanics, 1, 30–52.

Palaeostress analysis of a giant Holocene rockslide near Boaco and Santa Lucia (Nicaragua, Central America) ´ 3, ROMAN NOVOTNY ˘ 1,2*, MARKE´TA KERNSTOCKOVA ´ 4, IVO BARON 4 5 4 ´ ˘ ´ DAVID BURIANEK , PETR HRADECKY , PAVEL HAVLICEK & ROSTISLAV MELICHAR3 1

Czech Geological Survey, Leitnerova 22, 658 69 Brno, Czech Republic

2

Present address: Geologische Bundesanstalt, Neulingsgasse 38, 1030 Wien, Austria 3

Department of Geological Sciences, Faculty of Science, Masaryk University, Kotlarska 2, 612 00 Brno, Czech Republic 4

Czech Geological Survey, Leitnerova 22, 658 69 Brno, Czech Republic 5

Czech Geological Survey, Kla´rov 3, 658 69 Praha, Czech Republic *Corresponding author (e-mail: [email protected])

Abstract: A giant rockslide occurred on the southern side of an Upper Tertiary shield volcano in central Nicaragua in the Holocene. The failure caused tectonic-like deformation of rock masses and changed the local stress regime. The lower, compressional part of the rockslide produced a stress field with the axis of maximum stress (s1) parallel to the displacement vector of the main body. The upper part of the rockslide was gravity-driven with s1 vertical and s3 horizontal, and oriented SE– NW. The mass tended to move SE. In the crown, the stress field had a subvertical s1 steeply dipping towards the west. Data at the base of the Santa Lucia Depression, where eastand west-dipping reverse and thrust faults developed, showed that the compressional stress, s1, was nearly horizontal and east– west oriented, the horizontal s2 was north–south oriented, and the s3 was subvertical. These compressional conditions resulted from the collapse of the crown after the main slope failure phase. Simultaneously, along with the gravity relaxation of the main displaced mass, the slopes and mountain slopes along the main scarp depression underwent deep-seated sliding, sagging and flowing.

A giant rockslide occurred on the southern side of an ancient volcano near Boaco (central Nicaragua) in the Holocene. The rockslide was recognized during geological mapping in the area of Boaco and Santa Lucia (Hradecky´ et al. 2007). We used the term ‘giant’ to accentuate its extraordinary dimensions, which are in excess of 10 km3. Such slope failures are called ‘large’ if their volume is more than 1 km3 (Castelli et al. 2009) or classified as ‘extremely large’ (.5 km3) from a landslidemagnitude point of view (Fell 1994). Gravitational failures of stratovolcano edifices are quite common and they tend to be dramatically larger than non-volcanic terrestrial landslides (Reid et al. 2000, Siebert 2002) because the inherent instability of volcanoes is attributable to a large number of factors, for example, earthquakes, elevated fluid pressures, presence of unconsolidated pyroclastic rocks, hydrothermal alteration and dike or magma emplacement, acting individually or complexly (Siebert 2002). Thus, large volcanic landslides represent a major hazard in Central America. The rockslide near Boaco (nomenclature after Cruden & Varnes 1996) caused tectonic-like

deformation of rocks, produced or reactivated faults and changed the local stress field. Gomberg et al. (1995) noted the similarity between faults produced by large slides and those produced by regional tectonic processes. Some authors mentioned regional tectonic stress field as the principal controlling factor for the evolution and spatial distribution of landslides. Tectonic stresses, in particular, play an indirect role in landslide control through their direct impact on patterns of joints (Gupta 2005) and through their major role in shaping the mountain topography (e.g. Ollier 1981; Embleton 1987). The spatial orientation of valleys and drainage networks as a geomorphic factor in controlling landslide distribution is also closely related to tectonic joints (Scheidegger 1980; Abrahams & Flint 1983; Pohn 1983). However, the direct effect of regional tectonic stresses on near-surface stresses, changed by local topography, was given theoretically by Savage & Swolfs (1986). Contrary to the above-mentioned approaches, this paper describes how a giant rockslide can modify the local palaeostress field.

From: Jaboyedoff, M. (ed.) Slope Tectonics. Geological Society, London, Special Publications, 351, 133– 145. DOI: 10.1144/SP351.7 0305-8719/11/$15.00 # The Geological Society of London 2011.

134

I. BARON ET AL.

Geological setting of the giant slope failure The study area is located in mountainous terrain of central Nicaragua, northern Central America (Figs 1 & 2), which is a southern part of the Chortis block of the Caribbean plate. The Chortis block is situated in the northern part of Central America (Honduras, Nicaragua, El Salvador, Guatemala); it consists of Grenville –Palaeozoic and Mezosoic basement (phyllite, gneiss and orthogneiss metamorphoses in greenschist –amphibolite facies) partially modified by Tertiary metamorphism and intrusions (Rogers et al. 2007). The southern part of the Chortis block is almost completely covered by Miocene –Pliocene pyroclastic and effusive strata. Large accumulations of mainly Neogene volcanic rocks occur in the Nicaraguan Highlands region around Santa Lucia and Boaco (Weyl 1980). The volcanic material was produced at several volcanic centres. We anticipate that at least seven large volcanoes formed in the study area during the period of main volcanic activity; some of them have since been destroyed by erosion or buried by younger ones, while others are still quite

well preserved, such as the shield volcano complex of Santa Lucia its 4 km-wide caldera in the centre. The approximately 15 km-wide predominantly intermediate –basic volcanic structure of Santa Lucia is characterized by a large volume of pyroclastic rocks and andesitic lava flows that erupted probably during the Neogene. Recently the following five lithological units of volcanic rocks have been recognized in the Santa Lucia volcanic complex. † Acid ignimbrites, accompanied by surges, fine falls and debris flows (in lower position) are grey– brown. Some ignimbrite layers, several metres thick, have been altered (argilitization) and locally strongly affected by brittle deformation. † ‘Lower andesites’ are deeply altered sets of individual lava flows and clinkers mainly of basaltic andesite composition, alternating with layers of agglomerates. They are porous, which permits strong alteration and weathering processes. Chemical weathering (e.g. zeolitization, carbonatization, argilitization) alters primary unstable minerals and glass into a mixture of limonite,

Fig. 1. Digital terrain model (DTM) of Nicaragua, with the location of the study area near Boaco indicated. Co-ordinates for town of Boaco are: 128280 1500 N and 858390 3000 W (source of data: SRTM – USGS/NASA).

PALAEOSTRESS OF A GIANT ROCKSLIDE

135

Fig. 2. DTM of the entire rockslide and adjacent area with definition of distinct regions: I.A, zone of compression in the main body; I.B, block part of the slope failure, recently partly in tension; II.A, trench upslope of the main body, partly filled by subsequent deep-seated rockslides; II.B, relict of the original caldera of Santa Lucia; III.A and III.B, crown of the slope failure, now separated into blocks from subsequent very deep-seated rockslides; IV.A, eastern palaeo-dam, IV.B, SE palaeo-dam. The large arrow indicates the displacement vector of the giant rockslide; smaller arrows indicate subsequent deep-seated rockslide displacement vectors, which activated from unloading of the main rockslide crown. The numbers indicate the site location of palaeostress analysis.

chlorites and clay minerals, and makes the rock weak and sensitive to deformation. † The third unit represents mainly deposits of agglomerate pyroclastic flows, recorded by Garayar (1972) as agglomerates of the Coyol group. Total thickness reaches up to 250 m. † Upper andesitic basalts and olivine basalts predominate on plateaus of northern surroundings of the area. Total thickness reaches up to 70 m. Lateritic weathering is well and deeply developed. Younger basaltic flows and dykes are common throughout the area. † Coarse agglomerates dominate in the uppermost position along the rims of the caldera of Santa Lucia. They form imposing coulisses around the caldera. The base of the individual flows displays fine matrix, and reverse grading is common. This epiclastic unit is the youngest one in the area.

Methods The main aim of the geological project was a geohazard susceptibility assessment of the area of Santa

Lucia and the Boaco towns, an area of about 500 km2, which is a much wider area than that of the rockslide itself. When we came across the giant rockslide, we tried to survey it in the limited time and spectrum of methods available as complexly as possible. The main tools of the geomorphic analysis were: † satellite image interpretation: Landsat 7 and Aster; † stereoscopic interpretation of aerial photographs at a scale 1:25 000 and 1:40 000; † GIS-based analyses of a digital terrain model (DTM) of SRTM (Shuttle Radar Topography Mission) with resolution of 90 and 30 m; † field reconnaissance, geomorphic and geological, and engineering –geological mapping; † sedimentological logging, structural–geological measurement, and so on. Geological mapping was provided at scale of 1:50 000 and it was complemented with petrological analyses in the laboratory. For the purpose of dating, several trenches and probes were dug in order to take organic samples related to geodynamic

136

I. BARON ET AL.

events. The organic samples were later dated using a conventional 14C method at the Kyiv Radiocarbon Laboratory (Dr M. Kovalyukh, Ukraine) or by accelerator mass spectrometry (AMS) 14C method in the laboratories of Poznan University (Dr T. Goslar, Poland). The main goal of the palaeostress analysis was to analyse the change in local stress field caused by the rockslide. Faults with slickensides and striations, and with a known sense of displacement, were measured in natural outcrops, road-cuts and other artificial excavations along the rockslide body. For the analysis the data of total 146 faults were grouped to represent different typical parts of the rockslide. The palaeostress analysis was based on the multiple inverse method (Kernstockova 2005, Yamaji et al. 2006), which allows the process fault-slip data to be expressed as 9D vectors (Melichar & Kernstockova, 2008) and numerically separated into different deformational phases. Data were processed using computer program MARK (Kernstockova 2005). The software enables polyphase data analysis to be performed using a new numerical method presented by Kernstockova & Melichar (2008) and Melichar & Kernstockova (2008). The output results are represented by separated groups of slips on fault surfaces and palaeostress inversion may be visualized in contoured density plots of stress inversions for four-faults groups. Maxima of density are the best-estimated directions of the principal stresses s1, s2 and s3, respectively (Kernstockova & Melichar 2008).

Results Geomorphic and geological observations The rockslide body has an elliptic and slightly asymmetric shape with a length of 11 km, maximum width of 8 km, summit altitude of the body of about 940 m a.s.l. (metres above sea level) and summit altitude of the crown of about 1070 m a.s.l. The minimum altitude near the toe is about 190 m a.s.l. The slope failure was first recognized through the analysis of DTMs. We estimate the rockslide’s thickness at 400– 500 m. Thus, the approximate volume of the displaced mass would be at about 28–35 km3. Regarding the main trench dimension, we expect total maximum displacement of the main body to be about 1 km towards SSE, which is relatively short with respect to the mass dimensions (about 10% of the rockslide length). The basal shear zone was probably consequent along less competent tuff rocks related to a much older shield volcano now located below the volcano of Santa Lucia. The tuffs and related ignimbrites crop out along the base of the slope failure in

the NW and west part of the rockslide, as well as in erosive cuts near Boaco. However, most of the rockslide body is composed of andesites, agglomerates and epiclastic rocks. Southeast gently dipping stratification of the underlying older shield volcano was probably the dominant factor that controlled displacement of the rockslide to the SSE. The rockslide consists of different parts (Fig. 2): that is, the zone of compression, the main body, the trench and the crown. The compression zone (I.A) has hummocky relief developed on strongly deformed and weathered andesites (e.g. cuestas, pressure ridges and other elevations, hog-backs, deep erosional, cuts and scarps). At some places, relict basaltic dykes crop out locally. In contrast, the main body (I.B in Fig. 2) consists of relicts of old structural relief on agglomerates and andesites, which are variably inclined and bounded by scarps and rock cliffs. Typical scarps (surfaces of normal faults) run perpendicular to the longer axis of the main body and are inclined towards the trench area. These scarps were created from subsequent gravitational collapse of the main body towards the trench (Figs 3 & 4). The main trench developed at the location of mass detachment SE of the town of Santa Lucia (II.A in Fig. 2). Recently, the trench has been partly filled with deposits from subsequent gravitational collapse of its margins in the form of secondary deep-seated rockslides thicker than 300 m. The rest of the Santa Lucia Depression (the caldera) has a semi-circular shape in ground-plan view (II.B in Fig. 2) and is also affected by gravitational collapse of the depression margins. Relicts of the original volcanic palaeosurface around the margin of the Santa Lucia Depression (III.A and B in Fig. 2) represent the crown of the giant slope failure. The area was strongly affected by unloading, and blocks ranging in size from several hundreds of metres up to a kilometre have been sagging or sliding rotationally towards the depression (Fig. 5). The slopes of the depression have been the subject of other superimposed rather shallow sliding, falling and flowing, posing a great risk to the town of Santa Lucia.

History of the rockslide The history of the rockslide and the subsequent geodynamic events were studied in sedimentary deposits related to valley-blockage, faults, landslides and fluvial activity in the area. No sedimentary evidence of any of large-scale fragmentation of rock related to the failure (no debris fan, etc.) was recognized. The rockslide blocked two valleys, and dam-lakes developed along its eastern and SE margins (IV.A and IV.B in Fig. 2). Only remnants of the lake deposits have been preserved there up to the present because both dam fills have

PALAEOSTRESS OF A GIANT ROCKSLIDE

137

Fig. 3. Oblique view and interpreted cross-section along the line X– Y, depicted in Figure 2, with the expected subsidence caldera, the interpreted internal structure of the giant rockslide (dark grey) and the subsequent gravitational collapse of the main trench near Santa Lucia. Two dams are also depicted; one of them was dated.

Fig. 4. Oblique 3D view on an aerial photograph of the rockslide’s NW part and southern part of the Santa Lucia Depression approximately near to the place of the mass detachment. The mass (with the Cerro Grande Mt) has experienced two modes of movement; that is, first, towards the SE during the main mass movement (large arrow and dark dash line) and towards the NW, subsequently into the opened space of the trench (part of the Santa Lucia Depression). This latter motion is proved by secondary deep-seated rockslides (contoured) and normal faults (dot and dash line).

138

I. BARON ET AL.

Fig. 5. Oblique 3D view on an aerial photograph of the northern part of the Santa Lucia Depression. This depression has a complex evolution. Originally, it was an ancient volcanic caldera that was, in its SE part, enlarged as a trench during the main phase of the giant rockslide. Slopes of the depression were later affected by very deep-seated rotational rockslides up to 200– 500 m thick. These slides probably caused thrust and reverse faults (point SL 41) and hummocky relief that dominated in the centre of the depression. For scale: the length of the rotational rockslide left of the Cerro Chiscolapa Mt is about 3 km and its expected thickness is about 500 m.

been intensively eroded. The eastern and larger one comprises a flat area within the original valley, deeply modified by back-erosion in the west (erosion reached the ignimbritic bedrock there). The fill of the preserved part comprises blackbrownish– reddish clays with intercalations of fluvial fine gravels and sands. Proof of the second dam has been preserved in the form of two terracelike bodies 1 km south of Boaco, separated by gullies and back-erosion. They completely cover an area measuring 550 m long by 300 m wide. Their flat surface is related to the reddish and black organic-rich and stratified clayey lacustrine deposit. The clays are intercalated with coarse and partly weathered fluvial gravel and sand close to tributaries of the former lake. The blackish fine organic clay at the base of the lacustrine deposit was determined as 3410 + 35 14C years BP (Poz22073) using the AMS method. This is the approximate age of the main activity phase of the entire giant rockslide. The rockslide has undergone other failure phases since then. As evidenced in several outcrops within its body, it has been affected by several younger faults. Two activity phases of strike-slip faults

changed depositional conditions of colluvial and fluvial sediments on the ground surface within the zone of compression about 2 km SW of Boaco. Two of these phases were determined to be 240 + 30 (Poz-22072) and 115 + 30 14C years BP old (Poz-22071) by dating of the organic matter in the soil sediment at a base of each cycle. The interpretation of the faults is quite difficult: they could be of tectonic origin or, indeed, the result of the giant rockslide reactivation. The main body underwent subsequent gravitational failure in the back-scarp area in the form of secondary, very deep-seated, rockslides that moved towards the trench. One such activity phase was identified within the soil buried under a thrust of andesite debris near the trench base about 1 km SW of Santa Lucia (on the left-hand part of Fig. 4). The buried charcoal is 160 + 30 14C years BP old, as determined by the AMS dating (Poz22074). Several other reactivation phases, however, should also be expected there. The tension conditions caused, in the combination of heavy rainfall and earthquakes, several superimposed shallow landslides and earthflows near the backscarp. These rather shallow landslides could be

PALAEOSTRESS OF A GIANT ROCKSLIDE

relatively very active and dangerous. One of them was dated to 230 + 50 14C years BP (Ki-14374). However, most of the landslides have not yet been dated.

Tectonic features During field mapping we recognized that the area of Santa Lucia and Boaco is strongly affected by faults and joints. The faults have developed owing to both the regional tectonic stress and the giant rockslide activity. Specific studies on the current regional stress field in the Central Nicaraguan Highlands are not available in the literature. However, as recognized from focal mechanisms of recent earthquakes in the Nicaraguan trough (western Nicaragua), the prevailing regional tectonic stress field about 40 km to the west of the study site has maximum stress, s1, oriented horizontal and NE–SW to north–south, and the minimum stress, s3, oriented horizontal NW–SE to east –west (Caceres et al. 2005). The area of the rockslide has the characteristic occurrence of normal and strike-slip faults related to the regional tectonic processes. In general, NE–SW- and NW– SE-striking faults prevail in the southern and western parts of the study area. A system of distinct circular or radial normal faults is clearly visible on the DTM in the close vicinity of the Santa Lucia volcano (see Fig. 2). We expect that they were formed by caldera subsidence that was centred at the town of Santa Lucia. Folds and faults with striations, significant groove casts, strain shadows and spread-grains developed within the rockslide body. Some reverse and thrust faults occur in the compression zone. Small drag folds with duplexes and tension calcite veins also accompanied the faults occurring there. In addition, normal and strike-slip faults are present in the zone of compression. It was quite difficult to distinguish rockslide-related faults from the tectonic faults as both types occur in the hard rock and are, therefore, very difficult to differentiate by origin. However, younger faults, which developed near the ground surface, are often filled with striated clay or dissipated in the form of dense joint set towards the ground surface. It was also recognized that some older tectonic faults were reactivated during slope failure as documented by: (1) a set of two or more different striation systems on single fault plane at some outcrops; or (2) a cleavage of opaline fill in common older joints. In the crown of the entire rockslide, mainly normal faults were mapped in contrast to the base of the Santa Lucia Depression, where mostly compression structures were mapped. Thrust and reverse faults, attendant drag folds, and, in some cases, small open buckle folds developed in the base of the depression

139

resulting from the gravitational collapse of the crown towards the depression. The quality of the ‘slickenside record’ differed depending on the faulted material and on the depth in which the fault originated. Weathered andesites and basalts had the worst slickenside records, as did the faults that originated very close to the ground surface. These faults rather dispersed into joints. However, the faults in tuffs and ignimbrites, as well as faults that developed deeper below the ground surface, had the best slickenside records.

Palaeostress analysis For the palaeostress analysis 146 faults with kinematics markers were measured along all of the rockslide and surrounding area (Fig. 2). Data from close sites were grouped to represent each different part of the rockslide, and to give a good input. Data from outside the rockslide body served as reference data. The compressional part of the main rockslide, the area along the compressional margin of the main rockslide, the upper (block) part of the main rockslide, the crown of the main rockslide and the centre of the Santa Lucia Depression were the regions of interest because a different palaeostress field was expected. Area outside the giant rockslide. Two reference areas were analysed in the area outside of the giant rockslide to determine the regional tectonic stress field before the slope failure as a reference for our study. Only one tectonic phase was identified in the area north (site 1) and west (site 2) of the slope failure within the Oligocene volcanic terrain (Fig. 2). In both cases, the principal stress, s1, was vertical or subvertical, whilst s2 and s3 were subhorizontal and with similar values within this gravity-driven stress field (Fig. 6a). Compressional part of the main rockslide. The compressional part of the rockslide (area I.A on Fig. 2) is represented by a very complicated structural domain including all fault types (e.g. normal, reverse and strike-slip) and rough topography. In the western part of the zone of compression (sites 3 and 4), two separate tectonic episodes can be identified (Fig. 6b). The ‘A’ phase is more clearly developed, with s1 vertical; the s2 and s3 stresses are almost equal and oriented along the horizontal plane. However, the ‘B’ phase has s1 horizontal and NNW–SSE oriented, s2 is horizontal and directed to the ENE –WSW, and s3 is vertical. The ‘A’ phase is clearly gravity-driven, whilst the ‘B’ phase is compressional and accompanied by the evolution of reverse faults. Axis s1 in the ‘B’ phase is parallel to the displacement vector of the main rockslide.

140

I. BARON ET AL.

Fig. 6. Density maxima of the principal stress directions s1, s2, s3 plotted in an equal-area projection obtained from palaeostress analysis of data from (a) the area outside the giant rockslide, (b) the compressional part of the main rockslide and (c) the area along the compressional margin of the main rockslide.

Two palaeostress phases can also be identified in the area south and SW of Boaco City (sites 5–7 in Fig. 2). The ‘A’ phase is clearly gravity-driven, with s1 vertical; s2 and s3 horizontal and more diversified in contrast to sites 3 and 4, that is, s2 is NW–SE oriented and the s3 axis is perpendicular

to it (Fig. 6b). The ‘B’ phase is also very similar to sites 3 and 4. However, the axis of the maximum stress, s1, is inclined 58 –208 towards NW and the s2 axis is inclined about 708 –858 towards SE (Fig. 6b). As at sites 3 and 4, the axis of s1 is parallel to the displacement vector of the rockslide.

PALAEOSTRESS OF A GIANT ROCKSLIDE

Area along the compressional margin of the main rockslide. The area along the southern margin (sites 8 –10 in Fig. 2) performed different results (Fig. 6c). The s1 axis is subhorizontal and ENE– WSW oriented, which is almost perpendicular to data from other parts of the zone of compression (as presented earlier). The s2 axis is horizontal and directed to the NNW–SSE, whereas s3 is subvertical and steeply inclined towards SE. We expect that this stress field is rotated compared to other areas of compression because of the original local topography (the rockslide filled an ancient valley in this area), and from the large diversity and architecture of competent and weak rocks. Two stress phases were recognized in the area to the SW outside of the rockslide body (Fig. 2, sites 11– 14). The clearest phase (Phase ‘A’, Fig. 6c) had a vertical s1 axis, and s2 and s3 were oriented nearly horizontal with almost equal values. The second phase is represented by a subvertical s1 (538/728), steeply dipping towards the slide mass, a subhorizontal s2 (SW –NE oriented) and a s3 (SE –NW oriented). The s3 probably represents the gravity-driven and slope-failure-related phase (Fig. 6c). Principal stress s1 was vertical, and s2 and s3 were subhorizontal (Fig. 6c), in the area east of

141

the slope failure (Fig. 2, sites 15– 18) and clearly represent a gravity-driven phase. Upper block of the main rockslide. Data from this part of the slope failure are quite limited owing to abundant subsequent and relatively shallow landsliding, and the lack of relevant outcrops of fresh rock. However, two tectonic phases were distinguished in the eastern part of the slope failure (Fig. 2, sites 19 and 20). The most evident phase was characterized by a vertical s1, and a north– south-oriented subhorizontal s2 and s3 (Fig. 7a). The second, and less evident tectonic phase, had s1 subvertical and steeply inclined towards east. The s2 was slightly inclined towards the west, and s3 was horizontal and NNE– SSW oriented. The slope failure mass was displaced during this relatively poorly defined, and probably older, ‘B’ palaeostress phase. Only 10 fault surfaces were measured in the western upper part of the rockslide body (Fig. 2, site 21 and 22) owing to the lack of high-quality outcrops. However, two palaeostress phases were interpreted (Fig. 7a). In the both phases s1 was nearly vertical, and s2 and s3 were horizontal. In the first phase s2 was oriented east –west and s3 was oriented north –south. The second phase had s2

Fig. 7. Density maxima of principal stress directions s1, s2, s3 plotted in an equal-area projection obtained from palaeostress analysis of data from (a) the upper block of the main rockslide and (b) the crown of the giant rockslide.

142

I. BARON ET AL.

oriented NNW–SSE and s3 oriented ENE –WSW. Both phases are interpreted to be gravity-driven, most probably related to mass-wasting processes. Crown of the giant rockslide. Two palaeostress phases were recognized in the area of the crown of the giant rockslide (areas IIIA and III.B in Fig. 2, sites 23–25). The first phase has a subvertical s1 steeply dipping towards the west, subhorizontal s2 (east–west oriented), and horizontal north –southoriented s3 (Fig. 7b). These data suggest gravitydriven normal faulting with east –west-striking faults. The poorly defined older phase had s1 horizontal and north–south oriented, s2 subhorizontal and east –west oriented, and the s3 near vertical and steeply inclined towards the west. The ‘A’ phase was probably related to the unloading of the crown area, whilst the ‘B’ phase is very difficult to interpret. We suspect that the ‘B’ phase was related to subsequent deep-seated rockslides along the trench that moved northwards or southwards and influenced site 24, which was already situated in its zone of compression. We obtained surprising results from the base of the Santa Lucia Depression to the north of Santa Lucia town (Fig. 2) at site 26. This area is dominated by hummocky relief of north–south-oriented ridges. Stress s1 is nearly horizontal and east– west oriented, s2 is horizontal, and s3 is subvertically oriented (Fig. 7b). Reverse and thrust faults, both east- and west-dipping, developed under such conditions. These compressional conditions probably resulted from the collapse of the crown (block III.B on the Fig. 2) owing to rapid unloading after the main slope-failure phase.

Discussion The study presents a new point of view on the relationship between extremely large, deep-seated rockslides and the local stress field. As presented in the introduction, several papers have addressed regional and local palaeostress fields as one of principal direct or indirect factors controlling the spatial distribution of landslides (e.g. Scheidegger 1980; Ollier 1981; Abrahams & Flint 1983; Pohn 1983; Embleton 1987; Gupta 2005). However, our results show how a giant rockslide modified the local palaeostress field. The giant rockslide caused local rock deformation very similar to tectonic deformation. As we have no sedimentary evidence of large-scale fragmentation of rock during the failure, we expect that the rockslide movement was rather moderate –slow (velocity class 3–4 of Cruden & Varnes 1996). An extensional regime prevailed in the upper part of the rockslide body and near the head, accompanied by normal and strike-slip

faults. Quasi-intact blocks with original geological structures were preserved in the central part of the rockslide body. Compression with thrusts, reverse and strike-slip faults, and fault-related deposits, occur in the toe portion of the rockslide. The rockslide obscured evidence of older palaeostress regimes recorded in the volcanic rocks of the area. We used the original tectonic stress field outside the giant rockslide as a basis for comparison to data from our palaeostress analysis. In this stress field s1 was vertical –subvertical, whilst s2 and s3 were subhorizontal and with similar directions. This stress field was clearly gravity-driven in areas near the towns Santa Lucia and Boaco, and it supports the hypothesis of a large subsidence caldera near the centre of the Santa Lucia Depression (Fig. 8a). The compressional part of the giant rockslide was represented by two tectonic phases that could be identified; that is, a clearly gravity-driven phase and a compressional phase, with the axis of maximum stress (s1) parallel to the displacement vector of the main rockslide body (Fig. 8b). The gravity phase could be related to either the forming of a large subsidence caldera or to gravity relaxation after the occurrence of the main rockslide displacement (Fig. 8c). The upper part of the rockslide yielded quite limited data owing to extensive subsequent shallow –deep-seated landslides and, thus, a lack of relevant outcrops of fresh rock. However, two tectonic phases were distinguished in the upper part of the rockslide. Both of the phases were gravity-driven with s1 vertical. Both phases are probably related to mass-wasting processes. In the rockslide crown, two palaeostress modes were recognized. The first phase had a subvertical s1 steeply dipping towards the west, which suggests gravity-driven normal faulting with east –west-striking faults related to the unloading of the crown area. Results obtained from the base of the Santa Lucia Depression, where east- or west-dipping reverse and thrust faults developed, show that s1 was nearly horizontal and east –west oriented, the horizontal intermediate stress was north– south oriented and the minimum stress was subvertical. These local compression conditions probably resulted from the subsequent collapse of the crown and margins of the caldera. The displaced mass of the rockslide only slightly influenced the surrounding areas. The area along the southern margin of the rockslide except for the southernmost parts, in general, showed just small changes in the stress field, such as the slight tilting of originally vertical s1 outwards from the rockslide mass. These observations of the rockslide displacement vector and its related stress field completely differ from observations of the tectonic stress field control of landslide displacement in the Himalayas (Gupta 2005). There it was found that the

PALAEOSTRESS OF A GIANT ROCKSLIDE

143

Fig. 8. Overview of the palaeostress field registered by our analysis along the giant rockslide near Boaco: (a) original palaeostress outside of the rockslide body: (b) palaeostress field related to the main rockslide phase: (c) pre- or post-failure palaeostress field in the rockslide body. Black ellipse, maximum stress (s1) axis; grey ellipse, intermediate stress (s2) axis; white ellipse, minimum stress (s3) axis plotted in an equal-area projection (white circles). For an explanation of the regions see the Figure 2 caption.

144

I. BARON ET AL.

displacement vector of all landslides was either along the minimum stress axis (s3) or along the intermediate stress axis s2, but not in any case along the maximum stress axis (s1). The vector of displacement was either orthogonal to s1 or oblique to it (Gupta 2005). Interestingly, subsequent deep-seated–shallow landslides commonly occur in the block part of the rockslide, which was in extension, but subsequent landslides are rare in the zone of compression. This could probably be seen as an analogy of this giant mass movement and the tectonic effect on landslides analysed by Gupta (2005). Our analysis of the giant rockslide has helped to define the differences between tectonic and masswasting features. In contrast to endogenous-driven tectonics, for the stress analysis of mass movements it must be possible to distinguish both extensional and compressional features within the body of gravitational slope instabilities. This observation could be very helpful during regional geological studies to help determine what is a slope failure and what is a tectonic feature.

Conclusions A giant rockslide that occurred on the southern side of an ancient volcano near Boaco (central Nicaragua) in the Holocene was the focus of a detailed structural–geological and palaeostress study. We did discover that: † It was possible to define the palaeostress regime of a very large rockslide using techniques normally used for regional tectonic analysis, such as field measurements of slickensides and a palaeostress analysis † Results of the palaeostress analysis were generally consistent with the observed rockslide morphology and known displacement directions † The techniques used in this paper could be applied to other very large landslides where slickenside data are available and where it is important to distinguish ‘slope tectonic’ features from ‘regional tectonic’ features. The field work was carried out under the guidelines of a project of geological mapping of risky processes in the area of Boaco and Santa Lucia, which was provided by the Czech Geological Survey in co-operation with the Instituto Nicaragu¨ense de Estudios Territoriales (INETER) in 2007– 2009 with financial support from the Ministry of the Environment and the Ministry of Foreign Affairs of the Czech Republic. The authors would like to acknowledge field and technical support of our Nicaraguan colleagues from the INETER, especially W. Strauch, A. Alvarez, A. Castellon, M. Echeverry, T. Obando, A. Munoz and G. Chavez, and our Czech colleagues participating on the project.

Special thanks belong to M. Jaboyedoff, the chief editor of this volume, for his recommendations and advice, and N. Harland for reviewing the English. Last, but not least, we would like to express many thanks to J. Coe from the Colorado Geological Survey and another anonymous referee for their fruitful help and advice in improving this paper.

References Abrahams, A. D. & Flint, J. J. 1983. Geological controls on the topological properties of some trellis channel networks. Bulletin of the Geological Society of America, 94, 80– 91. Caceres, D., Monterroso, D. & Tavakoli, B. 2005. Crustal deformation in northern Central America. Tectonophysics, 404, 119– 131. Castelli, M., Scavia, C., Bonnard, C. & Laloui, L. 2009. Mechanics and velocity of large landslides – Preface. Engineering Geology, 109, 1– 4. Cruden, D. M. & Varnes, D. J. 1996. Landslide types and processes. In: Turner, A. K. & Shuster, R. L. (eds) Landslides: Investigation and Mitigation. Special Report 247, Transportation Research Board, National Research Council, Washington, DC, 36– 75. Embleton, C. (ed.) 1987. Neotectonics and morphotectonics. Zeitschrift fur Geomorphologie Supplement, 63, 240. Fell, R. 1994. Landslide risk assessment and acceptable risk. Canadian Geotechnical Journal, 31, 261– 272. Garayar, J. 1972. Geologı´a y depo´sitos de minerales de la regio´n de Chontales y Boaco. Informe, 11. Archivo de INETER, Managua. Gomberg, J., Bodin, P., Savage, W. & Jackson, M. E. 1995. Landslide faults and tectonic faults, analogs?: The Slumgullion earthflow, Colorado. Geology, 23, 41–44. Gupta, V. 2005. The relationship between tectonic stresses, joint patterns and landslides in the higher Indian Himalaya. Journal of Nepal Geological Society, 31, 51–58. Hradecky, P., Baron, I. et al. 2007. Estudio geologico para reconocimiento de riesgos naturales, area de Santa Lucia, Boaco. Project Stage Report, Archive of INETER, Managua. Kernstockova, M. 2005. Palaeostress analysis of fault/slip data sets. Bachelor of Science thesis, Faculty of Science, Masaryk University, Brno [in Czech]. Kernstockova, M. & Melichar, R. 2008. Numerical palaeostress analysis – the limits of automation. In: Poblet, J., Medina, M. G., Pedreira, D. & Ferna´ndez, C. L. (eds) International Meeting of Young Researchers in Structural Geology and Tectonics. YORSGET-08, Oviedo (Spain), 1 –3 July 2008. University of Oviedo, Oviedo, 317 –321. Melichar, R. & Kernstockova, M. 2008. 9D space – the best way to understand palaeostress analysis. In: Poblet, J., Medina, M. G., Pedreira, D. & Ferna´ndez, C. L. (eds) International Meeting of Young Researchers in Structural Geology and Tectonics. YORSGET-08, Oviedo (Spain), 1 –3 July 2008. University of Oviedo, Oviedo, 327 –331.

PALAEOSTRESS OF A GIANT ROCKSLIDE Ollier, C. D. 1981. Tectonics and Landforms. Longman, London. Pohn, H. A. 1983. The relationship of joints and stream drainage in flat lying rock of south central New York and northern Pennsylvania. Zeitschrift fur Geomorphologie N.F., 27, 375 –384. Reid, M. E., Christian, S. B. & Brien, D. L. 2000. Gravitational stability of three-dimensional stratovolcano edifices. Journal of Geophysical Research, 105, 6043–6056. Rogers, R. D., Mann, P. & Emmet, P. A. 2007. Tectonic terranes of the Chortis block based on integration of regional aeromagnetic and geologic data. Geological Society of America, Special Papers, 428, 65–88. Savage, W. Z. & Swolfs, H. S. 1986. Tectonic and gravitational stress in long symmetric ridges and

145

valleys. Journal of Geophysical Research, 91, 3677– 3685. Scheidegger, A. E. 1980. The orientation of valley trends in Ontario. Zeitschrift fur Geomorphologie N. F. 24, 19–30. Siebert, L. 2002. Landslides resulting from structural failure of volcanoes. In: Evans, S. G. & De Graf, V. J. (eds) Catastrophic Landslides: Effects, Occurrence, and Mechanisms. Reviews in Engineering Geology, 15, 209–236. Weyl, F., 1980. Geology of Central America. Gebruder Borntrager, Berlin. Yamaji, A., Otsubo, M. & Sato, K. 2006. Palaeostress analysis using the Hough transform for separating stress from heterogeneous fault-slip data. Journal of Structural Geology, 28, 980–990.

Complex landslide behaviour and structural control: ˚ knes rockslide, Norway a three-dimensional conceptual model of A MICHEL JABOYEDOFF1*, THIERRY OPPIKOFER1, MARC-HENRI DERRON1,2, ¨ HME2 & ALINE SAINTOT2 LARS HARALD BLIKRA2,3, MARTINA BO 1

Institute of Geomatics and Analysis of Risk, University of Lausanne, Amphioˆle, 1015 Lausanne, Switzerland

2

Geological Survey of Norway, Leiv Eirikssons vei 39, 7491 Trondheim, Norway 3˚ Aknes/Tafjord Beredskap IKS, Ødega˚rdsvegen 176, 6200 Stranda, Norway *Corresponding author (e-mail: [email protected]) ˚ knes is an active complex large rockslide of approximately 30–40 Mm3 located Abstract: A within the Proterozoic gneisses of western Norway. The observed surface displacements indicate that this rockslide is divided into several blocks moving in different directions at velocities of between 3 and 10 cm year21. Because of regional safety issues and economic interests this rockslide has been extensively monitored since 2004. The understanding of the deformation mechanism is crucial for the implementation of a viable monitoring system. Detailed field investigations and the analysis of a digital elevation model (DEM) indicate that the movements and the block geometry are controlled by the main schistosity (S1) in gneisses, folds, joints and regional faults. Such complex slope deformations use pre-existing structures, but also result in new failure surfaces and deformation zones, like preferential rupture in fold-hinge zones. Our interpretation provides a consistent conceptual three-dimensional (3D) model for the movements measured by various methods that is crucial for numerical stability modelling. In addition, this reinterpretation of the morphology confirms that in the past several rockslides ˚ knes slope. They may be related to scars propagating along the vertical folioccurred from the A ˚ knes slope is presented. ation in folds hinges. Finally, a model of the evolution of the A

The unstable slopes located in the Norwegian Fjord represent a major threat for populations because they can induce devastating tsunamis. Three major events are recorded in the last century in Norway (Blikra et al. 2005). The Storfjorden area in western Norway suffered several of these catastrophes in the past and, therefore, active landslides are an important concern in this region (Blikra et al. ˚ knes rockslide (Figs 1 & 2) is the 2006b). The A most active and thus most likely to provoke a catastrophic event. It is a large complex rockslide of approximately 30 –40 Mm3 (Derron et al. 2005; Ganerød et al. 2008), and is one of the most extensively monitored slope instabilities worldwide as it ˚ knes–Tafjord project that started belongs to the A in 2004. The final goal of the project is to set up an early warning system against rockslide-triggered tsunamis for the whole region (Blikra 2008). The understanding of the rockslide mechanism is crucial implementing a suitable monitoring system (Jaboyedoff et al. 2004; Blikra 2008; Froese et al. 2009). New data, such as airborne and terrestrial laser scanning digital elevation models (ALS-DEM and TLS-DEM), differential GPS and

high-resolution borehole inclinometers, allow the improvement of the conceptual model of this rockslide. In a conceptual view, the main types of rockslide mechanisms were described by Varnes (1978), Hoek & Bray (1981), Hutchinson (1988) and Cruden & Varnes (1996), but they all present idealized models. There are only few conceptual models integrating complex structures (Zaruba & Mencl 1969; Sjo¨berg 2000; Agliardi et al. 2001; Sartori et al. 2003; Eberhardt et al. 2004; Stead et al. 2006; Willenberg et al. 2008; Brideau et al. 2009a; Jaboyedoff et al. 2009). The modelling of rock failure mechanisms is still mainly performed in two dimensions (Eberhardt et al. 2004; Stead et al. 2006) and rarely in three (Ambrosi & Crosta 2006; Brideau & Stead 2010). Nevertheless, the development of a site-specific conceptual instability model is a requirement for reliable numerical modelling. Complex movements in rock slopes may be explained by mechanisms that involve pre-existing structures and the formation of new failure surfaces by breaking rock bridges that generate several

From: Jaboyedoff, M. (ed.) Slope Tectonics. Geological Society, London, Special Publications, 351, 147– 161. DOI: 10.1144/SP351.8 0305-8719/11/$15.00 # The Geological Society of London 2011.

148

M. JABOYEDOFF ET AL.

˚ knes rockslide and its limits on a hillshade model derived from ALS-DEM. The eastern Fig. 1. Location of the A rockslide limit corresponds to fault F1 and the SW limit of F2.

imbricated instabilities (Agliardi et al. 2001) with different movement directions (Brideau et al. ˚ knes rockslide is a good example of 2009). The A a complex slide (Blikra 2008; Ganerød et al. 2008; Kveldsvik et al. 2008). The structures, that is faults, joints and folds, are controlling the slope ˚ knes rockslide movements and the division of the A into several blocks. Even small variations of the main structural features have a strong influence on the slope tectonics and its morphology. The present study explores the impact of preexisting structures and the appraisal of new failure surfaces, and aims to show that complex slope movements need a good conceptual model in order to link observed movements to the model and design the accurate monitoring.

Geological and morphological settings ˚ knes rockslide is located on the NW flank of The A Sunnylvsfjorden, a branch of the Storfjorden in western Norway. During the Caledonian orogeny

(c. 400–500 Ma ago), Proterozoic granitic intrusions were metamorphosed to medium-grained, mica-rich orthogneisses with the creation of a welldeveloped mineral banding (Tveten et al. 1998). These orthogneisses make up the major part of the Storfjorden area. The area suffered several folding phases, which can be observed in several places. As most of the folds are tight– isoclinal, they are only observable ˚ knes area fold interferences in hinge areas. In the A of type 3 (Ramsay 1967) can be observed. Most of the instabilities in the Storfjorden area are caused by dip slopes along the main foliation planes that develop large potential slides (Braathen et al. 2004; Henderson et al. 2006). The topography of the Storfjorden area has been deeply cut by glacial erosion along weaker zones in the bedrock, such as foliation, faults and fractures. ˚ knes site itself has been described and anaThe A lysed in detail by Derron et al. (2005) for the main structural features, and by Ganerød et al. (2008) using geophysical data, structural analysis, displacement data and borehole data. The rockslide ranges

˚ KNES LANDSLIDE STRUCTURES OF A

149

˚ knes rockslide with predicted limits (from Derron et al. 2005). Three historical rockslide events Fig. 2. View of the A are outlined, including the timing of the events (modified from Kveldsvik et al. 2008).

from the toe at 150 m above sea level (m.a.s.l.) to the back scarp at 1300 m.a.s.l., and the average slope angle is 308–358 (Ganerød et al. 2008). But it is steepened locally with subvertical east –westtrending cliffs ranging from several metres to tens of metres in height. Borehole measurements reveal that the main foliation dips between 278 and 348, with an average value of 328 (Kveldsvik et al. 2006), which indicates that the gneiss foliation makes up the slope surface (Ganerød et al. 2008). This range of dip angles is related to the different

fold phases. The top of the rockslide is limited by a back-scarp zone that is identified by an open crevasse with a width of approximately 1 m in its eastern part to 30 m in its western part. Ganerød et al. (2008) and Kveldsvik et al. (2009a) showed that the landslide body is divided into several blocks by subvertical north–south fractures. From the surface morphology and geophysical data they deduced that the landslide is divided in four main bodies separated by sliding surfaces that overthrust each other and daylight at different

150

M. JABOYEDOFF ET AL.

levels (Blikra 2008). In that context, the water has been demonstrated to be flowing through a complex fractured zone (Lo¨w et al. 2008). The landslide is mainly moving south and downwards towards the fjord. The surface displacements are measured by several methods: GPS, photogrammetry, tachyometers, permanent laser distancemeter, extensometers, ground-based and satellite radar interferometry, and terrestrial laser scanning (Blikra et al. 2006a). These measurements show velocities of the order of a few centimetres per

year and up to 15 cm year21 in the upper western part (Fig. 3) (Eiken 2005; Kveldsvik et al. 2006; Blikra 2008; Ganerød et al. 2008), and confirm the division of the rockslide into several blocks moving in different directions at different velocities (Fig. 3). Three rockslides occurred in the recent past in ˚ knes landslide (Kveldsvik the western part of the A et al. 2008): (1) in the upper part in the second half of the nineteenth century; (2) in the lower part, in 1940; and (3) a 100 000 m3 rockslide in

˚ knes rockslide, displaying the measured displacements (data from Blikra 2008) and the Fig. 3. Hillshade model of the A main structural features. The two lateral faults (F1 and F2) are represented as blue lines, the JNS joint set as white lines and fold axis A2 are shown in green. The displacement vectors reveal south and downwards movement of the upper western part, with up to 10 cm year21, movement towards the SSE in the middle eastern part, and small upwards movements (,2 cm year21) in the toe zone.

˚ KNES LANDSLIDE STRUCTURES OF A

151

1960 in the mid western part. The latter event avalanched down to the sea (Fig. 2).

more complete structural description see Ganerød et al. (2008).

Methods

Main foliation

The morphological and structural analysis of the ˚ knes rockslide has been performed using an airA borne laser scanning digital elevation model (ALSDEM) acquired in 2004 from using a helicopter with a Fli-Mapw ALS equipment. The flight height was 50–60 m and the density of measurements was 15– 25 points per horizontal m2. The elevation accuracy on rock is 4– 5 cm. The reflections from the vegetation have been removed in order to obtain the ground elevation. A grid with a cell size of 1 m was then produced from the data points. Field surveys identified the principal structures and made structural measurements. Several orthophotographs were used coupled with the ALSDEM in order to make a precise link between field works and the ALS-DEM, and to confirm or discard the interpretations based on the ALS-DEM. The structural analysis based on ALS-DEM was made using the software Coltop3D (Jaboyedoff et al. 2007, 2009). This software allows the visualization and characterization of the structures by an orientation-specific colouring of the DEM. The full orientation of a topographical surface is displayed merging, in one colour, its dip angle and dip direction (Fig. 4). To do so, a lower-hemisphere Schmidt stereonet is combined with a Hue– Saturation–Intensity (HSI) wheel, which provides a unique colour for each spatial orientation. The colours are then attributed to each DEM cell by its spatial orientation of pole in the stereonet. This colour representation of the topography has the advantage of displaying the same slope orientations with the same colours. Assuming that the morphology of rock slopes is shaped by discontinuities, their orientation can be measured on the Coltop3D representation of the DEM. A terrestrial laser scanning DEM (TLS-DEM) was used to identify and characterize the main structures involved in the mass movement. It was also used to monitor the displacements of the upper part of the rockslide (Oppikofer et al. 2008, 2009). The surface displacements of the whole rockslide were extracted from the works of Eiken (2005), Kveldsvik et al. (2006), Ganerød et al. (2008) and Blikra (2008).

The main foliation (S1) is a fabric characterized by a mineralogical layering subparallel to the topography. This inherited structural heterogeneity controls most of the basal sliding surfaces of rockslides in the Storfjorden area (Henderson et al. 2006; Oppikofer & Jaboyedoff 2008; Oppikofer et al. in press). The variations in orientation of S1 induce changes in sliding direction, as can be seen by comparing Figures 3 & 4 (see the next subsection on ‘Fold hinges’). The foliation S1 undulates with a hectometric wavelength similar to large, gentle folds with the axis close to 1208/208. These undulations are termed U in the following section, ‘Fold hinges’. ˚ knes, At Rundefjellet hill, some 2.5 km SW of A the dip-slope along the gneiss foliation forms the topographical surface (Fig. 5). This surface displays a morphology in steps and is probably the basal sliding surface of a former rockslide (Oppikofer & Jaboyedoff 2007; Longva et al. 2009; Oppikofer et al. in press). Large, gentle folds or undulations (U) affect the gneiss foliation (S1) and may have led to changes in the sliding direction of the ancient rockslide. ˚ knes slope, the fold axis of the undulaIn the A tions U is subparallel to the SE-plunging axis of the isoclinal folds. Figure 6 shows that small orientation changes of the foliation direction lead to variations in the potential sliding directions.

Description of the relevant observed structures In this section the main structures that control the slope movements and which are relevant to the present conceptual model are described. For a

Fold hinges The fold hinges (with axis A2: 1208/208; Fig. 6) ˚ knes rockslide (Fig. 3) form ridges crossing the A in the slopes because they are more massive. The main foliation varies considerably in direction, thus avoiding a persistent weakness orientation. The S1 orientation changes direction over a short distance and is, therefore, is less prone to slip; hence, there is no possibility for the foliation to act as a sliding surface over long distances. The spacing between two successive ridges is fairly regular (approximately 250 m) and, therefore, reflects the wavelength of the folds A2. The folds A2 are upright or steeply inclined (Fig. 7). However, in these hinges zones the foliation orientation (S1) is very variable, including some locations where S1 is subvertical (Fig. 7b). Such zones are more prone to developing tension cracks or back scars than areas where sliding along the foliation is possible, provided that the foliation is roughly slope-parallel and fjord-dipping (Figs 7 & 8). The main back ˚ knes has a graben-like scarp on the western side of A geometry. It follows a fold hinge with an axis A2

152

M. JABOYEDOFF ET AL.

Fig. 4. Orientation-specific colouring of the ALS-DEM in COLTOP showing the rapid variation in the foliation’s orientation (white arrows). The slope orientations are displayed by a unique colour given by the pole in a lower-hemisphere Schmidt stereonet in association with a HSI wheel (a). The measurements on this COLTOP image were performed assuming that the slope follows the foliation, which is confirmed by field measurements. The poles of the measured foliation surfaces are shown in the density stereonet (b). The gneiss foliation S1 is displayed in different tones of magenta; joint sets JNS and JEW are shown in yellow-green and red, respectively. The limit between areas with south-dipping and SSE- to SE-dipping foliation is shown. This limit (white transparent line) corresponds to the separation line between two rockslide compartments: the western one moving to the south to SSW with high displacement rates (fast-moving ridge); the other moving to the SSE to SE.

(Figs 3 & 7a); the orientation of this axis is in good agreement with previous studies (Braathen et al. 2004; Oppikofer et al. 2009). The graben is continued as a 300 m-long succession of open cracks that follows the direction of the folds axis (Figs 7 & 8). Cracks (CA2) have developed along the flank of the folds where the foliation S1 is very steep and seems to have an en echelon geometry (Fig. 7a). These fold axes are noted A2 because at least one previous folding phase can be found (with axis A1), which is isoclinal (Fig. 7c). These isoclinal folds

make S1 parallel to the axial surface AS1 and, as a consequence, folds from this previous phase do not have a great influence on the orientation of the main foliation S1 except locally in the hinge zones of A2. A2 and U have similar axes, which suggest that they are related.

East – west subvertical joint set (JEW) The subvertical east –west-striking discontinuity set (JEW) has been measured in the field by Ganerød

˚ KNES LANDSLIDE STRUCTURES OF A

153

˚ knes displaying the undulation Fig. 5. View of the stepped basal sliding surface of a former rockslide, 2.5 km SW of A (U) of the foliation and the steps created by JEW.

et al. (2008). It may correspond either to newly formed extension cracks or to pre-existing joints acting as extension cracks. They probably form the steps disrupting the sliding surface (JEW). This can be interpreted from geophysical data (Fig. 9) and using observations about S1 in the vicinity of ˚ knes rockslide (Fig. 5), where an assumed the A former sliding surface displays steps and undulations (Oppikofer et al. in press). This interpretation is performed assuming that the JEW can be either saturated or not, giving opposite responses to electrical tomography. The spacings of these step fractures and of the foliation surfaces follow an exponential negative distribution (Oppikofer et al. in press).

North – south subvertical joint sets (JNS) The north –south-trending, subvertical joints (JNS) can be observed in the western part, close to the gully (Figs 3 & 7a). In the lower and central part of the rockslide these joints have a NNE–SSW orientation (Fig. 3). There is no evidence for distinct origins of these different joints orientations and, since they play similar roles in the rockslide’s mechanism, they are grouped together and called JNS. These joints shape the entire slope because they are found at different scales over the entire rockslide area (Ganerød et al. 2008). They create a network that defines the western limit of the landslide coupled with a regional subvertical fault crossing the area (F2) (Figs 2 & 3). They have also acted as auxiliary sliding surfaces within wedges for the 1940 and 1960 rockslides.

Lateral limiting faults of the rockslide The landslide is limited to the east by a fault (F1) dipping westwards with an angle of approximately 458 (Ganerød et al. 2008) (Fig. 3). There is little evidence at the slope surface for this fault, but its trace can be identified on the DEM. The fault forms the limit between areas that were affected by former rockslides and areas without large evident scars (Fig. 10). The trace of the F1 plane allows its orientation to be deduced, which is approximately 2358/458. The western boundary of the landslide is defined by a regional subvertical fault, F2, oriented NNW– SSE, characterized by a highly fractured zone. This structure became a preferential erosion zone and a deep gully developed along the fault. This permits the foliation to daylight in the gully and thus the landslides have been known about since the nineteenth century.

Landslide scars and back scarps In addition to the historical rockslide scars along the west-bounding fault several past rockslides can be observed on the present topographical surface, especially at the toe zone of the moving mass (Fig. 10). This indicates that the front of the landslide has already suffered several rockslides. The toe zone of the rockslide shows several scars in its frontal part, whose activity is evidenced by rockfall and overhangs (see fig. 6a in Ganerød et al. ˚ knes 2008). In the middle eastern part of the A rockslide the scars with a smoothed aspect may correspond to the old rockslides linked to the A2

154

M. JABOYEDOFF ET AL.

˚ knes hinge zones and involved the shallow parts of A rockslide. Two other recent scar locations are found on each side above the toe zone. Besides the scars at the toe zone, all ancient rockslide scars seem to be linked to weak zones defined by A2 fold hinges. ˚ knes rockslide is situThe most active part of the A ated in the upper western part, with the fast-moving ridge that is back-bound by the graben-like structure situated along the back scarp (Fig. 1). The graben opens by up to 10 cm year21 in some locations, but precise displacement measurements are very difficult to assess because the ridge is made of large unstable blocks. Monitoring by terrestrial laser scanning revealed that in addition to translational displacements to the SSW, most of these blocks topple towards the graben in the north, while others toppleslide towards the fjord (Oppikofer et al. 2009).

˚ knes Kinematics of blocks at A ˚ knes rockslide used for The movement data of the A this study are the recent differential GPS measurements (Ganerød et al. 2008). Comparisons between several series of aerial photographs permit estimates of movements back to the 1960s (Eiken 2005) and these reach several metres (up to 8 m) in some of the most active parts of the landslide. These photogrammetric results are very similar to those presented here for the upper part. The results obtained by GPS give two main directions of displacement: south to SSW in the upper part of the fast-moving ridge and the middle western part; and SSE to SE in the middle eastern part (Blikra 2008; Ganerød et al. 2008) (Fig. 3). These different movement directions coincide well with the different foliation dip directions, attesting once more to sliding along the foliation (Figs 4 & 6). The displacements and deformations at the toe zone of the rockslide remain enigmatic. Some GPS data indicate small upwards movements (Blikra 2008; Ganerød et al. 2008), but further measurements are necessary to confirm and clarify such kinematics. Moreover, the movements are strongly linked to water circulation with higher displacement rates occurring during snow melt when the water table rises by up to 4 m in

Fig. 6. Lower-hemisphere Schmidt stereonets of the poles to the foliation. (a) Density stereonet of field measurements, where the arrows indicate the trend and plunge of three possible movement vectors. These direction changes are mainly caused by the undulation (U). (b) Measurements made on the orientation-specific colouring of the ALS-DEM in the software COLTOP (see Fig. 4). (c) Detailed pole plot of field measurements in the A2 fold-hinge zones, including joints JNS. The fold axis A2 and the folded axial surfaces AS1 are shown.

˚ KNES LANDSLIDE STRUCTURES OF A

155

Fig. 7. (a) Outcrop in the main back scars showing the vertical foliation S1 folded within folds A2 and the cracks (CA2), the north– south-trending joints (JNS) are found everywhere and they cut quasi perpendicularly the folds. Note the variation in the S1 orientation due to the large-scale folding. (b) Fold forming a bulge in the middle eastern part; polyphase folding can be seen as early isoclinal folds. A1 were refolded in the open fold as shown in detail in (c).

156

M. JABOYEDOFF ET AL.

Fig. 8. Scheme explaining how the weaknesses (cracks CA2) are induced by folds A2 in combination with folds A1. ˚ knes rockslide, and the former rockslide The cracks CA2 and the subvertical foliation, S1, forms the back cracks of the A scars are detected on the morphology of the slope.

Fig. 9. Interpretation of the stepped failure surface (dashed black line) based on an inversion of a longitudinal two-dimensional electrical profile (electrical profile modified after Blikra et al. 2006a). The vertical dashed lines indicate that the vertical fractures are linked either to weaknesses (CA2) created by folds or to the JEW joint sets. It is assumed that they are creating highly permeable zones that are either saturated or free of water, and are therefore giving opposite electrical responses.

˚ KNES LANDSLIDE STRUCTURES OF A

157

Fig. 10. Three-dimensional representation of the hillshade indicating the former scars. Historical rockslide activity was located on the western boundary (next to fault F2), like the 1940 and 1960 scars. More ancient scars can be detected in the rockslide’s toe zone and in the middle eastern part.

the borehole located in the central part of the landslide (Blikra 2008).

Interpretation The relationship between the geometry of structures and the kinematics of blocks allows the mechanism ˚ knes rockslide to be better defined. Most of of the A the sliding surfaces are along the foliation S1. As shown in Figures 3, 4 & 6, small variations in the orientation of S1 are caused by U or A2 promoting different displacement directions. This results in several blocks moving down in different directions. The main WNW–ESE fold axis (A2) crosses the landslide body obliquely, with S1 found vertical near the hinge zones (Figs 3, 7 & 8). These zones with a vertical foliation are weak zones in the rock mass. They lead to the formation of extension failures acting as the back scars of the blocks. At the top of the landslide the opening of such a crevasse creates an approximately 30 m-wide trench (graben). Laterally, the landslide is limited by two regional faults crossing the mountain side (F1 and F2). The main blocks making up the landslide are laterally delimited by subvertical faults (JNS) acting as transfer surfaces or as auxiliary surfaces defining a wedge with S1. As regards to the movement direction, such wedges appear to be active only if JNS

are north–south trending and not when they are NNE –SSW trending (Figs 3 & 7). But wedge failure mechanisms formed by south-dipping S1 and NNE–SSW-trending JNS are possible when the wedges daylight in the western gully, which was the case for the 1940 and 1960 rockslides (Fig. 10) (Kveldvik et al. 2008). The sliding surfaces along the S1 foliation are not likely to be continuous. They are stepped by subvertical east –west-trending fractures (JEW) perpendicular to the main sliding direction or by WNW–ESE-oriented steep fractures (CA2) created along the A2 hinges cutting vertically the folds. These JEW fractures are clearly identified in the ˚ knes landslide topography in the vicinity of the A (Oppikofer et al. in press). The undulations of the ˚ knes slope are more or less parfoliation (U) in the A allel to the SE-plunging folds A2.

Proposed rockslide mechanism The rockslide has been continuously moving for more than 40 years (Kveldsvik et al. 2006). Water and snow melt play an important role in the activation of movements and in the destabilization at ˚ knes (Blikra 2008). Cyclic increases and decreases A of the groundwater table are sources of mechanical fatigue (Eberhardt et al. 2004; Jaboyedoff et al. 2009). This may be the main source of stability

158

M. JABOYEDOFF ET AL.

changes, with possibly some seismic activity (Kveldsvik et al. 2009b). The observations of (1) the back scars following the fold hinges and of (2) the steps in the basal failure surface render evident that the failure surfaces were initially less smoothed and a unique surface but interrupted by rock bridges. Nowadays, most of these rock bridges must be broken and a continuous failure surface has to exist as the rock masses have moved several metres and still continue to move downwards. This interpretation is based on the observation of the outcropping old failure surface at Rundefjellet (2.5 km SW of ˚ knes) that displays stepped surface following A the main foliation (Oppikofer et al. in press). It contrasts with the interpretation of Ganerød et al. (2008) that states that several circular or curved fail˚ knes ures are shaping the failure surface of the A rockslide. As the S1 foliation controls the sliding direction, its undulations provide several spread-out movements, which tend to break up the mass into several blocks. The latter are delimited or controlled by all the other structures (Fig. 11): † CA2: weakness zones in the WNW–ESE hinges of A2 folds; † JEW: east–west subvertical joints set; † JNS: north –south subvertical joints set; † F1 and F2: regional lateral faults. The past rockslides indicated by the scars have probably no more direct influence on the present stability because the basal failure surface of the

˚ knes rockslide is now determined and is entire A most probably through-going the rock mass. In the outcrops and in the boreholes the identified failure surfaces contain gouges and brecciated rocks (Ganerød et al. 2008; Grøneng et al. 2009). This fact avoids the problem of the estimation of the ‘apparent cohesion’ caused by rock bridges or joint roughness. Hence, it may be assumed that the dynamics of the landslide are mainly controlled by the geometry of its different parts and their interactions, but also by a worsening factor and, in the case of complete failure, by external triggering factors such as earthquakes. Analysing the movements (Fig. 3), it is obvious that the upper western part is the most hazardous area because of its high displacement rates (Blikra 2008); the eastern part moves more southeastwards. This reflects a divergent pattern of movements for the evolution of ˚ knes that tends towards collapse in several stages A of rockslides with a volume of several Mm3 each. The blocks forming the rockslide are not only delimited laterally but also vertically. As the S1 foliation is often less steep than the topography, it allows the basal failure surfaces to daylight in the slope. The basal failure surface of the upper blocks that move to the south probably daylights in the topography before reaching the JNS, otherwise the movement would have been controlled by the wedge intersection of S1 and JNS. In this case a displacement component to the west would be observed. Alternatively, the absence of westwards movement of this upper block signifies that it extends further to the east and thus completely

˚ knes rockslide before the rockslides occurred that created the observed scars. Fig. 11. Conceptual model of the A

˚ KNES LANDSLIDE STRUCTURES OF A

includes the JNS fracture. The wedge is operating only if the movement is compatible with the outcropping JNS, that is, mostly north–south, as was the case for most of the former slides and is so for the western part of the rockslide. The 1960 event was an exception: the NNE–SSW JNS was controlling the wedge S1 JNS (Figs 3 & 10). Furthermore, as demonstrated by borehole data (Ganerød et al. 2008) and the displacements recorded in the area of former scars, several superposed sliding surfaces have to be present. This multiplicity of movement vectors may create open fractures and zones of contraction at the limits of the blocks, giving rise to a dynamic evolution of the slope stability. The nature of the slope instability becomes increasingly more complex involving an ever-increasing number of landslide fragments.

Mechanism at the rockslide toe zone As the movement at the toe zone is unclear, a single large event remains possible. If the toe zone actually undergoes some vertical upwards displacements (as suggested by some GPS data), it means that the toe zone is somehow blocking the rockslide mass that is pushing from behind. The rockslide scars in the toe zone indicate that up to now slope movements have destabilized the front (Fig. 10), but the displacements were not sufficient to trigger the whole landslide. This

Fig. 12. A possible interpretation explaining the potential upwards movements at the toe: (1) the middle upper part of the rockslide moves down to the SSE along S1; (2) the eastern part slides down to west along F1; and (3) if the rock is already dismantled, as the scar at the toe (on the figure) may indicate, the movement can induce upwards movements by squeezing the rock.

159

is probably caused by particular structures at the toe zone. A large wedge structure may retain the mass, thus potentially leading to some retrogressive movements – as already proposed by Ganerød et al. (2008) – and creating a situation of contraction pushing up a block as ‘reverse faulting’ – as supported by the GPS measurements (Fig. 12) (Blikra 2008). To supplement these findings, our conceptual model provides an explanation of the mechanisms in the rockslide toe zone (Fig. 12). The toe of the ˚ knes rockslide can be analysed as a wedge comA posed of the eastern fault (F1) and the foliation S1 (Fig. 12). Rock masses moving in the wedge lead to compression. Some parts can be pushed up, indicating that the stress modifies the stability conditions and that the toe zone is acting as a buttress. This assumption is supported by the slide scars and the bulge in the toe area, which indicate that this destabilization activity has already started.

Conclusion Detailed field surveys and DEM analysis indicate that the kinematics and the geometry of the blocks ˚ knes rockslide are controlled by strucforming the A tures, such as the main foliation, folds, joints and faults. This interpretation gives a consistent framework for the movements observed by various ˚ knes. methods at A In addition, this model leads to a reinterpretation of the morphology, implying that in the past several rockslides have occurred. They can be related to scars occurring in fold-hinge zones where the gneiss foliation is locally vertical. This interpretation indicates that different parts of the landslide are moving. This will eventually lead to several individual failures involving several Mm3. However, as the interactions between blocks, along with groundwater circulation, rapidly change the stability conditions the possibility of a significant large event cannot be ruled out. This means that monitoring must focus on the toe zone and on the ˚ knes fast moving blocks in the upper part of the A rockslide. As the presented model is a complex threedimensional framework, standard mechanical models will probably be unsuitable. If the present model (or a future better one) is not taken into account then the computed stability models will probably be meaningless. ˚ knes– Tafjord This paper has been supported by the A project (IKS-Beredskap senter). The authors wish to thank Prof. A. Braathen for a very detailed review that helped us greatly to improve the manuscript and another anonymous reviewer who carried out very detailed editing. We are also grateful to J. Griffiths for his editorial work.

160

M. JABOYEDOFF ET AL.

References Agliardi, F., Crosta, G. B. & Zanchi, A. 2001. Structural constraints on deep-seated slope deformation kinematics. Engineering Geology, 59, 83–102. Ambrosi, C. & Crosta, G. B. 2006. Geomorphological and geotechnical constrains on large slope deformations. Geophysical Research Abstracts, 8, 04178. ˚ knes rockslide; monitoring, Blikra, L. H. 2008. The A threshold values and early warning. In: Chen, Z., Zhang, J.-M., Ho, K., Wu, F.-Q. & Li, Z.-K. (eds) Landslides and Engineered Slopes. From the Past to the Future: Proceedings of the 10th International Symposium on Landslides and Engineered Slopes, 30 June– 4 July 2008, Xi’an, China. CRC Press, Baton Rouge, FL. Blikra, L. H., Anda, E., Belsby, S., Jogerud, K. & Klempe, Ø. 2006a. Statusrapport for Arbeidsgruppe ˚ knes/Tafjord pros1 (Undersøking og overvaking). A jektet Report. Blikra, L. H., Anda, E., Høst, J. & Longva, O. 2006b. ˚ knes/Tafjordprosjektet: Sannsynlighet og risiko A ˚ knes og Hegknyttet til fjellskred og flodbølger fra A guraksla. Norges geologiske undersøkelse Report 2006.039, 20. Blikra, L. H., Longva, O., Harbitz, C. & Løvholt, F. 2005. Qualification of rock avalanche and tsunami hazard in Storfjorden, western Norway. In: Senneset, K., Flaate, K. & Larsen, J. O. (eds) Landslide and Avalanches ICFL 2005 Norway. Taylor & Francis, London. Braathen, A., Blikra, L. H., Berg, S. & Karlsen, F. 2004. Rock-slope failures of Norway; types, geometry, deformation mechanisms and stability. Norwegian Journal of Geology, 84, 67–88. Brideau, M.-A. & Stead, D. 2010. An investigation into the controls on block toppling using a 3-dimensional distinct element approach. Rock Mechanics and Rock Engineering, 43, 241–260. Brideau, M.-A., Yan, M. & Stead, D. 2009. The role of tectonic damage and brittle rock fracture in the development of large rock slope failures. Geomorphology, 103, 30–49. Cruden, D. M. & Varnes, D. J. 1996. Landslide types and processes. In: Turner, A. K. & Shuster, R. L. (eds) Landslides: Investigation and Mitigation. Special Report 247, Transportation Research Board, National Research Council, Washington, DC, 36–75. Derron, M. H., Blikra, L. H. & Jaboyedoff, M. 2005. High resolution digital elevation model analysis for ˚ kerneset, Norway). In: landslide hazard asessment (A Senneset, K., Flaate, K. & Larsen, J. O. (eds) Landslide and Avalanches ICFL 2005 Norway. Taylor & Francis, London, 101 –106. Eberhardt, E., Stead, D. & Coggan, J. S. 2004. Numerical analysis of initiation and progressive failure in natural rock slopes – the 1991 Randa rockslide. International Journal of Rock Mechanism, 41, 69– 87. Eiken, T. 2005. A˚knes Fotogrammetrisk undersøking av rørsle pa˚ grunnlag av flybilete 1961 og 1983 og ortho˚ knes/Tafjord Project Report. foto 2004. A Froese, C., Moreno, F, Jaboyedoff, M. & Cruden, D. 2009. 25 years of movement monitoring on South

Peak, Turtle Mountain: understanding the hazard. Canadian Geotechnical Journal, 46, 256 –269. Ganerød, G. V., Grøneng, G. et al. 2008. Geological ˚ knes rockslide, western Norway. Enginmodel of the A eering Geology, 102, 1– 18. Grøneng, G., Nilsen, B. & Sandven, R. 2009. Shear ˚ knes sliding area in western strength estimation for A Norway. International Journal of Rock Mechanics & Mining Sciences, 46, 479– 488. Henderson, I. H. C., Saintot, A. & Derron, M. H. 2006. Structural mapping of potential rockslide sites in the Storfjorden area, western Norway: the influence of bedrock geology on hazard analysis. Norges geologiske undersøkelse Report 2006.052, 1– 82. Hoek, E. & Bray, J. W. 1981. Rock Slope Engineering, revised 3rd edn. Institute of Mining and Metallurgy, London. Hutchinson, J. N. 1988. Morphological and geotechnical parameters of landslides in relation to geology and hydrogeology. In: Bonnard, Ch. (ed.) Landslides. Proceedings of the 5th International Symposium on Landslides, Volume 1. Lausanne. Balkema, Rotterdam, 3– 35. Jaboyedoff, M., Baillifard, F., Couture, R., Locat, J. & Locat, P. 2004. New insight of geomorphology and landslide prone area detection using DEM. In: Lacerda, W. A., Ehrlich, M., Fontoura, A. B. & Sayao, A. (eds) Landslides Evaluation and Stabilization. Balkema, Rotterdam, 199–205. Jaboyedoff, M., Metzger, R., Oppikofer, T., Couture, R., Derron, M.-H., Locat, J. & Turmel, D. 2007. New insight techniques to analyze rock-slope relief using DEM and 3D-imaging cloud points: COLTOP3D software. In: Eberhardt, E., Stead, D. & Morrison, T. (eds) Rock Mechanics: Meeting Society’s Challenges and Demands, Volume 1. Taylor & Francis, London, 61– 68. Jaboyedoff, M., Couture, R. & Locat, P. 2009. Structural analysis of Turtle Mountain (Alberta) using digital elevation model: toward a progressive failure. Geomorphology, 103, 5 –16. Kveldsvik, V., Eiken, T., Ganerød, G. V., Grøneng, G. & Ragvin, N. 2006. Evaluation of movement data and ˚ knes rock slide. In: Interground conditions for the A national Symposium on Stability of Rock Slopes in Open Pit Mining and Civil Engineering Situations, 3 April 2006. South African Institute of Mining and Metallurgy, Marshalltown, 279–299. Kveldsvik, V., Nilsen, B., Einstein, H. H. & Nadim, F. 2008. Alternative approaches for analyses of a 100,000 m3 rock slide based on Barton–Bandis shear strength criterion. Landslides, 5, 161–176; doi: 10.1007/s10346-007-0096-x. Kveldsvik, V., Einstein, H. H., Nilsen, B. & Blikra, L. H. 2009a. Numerical analysis of the 650,000 m2 ˚ knes rock slope based on measured displacements A and geotechnical data. Rock Mechanics and Rock Engineering, 42, 689– 728. Kveldsvik, V., Kaynia, A., Nadim, F., Bhasin, R., Nilsen, B. & Einstein, H. 2009b. Dynamic distinct˚ knes rock slope. element analysis of the 800 m high A International Journal of Rock Mechanics and Mining Sciences, 46, 686– 698.

˚ KNES LANDSLIDE STRUCTURES OF A Longva, O., Blikra, L. H. & Dehls, J. 2009. Rock avalanches – distribution and frequencies in the inner part of Storfjorden, Møre og Romsdal County, Norway. Norges geologiske undersøkelse Report 2009.002. Lo¨w, S., Frei, C., Tho¨ny, R., Evans, K. & Herfort, M. 2008. Complex groundwater hydraulics of a large active rockslide (Aknes, Norway). In: Abstracts of the Proceedings of the 33rd International Geological Congress, Oslo. http://www.cprm.gov.br/33IGC/ 1343823.html. Oppikofer, T. & Jaboyedoff, M. 2007. A˚knes/TafjordProject: DEM analysis of the Rundefjellet/Ta˚rnet area. University of Lausanne, Switzerland, Report IGAR-TO-005. Oppikofer, T. & Jaboyedoff, M. 2008. A˚knes/Tafjord project: Analysis of ancient rockslide scars and potential instabilities in the Tafjord area & Laser scanner monitoring of instabilities at Hegguraksla. University of Lausanne, Switzerland, Report IGAR-TO-009. Oppikofer, T., Jaboyedoff, M., Blikra, L. H. & Derron, M.-H. 2008. Characterization and monitoring ˚ knes landslide using terrestrial laser scanning. of the A In: Locat, J., Perret, D., Turmel, D., Demers, D. & Leroueil, S. (eds) 4th Canadian Conference on Geohazards: From Causes to Management, Que´bec, Canada, 20–24 Mai 2008. Presse de l’Universite´ Laval, Quebec, 211– 218. Oppikofer, T., Jaboyedoff, M., Blikra, L. H. & Derron, M.-H. 2009. Characterization and monitoring of ˚ knes rockslide using terrestrial laser scanning. the A Natural Hazards and Earth System Sciences, 9, 1003–1019. Oppikofer, T., Jaboyedoff, M., Pedrazzini, A., Blikra, L. & Derron, M.-H. in press. Detailed DEM analysis

161

of a rockslide scar to improve the basal failure surface model of active rockslides. Journal of Geophysical Research (Earth Surface). Ramsay, J. G. 1967. Folding and Fracturing of Rocks. McGraw-Hill Books, New York. Sartori, M., Baillifard, F., Jaboyedoff, M. & Rouiller, J.-D. 2003. Kinematics of the 1991 Randa rockfalls (Canton of Valais, Switzerland). Natural Hazards and Earth System Sciences, 3, 423– 433. Sjo¨berg, J. 2000. Failure mechanisms for high slopes in hard rock. In: Hustrulid, W. A., McCarter, M. K. & Van Zyl, D. J. A. (eds) Slope Stability in Surface Mining. Society for Mining, Metallurgy, and Exploration, Littleton, CO, 71– 80. Stead, D., Eberhardt, E. & Coggan, J. S. 2006. Developments in the characterization of complex rock slope deformation and failure using numerical modelling techniques. Engineering Geology, 83, 217– 235. Tveten, E., Lutro, O. & Thorsnes, T. 1998. Berg˚ lesund. 1:250,000 (A ˚ lesund, western grunnskart A Norway), NGU Trondheim (bedrock map). Varnes, D. J. 1978. Slope movement types and process. In: Schuster, R. L. & Krizek, R. J. (eds) Landslides: Analysis and Control. Special Report 176, Transportation Research Board, National Research Council, Washington, DC, 11– 33. Willenberg, H., Evans, K. F., Eberhardt, E., Spillmann, T. & Loew, S. 2008. Internal structure and deformation of an unstable crystalline rock mass above Randa (Switzerland): Part II – Three-dimensional deformation patterns. Engineering Geology, 101, 15–32. Zaruba, Q. & Mencl, V. 1969. Landslides and their Control. Elsevier, Amsterdam.

Structural analysis of Turtle Mountain: origin and influence of fractures in the development of rock slope failures ANDREA PEDRAZZINI1*, MICHEL JABOYEDOFF1, COREY R. FROESE2, C. WILLEM LANGENBERG3 & FRANCISCO MORENO2 1

Institut de Ge´omatique et d’Analyse du Risque – IGAR, AMPHIPOLE; University of Lausanne, CH-1015 Lausanne, Switzerland 2

Alberta Geological Survey/Energy Resources Conservation Board, 4999-98 Avenue, Edmonton, Alberta, T6B 2X3, Canada

3

Long Mountain Research Inc., 11439-56 Avenue, Edmonton, Alberta, T6H 0Y1, Canada *Corresponding author (e-mail: [email protected]) Abstract: Large slope failures in fractured rocks are often controlled by the combination of pre-existing tectonic fracturing and brittle failure propagation in the intact rock mass during the pre-failure phase. This study focuses on the influence of fold-related fractures and of postfolding fractures on slope instabilities with emphasis on Turtle Mountain, located in SW Alberta (Canada). The structural features of Turtle Mountain, especially to the south of the 1903 Frank Slide, were investigated using a high-resolution digital elevation model combined with a detailed field survey. These investigations allowed the identification of six main discontinuity sets influencing the slope instability and surface morphology. According to the different deformation phases affecting the area, the potential origin of the detected fractures was assessed. Three discontinuity sets are correlated with the folding phase and the others with post-folding movements. In order to characterize the rock mass quality in the different portions of the Turtle Mountain anticline, the geological strength index (GSI) has been estimated. The GSI results show a decrease in rock mass quality approaching the fold hinge area due to higher fracture persistence and higher weathering. These observations allow us to propose a model for the potential failure mechanisms related to fold structures.

The influence of structural features on the development of large rock slope instabilities has been investigated in several previous studies (Agliardi et al. 2001; Ambrosi & Crosta 2006; Brideau et al. 2009; Jaboyedoff et al. 2009). These authors emphasized the role of pre-existing fractures for the initiation of the large rock slope instabilities. Brideau et al. (2009) discussed the origin and the influence of rock mass damage and their roles on the development of instabilities, and highlighted the importance of tectonic structures not only on the three-dimensional (3D) geometry but also for the reduction in rock mass strength caused by induced damage. The damage is physically represented by weak zones in the rock mass, forming microcracks, mylonitic layers or void areas (Brideau et al. 2009). Turtle Mountain is located in the Crowsnest Pass in SW Alberta, approximately 5 km south of Blairmore (Fig. 1). This area became famous in 1903 after the destruction of the village of Frank by a large rock avalanche known as the Frank Slide.

McConnell & Brock (1903) described the event as a rockslide, and Cruden & Krahn (1973) reexamined the origin and the failure mechanisms and interpreted the Frank Slide as a rock avalanche. In the last few years the southern part of Turtle Mountain (Third Peak and South Peak areas) has become an important field laboratory where different techniques related to the characterization and monitoring of large slope mass movements are tested (Moreno & Froese 2007; Froese et al. 2009b). Moreover, a potential rockslide hazard has recently been re-emphasized in the South Peak area (Read et al. 2005; Froese et al. 2009a). The structural and the geological settings in the South Peak area were first studied in detail by Allan (1933). His work set the foundations for the present interpretation of the structure of the Frank Slide and South Peak area provided by Cruden & Krahn (1973). Langenberg et al. (2007) carried out a more detailed structural and geological investigation, and Cruden & Martin (2007) analysed the situation and the predisposing factors before the

From: Jaboyedoff, M. (ed.) Slope Tectonics. Geological Society, London, Special Publications, 351, 163– 183. DOI: 10.1144/SP351.9 0305-8719/11/$15.00 # The Geological Society of London 2011.

164

A. PEDRAZZINI ET AL.

Fig. 1. (a) Location of the Turtle mountain area. (b) Airborne LiDAR view of Turtle Mountain between North Peak and Hillcrest Mountain with the location of the geological cross-section presented in Figure 2.

1903 event. Others researchers have undertaken more specific analyses of the joint sets based on field survey (Couture 1998), as well as borehole logging analysis and ground penetrating radar (GPR) (Spratt & Lamb 2005) in the South Peak area. Recent studies (Sturzenegger et al 2007; Froese et al. 2009b; Jaboyedoff et al. 2009; Sturzenegger & Stead 2009) provide a structural analysis based on remote-sensing datasets such as airborne laser scanning (ALS), terrestrial laser scanning (TLS) and the derived digital elevation model (DEM). In this paper we first describe the structural characteristics of Turtle Mountain and then attempt to define the origin of the different detected discontinuity sets. Based on these data, we examine the influence of fold-induced joints on the 3D kinematic release of rock slope instability and their influences on the increase in rock mass damage near the fold hinge area.

Geology and geomorphology Turtle Mountain lithologies range in age from Devonian to Cretaceous (Norris 1993). The upper part of Turtle Mountain is composed of a fine- to coarse-grained, sparitic limestone. This important lithological unit is called the Livingstone Formation (Vise´an) and was the main unit involved in the 1903 event. The Palliser Formation constitutes the central part of the mountain and consists of fractured dolomitic limestone. The Banff Formation is transitional between the Livingstone and Palliser formations, and consists of shale, sandstone and cherty, argillaceous limestone. Outcrops belonging to the Mount Head Formation are present, in

particular in the western flank of the mountain and above the Turtle Mountain thrust (Langenberg et al. 2007). This formation is composed of an intercalation of coarse-grained sparitic limestone and fine-grained silty-dolomitic layers. Below the Turtle Mountain thrust, three Mesozoic geological units could be identified: the Fernie Formation, the Kootenay group and the Blairmore group. The Fernie Formation is mainly composed of shale and silty shale; the Kootenay group is composed of shale and siltstone, and contains the coal-bearing strata. The Blairmore group is composed of different detritic lithologies, especially sandstone and conglomerate. The main structure in the area is the Turtle Mountain anticline (Fig. 2). This fold can be described as a modified fault-propagation fold (Langenberg et al. 2007). In the North Peak area the Turtle Mountain fold is cut by the thrust, and only the western limb of the anticline remains. As the anticline has an asymmetrical shape plunging mainly to the west, the fold shape is not perfectly cylindrical. In the field its axis is difficult to follow owing to the important fracturing of the hinge area, and that is why, in two different zones, the fold axis has been deduced from the orientation of bedding planes (Fig. 3). According to Ramsay & Huber (1987), the fold is assumed to be at least locally cylindrical in order to reconstruct its axis orientation using the P-diagram method. From Drum Creek to the Third Peak area the axis plunges gently towards the south (38 towards 1818 in Drum Creek and 138 toward 1938 in the Third Peak area), but in the South Peak area it changes and plunges gently toward the north (28 towards 0248) (Langenberg et al. 2007). The fold axis trend also varies: in the Drum Creek area the trend

STRUCTURAL ANALYSIS OF TURTLE MOUNTAIN

165

Fig. 2. Cross-sections through the Turtle Mountain area. (a) Cross-section under the South Peak area. (b) Cross-section through the Frank Slide scar. Lithological units: 1, Livingstone Formation; 2, Mount Head Formation; 3, Fernie Formation; 4, Kootenay group; 5, Blairmore group; 6, Banff and Palliser formations.

direction is mainly north –south, whereas in the north it is predominantly oriented NW –SE. This observation is also supported by the large-scale view observed in the 1:50 000 geological map (Norris 1993).

Methods In this study the classical method of geological field survey (manual compass measurement of structures), is combined with semi-automated extraction of large-scale structures by means of a high-resolution DEM (HRDEM). The main advantage of using the HRDEM is that the time spent carrying out the field survey could be substantially reduced and investigations could be concentrated in the most representative areas. The structural analysis was performed in two separate areas corresponding to the two fold limbs (eastern and western fold limb). Using this approach, structural

features related to the two fold limbs could be differentiated.

DEM analysis Topographical analyses were performed using the software COLTOP 3D (Jaboyedoff & Couture 2003; Derron et al. 2005; Jaboyedoff et al. 2009). This software allows the representation of a DEM by a 3D shaded relief that displays the orientation of the slopes by means of a Schmidt –Lambert projection with one colour for a given dip and dip direction. It results in a coloured shaded relief map that combines slope and slope aspect in a unique representation where the slope orientation is coded by the Hue–Saturation– Intensity system (HSI).

Field survey The results of previous scanlines in the South Peak area (Spratt & Lamb 2005; Langenberg et al. 2007)

Fig. 3. Determination of the fold axis orientation using the P-diagrams method (after Ramsay & Huber 1987). (a) Results obtained for Drum Creek area. (b) Results obtained for the Third Peak area.

166

A. PEDRAZZINI ET AL.

were verified and completed. The field survey was performed in the southern part of Turtle Mountain (between the Third Peak and South Peak areas) where 25 structural stations have been performed. In addition, 20 structural stations were established in the Drum Creek –Hillcrest Mountain area. These were situated in an area where no rock instabilities or gravitational movements have been detected in order to minimize the potential influence of the Frank Slide movements on the identification of the initial fracturing. For each station the characteristics of the joint sets were measured and the rock mass condition described following the methodology suggested by the ISRM (1978).

by Cai et al (2004) that takes into account the block volume and the discontinuity spacing to describe the decreasing the interlocking of rock pieces and the joint condition factor to characterize the surface quality. However, as suggested by Marinos et al. (2005), the GSI quantification has to be applied with caution, and the visual observations and geological considerations still remain the best approaches. Each GSI value estimated in the field has been considered as the mean values for outcrops with an extension of 50 –100 m2.

Results Characteristics of the fold hinge area

Geological strength index In order to provide an objective estimation of the rock mass quality, the geological strength index (GSI) (Hoek & Brown 1997; Marinos et al. 2005) has been used. The GSI value is estimated taking into account the structural conditions (number of discontinuities, block geometry, persistent foliation) and the surface conditions (weathering, fracture infilling and roughness). As suggested by previous authors (Marinos et al. 2005), ranges of +5 were used for the GSI estimation in order to take account of the possible natural variations in the rock mass. The GSI estimation has been applied in 52 locations distributed in different parts of the anticline (Third Peak, South Peak, Drum Creek and Hillcrest Mountain). Even though the GSI is considered as a qualitative estimator of the rock mass quality based on geological observations, an attempt to quantify the GSI classification has recently been proposed (Cai et al. 2004, 2007). In this study, we employed the GSI chart proposed

The Turtle Mountain anticline is characterized by a disturbed hinge area (Fig. 4a). The general trend of the hinge could only be defined by distant observations, and it is difficult to follow clearly the bedding planes and the progressive transition from normal to inverted limb, especially close to the hinge. In addition, some local limb thrusts (Fig. 4a) along bedding planes close to the hinge area are suspected (Ramsay & Huber 1987). The inverted limb shows a steeper slope than the normal limb. Close to the hinge area, field survey indicates the presence of at least four discontinuity sets with frequently close spacing and medium– high persistence (ISRM 1978). Important local variations in spacing and persistence along the hinge zone are also recognizable. The fracturing is principally caused by compressive stress during the folding phase and by the presence of rock with different elastic characteristics (massive limestone v. silty-dolomite) in the stratigraphic sequence (Ramsay & Huber 1987). These differences in the

Fig. 4. (a) View of the Turtle Mountain anticline (Drum Creek area). The picture shows the asymmetric form of the Turtle Mountain anticline and the local uniformity in the anticline structure due to local thrusting. Differential erosion of less competent layers gives an important morphological signature. Evidence of rock instabilities located close to the hinge area are visible in the eastern mountain flank. (b) An example of rock weathering (Grade III– IV) and dissolution phenomena of limestone belonging to the Livingstone Formation in the Hillcrest area.

STRUCTURAL ANALYSIS OF TURTLE MOUNTAIN

mechanical behaviour also affect the general morphology of the mountain. Silty-dolomite layers belonging to the Mount Head Formation are generally more eroded and form significant depressions in both fold limbs, whereas competent limestone beds create positive relief. Dissolution phenomena are also abundant in the hinge area in coarse-grained limestone (Livingstone and Mount Head formations) as the numerous cavities and heavy weathered zones show (Fig. 4b).

Main lineaments The main lineaments extracted from the DEM and orthophotographs (283 observations) were entered into a database and represented in rose diagrams (58 interval). Turtle Mountain was divided into five sectors with the same lineament trends (Fig. 5). These lineaments are expressions of dipping planes and represent their apparent strikes. These planes correspond in the field either to composite tension cracks, trenches, scarps or transfer faults (following two or more tectonic joint sets) or to gravitational fractures following a predominant tectonic joint set. In all of the five areas a major lineament orientation NW– SE can be identified, whereas in the southern part the trend is mainly NNW–SSE, which changes progressively in the northern area to a WNW –ESE orientation. In the crest area, more particularly in the northern portion of Turtle Mountain, another predominant north–south direction is clearly defined.

Fractures in the eastern fold limb The slope morphology in the eastern limb of the anticline is mainly controlled by the orientation of the bedding planes, particularly those situated under Third Peak. Owing to the thick vegetation cover and extensive deposits of rockfall debris, it is difficult to identify structures in portions of this area. In this case the structural analysis using DEM became inefficient. Nevertheless, five main orientations were detected using COLTOP 3D between Drum Creek and South Peak (Fig. 6a). In the lower Third Peak area, field measurements are difficult and often not representative of the discontinuity sets owing to local fracture systems, toppled blocks, exfoliation and physical weathering (freeze –thaw cycles). Exfoliation and related weathering (that could reach locally grade IV) are much more developed in the dolimitic-siltstone of the Salter– Baril– Wileman members of the Mount Head Formation. In these rocks, exfoliation creates very–extremely closely spaced discontinuities parallel to the general outcrop orientation. Freeze– thaw cycles cause the joint to enlarge, especially the closely spaced exfoliation joints, and increase

167

the rock mass fragmentation (Matsuoka 2008). These effects are evident at the outcrop scale and affect the first 2–3 m of the rock mass. The typical block has columnar shape with a volume of between 100 and 500 cm3. Another factor is the occurrence of sedimentary anisotropy in the Livingstone Formation, such as cross-bed structures, that influence the measurements at local scale. Field surveys show the same trend as the DEM analysis (Fig. 6b), but with a greater dispersion. The majority of discontinuity sets are oriented toward the east, and J4 (2158/ 458) and S0 (1008/558) have been clearly identified in the two studies. Concerning discontinuity sets J1 (0208/458) and J2 (0608/558), the field data show an important variation, especially in dip angle. J3 is not visible on the outcrops but only locally on the DEM. Mean values of persistence, spacing, and primary and secondary roughness were measured on different structural stations following the ISRM (1978) suggestions (Table 1). Important variations in discontinuity characteristics have been observed in the different structural stations. During the field survey, no important water flows were observed along discontinuity sets, and the seepage rating for unfilled discontinuities could be described by classes II and III, but rarely IV (ISRM 1978). However, owing to the karstic hydrological system of the area, important water flow could be locally expected during important rainfall or during the snow-melting period. The most important characteristics visible in almost all outcrops are the presence of at least four different discontinuity sets, the medium– high persistence of the J1 set and the significant variation in the bedding thickness.

Fractures in the western fold limb The inclination and the position of the fold axis allow the outcropping of the western fold limb of the anticline, both in the upper eastern face (Frank Slide scar) and in the western face of Turtle Mountain. As suggested by Cruden & Krahn (1973), on the western face of Turtle Mountain the morphology is clearly influenced by the bedding orientation. Using the analysis performed by COLTOP 3D (Fig. 7a), the bedding plane fractures are visible in the western face and their orientations correspond principally to 2708/458. However, in the South Peak area, the bedding orientation changes according to the evolution of the fold axis. In this zone the orientation of the J6 set (3258/458) corresponds to a different discontinuity set parallel to the bedding, which could not be related to the bedding orientation elsewhere (Langenberg et al. 2007). In fact, in the North Peak area the bedding is mainly parallel to the J5–S0 (2708/458) discontinuity set. Two other discontinuities are well developed on

168

A. PEDRAZZINI ET AL.

Fig. 5. Main lineaments detected in different structural zones between North Peak and Third Peak.

STRUCTURAL ANALYSIS OF TURTLE MOUNTAIN

169

Fig. 6. Comparison between the discontinuity sets detected using (a) the COLTOP3D software and (b) field survey for the eastern fold limb (equal-area, lower-hemisphere projection).

the west side of the mountain: J4 (2108/508) and J5–S0. These are probably also encountered in the eastern face, but do not influence the slope morphology and cannot be detected using an airborne LiDAR (Light Detection And Ranging) dataset. The upper eastern face is dominated by the Frank Slide scar, which is controlled by three main orientations: J1 (0208/458), J2 (0558/658) and J3 (1158/ 508). Sets J3 and J4 show a fair amount of variation in dip and in dip direction (Fig. 7b). The mean characteristics of the different discontinuity sets derived from the compilation of all structural station are listed in Table 2. Between the Frank Slide scar and South Peak, discontinuity sets J3, J6 and J2 show high– very high persistence values (Sturzenegger & Stead 2009), which differ to the mean values reported in Table 2. This is probably related to the large gravitational movements that

occurred before and during the 1903 event. In fact, J2 and J3 have been actively involved in the progressive failure of the upper part of Turtle Mountain (Jaboyedoff et al. 2009). Discontinuity set J1 is the dominant set in the lower South Peak area (both in the eastern and western fold limbs) and also in the scar area of the 1903s slide where it shows medium– very high persistence. In the South Peak crown area, field measurements indicate that the J1 set is less frequent and shows a lower persistence. Primary and secondary roughness of the discontinuity seem to be fairly constant along the different structural stations, and could be defined as undulating (subordinate planar) for intermediate-scale and rough (only locally smoothed or slickensided) for small-scale observations. It is important to note that only a few structural stations (located in the Frank Slide scar area) record some slickenside on

Table 1. Summary of general discontinuity sets characteristics of the eastern fold limb based on field and DEM analysis. Spacing and persistence values in brackets could show important local variations Name (colour/ variation)

Dip direction (8)

Dip (8)

J1 (yellow, +108)

20

45

J2 (red, +108)

60

55

S0 (violet, +108)

100

55

J3 (light blue, +108)

135

50

J4 (dark blue, +108)

215

45

Persistence/ spacing (m)

Primary roughness

Secondary roughness

Medium – high persistence/1 – 2

Long wave undulation

Low –medium persistence/0.3– 1 High persistance/ 0.1– 3

Planar

Rough (dominant) Smooth (subordinate) Rough

Long wave undulation (folded) Planar

Rough (dominant) Slickensided (local) Smooth

Subplanar

Rough

Low persistence/ (0.5– 2) Low persistence/ (0.5– 1)

170

A. PEDRAZZINI ET AL.

Fig. 7. Comparison between the discontinuity sets detected using (a) the COLTOP3D software and (b) field survey for the western fold limb (equal-area, lower-hemisphere projection).

bedding planes (pitch 358W), which differ from the first observations on Turtle Mountain (Cruden & Krahn 1973), but it is in agreement with more recent studies (Langenberg et al. 2007). As suggested by Langenberg et al. (2007), slickensides are mainly localized along steep faults orientated north–south or NW–SE. As with the eastern fold limb, no important water flows were observed along the discontinuity sets. Infilling material of joint sets are often absent, especially for discontinuity sets J1 and J5. However, under the South Peak area, J1 fractures are unfilled or partially filled by frictional material (rock debris, coarse sand) indicating previous movements. Concerning discontinuity sets J2, J3 and J4, infilling is mainly composed of calcite mineralization. Locally,

approaching the hinge area, the karstification process progressively increases, especially along joint walls. It is also possible to observe the formation of calcite concretions. Infilling material along karstified joints is often absent or composed of residual insoluble material (iron deposit, clay).

Rock mass conditions On Turtle Mountain, the rock mass conditions change depending on local geological conditions (lithology, fracturation) and the location of the outcrops related to the anticline geometry. Figure 8 shows the location of the different GSI estimates along the Turtle Mountain anticline. Silty-dolomite outcrops of the Mount Head Formation generally

Table 2. Summary of general discontinuity sets characteristics of the western fold limb based on field and DEM analysis. Spacing and persistence values in brackets could show important local variations Name (colour/ variation)

Dip direction (8)

Dip (8)

Persistence/ spacing (m)

Primary roughness

J1 (yellow, +108)

20

45

Medium –high persistence/(1 – 2)

Long wave undulation

J2 (red, +108)

55

65

Planar

J3 (light blue, +158)

115

55

Stepped

Rough

J4 (dark blue, +158)

210

50

Planar

Rough

J5– S0 (violet, +108)

270

45

Planar

J6 (green, +108)

325

45

Low –medium persistence/ (0.1 –0.5) Medium persistence/ (0.5 –2) Low persistence/ (0.5 –1) High persistence/ (0.5 –1) Low persistence/ (0.2 –1)

Rough (dominant) Smooth (subordinate) Rough

Rough (dominant) Slickensided (local) Rough

Long wave undulation

Secondary roughness

STRUCTURAL ANALYSIS OF TURTLE MOUNTAIN

171

Fig. 8. Rock mass characterization using the GSI estimate: the location of the structural stations where GSI values were estimated (the red point corresponds to silty-dolomite outcrops).

172

A. PEDRAZZINI ET AL.

Fig. 9. Histogram of the GSI estimates obtained in the Hillcrest –South Peak area divided into three categories based on the structural location (only the limestone lithology has been used).

show a more important grade of weathering (generally between II and IV) and a closer spacing (close to extremely close) of bedding planes than the limestone belonging to the Livingstone Formation. The typical GSI values for the silty-dolomite (estimated in the eastern fold limb only) range between 35 and 45. Figure 9 shows the histogram of GSI estimates obtained for the western and the eastern fold limbs and for the hinge area of the Turtle mountain anticline. Along the two fold limbs limestone outcrops belonging to the Livingstone Formation show local important variations in rock mass quality. These variations are often related to the granulometry, to the presence of sedimentary anisotropy (mainly cross-bedding) and to previous gravitational movements. Fine-grained limestone typically exhibits higher GSI values than coarse-grained, well-graded limestone. This difference concerns the surface conditions and the joint condition factor (Jc), and it is probably linked to the porosity of the coarse-grained limestone – allowing faster water infiltration and increasing the surface dissolution. Block shape and size were also observed to be locally influenced by large gravitational structures (tension cracks, reactivated joints), and, as a consequence, the general rock mass conditions are also reduced. A clear example is represented by outcrops located between South Peak and North Peak. In this area, especially close to large cracks, the mean block volume is often smaller when compared

to outcrops where gravitational movements are less prevalent. Approaching the hinge area, both limestone lithologies show a decreasing trend in their GSI values. Common values for this zone range between 30– 40 and 45–55. In Figure 10 some examples of rock mass conditions for different outcrops in Turtle Mountain are reported. The potential extension of the hinge area was chosen according to previous work (Baillifard et al. 2003), indicating that large-scale structures, like faults or thrusts, induce local perturbations in the fracturing density and in the underground water flow within 75 –100 m of each side of the structural feature. For these reasons GSI stations within this distance were used to create the histogram representing the characteristics values for the hinge area.

Rock instabilities Several past and present-day instabilities involving limited volumes (generally ,100 000 m3) are observed along the hinge zone of the Turtle Mountain anticline. Similar processes involving foldrelated joint sets and related to slope failures have been investigated in the Rocky Mountains by different previous authors (Jones 1993; Badger 2002; Jackson 2002). Field observations of the ancient scars show that an intense fracturing related to the hinge area would have limited instabilities to relatively small

STRUCTURAL ANALYSIS OF TURTLE MOUNTAIN

173

Fig. 10. Examples of rock mass characteristics of different outcrops across Turtle Mountain. Each outcrop has been is described using the parameters proposed by Cai et al. (2004). For each outcrop co-ordinates are given in NAD 1983 UTM zone 32: (a) 687440/5494953; (b) 687027/5494964; (c) 687441/5494984; (d) 687029/5494987; (e) 687137/ 5493067; (f ) 687046/5492282.

volumes. Depending on the structural position on the anticline, different instability mechanisms could be observed. † Sliding–toppling (on extensional fracture J3) and wedge failures (essentially J2^J4) affecting

the western limb of the anticline but only in the eastern flank of the mountain. Toppling on bedding planes is kinematically possible but it was not observed during the field survey. † Step-like planar failure occurs along the bedding planes in the fold hinge zone. The prone areas for

174

A. PEDRAZZINI ET AL.

this mechanism are limited and are concentrated where the bedding is parallel to the main topography (dip-slope). In the upper part of the Third Peak area, large gravitational movements, such as toppling or wedge failures, are not kinematically feasible owing to rapid changes in the bedding orientation and the relatively small fold interlimb angle. Potential shallow instabilities related to the hinge zone are evident below Third Peak. Presently, the main unstable zone characterized with continued displacement, open gravitational cracks and rockfall activities is limited to the South Peak area (Moreno & Froese 2007).

Discussion Origin of detected discontinuities In order to understand the possible origin of detected discontinuity sets, the main tectonic phases affecting the studied area are reported below (Price & Carmichael 1986; Cooper 1992): † folding and thrusting phase, mainly east– west (Late Jurassic –Early Cretaceous); † right-lateral transpression phase, mainly NE– SW (Late Cretaceous – Paleocene); † left-lateral extension phase, mainly NW–SE (Early and Middle Miocene), and stress release during unloading between the Early Miocene and present. In addition to the chronological distinctions, it is possible to have the superposition of joints related to regional structural trends or to an individual geological structure. Currie & Reik (1977) showed that in the Rockies foothills it is possible to have the formation of fracture systems due to the local structural conditions (folds, faults) in addition to the regional trending fractures imposed by the large tectonic structure (anticlines, synclines, thrusts). The compilation of the data obtained by use of COLTOP3D and those measured in the field shows that four discontinuity sets (J1, J2, J3 and J4) appear in both limbs of the Turtle Mountain anticline. J6 and J5– S0 are only visible in the western fold limb. No discontinuity sets are folded, meaning that no pre-folding discontinuity sets appear to be identifiable in this area. Several previous studies (Stearns 1968; Price & Carmichael 1986; Cooper 1992) showed that the main fracturing can be related to the folding and thrusting. The development of fold-induced fractures depends on rheological conditions during the folding phase, such as strain rate and temperature (Price & Cosgrove 1990). In addition, local heterogeneities in the rock composition could induce fractures in

different orientations to those of the main foldinginduced fractures. In order to understand their possible origins, the orientation of discontinuity sets observed in the two fold limbs was compared to the classical relationship between fold and fractures proposed by Cooper (1992) for the Alberta foothills. The results concerning the western fold limb are presented in Figure 11a. It is possible to identify some similarities between the conceptual model proposed by Cooper (1992) and the DEM or field measurements, where we can observe that three joint orientations (J2, J3 and J4) are often related to fracturing during the folding phase (Fig. 11b). J3 could be interpreted as an extensional joint subparallel to the fold axis, and J2 and J4 could be interpreted as strike-slip conjugate faults with an acute bisectrix perpendicular to the fold axis (Muecke & Charlesworth 1966). In the western fold limb the strike-slip conjugate faults are only developed at a local scale; often only one of the two joint sets could be identified and this set tends to be subperpendicular to the fold axis. J5 and J1 are difficult to interpret using the fracturing models from Stearns (1968) or Cooper (1992). Concerning the eastern fold limb, a simple interpretation is more difficult to obtain owing to the low rock mass quality and the strong influence of gravitational movements affecting the true fracture orientation. The same problems concerning the interpretation of the fractures of the forelimb zone were pointed out by Cooley (2007) during his studies of the southern Livingstone range. We can speculate that conjugate strike-slip faults occur, but large-scale extensional fractures are probably missing owing to the compression stress encountered in the lower fold limb, which is overturned in the southern part of Turtle Mountain. The same origin could also be proposed for discontinuity set J5. This joint is well developed in the western flank and controls the erosion, as well as the channel incision. In general, the J5 set is very close to bedding and its origin could also be linked to postfolding movements and be associated to the J1 set as a conjugate joint. This assumption is in agreement with the reconstruction of the mean stress direction influencing the area after the folding phase. The NW–SE extension phase and the stress released during the thermal relaxation period have probably played an important role for the reactivation of syn-folding structures, like J2 and J3. This extension has most likely enabled the development of the fractures in the two fold limbs. Other indications of post-folding movements can be seen in the lower part of Third Peak. In this area the presence of a post-folding fault oriented approximately north– south following the pre-existing bedding plane orientation with an estimated displacement

STRUCTURAL ANALYSIS OF TURTLE MOUNTAIN

175

Fig. 11. Comparison between the relationship of fold and fractures (a) observed in the western Turtle Mountain anticline and (b) the general model proposed by Cooper (1992).

of about 10 m can be assumed. Normal and strikeslip faults with similar directions have been observed around South and North Peak (Langenberg et al. 2007). This suggests that the presence of different phases of movement in the same area can induce a reactivation and reorientation of the preexisting joints. Slickensides that could confirm or reject previous hypotheses are often absent owing to the subsuperficial pressure–temperature regime during the last movement phases and to the dissolution phenomenon along the calcareous joint walls. Another important reactivation of pre-existing discontinuity sets is probably related to recent gravitational movements. In fact, the orientation and the opening of cracks present in the Turtle Mountain area are influenced by tectonic discontinuity sets such as those suggested in Figure 12. The comparison of the dip directions of the discontinuity sets (Fig. 12a) with the lineament directions (Fig. 12b) for the whole Turtle Mountain area shows a close fit between the main discontinuity dip direction and the lineament strike, confirming the significant structural control of tectonic-related joints on the slope deformation. It is interesting to note that lineaments follow the NW–SE discontinuity (generally

in the J1 or J2 dip direction) and also the NE–SW direction corresponding to the J3 discontinuity set. In the upper South Peak the lineament orientation highlights the irregular (saw-tooth) tension crack disposition following two or more pre-exiting discontinuity sets. In the field, manual estimations of displacement directions have been attempted along large cracks in the lower South Peak area and also indicate a general movement plunging 108 toward NE.

Interactions between fold structure and slope instabilities In this section we analyse the structure and rock mass conditions in the hinge area of the Turtle Mountain anticline in relation to the past and potential future instabilities. The primary influence of joints sets in rock slope instabilities is defined by the failure mechanism and the limits of the unstable volume (Hoek & Bray 1981). As a result of tectonic features (shear zones, fault zones and fold hinge areas), the fracturing increases and the rock mass strength decreases (Brideau et al. 2009).

176

A. PEDRAZZINI ET AL.

Fig. 12. Rose diagrams (58 interval) comparing (a) the discontinuity set orientation and (b) the crack orientation between South Peak and the central part of the Frank Slide crown area.

The fold hinge area represents an important structural feature where strain and related fracturing are often important (Ramsay & Huber 1987; Price & Cosgrove 1990; Badger 2002). As a consequence, significant damage induced by extensional stress in the fold hinge could be expected. In these conditions the deformed rock mass is more susceptible to weathering and the development of varied types of failure mechanisms. Influence of fold-related fractures in the 3D kinematic release. The morphology of Turtle Mountain is dominated by the Frank Slide scar. The failure mechanisms of the 1903 event have been studied by several different authors (Cruden & Krahn 1973; Benko & Stead 1998; Jaboyedoff et al. 2009). The influence of fold-related fractures on failure mechanisms has been indirectly suggested by Jaboyedoff et al. (2009). In their model the fold-induced joint J2 is the rear limit of the toppled wedge formed by regional fault J1 and fold-induced fracture J3. Kinematic analyses of sliding, toppling and wedge failure mechanisms were performed in order to verify the failure mechanisms observed in the field (Fig. 13). A topographical attitude of 1108/508corresponding to the mean value near the hinge area has been used as reference topography. A conservative friction angle of 308 has been chosen according to the worst discontinuity conditions observed in the field (roughness and weathering) and assuming that locally previous movements occurred along the discontinuity sets; the resulting shear strength will be close to the residual. The kinematic tests confirm these field observations concerning planar sliding on bedding,

sliding–toppling on J3 and wedge sliding (J2^J4 but locally also J2^J3, J1^J3) mechanisms affecting the western fold limb, essentially in the eastern flank of the mountain. In addition, the presence of regional-scale post-folding fractures (J1, J5) showing locally a very high persistence and crossing the entire anticline introduces another important instability factor not directly related to the presence of fold-related fractures. In the hinge zone the presence of wedges, formed by J1 and J5 sets, allow the back and lateral release required for planar sliding along bedding planes. The influence of J1 set is also very important in the lower portion of South Peak, where it defines six large unstable areas, with volumes varying between 0.15 and 5 Mm3 (Froese et al. 2009a). Influence of fold-related fractures and karstification in the reduction of the rock mass strength. The evolution of rock mass quality along the Turtle Mountain anticline has been investigated by analysing the variations in GSI values obtained in the different structural stations. Figure 9 shows that GSI values in the two fold limbs are quite similar. However, values for the eastern fold limb have a tendency to be more dispersed compared to the values in the western fold limb. This is probably related to the post-folding movements (a fault subparallel to the bedding planes have been detected under the Third Peak area) and recent gravitational movements. The GSI values in the hinge area show a large dispersion and mean values clearly lower than the GSI values in the two fold limbs. In order to explain this trend, two main factors must be taken into account. First, the fold-related joint sets, especially the extensional joint (J3), tend to be

STRUCTURAL ANALYSIS OF TURTLE MOUNTAIN

177

Fig. 13. Instability mechanisms related to the fold hinge area. (a) Kinematic analysis of planar sliding for discontinuity sets detected close to the fold hinge (equal-area, lower-hemisphere projection). (b) Kinematic analysis of wedge sliding involving J2 and J4 discontinuity sets, affecting the western fold limb (equal-area, lower-hemisphere projection). (c) Kinematics analysis of the toppling mechanism involving the S0 discontinuity set affecting the western fold limb (equal-area, lower-hemisphere projection). (d) Schematic representation of failure mechanisms and their location along the Turtle Mountain anticline.

more persistent and open approaching the hinge. Referring to the GSI classification, there is a progressive evolution from a very blocky structure with multifaced angular blocks towards a blocky/ disturbed rock mass. This evolution can be followed in Figure 8, in particular in parts (a) –(d). Secondly, a decrease in structure quality also involves a decrease in surface conditions. In fact, along open joints in the hinge area it is possible to see a significant karstification involving the creation of cavities (Figs 4b and 8). Moreover, along joints an important calcite recrystallization can be observed indicating previous fluid flows. In addition, a general decrease in surface condition (reaching locally grade III –IV) could be observed approaching the hinge area (Fig. 4b). Using the GSI chart (Fig. 14), the influence of the increase in joint persistent and decreasing surface conditions in the hinge area can be illustrated. The vector A could be decomposed in two components: the vertical vector B corresponding to an augmentation of fracturing (in our specific case, the increasing of persistence); and the horizontal vector C corresponding to a decrease in surface

conditions such and an increasing of joint weathering related to karstification. Persistence is not directly taken into account by the GSI chart proposed by Cai et al. (2004). However, assuming the presence of the same number of joint sets in the fold limb and in the hinge area, an increase in the joint persistence could decrease the block size. The relationship between joint set persistence and block size is more complex, as suggested by Kim et al. (2007), and the correlation with vector B (Fig. 14) must be interpreted in a qualitative manner. The influence in the strength reduction induced by karstification cavities is difficult to characterize using the GSI chart, as both GSI parameters are directly or indirectly affected. Concerning the joint condition factor (Jc) factor, karstification influences mainly the joint alteration factor (Ja). In fact, when karstification occurs in calcareous rock, calcite is dissolved but often impurities remained (clay, quartz particles) resulting in a local decrease in rock mass strength. Joint walls are also influenced (macroscopic voids, calcite) and often show a higher alteration than that of the intact rock.

178

A. PEDRAZZINI ET AL.

Fig. 14. Factors influencing the decreasing GSI values approaching the hinge area plotted in the GSI evaluation chart (after Marinos et al. 2005). Vector B corresponds to an increase in fracturation and vector C corresponds to an increase in rock weathering. The composition of these two vectors corresponds to vector A.

STRUCTURAL ANALYSIS OF TURTLE MOUNTAIN

Large- (Jw) and small-scale roughness (Js) are only marginally influenced because, during karstification, no shear movements occurred along joint sets. At the same time, especially in porous coarsegrained limestone, water infiltration and related karstification decrease the rock strength promoting the rock mass disintegration and decreasing the stability. Cruden & Martin (2007) estimated the GSI value for each geological unit based on Norris’s map (Norris 1993) and obtained a GSI value of 80 for the Livingstone Formation, which appears much higher compared to the ones in the present study. This discrepancy is principally due to the difference in the analysis scales. Our work focused on variations of GSI along the anticline and we describe the rock mass conditions at the outcrop scale (50–100 m2). At this scale the rock mass characteristics are influenced by local heterogeneities (lithological and structural) and by the climate effects (freeze –thaw cycles, exfoliation). This means that outcrops used for the estimation of GSI often show lower values than estimations performed in excavated slope and tunnel (Marinos et al. 2005). However, based on the outcrop conditions and suggestions proposed by Marinos et al. (2005) to project GSI values into the ground, a mean GSI value for the Livingstone Formation of around 60– 70 seems to be more realistic for describing the general condition of the Turtle Mountain anticline, especially the fold limb conditions (Fig. 14). For the hinge area the projected GSI values into the ground will probably be lower (40–60) and more scattered depending on the relative position of the neutral surface and on the behaviour of the involved lithology. The suggested projection of GSI information from observations in outcrops to depth is reported in Figure 14.

General model for fold-related instabilities Based on studies performed on the Turtle Mountain anticline, it is possible to propose a general relationship between the localization and the specific characteristics of the fold and the potential failure mechanisms affecting the different portions of the mountain. Two main parameters have been chosen for the failure mechanism description in the different portions of the mountain: the interlimb angle; and the angle between the fold axis and the mountain axis. The interlimb angle describes the minimum angle between two fold limbs measured in the profile section (Price & Cosgrove 1990). In other words, the interlimb angle gives an indication of the tightness of the fold. The other important parameter is defined by the angle between the mean vertical axis of the mountain and the axial plane.

179

This angle determines the relative position of the hinge zone in the mountain limb. According to the relative position of the hinge for different stability situations, we define three limit angles of 158, 908 and 1808. The mountain slope angle plays a fundamental role in the effective kinematic feasibility of the different failure mechanisms (Cruden 2003). In order to simplify the description of the theoretical potential failure mechanisms, we assume a mountain flank slope that is greater than the basal friction angle along the different discontinuity sets. Based on previous studies (Cruden & Hu 1999), an angle of 308 was selected as an appropriate value for the basal friction angle of carbonate rocks. Figure 15 presents the nine different geometrical situations that could be obtained by combining the tightness of the fold (interlimb angle) and the different angles between the mountain’s mean axis and the fold axis. For each situation the potential failure mechanisms involving the bedding planes and the extensional fractures have been analysed. Conjugate faults related to folding have not been taken into account in the general model because of their local extension and development. The presence of extensional joints is related to the fold mechanism (flexural slip, flexural flow and tangential longitudinal plane), the fold type, and the strain and temperature conditions during folding (Ramsay & Huber 1987; Price & Cosgrove 1990). In this study we focus on parallel folds where the bed thickness measured perpendicularly remains constant throughout the fold and assume analogous mechanical behaviour for the different beds involved during the folding phase. Parallels folds correspond to the typical shape issued from a flexural-slip mechanism (Ramsay & Huber 1987), and this appears to be the prominent process in the Rocky Mountain foothills and front ranges (Langenberg 1992). Other fold shapes, such as similar folds, imply different folding mechanisms/conditions and the development of other types of discontinuities (cleavage schistosity, crenulations, etc.). From a qualitative point of view, the failure mechanisms affecting the different portions of the mountain could easily be assessed using the chart proposed by Cruden (2003) for anaclinal and cataclinal slopes. This chart could be applied to the different configurations presented in Figure 15, assuming a correlation between kathetal (normal to bedding) joints and extensional joints related to the folding phase. Generally, a more stable configuration is reached when the angle between the axial plane and the mountain axis is close to 1808. In this configuration the mountain is only affected by a potential toppling mechanism. According to previous observations in the Canadian Rockies (Gadd 1986), another favorable situation is represented by an angle close to

180

A. PEDRAZZINI ET AL.

Fig. 15. General model of potential fold-related instabilities. The potential failure mechanisms in the different mountain areas depend on the interlimb angle and the angle between the fold axis and the mountain axis. F symbolizes the basal friction angle.

STRUCTURAL ANALYSIS OF TURTLE MOUNTAIN

181

908 for tight folds. In this configuration bedding is near horizontal. Gadd (1986) indicated that the higher peaks in the Canadian Rockies have this type of configuration. The folds with a tight interlimb angle present a very limited hinge area where instabilities seem to affect only limited volumes. In addition, failures mechanisms tend to be the same along a given mountain flank because of subparallel bedding in the two fold limbs (constant anaclinal or cataclinal configuration). Folds with a more open interlimb angle show different failure mechanisms in the same flank. In fact, within the same flank, it is possible to have a change between a cataclinal and an anaclinal situation. Larger interlimb angles also signify a larger hinge area with its associated failure mechanisms.

The authors wish to thank F. Humair for his assistance in the field. C. Longchamp, T. Oppikofer (University of Lausanne) and M.-A. Brideau (Simon Fraser University) improved the manuscript and shared some interesting discussion. The manuscript benefited from the careful review by two anonymous reviewers.

Conclusions

References

The combination of a structural study performed using high-resolution DEM analysis with a detailed field survey has determined the main structural sets present at Turtle Mountain. At least three discontinuity sets (J2, J3 and J4) seem to be related to the folding phase. Regional post-folding movements must also be taken into account in order to understand the origin of the J1 and J5 sets. The fold-induced fracturing and the related rock mass damage in the hinge area seem to be the most significant elements to explain the presence of frequent small instabilities. The influence of fold-related damage is an important factor not only for local-scale instabilities but also in the development of large rock slope failures. Particularly in a carbonate environment, the dissolution along fractures could play an important role in decreasing the rock mass strength and in increasing the water pressure along discontinuities. As suggested by Cruden & Martin (2007), water infiltration along karst-widened joints represented a major triggering factor for the Frank Slide. The GSI classification represents an interesting tool in characterizing the variation in rock mass quality variation along large structural features, such as a fold thrust or large fault. However, in order to have a reliable quantitative description of the rock mass strength variation more rigorous field investigations are necessary. In particular, a clear definition of the outcrop’s extension considered for the GSI estimation and a homogeneous coverage of the study area (regular distribution of the GSI estimation along the slope) represent the main challenges when using this technique. Unfortunately, at Turtle Mountain, homogenous cover is difficult to perform owing to the difficulty in the field access. A general model based on the location and the tightness of the fold in the mountain has been

Agliardi, F., Crosta, G. B. & Zanchi, A. 2001. Structural constraints on deep-seated slope deformation kinematics. Engineering Geology, 59, 83–102. Allan, J. A. 1933. Report on stability of Turtle Mountain, Alberta and survey fissures between North and South Peak. Department of Public Works, Edmonton, Alberta. Provincial Archives. Ambrosi, C. & Crosta, G. B. 2006. Large sackung along major tectonic features in the Central Italian Alps. Engineering Geology, 83, 183–200. Badger, T. C. 2002. Fracturing within anticlines and its kinematic control on slope stability. Environmental and Engineering Geoscience, 8, 19–33. Baillifard, F., Jaboyedoff, M. & Sartori, M. 2003. Rockfall hazard mapping along a mountainous road in Switzerland using a GIS-based parameter rating approach. Natural Hazards Earth Systems Science, 3, 435– 442. Benko, B. & Stead, D. 1998. The Frank Slide: a re-examination of the failure mechanism. Canadian Geotechnical Journal, 35, 299–311. Brideau, M-A., Ming, Y. & Stead, D. 2009. The role of tectonic damage and brittle rock fracture in the development of large rock slope failure. Geomorphology, 103, 30–49. Cai, M., Kaiser, P. K., Tasaka, H. & Minami, M. 2007. Determination of residual strength parameters of jointed rock masses using the GSI system. International Journal of Rock Mechanics & Mining Sciences, 44, 247–265. Cai, M., Kaiser, P. K., Uno, H., Tasaka, Y. & Minami, M. 2004. Estimation of rock mass deformation modulus and strength of jointed hard rock masses using the GSI system. International Journal of Rock Mechanics & Mining Sciences, 41, 3 –19. Cooley, M. A. 2007. The structural, thermal, and fluids evolution of the Livingstone Range Anticlinorium, and its regional significance to the Southern Alberta foreland thrust and fold belt. PhD thesis, Queen’s University. Cooper, M. 1992. The analysis of fracture systems in subsurface thrust structures from the Foothills of the

proposed to show the potential rock slope failures induced by the presence of fold and the fold-related fractures. This model represents a specific evolution of the model proposed by previous studies (Cruden & Hu 1999; Cruden 2003). The localization and opening of the anticline on the mountain influence the development of small or larger rock slope failures, and represents an important parameter in the assessment of hazards resulting from large rock slope failure.

182

A. PEDRAZZINI ET AL.

Canadian Rockies. In: McClay, K. R. (ed.) Thrust Tectonics. Chapman & Hall, London, 391–405. Couture, R. 1998. Contributions aux aspects me´caniques et physiques des e´croulements rocheux. PhD thesis, Laval University. Cruden, D. M. 2003. The shapes of cold, high mountains in sedimentary rocks. Geomorphology, 55, 249– 261. Cruden, D. M. & Hu, X.-Q. 1999. The shapes of some mountain peaks in the Canadian Rockies. Earth Surface Processes and Landforms, 24, 1 –13. Cruden, D. M. & Krahn, J. 1973. A re-examination of the geology of the Frank Slide. Canadian Geotechnical Journal, 10, 581–591. Cruden, D. M. & Martin, D. K. 2007. Before Frank slide. Canadian Geotechnical Journal, 44, 765– 780. Currie, J. B. & Reik, G. A. 1977. A method of distinguishing regional directions of jointing and of identifying joint sets associated with individual geological structures. Canadian Journal of Earth Sciences, 14, 1211–1228. Derron, M.-H., Jaboyedoff, M. & Blikra, L. H. 2005. Preliminary assessment of rockslide and rockfall hazards using a DEM (Oppstadhornet, Norway). Natural Hazards and Earth System Science, 5, 285– 292. Froese, C., Jaboyedoff, M., Pedrazzini, A., Hungr, O. & Moreno, F. 2009a. Hazard mapping for the eastern face of Turle Mountain, adiacent to the Frank Slide, Alberta, Canada. In: Malet, J. P., Remaıˆtre, A. & Bogaard, T. (eds) Proceedings of the International Conference on Landslide Processes. Geomorphologic Mapping to Dynamic Modeling Conference. A tribute to Prof. Theo van Asch. CERG Editions, Strasbourg. Froese, C., Moreno, F., Jaboyedoff, M. & Cruden, D. M. 2009b. 25 years of movement monitoring on the South Peak of Turtle Mountain: understanding the hazard. Canadian Geotechnical Journal, 46, 256– 269. Gadd, B. 1986. Handbook of the Canadian Rockies. Corax Press, Alberta. Hoek, E. & Bray, J. 1981. Rock Slope Engineering. Spon Press, London. Hoek, E. & Brown, E. T. 1997. Practical estimates of rock mass strength. International Journal of Rock Mechanics and Mining Sciences, 34, 1165– 1186. INTERNATIONAL SOCIETY FOR ROCK MECHANICS (ISRM). 1978. Suggested methods for the quantitative description of discontinuities in rock masses. International Journal of Rock Mechanics and Mining Sciences & Geomechanics Abstracts, 15, 319– 368. Jaboyedoff, M. & Couture, R. 2003. Report on the project COLTOP3D for March 2003: stay of Michel Jaboyedoff at GSC - Ottawa. Quanterra Administrative Document – Activity Report, RA01. Jaboyedoff, M., Couture, R. & Locat, P. 2009. Structural analysis of Turtle Mountain (Alberta) using digital elevation model: toward a progressive failure. Geomorphology, 103, 5– 16. Jackson, E. L. 2002. Landslide and landscape evolution in the Rocky Mountains and adjacent Foothills area, southwestern Alberta, Canada. In: Evans, S. G. &

DeGraff, J. V. (eds) Catastrophic Landslides: Effects, Occurrence, and Mechanisms. Geological Society of America, Reviews in Engineering, 15, 325–344. Jones, P. B. 1993. Structural geology of the modern Frank Slide and ancient Bluff Mountain Slide, Crownest, Alberta. Bulletin of Canadian Petroleum Geology, 41, 232– 243. Kim, B. H., Cai, M., Kaiser, P. K. & Yang, H. S. 2007. Estimation of block sizes for rock masses with nonpersistent joints. Rock Mechanics and Rock Engineering, 40, 169– 192. Langenberg, W. 1992. Styles of deformation in the Rocky Mountain foothills and front ranges, Alberta, Canada. AAPG Bulletin, 76. Langenberg, C. W., Pana, D., Richards, B. C., Spratt, D. A. & Lamb, M. A. 2007. Structural Geology of the Turtle Mountain area near Frank, Alberta. EUB/ AGS Science Report, 2007– 01. World Wide Web Address: http://www.ags.gov.ab.ca/publications/ ABSTRACTS/ESR_2007_03.html. Marinos, V., Marinos, P. & Hoek, E. 2005. The geological strength index: applications and limitations. Bulletin of Engineering Geology and the Environment, 64, 55–65. Matsuoka, N. 2008. Frost weathering and rockwall erosion in the southeastern Swiss Alps: Long-term (1994– 2006) observations. Geomorphology, 99, 353–368. Mcconnell, R. G. & Brock, R. W. 1903. Report on the Great Landslide at Frank, Alberta, Department of the Interior, Annual Report for 1903. Ottawa, Part 8. Edmonton Geological Society, 2003. Moreno, F. & Froese, C. R. 2007. Turtle Mountain Field Laboratory Monitoring and Research Summary Report, 2005. Earth Sciences Report, 2006. World Wide Web Address: http://www.ags.gov.ab.ca/ publications/ABSTRACTS/ESR_2006_07.html. Muecke, G. K. & Charlesworth, H. A. H. 1966. Jointing in folded Cardium Sandstones along the Bow River, Alberta. Canadian Journal of Earth Science, 3, 579–596. Norris, D. K. 1993. Geology and Structure Crosssections, Blairmore (West Half), Alberta. Geological Survey of Canada, Map 1829A, scale 1:50,000. Geological Survey of Canada, Ottawa. Price, R. A. & Carmichael, D. M. 1986. Geometric test for Late Cretaceous– Paleogene intracontinental transform faulting in the Canadian Cordillera. Geology, 14, 468–471. Price, N. J. & Cosgrove, J. W. 1990. Analysis of Geological Structures. Cambridge Press, Cambridge. Ramsay, J. G. & Huber, M. I. 1987. The Techniques of Modern Structural Geology, Volume 2. Academic Press, London. Read, R. S., Langenberg, W. et al. 2005. Frank Slide a century later; the Turtle Mountain Monitoring Project. In: Hungr, O., Fell, R., Couture, R. R. & Eberhardt, E. (eds) Proceedings of the International Conference Landslide Risk Management. Vancouver, Balkema, Leiden, 713 –723. Spratt, D. A. & Lamb, M. A. 2005. Borehole data interpretation and orientation, Turtle Mountain

STRUCTURAL ANALYSIS OF TURTLE MOUNTAIN Project. Alberta Municipal Affairs, Internal Report of Work Package, WP15B. Stearns, D. W. 1968. Certain aspects of fractures in naturally deformed rocks. In: Riecker, R. E. (ed.) Rock Mechanics Seminar. Terrestrial Sciences Laboratory, Bedford, MA, 97–118. Sturzenegger, M. & Stead, D. 2009. Quantifying discontinuity orientation and persistence on high mountain rock slopes and large landslides using

183

terrestrial remote sensing techniques. Natural Hazards and Earth System Sciences, 9, 267– 287. Sturzenegger, M., Stead, D., Froese, C., Moreno, F. & Jaboyedoff, M. 2007. Ground based and airborne LiDAR for structural mapping of the Frank slide. In: Proceedings of the 1st Canada– US Rock Mechanics Symposium. Vancouver, Canada, 27–31 May 2007. Taylor & Francis, London, 925–932.

A structural, geomorphological and InSAR study of an active rock slope failure development I. H. C. HENDERSON1*, T. R. LAUKNES2, P. T. OSMUNDSEN1, J. DEHLS1, Y. LARSEN2 & T. F. REDFIELD1 1

Geological Survey of Norway, Leif Eirikssons Veg 39, 7491 Trondheim, Norway 2

Norut, P.O. Box 6434, 9294 Tromsø, Norway

*Corresponding author (e-mail: [email protected]) Abstract: Few studies of rockslides have addressed the relationships between structures, geomorphological expression and direct evidence for movement. We employ structural geology, geomorphology and interferometric synthetic aperture radar (InSAR) to investigate the evolution of the surface features developed in response to movement of the Gamanjunni rockslide site in Troms County in northern Norway. The slide is located on a west-facing mountainside, and is bounded by two angled back scarps and a 208 –308 basal sliding plane. The volume is estimated at 24 Mm3 and is therefore among the largest potential rockslides in Norway. InSAR provides a new method to measure the movement of potential rockslides, and thus provides a direct link between qualitative movement data and field observations. We document the relationship between variations in ground movement rates and changing back-scarp geomorphology at the Gamanjunni site as well as movement patterns within the incipient rockslide. We demonstrate that variations in InSAR documents millimetre variations in scarp displacement and that this is reflected in the evolving back-scarp geometry. We conclude that InSAR can provide important information to complement field observations. The ability of InSAR to document landslide movement patterns greatly extends our knowledge of back-scarp evolution and active landslide processes.

In recent years structural geology has been applied to investigate the evolution of potential rockslides (e.g. Chigira 1992; Agliardi et al. 2001; Braathen et al. 2004; Ambrosi & Crosta 2006). Recent studies have concentrated on documenting rockslide kinematics and identifying geometric configurations particularly susceptible to sliding (e.g. Braathen et al. 2004; Henderson & Saintot 2011; Saintot 2011). Less emphasis has been placed on the observed movements, and on the evolution of the structural and geomorphical architecture (see Clifton et al. 2003; Tarchi et al. 2003; Canuti et al. 2005; Catani et al. 2005; Saroli et al. 2005; Colesanti & Wasowski 2006; Ganerød et al. 2008). Integrated studies using geological, geomorphological and newly developed remote-sensing techniques are relatively few (e.g. Berardino et al. 2003; Jaboyedoff et al. 2004; Squarzoni et al. 2005). Reconciliation between field observations and conventional measuring techniques such as global positioning satellite (GPS) and total station measurements is often ambiguous. Such quantification is, however, a necessary step in hazard and risk assessment (Solheim et al. 2005). The interferometric synthetic aperture radar (InSAR) satellite technology (Gabriel et al. 1989) provides a new technique to determine rockslide movement,

and, therefore, a direct link between quantitative ground movement data and structures, kinematics and changes of slope. The direct linkage of InSAR data to structural geology and sliding processes is a novel niche in landslide research. Below, we describe a case study where we combine InSAR with structural geology and geomorphology in the assessment of sliding processes on the Gamanjunni site in Troms County, Norway.

InSAR methodology The relatively recent methodology of surveying land deformation by using the space-borne synthetic aperture radar (SAR) instrument has gained wide interest. The satellites carrying SAR instruments have a repeat cycle of 25 –35 days, giving a unique opportunity to perform long-term monitoring of selected areas. SAR interferometry (InSAR) is a technique that exploits the phase difference between two complex SAR images, captured at different times (e.g. Fruneau et al. 1996; Massonnet & Feigl 1998; Rosen et al. 2000; Lauknes et al. 2005). By using the phase information, relative distances can be measured with an accuracy that is within a fraction of a wavelength. Deformation

From: Jaboyedoff, M. (ed.) Slope Tectonics. Geological Society, London, Special Publications, 351, 185– 199. DOI: 10.1144/SP351.10 0305-8719/11/$15.00 # The Geological Society of London 2011.

186

I. H. C. HENDERSON ET AL.

mapping using InSAR has been proven capable of giving relative ‘line of sight’ (LOS) deformation measurements with accuracy of the order of millimetres– centimetres (Lauknes 2004). Lately, the concept of combining InSAR information from a large number of SAR images, thus allowing the creation of a deformation time series, has been introduced (e.g. Ferretti et al. 2001; Berardino et al. 2002). By using a multi-temporal technique an estimated time series of the LOS deformation in each location is created. In addition, by using large data stacks, artefacts due to atmospheric water vapour and orbital errors may be mitigated (Zebker et al. 1997; Hanssen 2001). We have used the InSAR as a tool to identify new rockslide sites as part of a government-funded regional programme (Henderson et al. 2007). To determine the detailed activity, evolution and magnitude of displacement of the Gamanjunni slide in Troms, Norway (Fig. 1), we conducted a detailed InSAR measurement and analysis campaign. Below we briefly describe the important processing steps relevant for this study. Data were extracted from ERS-1/2 SAR in the 1992–1999 timeframe. For the InSAR analysis we used a total of 19 (snow-free) ERS-1/2 SAR scenes from a descending orbit (Track 251, Frame 2196), covering the period 1992–1999 (Larsen et al. 2005). From these scenes we computed 60

interferograms with a maximum baseline of 300 m and a maximum temporal separation of 4 years. The overall phase coherence is very good as a result of the limited vegetation cover at elevations higher than 600–700 m above sea level. For each interferogram an orbital phase ramp, as well as phase delay due to tropospheric stratification, was estimated and removed. Finally, based on the 60 InSAR pairs, we applied the small-baseline subset (SBAS) algorithm in order to estimate and separate a mean displacement velocity field from phase contributions caused by errors in the digital elevation model (DEM) and heterogeneous atmospheric path delay effects (Berardino et al. 2002). We applied a complex multi-look operation using two and eight looks in range and azimuth, respectively. The ground pixel resolution is approximately 40  34 m in the range and azimuth directions, respectively. In order to exclude decorrelated areas from the study, we selected only the common coherent pixels in all interferograms. Atmospheric contributions were estimated and filtered out before estimating a mean displacement velocity. In all of the InSAR data documented here we present the estimated mean displacement velocity. After filtering out atmospheric disturbances, the precision in the mean velocity measurements is of the order of a few millimetres per year. Areas of apparent subsidence are shown in a red colour (e.g. see Fig. 2).

Fig. 1. Location map of Troms County showing the location of rockslide sites as black dots. Most of the identified rockslide sites are clustered around the presently active (Osmundsen et al. 2009) Lyngen Fault Complex (shown by the dashed lines with the downthrow indicated). The location of the Gamanjunni slide is shown by the grey box. The location of Troms County in Norway is shown in the inset.

ACTIVE LANDSLIDE DEVELOPMENT

Areas with no points are areas where the InSAR technique fails to provide information.

Regional setting The location of potential rockslide sites was initially determined from conventional aerial photograph analysis, historical data, local knowledge and field observations (Fig. 1). Figure 2 shows an orthophotograph montage of the eastern part of Troms County, with the mean displacement values from the regional InSAR data and the location of potential rockslide sites mapped by other means apart from InSAR (aerial photographs and field mapping). Two observations can be made from a comparison of these two regional data sets. First, the red areas that are moving with the greatest velocity, according to the InSAR, were moving away from the satellite at rates of up to 10 mm year21 during the 1992 – 1999 InSAR data coverage window (see the previous section). Considering the angle of the satellite with respect to the topography, we interpret these areas as being localized areas of maximum subsidence. Spatially, the areas of high ground movement rates, as identified by the InSAR, coincide remarkably well with the potential rockslide

187

sites identified by a combination of aerial photographic analysis and structural field mapping. We will go on to show later that InSAR also has the potential to resolve the movement pattern of individual rockslides, but initially the correlation shown in Figure 2 demonstrates a close relationship between InSAR pattern and rockslide distribution. Secondly, on a regional scale, there is an obvious clustering of the identified rockslide sites in the vicinity of the Lyngen peninsula, within an area identified as regionally subsiding from the InSAR data. Eighty per cent of all rockslides identified within the county occur within this area. The reason for this clustering appears to be a regionalscale pattern of active normal faults (Osmundsen et al. 2009). Therefore, from the outset, the InSAR appears to be able to resolve ground movements on both a site scale and a regional scale. The close relationship between InSAR results, geological and geomorphical features that we present below from the Gamanjunni site is by no means atypical of the 60 other rockslide sites that we are currently studying in the county of Troms. A nationwide programme of rockslide investigations has been launched in Norway that is documented elsewhere (Henderson et al. 2007, 2008).

Fig. 2. Orthophotomontage of the eastern part of Troms County showing the location of rockslide sites as black dots, the mapped structures and the overlain InSAR data. The sites that are moving the most (up to 10 mm year21 over the last 10 years) are shown in red.

188

I. H. C. HENDERSON ET AL.

All the potential rockslide localities were identified during an initial aerial photographic analysis of Troms County that was designed to identify lineaments with obvious vertical and horizontal separations, and direct geomorphological evidence for ongoing slope processes (e.g. debris fans). Subsequent work has involved comparison of the geometrical pattern of these structures with the pattern shown by the InSAR data. This procedure was followed by helicopter-aided field reconnaissance of the sites of highest priority. Selected sites were mapped in detail with respect to structures, kinematics, geomorphology and evidence for movement. Sites in which recent movement is suspected are selected for the establishment of GPS measuring points.

Geological setting of the rockslide site The Gamanjunni slide is composed of garnetbearing, quartz-mica schist belonging to a suite of the Cambro-Ordovician rocks of the Caledonian Nappe Stack. This is a grey–red coloured, fine – medium-grained rock that cleaves along mica-rich foliation planes. This thrust package is infolded with a Silurian nappe along tight- to open-folds on north–south axial traces (Fig. 3a). The eastern flank of the north–south valley of Manndalen, in which the Gamanjunni site is located, straddles one such north –south-trending anticline. Here the fold has an open, flat-lying geometry. The Gamanjunni site lies almost directly on the crest of the anticline; the back-wall of the rockslide is on the immediate east-dipping limb of the anticline, whereas the active block is on the downslope, westdipping limb. This regionally mapped structure is apparent in the sparse foliation data from the active block, which shows a north– south-trending and open anticline with a flat-lying axial hinge (Fig. 3b). On top of the bedrock rests a block field, up to 1 m thick, with individual blocks varying in size from a few centimetres up to a few tens of centimetres. A thin layer of soil, not more than a few centimetres thick, lies on top of the block field. The soil layer is patchily developed across the whole surface area of the active block, and is partly vegetated with moss, grass and heather.

Results Below we focus on the use of satellite-based InSAR at the rockslide-site scale in unravelling structural architecture and active movement of individual rockslides. We present a combination of field observations and observations from the InSAR data at the deposit scale, and demonstrate the relationship between rockslide geometry, the magnitude of

movement, the nature of the back scarps and how these are reflected in the surface geomorphological response.

InSAR The Gamanjunni slide was identified as a high displacement area from the regional InSAR (Fig. 2), with a well-defined pattern of high-magnitude displacement on the InSAR correlating with mapped structures in the field. On the site scale, there is a remarkable spatial coincidence of the structures observed in the suspected active block independently of the aerial photographic analysis and the detailed pattern of the InSAR data (Figs 3 & 4). The spatial extent of the InSAR movement area fits very well with the area delimited by Scarp A and Scarp B (Fig. 4a, b). The orthophotography and InSAR (Fig. 4a, b, respectively) show limited evidence of segmentation of the sliding block into smaller blocks along secondary extensional fractures. Figure 4a shows a rare NNW–SSE-trending structure, subparallel to Scarp A, in the southern part of the block. Although, this structure dies out 100 m in to the block towards the north and probably does not take up much displacement. However, the elongated NNW–SSE pattern in the variation of InSAR movement mimics this structural trend suggesting that the detailed pattern of InSAR data within the block reflects ongoing geological processes. The InSAR data show that the block is subsiding and moving towards the west (Fig. 4b) at a rate of approximately 5–10 mm year21 relative to the surrounding mountainside. The InSAR satellite makes an angle of 228 from the vertical with the surface of the Earth and, therefore, this degree of movement is the direction measured directly away from the satellite. The true horizontal and vertical components of a combined vector are not known. This area is coloured red–orange in Figure 4b, reflecting the rate of motion away from the satellite. The two extensional fractures marking the back of the block delimit the back of the area that is subsiding. The area marked ‘X’ at the front of the block (Fig. 4a) is subsiding less rapidly than the backmost part of the block. This frontal area is subsiding at between 0 and 4 mm year21 (green –yellow pixels on Fig. 4b) and is elongated subperpendicular to the inferred movement direction of the block (i.e. parallel to the slope). Because InSAR gives no information of absolute movement direction, we are unable to determine the horizontal component of displacement. However, a cumulative vector of the two back scarps gives an approximate movement direction towards 2708N. This movement direction is subperpendicular to the NNW– SSE elongation axis of this less-subsiding area and perpendicular

ACTIVE LANDSLIDE DEVELOPMENT

189

Fig. 3. (a) Bedrock map of the area in the vicinity of the Gamanjunni rockslide. (b) Three-dimensional composite image of the Gamanjunni site seen from the SE, which shows the transparent InSAR data draped over the digital orthophotograph and DEM. Areas in red are subsiding at up to 9 mm year21. Those areas stable or moving slightly upwards are in blue. Areas with no InSAR data either have no reflectors (are vegetation-covered) or are in the satellite shadow (for example, in front of the sliding block). The three stereonets are: (1) foliation in the block; (2) Scarp B geometry and associated joints (dashed lines); (3) Scarp A geometry and associated joints (dashed lines).

190

I. H. C. HENDERSON ET AL.

Fig. 4. (a) Simplified geological interpretation overlain on the orthophotograph for the Gamanjunni rockslide. The annotation shows that the two different scarps have different morphologies and that Scarp A has a segmented morphology. Very few structures are observed internally within the sliding block except for a NNW– SSE-trending

ACTIVE LANDSLIDE DEVELOPMENT

to the elongated pattern of InSAR within the block shown in Figure 4b.

Structures The Gamanjunni slide is located on a west-facing mountain at 1200 m above sea level, near to the summit of the mountain, and is bounded by two back scarps (Figs 3 & 4) that join at the back of the block at an angle of approximately 1208. Scarp A strikes NW–SE at 3208N and dips approximately 708 to the SW (Fig. 5a). Scarp B strikes SW –NE at 2208N and dips approximately 758 to the NW (Fig. 5b). These two orientations correspond to those of major regionally mapped lineament sets (Gabrielsen et al. 2002). These lineament sets form some of the main fjord and valley orientations (Fig. 1). The back scarps, however, appear to be newly formed structures as we observe no evidence of reactivation of pre-existing fault structures, such as the presence of slickensides or gouges. Where the scarp surfaces can be examined, we observe that the steeply dipping back scarps cut through the shallowly westwards- (i.e. valleywards-) dipping foliation. Scarp A has a total maximum vertical throw of approximately 2 m. The scarp surface is not fully exhumed and this displacement estimate is based on the apparent displacement of the gently west-dipping topographical surface on either side of the scarp, which forms a surface monocline (Fig. 5a). Scarp B, on the other hand, records a much higher maximum total displacement (up to 5 m) and breaches the topography along the whole of its length (Fig. 5b). However, disintegration of Scarp B by minor block toppling along much of its length has resulted in a poorly preserved scarp surface. Scarp B shows a minimum displacement at its NE tip, where the displacement is approximately 1 m and increases gradually in displacement along its length towards the SW to a maximum displacement of 5 m. Scarp A displays a more complicated along-strike geometry as it is divided up into a number of fault segments with intervening relay zones (Fig. 4a). The segments where the fault plane breaches the monocline surface show the most displacement but the overall displacement generally decreases towards the NW extremity where it appears to die out. A basal sliding plane (BSP) is observed with a 208 –308 dip towards the valley floor to the west

191

(Figs 4 & 6d). The BSP is subparallel to, and appears to utilize, the shallow west-dipping foliation. The presence of fresh talus slopes at the front and sides of the block, and particularly in the vicinity of extensional structures within the block, suggests that active shearing occurs along this basal plane.

Geomorphology The two back scarps reveal morphological variations that are interpreted to reflect the structural variations described above. Large parts of Scarp A are expressed as a monocline that deflects the 0.5 –1.0 m-thick block-field layer described earlier. Fresher gravel indicates some movement of surface material, probably by running water across the front of the monocline. Where the height of the scarp becomes larger, redistribution of block-field material becomes more evident in the form of fan-like deposits that partly cover the scarp (see Fig. 5b) but that are largely deposited on the downthrown side. Locally, it can be demonstrated that the deposits rest on top of a thin layer of soil and vegetation, which is not more than a few centimetres thick. This demonstrates that parts of these fan-like deposits are relatively young (an example of this can be seen later in the section on ‘Synthesis’ and the figures therein). Where the monocline is breached and the scarp reaches its maximum height, a bedrock scarp is intermittently exposed. Alluvial deposits in the form of redistributed blocks coalesce into a 1–2 m-thick semicontinuous skirt along the scarp. At its NW end, Scarp A splays into two separate branches. The NE branch is expressed mostly as a surface monocline, whereas the SW branch, further into the block, is expressed predominantly as a breached monocline. Sinkholes (circular depressions in the terrain) crop out intermittently along the surface trace of Scarp A, and appear to be preferentially localized on the hanging-wall side of the breached monoclines (Fig. 4a). Locally, sinkholes occur adjacent to the bedrock scarp, draining material deposited in the alluvial skirt described earlier. The total vertical displacement on Scarp B is up to 5 m and its total horizontal displacement is up to 5 m. A monocline also appears to have formed on Scarp B but was completely breached along the

Fig. 4. (Continued ) fracture in the southern part of the block that dies out approximately 100 m to the north. (b) Detailed InSAR image overlain on the orthophotograph. The detaching block is clearly shown in red and orange. Although there is very little seen geologically within the sliding block, the InSAR image shows a definite NNW– SSE structural grain (shown by the double-headed arrow). The area marked X is moving upwards relative to the rest of the block and we interpret this as representing back rotation of the block, where the part of the block furthest up the slope is moving down relatively faster than the front of the block. This area appears to contain more sinkholes than the rest of the block.

192

I. H. C. HENDERSON ET AL.

Fig. 5. Examples of the two different fault scarps with their differing geometry, displacement and geomorphology. (a) Scarp A looking towards the SE. The displacement becomes less towards the NW (arrowed). Rubble fans that have covered over the partially exhumed scarp are clearly seen at the 2 m arrow. (b) Scarp B looking towards the east. This scarp has a higher total displacement than Scarp A (up to 5 m) and the scarp surface is fully exposed.

ACTIVE LANDSLIDE DEVELOPMENT

193

Fig. 6. Examples of geomorphological response to movement. (a) Very active sinkhole where the vegetation and earth has been disturbed. (b) Extremely active sinkhole where mud deposited from the previous seasonal snow melt has been disturbed. (c) A 15 m-diameter sinkhole, among the largest we have seen anywhere in Troms County. (d) Profile through the sliding block, which shows the thickness at the front of the block and the presence of a rare extensional fracture internal to the block. This fracture is coincident with an area of high rockfall activity.

entire fault length. Alluvial material distributed along the scarp comprises both redistributed surface material from the block field and large blocks derived from the disintegrating bedrock scarp. This is particularly well illustrated where large blocks deposited on the downthrown side of the scarp reflect local variations in the bedrock lithology exposed in the scarp. Thus, rockfall from the scarp itself appears to be a dominating depositional process along Scarp B. This contributes to the jagged scarp morphology. In summary, Scarp A and Scarp B show quite different geomorphical characteristics. Scarp B consists exclusively of exhumed footwall rocks along its entire length. Scarp A shows less continuously exposed footwall, consisting rather of segments of alternating surface monoclines and breached monoclines with intermittently exposed bedrock scarps. The varying geomorphical characteristics along Scarp A can be divided into three types, interpreted to reflect stages of scarp evolution: (1) initial offset causes the initial development of a surface monocline, where the back scarp has not reached the surface but offset has nevertheless created a welldefined deflection of the topographical surface; (2)

further offset has caused the formation of a breached monocline, where the fault has penetrated to the surface and the back scarp is partially exposed. Here the back scarp surface is partially exhumed but often covered in rubble; and (3) extensive and continued offset along the fault produces a fully exhumed footwall where the fault scarp is continuously exposed. Figure 4a shows that some of these variations are visible on orthophotographs, as the bedrock scarps are commonly more visible than the monocline areas. This structural –geomorphological evolution can be compared to that observed along tectonic faults (e.g. Gawthorpe et al. 1997 and discussion below). Sinkholes occur commonly on this site and others studied in the county (Henderson et al. 2007). Those that cut through vegetation are considered the most active. At the Gamanjunni site, sinkholes cover the range from inactive to very active. Some examples are shown in Figure 6. Figure 6a shows a sinkhole that has an approximately 1 m-deep depression, and is filled with grass and earth that has been dislodged from the surface level. Figure 6b shows a similar sinkhole that displays dislodged grass and earth. A disrupted

194

I. H. C. HENDERSON ET AL.

mud layer is also observed that was presumably deposited after the last snow melt. We interpret this as a sign of very recent movement. The position of these sinkholes is shown on Figure 4 and shows that some are so large, up to 15 m across (Fig. 6c), that they are resolved on the orthophotograph. The sinkholes appear to be associated specifically with two structural locations. First, where the breaching occurs along Scarp A, it appears to coincide with the presence of active sinkholes in the immediate hanging-wall block. In some locations sinkholes drain material into the alluvial skirt deposited along the scarp. Secondly, large sinkholes occur in the uplifting (as evinced by the InSAR data) area at the front of the block. This is also reflected in the details of the InSAR data, which will be discussed in more detail below. By considering the combined structural and geomorphological observations, we suggest that Scarp B appears to be more active and displays a greater active displacement rate than Scarp A. This is also reflected on the orthophotograph: Scarp B is more topographically incised than Scarp A (Fig. 4). Scarp A is, however, oriented more normal to the dip of the topographical slope than Scarp B. Thus, theoretically, Scarp A may display more phenomena related to dilation (extension fractures, sinkholes) than Scarp B. This may explain the relative scarcity of sinkholes along the better-developed Scarp B. Using these bounding structures to delineate the main compartment, the surface area of the block between these bounding structures and delimited by the scarp at the front of the block defines an area of approximately 780 000 m2. The estimated depth to the sliding plane from the surface is 30 m (Fig. 6d). This gives an approximate volume for the unstable block of approximately 24 Mm3, making this site one of the largest potential rockslides in Norway. The paucity of extensional structures within the block (we have only observed one minor structure shown in Fig. 6d) strongly suggests the possibility of a single catastrophic failure. A large river runs along the bottom of the valley that could be blocked and dammed by a rockslide of this size. Such temporary dams may later be breached, often causing catastrophic flooding downstream (Crosta et al. 2004). In Norway, the identification of such sites through different techniques is becoming vitally important from a land-use planning perspective (Blikra 1990).

Discussion Much of the ground motion resolved in the InSAR data correlates with actual structures observed on the ground. The geometry of the InSAR pattern presented in Figure 4 shows a well-defined area

subsiding at up to 10 mm year21. This area on the InSAR is bordered by well-defined structures that have been mapped in orthophotographs and on the ground. Furthermore, the pattern of internal variation in the InSAR data within the active block demonstrates a close correlation with the trend of potentially active structures, suggesting that the nuances of the InSAR data on a site scale reflect real geological processes. In addition, it appears that the InSAR is able to resolve displacement variations along the back scarps. For example, the segmental geometry of Scarp A (Fig. 4a) is reflected in the InSAR data as an along-strike variation in differential ground movement (see later). Figure 7a shows five east –west profiles across the Gamanjunni site. Each profile is subparallel to the direction of movement. Profiles 1 –3 intersect Scarp A, and profiles 4 and 5 intersect Scarp B. All profiles, irrespective of the scarp across which they are located, show a similar displacement curve geometry but with a varying magnitude of movement rate. This suggests that it is valid to compare the relative displacements of the different profiles as the displacement curves probably reflect real geological processes. The profiles intersecting Scarp A display an average yearly displacement of approximately 5.5 mm (Fig. 7b). The profiles intersecting Scarp B (profiles 4 and 5) show a yearly displacement average of approximately 9 mm (Fig. 7b). This appears to correlate with the different geomorphical characteristics of the two scarps. Whereas Scarp B is constituted by a bedrock scarp along its entire length, the morphology of Scarp A varies between monoclinal warp, breached monocline and bedrock scarp. Thus, we suggest that the highest displacement rates are reflected by scarp segments where the bedrock has been exhumed in the footwall, whereas lower displacement rates may result in the lessened ability of the back fracture to breach the block-field cover because of less cumulative displacement. An alternative explanation is that the rate of sedimentation of redistributed block-field material and rockfall across scarps with low displacement rates is able to keep up with the accommodation creation rate in such a way that bedrock exposure becomes scarce. In conclusion, our observations indicate a relationship between scarp geomorphology and millimetre variations in the yearly displacement rate detected by the InSAR. In all our profiles in Figure 7, a maximum displacement rate appears to be achieved in the middle of the block (Fig. 7b), irrespective of the magnitude of displacement. This can also be seen in the red area perpendicular to the movement direction and parallel to area X in Figure 4. All profiles also show a relative uplift in their westernmost parts. A possible interpretation is the presence of

ACTIVE LANDSLIDE DEVELOPMENT

195

Fig. 7. Detailed profiles through the InSAR data at Gamanjunni. Three profiles (1, 2 and 3) intersect the 3208N back fracture, whereas profiles 4 and 5 intersect the 2208N structure. The profiles on Scarp A give an average displacement of 5.5 mm year21, whereas Scarp B gives an average displacement of 9.0 mm year21. All profiles show a relative uplift in their western parts consistent with back rotation of the block.

196

I. H. C. HENDERSON ET AL.

a restricting structure at the front of the block causing toe-zone compression that stops the block from failing catastrophically. This would have the effect of uplifting the front of the block at the same time as the block as a whole is subsiding and moving towards the west. This would be reflected in either an apparent upwards or downwards relative vertical component at the front of the block. Similar toe-zone compressional structures have been inferred, observed (Braathen et al. 2004) and measured from GPS data elsewhere (Ganerød et al. 2008). However, we believe that this is the first time that such aspects of rockslide failure mechanisms have been resolved by InSAR data. We also note that there is a displacement variation difference in the InSAR data between the two sets of profiles on Scarp A and Scarp B. For example, there is a wider range of maximum displacement on Scarp A than there is on Scarp B (Fig. 7b), even though the maximum average displacement is much less on Scarp A. This may reflect the complex segmentation observed on Scarp A (Fig. 4a), where each ‘mini-scarp’ probably has different displacement rates and a complex movement pattern. On Scarp B, only a single scarp surface is observed with no relay zones and no obvious scarp segmentation. Here there is a much narrower range of displacement relative to the total maximum displacement.

Synthesis The geomorphology of monoclinal back scarps described here bears remarkable resemblance to the geometries that are produced where normal faults propagate upwards through an overlying sedimentary cover (Gawthorpe et al. 1997). Grant &

Kattenhorn (2004) also demonstrated that shallowlevel, newly formed tectonic faults propagate upwards to the surface and are initially blindly associated with surface monoclines. Kaven & Martel (2007) presented remarkably similar fault scarps and monoclinal scarp evolution from active normal faults related to volcanism in Hawaii. The Gamanjunni rockslide site displays a complete progression in the development of a back-wall scarp from monocline, through breached monocline, to fully exhumed fault plane as the extensional structures propagate upwards from the sliding plane towards the surface topography, all within the context of a large volume rockslide. Thus, even though the Gamanjunni slide itself may be gravitational in origin, its bounding structures have produced a suite of morphologies that are related to each other in a time–evolution sequence and that are remarkably similar to those that are welldocumented in actively growing normal fault tectonic environments. In the schematic model in Figure 8 we envisage a time-dependent evolution of such structures in the Gamanjunni rockslide site, but this sequence of events could just as well represent those observed in actively forming normal faults in a range of tectonic environments. In Figure 8a the incipient extensional fault begins to develop at the back of the moving block and begins propagating upwards from the BSP (0). As the active block begins to displace and the extensional structure begins to approach the surface, the surface topography will be displaced on both sides of the extensional structure to produce a surface monocline (1). Further movement will lead to exposure of the scarp surface as the extensional structure breaches the topography as it propagates fully to the surface. This produces a breached monocline (2). The difference between stages 1 and 2 may

Fig. 8. Schematic model of fault zone evolution through time. (a) The different stages of fault evolution. (0) Incipient fault begins to develop. (1) As the fault begins to propagate a monocline develops in the surface topography. (2) The fault breaks through but is not visible owing to the amassed rubble on the surface. (3) The footwall becomes fully exhumed exposing the fault surface. (b) Summary diagram showing the different geomorphological responses relative to fault evolution. The numbers refer to stages in (a) and to the photographs in Figure 9.

ACTIVE LANDSLIDE DEVELOPMENT

197

Fig. 9. Images of the different stages of development of the fault scarps. On the left lie the least active parts of Scarp A, where only a surface monocline is developed above the scarp (1 in Fig. 8). The middle image shows the footwall scarp starting to exhume (2 in Fig. 8). The image on the right shows the fully developed scarp (3 in Fig. 8). These geomorphological features can be directly related to InSAR displacement rate.

be difficult to distinguish dependent on the amount of rubble build up from immediate disintegration of the footwall scarp, which then produces rubble fans that mask the sharp scarp geometry. Continued movement on the extensional structure will lead to full exposure of the scarp surface as the sliding block begins to fully detach from the back wall. This produces a full footwall exhumation (3). In Figure 8b we envisage a spatial, as well as a temporal, relationship between these structures. In addition to the scarp propagating upwards from the BSP, it will also propagate along strike, such that these three phases of gradual scarp development and exhumation will have an along-strike relationship, the most geomorphologically evolved representing the most mature parts of the structure. The topographical relationships to the fault mechanics described by Kaven & Martel (2007), which are identical those structures described here, suggest that the structures we observe on Gamanjunni are actively forming and interacting with the topography and geomorphology at rates that can be resolved by the InSAR technique. Furthermore, variation in displacement rate on the two faults described here, within the same rockslide site, appears to occur as a result of the differential obliquity of the two faults to the transport direction. Fundamentally different geomorphological signatures reflect the apparent displacement rate difference of several millimetres per year between the two faults. Figure 9 shows a montage of examples from the Gamanjunni rockslide showing a timedependent evolution of these fault scarps, relating the observed geomorphology to millimetre-scale differences in the InSAR response. The lowest displacement rates of 1–2 mm year21 are represented by weak surface monoclines with no scarp exhumation and a low total displacement of less than 1 m (1). Medium displacement rates of between 2 and 5 mm year21 produce segments of partially exhumed scarps that partially breach the growing

surface monocline (2). These structures have moderate total displacement of up to 2 m. The highest rates of displacement of between 5 and 9 mm are represented by fully exhumed footwall scarps, with a total maximum displacement of up to 5 m (3). We believe that the back scarp evolution we observe here, its interaction with the surface geomorphology and how this is reflected in the InSAR data, is a new and interesting contribution to the field of landslide research. Bu¨rgmann et al. (2006) demonstrated that a close relationship can exist between fault zone architecture, kinematics and the detailed pattern of geomorphological response from InSAR data along the San Andreas Fault, and Hilley et al. (2004) documented nuances in displacement rate from InSAR data from several different landslides with similar displacement rate variations as those observed here. There is a growing body of evidence for post-glacial tectonic faulting in Norway (Dehls et al. 2000; Olesen et al. 2004), and recent work in Troms County suggests that both the regional topographical pattern and the distribution of rockslides are controlled by active tectonics (Osmundsen et al. 2009). It has also been suggested that inherited tectonic structures may influence the development of rockslide fabrics (Anda et al. 2002; Blikra et al. 2002; Braathen et al. 2004; Redfield & Osmundsen 2009) even if individual rockslides are gravitational in origin. Other studies have also looked closely at the relationship between active faults and InSAR response but we believe ours is an important contribution to both an understanding of the geological implications of the InSAR data and the processes that create them.

Conclusions Thus study demonstrates that InSAR is a powerful tool that can be used to identify the relative magnitude and spatial pattern of active rockslide sites on

198

I. H. C. HENDERSON ET AL.

both a regional scale and the individual rockslidesite scale. The correlation between the geometry of structures developed during incipient rocksliding and the pattern of subsidence recognized from InSAR on the Gamanjunni active rockslide is remarkable. Ground observations of fault geometry and developing footwall profile are precisely reflected by the different magnitudes of relative vertical motion recorded in profiles through the InSAR data. Differences of a few millimetres displacement per year provoke a definite response in scarp propagation and behaviour. These differences in geomorphological response are resolved in the InSAR data. The InSAR pattern is able to discern internal structures within the sliding block that can contribute to a detailed interpretation of the failure mechanism. InSAR can thus be a valuable tool in providing quantitative data of the activity in rockslides with exposed rock surfaces, and an important confirmation of the field observations in areas of movement. We are grateful for positive and helpful reviews from two anonymous reviewers that greatly improved the manuscript. This paper has been partly funded by the Norwegian Government through the inter-municipality project ‘Fjellskred i Troms’. We would particularly like to thank R. Krag from Lyngen Kommune and R. Elvenes from Ka˚fjord Kommune for their enthusiastic contribution. Partial funding was also received from the International Centre for Geohazards. This work is ICG Publication No. 272.

References Agliardi, F., Crosta, G. & Zanchi, A. 2001. Structural constraints on deep-seated slope deformation mechanisms. Engineering Geology, 59, 83– 102. Ambrosi, C. & Crosta, G. B. 2006. Large Sackung along major tectonic features in the central Italian Alps. Engineering Geology, 83, 183– 200. Anda, E., Blikra, L. H. & Braathen, A. 2002. The Berill Fault-First evidence of neotectonic faulting in southern Norway. Norwegian Journal of Geology, 82, 175–182. Berardino, P., Costantini, M., Franceschetti, G., Iodice, A., Pietranera, L. & Rizzo, V. 2003. Use of differential SAR interferometry in monitoring and modelling large slope instability at Maratea (Basilicata, Italy). Engineering Geology, 68, 31–51. Berardino, P., Fornaro, G., Lanari, R. & Sansosti, E. 2002. A new algorithm for surface deformation monitoring based on small baseline differential SAR interferograms. IEEE Transactions on Geoscience and Remote Sensing, 40, 2375–2383. Blikra, L. H. 1990. Geological mapping of rapid mass movement deposits as an aid to land-use planning. Engineering Geology, 29, 365– 376. Blikra, L. H., Braathen, A., Anda, E., Stalsberg, K. & Longva, O. 2002. Rock avalanches, gravitational bedrock fractures and neotectonic faults onshore

northern west Norway. Norges geologiske undersøkelse Report 2002.016. Braathen, A., Blikra, L. H., Berg, S. S. & Karlsen, F. 2004. Rock-slope failures in Norway, type, geometry, deformation mechanisms and stability. Norwegian Journal of Geology, 84, 67– 88. Bu¨rgmann, R., Hilley, G., Ferretti, A. & Novali, F. 2006. Resolving vertical tectonics in the San Francisco Bay area from permanent scatter InSAR and GPS analysis. Geology, 34, 221– 224. Canuti, P., Margottini, C. et al. 2005. Preliminary remarks on monitoring, geomorphological evolution and slope stability of Inca Citadel, Machu Picchu. In: Sassa, K., Fukuoka, H., Wang, F. & Wang, G. (eds) Landslides: Risk Analysis and Sustainable Disaster Management. Springer, Berlin. Catani, F., Farina, P., Moretti, S., Nico, G. & Strozzi, T. 2005. On the application of SAR interferometry to geomorphological studies: estimation of landform attributes and mass movements. Geomorphology, 66, 119–131. Chigira, M. 1992. Long term gravitational deformation of rocks by mass rock creep. Engineering Geology, 32, 157–184. Clifton, A. E., Pagli, C., Jo´nsdo´ttir, J. F., Eythorsdo´ttir, K. & Vogfjo¨rð, K. 2003. Surface effects of triggered fault slip on Reykjanes Peninsula, SW Iceland. Tectonophysics, 369, 145–154. Colesanti, C. & Wasowski, J. 2006. Investigating landslides with space-borne Synthetic Aperture Radar (SAR) interferometry. Engineering Geology, 88, 173–199. Crosta, G. B., Chen, H. & Lee, C. F. 2004. Replay of the 1987 Val Pola Landslide, Italian Alps. Geomorphology, 60, 127– 146. Dehls, J. F., Olesen, O., Olsen, L. & Blikra, L. H. 2000. Neotectonic faulting in northern Norway; the Stuoragurra and Nordmannvikdalen postglacial faults. Quaternary Science Reviews, 19, 1447– 1460. Ferretti, A., Prati, C. & Rocca, F. 2001. Permanent scatterers in SAR interferometry. IEEE Transactions on Geoscience and Remote Sensing, 39, 8– 20. Fruneau, B., Achache, J. & Delacourt, C. 1996. Observation and modelling of the Saint-Etiennede-Tinee landslide using SAR interferometry. Tectonophysics, 265, 181 –190. Gabriel, A. K., Goldstein, R. M. & Zebker, H. A. 1989. Mapping small elevation changes over large areas: differential radar interferometry. Journal of Geophysical Research, 94, 9183– 9191. Gabrielsen, R. H., Braathen, A., Dehls, J. F. & Roberts, D. 2002. Tectonic lineaments in Norway. Norwegian Journal of Geology, 82, 153–174. Ganerød, G. V., Grøneng, G. et al. 2008. A geological ˚ knes rockslide, western Norway. Engimodel of the A neering Geology, 102, 1 –18. Gawthorpe, R. L., Sharp, I. R., Underhill, J. R. & Gupta, S. 1997. Linked sequence stratigraphic and structural evolution of propagating normal faults. Geology, 25, 795 –798. Grant, J. V. & Kattenhorn, S. A. 2004. Evolution of vertical faults at an extensional plate boundary, southwest Iceland. Journal of Structural Geology, 26, 537–557.

ACTIVE LANDSLIDE DEVELOPMENT Hanssen, R. F. 2001. Radar Interferometry. Kluwer Academic, Dordrecht. Henderson, I. H. C. & Saintot, A. 2011. Regional spatial variations in rockslide distribution from structural geology ranking: an example from Storfjorden, western Norway. In: Jaboyedoff, M. (ed.) Slope Tectonics. Geological Society, London, Special Publications, 351, 79–95. Henderson, I. H. C., Osmundsen, P. T. & Redfield, T. F. 2008. ROS Fjellskred i Troms: statusrapport 2007. Norges geologiske undersøkelse Report 2008.025 (in Norwegian). Henderson, I. H. C., Saintot, A., Venvik-Ganerød, G. & Blikra, L. H. 2007. Rockslide mapping in Troms County, Norway. Norges geologiske undersøkelse Report 2007.041 (in Norwegian). Hilley, G. E., Bu¨rgmann, R., Ferretti, A., Novali, F. & Rocca, F. 2004. Dynamics of slow-moving landslides from Permanent Scatterer Analysis. Science, 304, 1952–1955. Jaboyedoff, M., Ornstein, P. & Rouiller, J. D. 2004. Design of a geodetic database and associated tools for monitoring rock-slope movements: the example of the top of Randa rockfall scar. Natural Hazards and Earth System, 4, 187– 196. Kaven, J. O. & Martel, S. J. 2007. Growth of surfacebreaching normal faults as a three-dimensional fracturing process. Journal of Structural Geology, 29, 1463–1476. Larsen, Y., Engen, G., Lauknes, T. R., Malnes, E. & Høgda, K. A. 2005. A generic differential interferometric SAR processing system, with applications to land subsidence and snow-water equivalent retrieval. In: Proceedings of the Fringe 2005 Workshop, Frascati, Italy. ESA-ESRIN, Frascati, Rome, Italy, 28 November –2 December. ESA Publications, on CD-ROM. Lauknes, T. R. 2004. Long-term surface deformation mapping using small-baseline differential SAR interferograms. Master’s thesis, University of Tromsø, Norway. Lauknes, T. R., Dehls, J., Larsen, Y., Høgda, K. A. & Weydahl, D. J. 2005. A comparison of SBAS and PS ERS InSAR for subsidence monitoring in Oslo, Norway. In: Proceedings of the Fringe 2005 Workshop, Frascati, Italy. ESA-ESRIN, Frascati, Rome, Italy, 28 November–2 December. ESA Publications, on CD-ROM.

199

Massonnet, D. & Feigl, K. L. 1998. Radar interferometry and its application to changes in the earth’s surface. Reviews of Geophysics, 36, 441– 500. Olesen, O., Blikra, L. H. et al. 2004. Neotectonic deformation in Norway and its implications: a review. Norwegian Journal of Geology, 84, 3–34. Osmundsen, P. T., Henderson, I. H. C., Lauknes, T. R., Larsen, Y., Redfield, T. F. & Dehls, J. 2009. Active normal fault control on landscape and rockslope failure in northern Norway. Geology, 37, 135– 138. Redfield, T. F. & Osmundsen, P. T. 2009. The Tjellefonna fault system of Western Norway: Linking late-Caledonian extension, post-Caledonian normal faulting, and Tertiary rock column uplift with the landslide-generated tsunami event of 1756. Tectonophysics, 474, 106– 123. Rosen, P. A., Hensley, S., Joughin, I. R., Li, F. K., Madsen, S. N., Rodrı´guez, E. & Goldstein, R. M. 2000. Synthetic aperture radar interferometry. Proceedings of the IEEE, 88, 333 –382. Saintot, A., Henderson, I. H. C. & Derron, M.-H. 2011. Inheritance of ductile and brittle structures in the development of large rock slope instabilities: examples from western Norway. In: Jaboyedoff, M. (ed.) Slope Tectonics. Geological Society, London, Special Publications, 351, 27– 78. Saroli, M., Stramondo, S., Moro, M. & Doumaz, F. 2005. Movements detection of deep seated gravitational slope deformations by means of InSAR data and photogeological interpretation: northern Sicily case study. Terra Nova, 17, 35– 43. Solheim, A., Bhasin, R. et al. 2005. International Centre for Geohazards (ICG): assessment, prevention and mitigation of geohazards. Norwegian Journal of Geology, 85, 45–62. Squarzoni, C., Delacourt, C. & Allemand, P. 2005. Differential single-frequency GPS monitoring of the La Valette landslide (French Alps). Engineering Geology, 79, 215–229. Tarchi, D., Casagli, N. et al. 2003. Landslide monitoring by using ground-based SAR interferometry: an example of application to the Tessina landslide in Italy. Engineering Geology, 68, 15–30. Zebker, H. A., Rosen, P. A. & Hensley, S. 1997. Atmospheric effects in interferometric synthetic aperture radar surface deformation and topographic maps. Journal of Geophysical Research, 102, 7547–7563.

Numerical modelling of Plio-Quaternary slope evolution based on geological constraints: a case study from the Caramanico Valley (Central Apennines, Italy) G. BIANCHI FASANI1, E. DI LUZIO2, C. ESPOSITO1*, S. MARTINO3 & G. SCARASCIA-MUGNOZZA3 1

‘Sapienza’ University of Rome – CERI Research Centre on Geological Risks, Piazza U. Pilozzi, 9-00038, Valmontone (Roma), Italy

2

Italian National Council for Research CNR-IGAG, Area della Ricerca di Roma 1, Via Salaria Km 29.3-00016, Monterotondo Scalo (Roma), Italy 3

‘Sapienza’ University of Rome – Department of Earth Sciences, P.le A. Moro, 5-00185, Roma, Italy *Corresponding author (e-mail: [email protected]) Abstract: Evidence of deep-seated gravitational slope deformations (DSGSD) and of large prehistoric landslides is fairly widespread within the Central Apennines (Italy). These gravityinduced processes accompanied the intense Plio-Quaternary uplift phases that affected the mountain chain. In this study a multidisciplinary approach has been adopted in order to better constrain the relationship between the tectonic evolution and the gravitational morphogenesis of a typical Apennine morphostructure, such as the Caramanico Valley. For this purpose a conceptual model of the morphostructural evolution of the area has been reconstructed, on the basis of geological constraints derived by the integration of detailed geological– structural and geomorphological surveys with available literature data. Based on this evolutionary model, a multistage numerical modelling using the finite difference method code FLAC 6.0 has been performed in order to: (i) evaluate the effect of the uplift-related morphological changes of the valley–slope system; and (ii) assess the role of the horizontal/vertical stress ratio variations due to geodynamic regime shifts. The results of the numerical model show a good fit with the actual geomorphical evidence and also confirm the presence during some evolutionary stages of stress –strain conditions compatible with those necessary to produce the massive rock slope failures testified by the presence of large palaeo-landslide deposits.

The Central Apennines are a quite young mountain chain that formed in Neogene and Early Pleistocene times, and experienced several and significant phases of regional tectonic uplift during Quaternary. Evidence of deep-seated gravitational slope deformations (DSGSD), as well as of large prehistoric landslides, is quite diffuse within mountain ridges mainly made of carbonate and siliciclastic rocks. Most of this geomorphical evidence can be dated back in a wide time interval spanning from a generic Lower Pleistocene, which is the indirect dating of the Mt Arezzo landslide (Cinti et al. 2001), to 4800 + 60 years BP, which is the radiocarbon dating of the Lettopalena landslide (Paolucci et al. 2001) – the most recent, directly dated, catastrophic failure event. The presence of thick beds of breccias at the base of the continental sequences filling the intramontane basins, a typical morphostructure of the Apennine chain, is also commonly recognized and related to the opening

and deepening of such structures in the Middle Pliocene –Lower Pleistocene time interval (Cavinato & De Celles 1999; Bosi et al. 2003; Centamore & Nisio 2003). The morphostructural setting of the Central Apennines is not characterized by significant topographical stress in terms of both elevation difference and slope angle if compared to other mountain chains. In this setting, the onset and development of slope-scale gravitational processes should be regarded as the effect of an ‘evolving’ disequilibrium related to the Plio-Quaternary intense uplift history of the Apennines rather than as a consequence of an ‘inherited’ gravitational disequilibrium related to significant topographical stresses. The strict relationship existing between tectonic uplift, DSGSD and large landslide events are being studied within the Central Apennines (Di Luzio et al. 2004a, b; Galadini 2006; Scarascia-Mugnozza et al. 2006; Esposito et al. 2007), with particular

From: Jaboyedoff, M. (ed.) Slope Tectonics. Geological Society, London, Special Publications, 351, 201– 214. DOI: 10.1144/SP351.11 0305-8719/11/$15.00 # The Geological Society of London 2011.

202

G. BIANCHI FASANI ET AL.

focus on the influence of the inherited structural pattern on the mechanisms and kinematics of massive rock slope failures and DSGSD processes. In this study a multidisciplinary approach has been adopted in order to better constraint the onset and development of slope-scale gravitational processes in the ‘dynamic’ frame of the Quaternary morphostructural evolution. Using geological, structural and geomorphical field evidence, and the data available in the literature, in this paper we try to reconstruct the morphostructural evolution of the Caramanico Valley, with particular focus on the role of the tectonically growing eastern border (Maiella Massif ). This reconstructed evolutionary model is the basis on which we performed the numerical modelling of the multi-stage gravitational morphogenesis of the slope –valley system by simulating the stress –strain ‘history’ due to the Quaternary tectonic uplift of the Maiella western slope, as well as to the variations in the geodynamic frame in terms of changing stress ratios. Finally, the Caramanico Valley was chosen as a reference study area because it can be considered as a ‘pilot’ area in the frame of the Apennine chain owing to the facts that: (1)

(2)

(3)

it is located in the external domain of the chain, that is, the sector that experienced the most recent morphostructural evolution and significant uplift rates due to Quaternary tectonic activity; it can be regarded as a typical Apennine intramontane basin dominated by erosive processes in its youthful stage of evolution, thus representing a possible key to understanding the evolution of other, older basins; its Plio-Quaternary evolution has been characterized by the persistence of a continental depositional environment, where even Middle Pliocene continental deposits and landforms have been preserved.

Overview of the geology and geomorphology of the Caramanico Valley and surrounding mountain ridges Geological and geomorphological setting The Caramanico Valley is a north –south-oriented narrow tectonic depression located about 40 km west of the Adriatic Sea in the easternmost part of the Central Apennines, which are a mountain belt formed between the Neogene and the Early Pleistocene, and made up of several thrust sheets piled on each other with a NE and east tectonic vergence (Mostardini & Merlini 1986; Patacca et al. 1991, 2008).

In the study area there are, from west to east, three main tectonic units composed of different thrust sheets: the Morrone, the La Queglia and the Maiella units (Fig. 1). The Morrone Unit bounds the Caramanico Valley to the west. The mountain ridge reaches elevations of about 1800–2000 m a.s.l. (metres above sea level), and is made up of Jurassic–Lower Cretaceous limestones deformed in a NNW – SSE-oriented, east-vergent fault-propagation fold (with a steep and locally reverse forelimb). The Morrone Unit is over-thrusted on the Upper Messinian–Lower Pliocene terrigenous deposits of the La Queglia Unit filling the Caramanico Valley, a tectonic depression with elevations ranging from 1200 to 600–700 m a.s.l., and a maximum width of about 4 km. The La Queglia Unit lies in the footwall of the Morrone Unit basal thrust and in the hanging wall of the Caramanico Fault System (hereafter CFS) (Fig. 1).

Fig. 1. Geological map of the Morrone–Caramanico– Maiella area, modified after Patacca et al. (2008). The dark grey dashed line encompasses the study area. 1, Middle Pleistocene– Holocene alluvial deposits; 2, Upper Pliocene–Lower Pleistocene sedimentary sequence (Upper Pliocene–Emilian marine clays and Sicilian regressive conglomerates); 3, Upper Pliocene (Lower Villafranchian)– Quaternary rock avalanche deposits; 4, Pliocene marine deposits (including the Turrivalignani conglomerates); 5, Lower Pliocene foredeep deposits of the Maiella Unit (Fara Formation); 6, Messinian foredeep deposits; 7, La Queglia Unit deposits (Cretaceous– Miocene basin and ramp carbonate and Messinian– Lower Pliocene foredeep deposits); 8, shallow-water, carbonate shelf (Jurassic– Cretaceous) and ramp deposits (Cretaceous–Miocene); 9, Maiella and Morrone units slope-to-basin carbonates (Cretaceous) and ramp deposits (Cretaceous– Miocene); 10, Molise pelagic deposits (Cretaceous– Tortonian); 11, thrust fault; 12, normal fault; 13, Caramanico Fault; 14, fold axis and plunge direction.

SLOPE EVOLUTION IN THE CARAMANICO VALLEY

203

Fig. 2. Digital elevation model of the Maiella Massif, the Caramanico Valley and the Morrone Ridge (a), where the most significant gravity-driven deposits and tectonic elements are reported (b).

The flysch deposits, which represent the bedrock of the valley floor, are directly covered by a discontinuous, clastic continental sequence (Fig. 2), whose maximum thickness does not exceed 200 m. According to Cavinato & De Celles (1999), the overall geomorphical features of the Caramanico Valley are typical of a youthful, erosion-dominated stage of Plio-Quaternary intramontane basins within the Central Apennines. Finally, the Maiella Unit lies in the footwall of the CFS and is the lowermost tectonic unit of the local thrust system. It is composed of a Lower– Upper Cretaceous shallow-water carbonate sequence in the southern part of the massif, and a coeval slope-to-basin sequence in its northern part (Fig. 1). Uppermost Cretaceous –Upper Miocene ramp deposits, Messinian evaporites, hypoaline

clays and Lower Pliocene carbonatic conglomerates feature the uppermost part of the Maiella sequence and underlie Lower Messinian flysch deposits (Crescenti et al. 1969; Accordi & Carbone 1988; Patacca et al. 1991, 2008; Bernoulli et al. 1996; Vecsei & Sanders 1997). The Maiella sequence is deformed in a wide, north– south-oriented anticline, whose morphostructural configuration is featured by a large flat crest area (at an average elevation of about 2300 m) and by an evident asymmetry between the eastern and the western slopes. The eastern slope of the Maiella Massif is a dip-slope following the curvature of the anticline forelimb. The western slope is a reverse-slope featured by a steep morphotectonic escarpment related to the CFS that truncates the anticline back-limb (Fig. 2).

204

G. BIANCHI FASANI ET AL.

Geological evolution of the Caramanico area The geological setting of the Caramanico area is the result of a complex tectonic evolution featured by the emplacement of the aforementioned tectonic units and by the subsequent collapse of the thrustrelated structures. The Morrone tectonic unit formed at the Miocene–Pliocene time boundary when the La Queglia Unit hosted a foredeep sedimentation and the Maiella domain was still part of the foreland. The Maiella Unit hosted in its turn a foredeep sedimentation in the Middle Pliocene, and began to form between the end of the Middle Pliocene and the lower part of the Late Pliocene (Patacca et al. 1991, 2008; Bigi et al. 1995). With regard to the Quaternary evolution of the Caramanico Valley, it is linked to the structuring of the Maiella Massif, and in particular to the activity of the CFS. Different tectonic models were proposed in the last decade to explain the tectonic configuration of the Maiella anticline at depth and to investigate its tectonic evolution with a particular interest for the role played by the CFS. Scisciani et al. (2001) considered the Maiella anticline as a para-authoctonus unit bordered in the Late Messinian –Early Pliocene by major normal faults, including the Caramanico basin, accommodating the flexure of the Lower Pliocene foredeep. The Caramanico basin would have, therefore, hosted a thickened wedge of terrigenous deposits belonging to the Maiella Unit in its hanging wall, while a thinned sequence was deposited in the footwall. Successively, the Caramanico Fault would have been cut and passively transported in the Middle Pliocene by a major thrust cutting at a high angle through the limit between the Adriatic sedimentary cover and the crystalline basement. Ghisetti & Vezzani (2002) entirely attributed the tectonic activity of the Caramanico Fault to the post-thrusting Quaternary evolution of the Maiella area. According to these authors the CFS cut the pre-existing thrust-related tectonic features and would then form the easternmost normal fault system originated from the collapse of the overthickened Apenninic wedge. Finally, Patacca et al. (2008) presented a tectonic model based on the interpretation of the CROP 11 crustal profile and of an original reflection seismic line, both passing through the Maiella structure. They reconstructed, after also having studied the local Plio-Pleistocene sedimentary sequence, a multiple-stage tectonic evolution of the Maiella anticline with two main thrusting and related uplift phases in the Middle Pliocene and at the end of the Early Pleistocene. During both of these thrusting phases the CFS would have accommodated the

gravitational disequilibrium of the growing structure, so determining the collapse of the Maiella anticline along the back-limb zone.

Evidence of large palaeo-landslides Among the thick cover of continental deposits of the Caramanico Valley it is possible to distinguish some clastic deposits with peculiar morphological and sedimentological features, as in the northern sector of the valley where several scattered deposits of continental breccias are present (Fig. 2). These clastic deposits are named Caramanico Breccias in the geological literature, and are generally defined as chaotic calcareous breccias and dated to the Lower Villafranchian (Centamore & Nisio 2003 and references therein). The analysis of the internal features and composition suggests that the debris could be the remnants of eroded large landslides transported as high-energy, granular flows related to massive rock slope failures. Debris is intensively fractured and angularly shaped, clast- to bouldersized and immersed in a sandy matrix (Fig. 3). Moreover, the stratigraphic analysis of the deposits showed that for each outcrop of the Caramanico Breccias the clastic material is basically monogenic. The sedimentary features of the Caramanico Breccias therefore suggest that their deposition was due to massive rock slope failures rather than glacial origin or undistinguished tectoniccontrolled processes that would have originated much more heterogenous debris. The identification of their source areas is complicated by the significant age of the deposits, as erosion has surely bevelled the original scars of detachment areas along the collapsed mountain slopes since the failure events. Furthermore, these clastic deposits rest on a subhorizontal erosional surface (Fig. 2) carved within the Caramanico Valley in the Neogene calcareous and terrigenous deposits of the Maiella and La Queglia units. Other remnants of erosion surfaces are present within the Maiella Massif, occurring at different elevations until about 1900 m a.s.l. Based on the above-mentioned evidence, the deposition of the ‘Caramanico Breccias’ can be ascribed to catastrophic rock slope failures that have occurred in an embryonic continental environment since the Lower Villafranchian (Middle Pliocene). Another important clastic deposit is located in the southern sector of the valley and corresponds to the so-called ‘Campo di Giove’ rock avalanche (Di Luzio et al. 2004a), which is the result of the failure of about 200 Mm3 of limestone from the Maiella Massif (Fig. 4). The debris travelled a distance of about 7 km, up until the base of the Morrone Ridge on the opposite side of the valley; the timing of the event is still uncertain

SLOPE EVOLUTION IN THE CARAMANICO VALLEY

205

Fig. 3. Outcrop of the ‘Caramanico Breccias’ along a road-cut exposure.

Fig. 4. The Campo di Giove rock avalanche: view from the opposite slope (Morrone Ridge) of the detachment area and related accumulation. The picture in the box is a detail of the landslide debris, taken in the outcrop encompassed in the black circle. 1, boundary of the detachment area; 2, direction of the run-out; 3, landslide debris; 4, landslide-related lacustrine deposit.

206

G. BIANCHI FASANI ET AL.

and it is indirectly ascribed to a post-Middle Pleistocene age.

DSGSD-related landforms Geomorphical evidence of DSGSD are quite pervasive in the slope systems of both the Maiella (Di Luzio et al. 2004b) and Morrone ridges. With regard to the former, it is possible to observe many gravity-induced landforms that are mainly featured by linear trenches and fractures, with length and width of the order of 101 and 100 m, respectively (Fig. 5). The widespread network of such linear landforms mainly involves the westernmost sector of the crest area and suggests deformation processes under a tensile stress regime, as also confirmed by the presence of karst-like small depressions along the bottom of some trenches, and according with findings by Casini et al. (2006). In this context, a peculiar piece of evidence of a slope-scale, gravity-induced process is represented by a section of the slope displaced by the surrounding rock mass: the displacement has a vertical component of about 30 m, as inferred by geological evidence, while the estimated volume of the displaced rock mass is about 50 Mm3. This slope sector partially corresponds to the right shoulder of the Campo di Giove rock avalanche, thus suggesting a common origin of the two phenomena; in fact, the latter could be the result of a localized failure in a wider frame of the DSGSD process. With regard to the Morrone Ridge, the main DSGSD-related feature is represented by a huge trench, particularly wide and long, that separates an entire section of the slope from the main ridge (Fig. 2). The section of the slope bounded by the

valley bottom at the toe and by the trench at the top is, in its turn, dissected by other minor trenches and fractures. The geomorphical setting, as well as the geological–structural framework, suggests a rock-block slide mechanism for the deformational processes. The tectonic superimposition of a rigid rock mass (such as the outcropping calcareous sequence) over a more deformable material (such as the flysch deposits), together with the dip-slope attitude of the bedding planes, could in fact represent the onset factors of such a DSGSD process, typical at the edge of thrust-related structures (Esposito et al. 2007).

Quaternary gravitational morphogenesis and morphostructural evolutionary model The presence in the Caramanico Valley and the surrounding ridges of both DSGSD-related landforms and landslide-related deposits, whose emplacement started in the Lower Villafranchian and spans a large period of time, highlights the relationship between the gravity-induced processes and the morphostructural evolution. The reconstruction of the Plio-Quaternary phases of the morphostructural evolution is a ‘key’ to relating the gravity-driven processes in a wider geological context. For this purpose, a morphostructural evolutionary model of the valley is hereby proposed and is based on the reinterpretation of available data in the literature, constrained by geological and geomorphological data collected for this study. The Morrone structure was already formed at the Miocene –Pliocene boundary (Patacca et al. 1991; Bigi et al. 1995). Then, since the Middle Pliocene onwards, the morphogenesis of the Caramanico

Fig. 5. Panoramic view (a) and detail (b) of one of the most significant trenches present in the southern sector of the Maiella crest zone.

SLOPE EVOLUTION IN THE CARAMANICO VALLEY

Valley was mainly controlled by the tectonic uplift of the Maiella structure and by the CFS. As previously mentioned, three different tectonic models have been hypothesized for the tectonic evolution of the Maiella Massif: Scisciani et al. (2001) considered Late Messinian–Early Pliocene activity of the Caramanico Fault, passively transported in the Middle Pliocene by a major thrust; Ghisetti & Vezzani (2002) entirely attributed the tectonic activity of the Caramanico Fault to the post-thrusting Quaternary evolution of the Maiella area; Patacca et al. (2008) considered the CFS as a collapse-normal fault accommodating the Middle Pliocene–Early Pleistocene thrusting and related uplift phases of the Maiella anticline. In this paper we refer to the model proposed by Patacca et al. (2008) because, in our point of view, it is better constrained by local and regional geological and geophysical data. The hypothesis of the CFS as a collapse-normal fault system linked to the tectonic uplift of the Maiella structure, as proposed by Patacca et al. (2008), is in agreement with the evidence of the along-strike decrease in the down-throw moving from the axial culmination zones of Mt Amaro and Mt Macellaro, about 3.8 km, to the northern (and even southern) plunge termination zone, where the tectonic displacement is of few hundreds metres. Moreover, a multiple-stage tectonic process accounts for the features of the local sedimentary sequences (Fig. 2). In this hypothesis, the end of the first tectonic phase was marked by the deposition of fandelta conglomerates (the Turrivalignani conglomerates: Crescenti 1970) resting on the Lower Pliocene terrigenous deposits of the Maiella Unit with an angular unconformity. After a quiescence phase during the Lower Pleistocene, testified by the deposition of Emilian shelf clays (sub-Apenninic clays Auctt.), a rejuvenation of the local relief occurred due to tectonic uplift as shown by the deposition of Sicilian regressive sands and conglomerates. Therefore, we adopt the model of Patacca et al. (2008) as a constraint to our morphotectonic model of the Maiella western slope and to the numerical modelling. Reassuming, according to Patacca et al. (2008), the formation of the Maiella anticline was a multiple-stage tectonic process that began at the end of the Early Pliocene and persisted until the end of the Early Pleistocene. The first tectonic phase (latest Early Pliocene –earlier Late Pliocene) led to the formation of the Maiella anticline within the Apulian foreland geological domain. Still according to the same authors, the Maiella structure experienced a second tectonic phase (end of Early Pleistocene), characterized by passive uplift and rotation in the hanging wall of a buried back-thrust developed in the Adriatic foreland.

207

Afterwards, the Maiella structure, as well as the Peri-Adriatic region, underwent a generalized uplift linked to still active geodynamic processes (D’Agostino et al. 2001; Pizzi 2003). Based on these assumptions, and on geological and geomorphological evidence, we discern three different phases for the evolution of the Caramanico Basin. † At the end of the first tectonic phase the tectonic uplift determined an embryonic continental environment between the western edge of the Maiella structure and the Caramanico Basin, as evidenced by the presence of Lower Villafranchian clastic deposits resting above a low relief palaeo-landscape (Fig. 6a). At this time the CFS was already developed within the hanging wall of the Maiella structure along the crest-to-backlimb transition zone (Fig. 6a). Such an ‘original’ fault preserved the deposits of the La Queglia Unit in its hanging wall (the maximum thickness of these deposit is more than 2000 m based on well data by Patacca et al. 2008). However, neither the thrust-related uplift nor the Caramanico Fault determined an elevated morphostructure, as testified by the absence of Maiella-derived pebbles in the clastic sedimentary record along the eastern side (Turrivalignani conglomerates). It follows that potential source areas for landslide events (deposits marked as CA2–CA3, CSM in Fig. 2) in this phase should have originated from the Morrone Ridge. † The passive deformation and further uplift of the Maiella structure and, in particular, of its western edge resulting from back-thrusting processes was accommodated at the end of the Early Pleistocene by the complete development of the CFS. Newly generated shear surfaces allowed the significant uplift of the footwall block with respect to a steady hanging-wall block (Fig. 6b). This hypothesis is supported by the preservation of continental conditions in the Caramanico Basin and by the evidence of a renewed clastic sedimentation fed by the Maiella western slope (deposit marked as CA1 in Fig. 2). According to Patacca et al. (2008), the second tectonic phase occurred at the Emilian –Sicilian boundary and before the end of the Lower Pleistocene. Considering that about 2.0 km of uplift is indicated by the same authors for the entire time interval spanning the Sicilian, we can infer a remarkable mean uplift rate of about 4.5 mm year21, which is consistent with thrustingrelated uplift rates in the Central Apennines. † Since the beginning of the Early Pleistocene, the uplift of the Maiella structure and the Caramanico Basin (with a mean rate of 1.0 mm year21, as stated by Pizzi 2003) have been

208

G. BIANCHI FASANI ET AL.

Fig. 6. Scheme of the morphostructural evolution of the Caramanico Valley and surrounding slopes. (a) Late Pliocene (Lower Villafranchian): erosive phase with the formation of a low-relief landscape carved within the deposits of the La Queglia and Maiella units; rock avalanche and landslide events from the Morrone Ridge. (b) Emilian–Sicilian (Early Pleistocene): back-thrusting related uplift (about 2.0 km); genesis of shear surfaces within the CFS and deformation of the Maiella anticline; deposition of the CA1 rock avalanche from the Maiella slope over older landslides. (c) Post-Early Pleistocene: doming-like uplift (at an average rate of 1 mm year21) and related gravitational collapses with active DSGSD along inherited faults and newly formed shear surfaces (MMFZ). Erosion of landslide deposits.

controlled by the regional doming rather than by small-scale tectonic-controlled kinematics. Within this scenario, second-order fault zones (Mt Macellaro Fault Zone – MMFZ) developed within the Maiella culmination zone (Fig. 6c).

Back-analysis of the gravity-induced conditions Based on the above-mentioned evolutionary model, we developed a numerical modelling approach to simulate the stress –strain variations in the growing eastern bound of the valley (Caramanico Basin). In this framework the analysis was aimed at evaluating the effect of the morphological changes of the valley –slope system resulting from the uplift of the structure and to assess the role of the horizontal/vertical stress ratio variations due to geodynamic regime shifts; the values of the latter were indicatively attributed on the basis of available data (Montone et al. 2005; Mariucci et al. 2006). Finally, even if the lack of strong time constraints did not allow the introduction of time-dependent behaviour of the rock mass, the analysis was also focused on an attempt to find a

sort of relative chronology for the gravitational morphogenesis in the wider morphoevolutionary framework.

Engineering-geology model In order to complete an analysis, the results of which could be generalized, the reference engineeringgeology model for the simulation was built starting from a geological cross-section that can be considered to be representative of average structural (in terms of bedding attitude and jointing conditions) and geomorphological (in terms of slope height and slope angle) conditions of a significantly long ridge, such as the Maiella Massif (Fig. 7). Regarding the attribution of geotechnical parameters to the geological section, a simple equivalent continuum approach has been adopted. The intact rock properties (UCS, tension cut-off, unit weight of volume) for each considered geological unit were derived by available laboratory test data determined on the same materials (ScarasciaMugnozza et al. 2006). The rock mass properties were assessed by applying the geological strength index (GSI) value (Hoek & Brown 1997) determined by accurate field surveys (Fig. 8). The

SLOPE EVOLUTION IN THE CARAMANICO VALLEY

209

Fig. 7. Engineering-geology section of the slope-to-valley system (see Tables 1 & 2 for the corresponding parameters).

generalized Hoek–Brown strength and deformability parameters were then determined according to the Hoek & Brown criterion (Hoek et al. 2002) by using the RocLab software (Rocscience 2007), which also allowed for the calculation of equivalent Mohr– Coulomb parameters over appropriate stress ranges.

By integrating geological and geomechanical parameters, three units were differentiated in the engineering-geology model (Fig. 7): † Unit 1: thick-bedded micritic limestone and calcarenites; † Unit 2: massive and thick-bedded limestone;

Fig. 8. GSI chart, with the positioning of the average GSI values assessed for the rock masses in stages 0 –1 (grey circle and related representative photograph in the top right-hand corner) and 2 –3 (black circle and related representative photograph in the bottom right-hand corner), according to the criteria explained in the text.

210

Table 1. Geomechanical parameters adopted for the limestone deposits of the Maiella western slope Rock mass properties

Lithotechnical units: lithology

g (kg m23)

sci (MPa)

Elastic modulus (MPa)

Numerical model step

GSI

Elastic modulus (MPa)

Bulk modulus (MPa)

Shear modulus (MPa)

Cohesion (MPa)

Friction angle (8)

Tension cut-off (MPa)

Unit 1a: thinly bedded limestone Unit 1b: thinly bedded limestone Unit 2a: thick bedded limestone Unit 2b: thick bedded limestone

2699

90

3270

2690

90

3270

2699

100

5620

2690

100

5620

0, 1 2, 3 0, 1 2, 3 0, 1 2, 3 0, 1 2, 3

77 65 77 65 77 65 77 65

2760 2070 2760 2070 4740 3550 4740 3550

1840 1380 1840 1380 3160 2370 3160 2370

1100 8260 1100 8260 1900 1420 1900 1420

4.43 2.11 4.67 2.26 9.67 5.07 5.09 2.45

48 51 46 49 36 38 47 49

1.99 0.80 1.99 0.80 1.77 0.71 2.21 0.89

G. BIANCHI FASANI ET AL.

Intact rock properties

SLOPE EVOLUTION IN THE CARAMANICO VALLEY

211

Table 2. Geomechanical parameters adopted for the flysch deposits filling the bottom of the Caramanico Valley Lithotechnical units: lithology

g (kg m23)

Unit 3a: lower Flysch Unit 3b: upper Flysch

2000 2000

Bulk modulus (MPa) 667 66.7

Shear modulus (MPa)

Cohesion (MPa)

Friction angle (8)

Tension cut-off (MPa)

40

1

24

8

4

1

24

8

Fig. 9. (a) – (e) Plots showing the values of the shear moduli for each evolutionary stage of the numerical modelling (see the text for a detailed explanation).

212

† Unit 3: flysch sandstones).

G. BIANCHI FASANI ET AL.

deposits

(mudstones

and

Each unit was, in its turn, subdivided into subunits based on the variation of the equivalent strength and deformability parameters related to the position in the stratigraphic sequence. The geomechanical parameters adopted for each unit are reported in Tables 1 and 2.

Multi-stage numerical model The numerical modelling approach was defined to back-analyse the onset and development of slopescale gravitational processes in an evolving morphostructural setting during a wide timespan (from Lower Pleistocene onwards). For this reason, based on the morphostructural evolutionary model, different stages were considered to simulate the transition from a laterally confined slope to the present configuration. The model geometries adopted for the evolutionary stages that led to the present topography were assumed by considering the evolution of the slope-to-valley system as the development of a tectonic escarpment roughly coincident with the main fault trace. The numerical modelling was performed by the finite difference code FLAC 6.0 (ITASCA 2006) in static gravitational condition and using a 20 m-square mesh. The geomechanical behaviour and the constitutive model considered for each engineering-geology unit can be summarized as follows. † Unit 1: anisotropic homogeneous behaviour; Mohr– Coulomb elastic– perfectly plastic constitutive model. † Unit 2: isotropic homogeneous behaviour and ‘fault-controlled’ heterogeneous behaviour for the sector near the CFS (more in particular, random Gauss-normal statistically distribution of parameter values, i.e., varying within a statistically controlled range but without a spatial control); Mohr –Coulomb elastic –perfectly plastic constitutive model. † Unit 3: heterogeneous behaviour depending on the confining stresses (i.e. on depth); Mohr – Coulomb elastic –perfectly plastic constitutive model with ubiquitous joints. For the sequential modelling the following solutions were applied in order to take into account both the stress tectonic field at a regional scale and the stiffness decay of the rock mass after reaching yield conditions: † the horizontal/vertical stress ratio (K ) was fixed to a defined value at each node of the mesh; † the shear and bulk modulus were reduced by one order of magnitude after reaching yield conditions.

Stage 0. Initial conditions representative of stage a in the morphostructural model (Fig. 7). The GSI values of the rock mass, and the related strength and deformability parameters, have been assumed for this phase based on observations of the same lithologies that crop out in the less disturbed eastern margin of the Maiella Massif. The geodynamic-related stress ratio is assumed to have a value of 1.5, considering the compressive tectonics acting in the area: the Morrone Ridge and the Maiella Massif were, in fact, in an advanced phase of deformation under east-verging thrust tectonics. Stage 1. This phase simulates the uplift of the Maiella western slope caused by the activation of the CFS as a response to the activation of a backthrust system (stage b of the morphostructural model) (Fig. 9a– c). The present maximum elevation difference between the crest and the base of the carbonate slope is reached in this phase: in order to take into account the impulsive character of the tectonic uplift and to evaluate the model sensitivity to the alternation of activity and steady-state phases, the total vertical displacement of the fault has been divided into three sub-stages (1a, 1b, 1c), with the value of the stress ratio, K, kept to 1.5 because of the back-thrust regime. Stage 2. This stage takes into account the effect of the regional doming-like uplift (stage c of the morphostructural model) (Fig. 9d) in terms of variation of the stress ratio, K, that is now assumed to be equal to 0.5 as a consequence of the transition to an extensional geodynamic regime, as no significant elevation difference is generated in this phase. The flysch deposits filling the valley bottom are now reshaped, thus simulating the effects of post-uplift linear erosion that is particularly intense on such sandy–clayey deposits. Another significant variation in this stage is represented by the enlargement of the ‘fault-controlled’ lithotechnical zone in order to consider the formation of new tectonic shear zones such as the MMFZ, while the equivalent rock mass parameters of the other zones are now derived from the GSI observed at the present time. Stage 3. The engineering-geology section is now reshaped, until the present configuration, by removing a slice of rock mass from the slope roughly coincident with the most stressed and strained rock wedge (Fig. 9e).

Preliminary results and interpretation The analysis of the results obtained by the sequential numerical modelling in correspondence with stages 1a and 1b (i.e. back-thrust stage) shows a wide released zone that reaches yield conditions, with a

SLOPE EVOLUTION IN THE CARAMANICO VALLEY

depth of about 100 m below ground level and an extension of about 2 km far from the flat-top boundary. In the following stage, stage 1c, a wedge-like released zone, which reaches yield conditions, involves the Caramanico fault slope and corresponds well to the portion of the rock mass assumed to have failed to create the present topography. Moreover, the resulting released zones at the end of the back-thrust stage are associated with vertical failures within the rock mass, with a depth of up to 100 m. The whole released areas, as well as the above-mentioned vertical failures, correspond well to the rock mass presently involved in intense karst processes associated with high-elevation springs (about 1800 m a.s.l.). The released rock mass, which is assumed to have been removed by erosional processes to create the present topography, can be related to landslide phenomena, such as the Campo di Giove rock avalanche that could have occurred within stage 1c and 2 – that is, in the time interval between the end of the back-thrust stage and the beginning of the doming-like uplift stage. In the final conditions, corresponding to the present topography (i.e. stage 3), as a consequence of the widening of the ‘fault-controlled’ zone, vertical failures caused by a released rock mass that reaches yield conditions occur at distances of up to 1 km from the flat-top boundary; in agreement with field evidence of opened trenches and cracks parallel to the Caramanico Fault slope direction.

Conclusions Most of the large landslides identified and studied in the Central Apennines can be regarded as palaeolandslides: often the present morphostructural setting of many Apennine ridges seems to be inconsistent with stress –strain conditions able to produce such DSGSD or catastrophic rock slope failures as those implied from the geomorphical evidence (landforms and deposits) widely collected. The multidisciplinary approach adopted for this ‘pilot’ case highlights the importance of the ‘dynamic disequilibrium’ linked to intense tectonic activity phases. The preliminary sequential stress– strain analysis confirmed the relevance of the tectonic activity: if the extremely high rates of tectonic uplift in a short (from the geological point of view) time interval are considered, the development of the Maiella western slope can be regarded as the sum of quite instantaneous growth stages, with their related releasing effects. The latter are able to produce local stress regimes that lead to large-scale slope instabilities and slope movements. Finally, DSGSD mapping is important even in a ‘quiescent’ environment, such as the Central Apennines: the overall fabric inherited by the

213

superimposition of tectonics and gravity of some Apennine ridges represents a potential predisposing factor of rock slope catastrophic failures triggered by probable high-magnitude seismic events and/or other conventional triggers (ice/snow melting, intense rainfalls etc.).

References Accordi, G. & Carbone, F. 1988. Carta delle litofacies del Lazio-Abruzzo ed aree limitrofe, con note illustrative: C.N.R. – Progetto Finalizzato Geodinamica, scale 1:250 000, 1 sheet. Quaderni della Ricerca Scientifica, 114. Bernoulli, D., Anselmetti, F. S., Eberli, G. P., Mutti, M., Pignatti, J. S., Sanders, D. G. K. & Vecsei, A. 1996. Montagna della Maiella: the sedimentary and sequential evolution of a Bahamian-type carbonate platform of the south-Tethyan continental margin. Memorie Societa` Geolologica Italiana, 51, 7–12. Bigi, S., Calamita, F. & Paltrinieri, W. 1995. Modi e tempi della strutturazione della catena centroappeninica abruzzese dal Gran Sasso alla costa adriatica. Studi Geologici Camerti (Special Volume), 2, 77–85. Bosi, C., Galadini, F., Giaccio, B., Messina, P. & Sposato, A. 2003. Plio-Quaternary continental deposits in the Latium-Abruzzi Apennines: the correlation of geologic events across different intramontane basins. Il Quaternario, 16, 55– 76. Casini, S., Martino, M., Petitta, M. & Prestininzi, A. 2006. A physical analogue model to analyse interaction between tensile stresses and dissolution in carbonate slope. Hydrogeology Journal, 14, 1387– 1402. Cavinato, G. P. & De Celles, P. G. 1999. Extensional basins in the tectonically bimodal central Apennines fold-thrust belt, Italy: response to corner flow above a subducting slab in retrograde motion. Geology, 27, 956– 959. Centamore, E. & Nisio, S. 2003. Effects of uplift and tilting in the Central– Northern Apennines (Italy). Quaternary International, 101/102, 93–101. Cinti, G., Donati, A. & Scarascia-Mugnozza, G. 2001. La grande frana di Monte Arezzo (Abruzzo). Memorie Societa` Geolologica Italiana, 56, 41– 50. Crescenti, U. 1970. Osservazioni sul Pliocene degli Abruzzi settentrionali: la trasgressione del Pliocene medio e superiore. Bollettino Societa` Geolologica Italiana, 90, 3– 21. Crescenti, U., Costella, A., Donzelli, G. & Raffi, G. 1969. Stratigrafia della serie calcarea dal Lias al Miocene nella regione marchigiano-abruzzese. Memorie Societa` Geolologica Italiana, 8, 343– 420. D’Agostino, N., Jackson, J. A., Dramis, F. & Funiciello, R. 2001. Interactions between mantle upwelling, drainage evolution and active normal faulting: an example from the Central Apennines (Italy). Geophysical Journal International, 147, 475–497. Di Luzio, E., Bianchi-Fasani, G., Esposito, C., Saroli, M., Cavinato, G. P. & Scarascia-Mugnozza, G. 2004a. Massive rock-slope failure in the Central Apennines (Italy): the case of the Campo di Giove

214

G. BIANCHI FASANI ET AL.

rock avalanche. Bulletin of Engineering Geology and the Environment, 63, 1– 12. Di Luzio, E., Saroli, M., Esposito, C., Bianchi-Fasani, G., Cavinato, G. P. & Scarascia-Mugnozza, G. 2004b. Influence of structural framework on mountain slope deformation in the Maiella anticline (Central Apennines, Italy). Geomorphology, 60, 417 –432. Esposito, C., Martino, S. & Scarascia-Mugnozza, G. 2007. Mountain slope deformations along thrust fronts in jointed limestone: An equivalent continuum modelling approach. Geomorphology, 90, 55–72. Galadini, G. 2006. Quaternary tectonics and large-scale gravitational deformations with evidence of rock-slide displacements in the Central Apennines (Central Italy). Geomorphology, 82, 201–228. Ghisetti, F. & Vezzani, L. 2002. Normal faulting, extension and uplift in the outer thrust belt of the Central Apennines (Italy): role of the Caramanico fault. Basin Research, 14, 225– 236. Hoek, E. & Brown, E. T. 1997. Practical estimates of rock mass strength. International Journal of Rock Mechanics Mining Sciences & Geomechanics Abstracts, 348, 1165– 1186. Hoek, E., Carranza Torres, C. & Corkum, B. 2002. The Hoek–Brown failure criterion – 2002 edition. In: Proceedings of the 5th North American Rock Mechanics Symposium and 17th Tunneling Association of Canada Conference NARMS-TAC. University of Toronto Press, Toronto, 267– 271. ITASCA. 2006. FLAC, Fast Lagrangian Analysis of Continua, Version 6.0. Itasca Consulting Group, license: DST – ‘La Sapienza’, Roma (serial number: 213-039-0127-16143). Mariucci, M. T., Montone, P. & Pierdominici, S. 2006. New insights on active stress field in Italy and its implications with tectonics. Geophysical Research Abstracts, 8, 04112. Montone, P., Mariucci, M., Pierdominici, S. & Amicucci, L. 2005. Stress Regime in Italy: State of the art. American Geophysical Union. Fall Meeting 2005, abstract #T51C-1357. Mostardini, F. & Merlini, S. 1986. Appennino centromeridionale: sezioni geologiche e proposta di

modello strutturale. Memorie Societa` Geolologica Italiana, 35, 177–202. Paolucci, G., Pizzi, R. & Scarascia-Mugnozza, G. 2001. Analisi preliminare della frana di Lettopalena (Abruzzo). Memorie Societa` Geolologica Italiana, 56, 131– 137. Patacca, E., Scandone, P., Bellatalla, M., Perilli, N. & Santini, U. 1991. La zona di giunzione tra l’arco appenninico settentrionale e l’arco appenninico meridionale nell’Abruzzo e nel Molise. Studi Geologici Camerti (Special Volume), 1991/2, CROP 11, 417–441. Patacca, E., Scandone, P., Di Luzio, E., Cavinato, G. P. & Parotto, M. 2008. Structural architecture of the Central Apennines. Interpretation of the CROP 11 seismic profile from the adriatic coast to the orographic divide. Tectonics, 27, TC3007; doi: 10.1029/ 2005TC001917. Pizzi, A. 2003. Plio-Quaternary uplift rates in the outer zone of the Central Apennines fold-and-thrust belt, Italy. Quaternary International, 101–102, 229– 237. ROCSCIENCE. 2007. RocLab v. 1.031. Rocscience Inc., Toronto (free license). Scarascia-Mugnozza, G., Bianchi-Fasani, G., Esposito, C., Martino, S., Saroli, M., Di Luzio, E. & Evans, S. G. 2006. Rock avalanche and mountain slope deformation in a convex, dip-slope: the case of the Majella Massif (Central Italy). In: Evans, S. G., Scarascia-Mugnozza, G., Ermanns, R. & Strom, A. (eds) Massive Rock Slope Failure. Nato Science Series Book. Kluwer Academic, Dordrecht, 357–376. Scisciani, V., Calamita, F., Bigi, S., De Girolamo, C. & Paltrinieri, W. 2001. The influence of synorogenic normal faults on Pliocene thrust system development: the Maiella structure (Central Apennines). Memorie Societa` Geolologica Italiana, 55, 193–204. Vecsei, A. & Sanders, G. K. D. 1997. Sea-level highstand and lowstand shedding related to shelf margin aggradation and emersion, Upper Eocene– Oligocene of Maiella carbonate platform, Italy. Sedimentary Geology, 112, 219–234.

Valley shape influence on deformation mechanisms of rock slopes CHRISTIAN AMBROSI & GIOVANNI B. CROSTA* Dipartimento di Scienze Geologiche e Geotecnologie, Universita` degli Studi di Milano-Bicocca, Piazza della Scienza 4, 20126 Milano, Italy *Corresponding author (e-mail: [email protected]) Abstract: Stress distribution in mountainous areas is influenced by local morphology. Valley morphology and the relationship between main and tributary valleys strongly depend on geological characteristics and evolution. They may control the evolution of slope instabilities, especially when interacting with pervasive structural features. We performed parametric three-dimensional (3D) numerical modelling of simplified slope geometries with variable slope angle (from 218 to 358), length, combining different orientations for different slope sectors and changing attitude of pervasive planes of anisotropy (foliation, schistosity, bedding). Data used in the 3D models are the initial slope geometry, rock mass properties and internal anisotropy. We assumed Mohr– Coulomb behaviour, with the presence of ubiquitous joints and different piezometric levels. The model results show that plastic deformation initiates near the highest ridge just after deglaciation commences. A shear zone develops and propagates toward the toe of the slope, and its shape is strongly controlled by slope geometry, anisotropy and in situ stresses. The thickness of the failing mass, for model slope reliefs up to 3200 m, increases from 50 m to some hundreds of metres during glacier retreat, and it depends on geometry of slopes, anisotropy and in situ stresses. Results are compared to examples of deep-seated slope deformations from the Alps, which helps in the interpretation of such phenomena and in the understanding of their influence on valley evolution.

Evolution of landscape and slope topography, and of slope stability, is the result of the combined effects of multiple variables. These variables can be the lithology and its association with different geometries, anisotropies, structural features and their activity, state of stress, uplifting, and action of external or internal agents or processes (e.g. ice, glacier, rainfall, groundwater). Large rock slope movements, commonly described as sackung or deep-seated slope gravitational deformation, are common in areas of high relief (Zischinsky 1966; Beck 1967; Radbruch-Hall et al. 1976; Bovis 1982, 1990; Savage & Varnes 1987; Varnes et al. 1989; Chigira 1992; Agliardi et al. 2001). The movements can produce conspicuous geomorphical features (also called bedrock linears, and including tension cracks, double ridges, ridge top depressions, scarps and anti-slope scarps, trenches, toe bulging) and can induce secondary slope failures. Slow and continuous movements, over long time periods (Cruden & Hu 1993; Ballantyne 2002), can produce large cumulative displacements (Varnes et al. 1989; Bovis 1990; Ambrosi & Crosta 2006). The surface displacements range from a few millimetres to several centimetres per year (Colesanti et al. 2006; Crosta et al. 2008b), and can reach large values (tens to hundreds of metres) over a large timespan.

The distribution of large slope deformations can be related not only to relief and gravitational stresses, but also to glaciation and deglaciation cycles (Whalley 1983: e.g. glacial debuttressing), weathering, groundwater flow (Crosta 1996; Ballantyne 2002; Holm et al. 2004; Arsenault & Meigs 2005), tectonic and locked-in stresses (Savage et al. 1986; Miller & Dunne 1996), earthquake shaking and coseismic displacements, structural fabric, tectonic uplift, fluvial erosion of the slope toe, fatigue and time effects (Molnar 2004), and subsidence. Large rock slopes can evolve differently under the control of different geological and geomechanical factors, and of different boundary conditions. Slope geometry or local and large-scale topography, including the position of the unstable slope with respect to the main and tributary valleys, may have important implications for stress distribution and localization causing sackung development. Miller & Dunne (1996) and Molnar (2004) suggest positive feedback among accelerated, increased stress concentration and fracture of bedrock on the floor of the streambed. Savage et al. (1986) showed that differential topographical stress scales with the magnitude of relief. Deeper valleys are associated with greater stresses than do shallower ones with the same depth-to-width scale.

From: Jaboyedoff, M. (ed.) Slope Tectonics. Geological Society, London, Special Publications, 351, 215– 233. DOI: 10.1144/SP351.12 0305-8719/11/$15.00 # The Geological Society of London 2011.

216

C. AMBROSI & G. B. CROSTA

However, the state of stress at shallow depths within the valley slopes depends strongly on the particular form of the topography. As rock is much weaker under tensile stress than under purely compressive stress, failure is much more likely in those places where tension prevails. Furthermore, the presence of pervasive weakness planes and tectonic stresses (Pan et al. 1995; Miller & Dunne 1996) can control in situ stress distribution, the type and amount of deformation that a slope can undergo, and also the possible evolution from slow to rapid movements. Large-scale toppling-related slope deformations (Hutchinson 1988) are frequent in Alpine areas. Bovis (1990) suggested toppling as a possible mechanism for anti-slope scarp formation, and Nichol et al. (2002) modelled the development of such features and of related slope instabilities (ductile self stabilizing flexural toppling and brittle catastrophic block toppling) by distinct element models. Two-dimensional modelling of stress distributions have been previously examined for symmetric and asymmetric ridges with assumed elastic (isotropic properties: McTigue & Mei 1981, 1987; Savage et al. 1986; Savage & Varnes 1987; anisotropic properties: Liao et al. 1992; Pan et al. 1994, 1995; Savage 1994; Martel 2000; Hasebe & Wang 2003) and elasto-plastic properties (Kinakin & Stead 2005), including tectonic stresses (Augustinus 1995a; Miller & Dunne 1996) and static fatigue (Molnar 2004). Closed form solutions have been presented to evaluate the elastic stress distribution for symmetric and asymmetric ridges in a half plane by conformal mapping. Numerical methods have been used to evaluate failure in the presence of elastic, elasto-plastic and elasto-visco-plastic materials (Radbruch-Hall 1978; Sjoberg 1999; Martel & Muller 2000; Nichol et al. 2002; Ambrosi & Crosta 2006 just to quote a few). Numerical models allow the verification of ‘model’ sensitivity to some factors such as slope orientation, height, slope angle, orientation of structural features relative to slope orientation and geometry, and groundwater table. Kinakin & Stead (2005) demonstrated through an elasto-plastic constitutive model the link between slope geometry in two dimensions and stress field locations. Different distributions of horizontal, vertical and shear stresses, and that of modes of deformation, are characterized by different ridge types. Three-dimensional numerical models allow the simulation of slopes characterized by relatively complex geometries resulting in more realistic simulations of Alpine slopes affected by glacial and fluvial erosion processes. We studied the effects of slope and rock mass characteristics on their stability and their possible

evolution. To undertake such a scope, a series of idealized slope models were generated and simulated. Our aims in this research have been: to evaluate the sensitivity, in terms of stress and strain distributions, to the main boundary conditions (e.g. topography, slope angle and length, glacial unloading, water table geometry); to run 3D numerical models of slopes characterized by complex geometries similar to real Alpine slopes affected by glacial and fluvial erosion processes; and to verify the passive constraints applied by topography, strength and geometry of rock mass features on the pattern of slope deformations.

Constraints on Alpine valleys morphology As stated earlier, we suggest that valley morphology and glacial deepening, together with physical mechanical rock properties, groundwater and in situ stress conditions, can control the local stress distribution and, consequently, the evolution of rock failure and slope instabilities. Different approaches have been adopted to study the evolution of glacial valleys (Hirano & Aniya 1988, 1989; Harbor 1992a, b; Augustinus 1992, 1995a). Harbor (1992a, b, 1995) showed by numerical modelling that, during glacier flow through a V-shaped valley, erosion dominates towards the base of the valley because of the topographically induced stresses. This condition leads to steeper valley flanks in the lower slope sectors, and can be further controlled by the interaction between glacial erosion processes and erosion resistance, and by rock mass strength. Augustinus (1992) showed the relevance of rock erosion resistance and strength properties on glacial trough evolution. These constraints assume a different role during alternating glacial and interglacial periods because of the constraining action (glacial), the subsequent debuttressing (glacial –interglacial transition), the increased degrees of freedom (interglacial) and the removal of weakened slope materials by successive glaciations (glacial). Glacial erosion measurements are rare and only partial, and they suggest (Drewry 1986) that erosion rates range from 0.07 to 30 mm year21. Harbor (1992a, 1995) showed numerically that erosion rates increase away from the margins and reach a peak on the valley sides, causing the valley bottom to broaden and the valley walls to steepen into the classical U-shaped form. In addition, localization of glacial erosion or accentuated erosion can result from the occurrence of some controlling factors, in single or combined form, including: weak lithologies, structural features with specific orientation, the presence of antecedent slide deposits, in situ stress concentrations,

VALLEY SHAPE AND SLOPE DEFORMATION

groundwater pressure, glacier flow velocity and thickness. For example, Augustinus (1995a) suggested that where valley flanks are composed of high-strength rock masses, the stress concentration occurs at the slope toe. As a consequence, narrow valley deepening takes place because of localized failure, and the wall development is increased by high horizontal tectonic stresses. In the presence of low-strength rock masses, topographical stress concentration (increased by horizontal tectonic stresses) along the entire slope occurs with consequent large slope failures developing after deglaciation (debuttressing), leaving slopes more easily erodible by glaciers.

Field examples In order to show and verify the effect of topography and geological setting on the development of large slope instabilities, we analysed a landslide inventory available for the Western –Eastern Alps, covering Italy, France, Switzerland, Austria and Slovenia (Crosta et al. 2008a; see Fig. 1 for the full extent

217

of the study area and the location of the five case histories discussed herein). The different mapped phenomena present various characteristics (i.e. geometry, extent, morphostructures, etc.) that are not always explained by slope relief and geometry. Many phenomena affect only one valley slope, usually with the highest relief, but some affect contemporaneously slopes pertaining both to contiguous main and to tributary valleys, and show a control by structural features at different scales (e.g. foliation, schistosity, layering, pervasive jointing, master joints, fault planes). Here we present five representative case studies, taken by this inventory, by showing the general settings and the mapped structures (Fig. 2), and by introducing only the main geological features. These five cases have been chosen because of the relatively simple geometrical and geological conditions, thus avoiding the introduction of too many complex elements and interactions into the numerical models. In fact, each case study presents clear lateral boundaries, involving a main valley and one or

Fig. 1. Location map of the five examples from the Alps used for the comparison with model results. These examples (see dots in the figure) have been chosen from an inventory of DSGSDs (see the yellow polygons in the figure) compiled (Crosta et al. 2008a) for the Alpine mountain belt: (1) La Vachey, Ferret Valley, Mt Blanc, Italy; (2) Bosco del Conte, Valtellina, Italy; (3) Motto d’Arbino, Ticino, Switzerland; (4) Curon Venosta, Alto Adige, Italy; (5) Mt Ganda, Livigno, Italy.

218

C. AMBROSI & G. B. CROSTA

Fig. 2. Details of the five examples located in Figure 1. Main morphostructures are drawn on the hillshade models, together with the mean representative orientation of the pervasive/ubiquitous features (i.e. foliation, schistosity, layering). The corresponding simplified geometry and orientation of the structural planes adopted in the successive modelling phase are shown in the insets. Representative profiles of the slopes affected by the DSGSDs are shown in the lower left-hand corner.

two tributary valleys, a homogeneous main discontinuity attitude, and the presence of clear morphologies suggesting the occurrence of slope instabilities. All of the cases are located in glacial valleys where the average ice thickness during the Last Glacial Maximum (LGM) ranged between 1200 and 2000 m. The involved lithologies are metamorphic. The slope relief ranges between 800 and 1300 m. The simplified slope profiles, ranging from linear to convex and broken, are reasonably comparable with those modelled in this study. We observe that at some of the sites (Ferret Valley, at La Vachey, and the Venosta–Valle Lunga Valley, at Curon Venosta) the structural control and the local morphology generate conditions allowing the development of more than one slope deformation along the same valley flank. These two slopes are oriented in the same way, and are characterized by similar structural features and morphologies. Field data have been collected at the different sites and aerial photographic interpretation was performed (Crosta et al. 2008a); however, more details are not considered directly relevant to the subject of this paper.

flank of the Ferret Valley on the SE margin of the Mont Blanc Massif within the Italian territory. The instability affects slightly metamorphosed rocks, including ‘schists lustre´e’ and marbles, with subvertical –vertical layering moving from the slope toe to the crest and separated from calcschists by a NE-trending tectonic lineament. Active landslides (e.g. Plan Cereux, which reaches about 70 m in depth and velocities of up to 1 cm day21) are observed along the left-hand valley flank and within the DSGSDs, and frequently occur in the form of toppling of the more weakened superficial material.

Bosco del Conte – Upper Valtellina The Bosco del Conte DSGSD takes place in the Bormio phyllites, a few kilometres SE of the Livigno DSGSD, which is described later. This slope is formed prevalently in phyllites and paragneiss, and is limited in the upper slope by a thrust plane dipping within the slope, thus putting the phyllites into contact with the overlaying orthogneiss. The slope is affected by a series of scarps, counterscarps and trenches, and by a large successive landslide superimposed on the DSGSD (Agliardi et al. 2009).

La Vachey, Ferret Valley – Mt Blanc

Motto d’Arbino

This series of deep-seated gravitational slope deformations (DSGSDs) occurs along the left-hand

The Motto d’Arbino DSGSD is located in Switzerland close to the Insubric Line, a large transcurrent

VALLEY SHAPE AND SLOPE DEFORMATION

fault of regional relevance. Intensely deformed and fractured heterogeneous gneisses, marbles and amphibolites form the bedrock. The schistosity dips at a high angle within the slope and parallel to the geological sequence. The strike trends ENE parallel to the Insubric Line. Numerous NW– SE-trending faults and master joints have been observed, and are interpreted as secondary faults belonging to a Riedel system associated with a righthand motion along the Insubric Line (Seno 1992). According to in situ observations, toppling instability affects the intensely weathered rock mass. In 1928 a 30 Mm3 rock avalanche detached from the front of the Motto d’Arbino DSGSD, damming the Arbedo River and creating the Orbello Lake. The Arbino Landslide is currently active, with movement rates of the order of several millimetres per year.

Curon Venosta – Vallunga Valley The Vallunga Valley starts at Curon Venosta and trends to the NE. A series of DSGSDs occur along the northern valley flank, whereas some large more isolated ones affect the southern valley flank. DSGSDs involve biotitic paragneiss with amphibolitic intercalations and foliation dipping at high angles within the slope. All of the upper Venosta Valley is characterized by abundant DSGSDs, as presented by Crosta & Zanchi (2000) and Agliardi et al. (2009), and most of them occur in the same lithologies as the Vallunga Curon DSGSD.

Livigno – Mt Ganda – Upper Valtellina This slope is located in the Italian Central Alps (Lombardy) close to the Italy–Switzerland boundary. The instability takes place in orthogneiss and micaschists with downslope foliation. Two eastward downslope dipping structural features (faults and thrusts) are located close to the DSGSD, and the lower one forms the upslope limit of the instability. The DSGSD affects both the main valley slope and a tributary valley slope, characterized by different relief.

Modelling procedure These examples show that different constraints can control the evolution of DSGSDs, and especially their geometry, the type and distribution of morphostructures, and the associated displacements. To test these observations, and to help in the understanding and generalization of the observations, we decided to perform some parametric analyses by means of a three-dimensional (3D) numerical code. Three-dimensional continuum numerical models, based on the finite difference method

219

(FDM – Flac3D) and on elasto-plastic rock mass response, have been prepared. FLAC3D (Itasca 2002) is a commercially available code that can simulate the behaviour of 3D soil or rock structures. The object to be modelled is discretized through a 3D grid. Polyhedral zones of different shapes are used to build the models and are subdivided into tetrahedral subzones. Each element can behave according to a given linear or non-linear stress – strain law in response to applied forces or boundary restraints. The material can yield and flow, and the grid can deform and move with the material that is represented. The explicit Lagrangian calculation scheme and the mixed-discretization zoning technique ensure that plastic collapse and flow are modelled very accurately. The aim of the modelling study is to evaluate the sensitivity of the models, in terms of stress and strain distributions, to the main boundary conditions (e.g. topography, in situ stress distribution, slope angle and length, glacial unloading, water table geometry). As a consequence, we controlled topography, strength and geometry of rock mass features (i.e. pervasive discontinuities or foliation planes) because of the passive control they can exert on the pattern of slope deformations. Because of the large variety of landscape forms recognizable in nature, we have identified a simple topographical condition to perform numerical modelling. We generated a model represented by a main valley intersected at a 908 angle by a tributary valley. To verify the sensitivity of the model deformation to the structural setting, we performed a parametric analysis by considering a transversally isotropic material (to include the effect of anisotropies on in situ stress distribution) and the presence of pervasive discontinuities with 10 different attitudes (e.g. horizontal, vertical or dipping steeply in or out of the slopes). Finally, the influence of tectonic or abnormal in situ stresses on failure mechanisms has been tested by applying horizontal stresses (in this case, those acting only perpendicular to the main valley axis).

Model slope geometry Five idealized slopes have been set up with a similar basic geometry (Fig. 3), including both a main and a tributary valley. The main valley is characterized by a lower constant elevation with respect to the tributary (suspended) and they intersect at a right angle. Slope angles range between 218 and 458 in the different models. The maximum slope relief in the main valleys ranges from 1800 (218 slope), 2100 (288 slope), 3200 (358) and 3300 m (deepened valley, 358 and 458). Relief in the tributary valleys is slightly lower between 1500 (218 slope) and 1600 m (288 slope).

220

C. AMBROSI & G. B. CROSTA

Fig. 3. Geometrical description of the generated 3D models showing the main and the tributary valleys crossing each other at 908.

Fig. 4. Thirty-eight transversal valley profiles showing both the valley flanks affected by the DSGSDs (left-hand side) and the unaffected opposite valley flank (right-hand side). Slope geometry, relief and inclination are reported.

VALLEY SHAPE AND SLOPE DEFORMATION

However, different profiles could be adopted when considering the valley shape both above and below the present valley bottom, and the effect of sediment infilling. To show the average profile geometry and slope inclination for DSGSDs in the Alps and, in particular, in the study area, we extracted a series of 38 profiles along slopes affected by such phenomena (Fig. 4). The mean slope angle is 278 (+58) and the slope profile can be linear or broken in segments of different inclination, with relief varying from a few hundred metres to more than 1000 m, and valley bottoms ranging from large and flat, filled or not by alluvial deposits, to narrow and incised in bedrock. Conversely, slopes unaffected by DSGSDs have similar relief but are usually steeper (338 + 88) and with linear profiles. We represent in Figure 5 the main morphological features for models with constant slope angles (288– 358) and models with a broken profile. The lower part of the main slope has been considered flat as is reasonable for a large Alpine valley, and no valley infilling has been included. The slope toe has been rounded to avoid high stress concentrations. These concentrations are also avoided thanks to the adopted grid size. Broken profile models are obtained from the 358 slope model by increasing the slope angle of the lower half up to a maximum 458. This model describes the case of a U-shaped glaciated valley, with a tributary hanging valley, successively reshaped over the long term by further glacial and fluvial erosion, as well as by weathering and progressive superficial slope instabilities during paraglacial rock slope adjustment (Augustinus 1995a). Furthermore, Harbor (1995) suggests that erosion dominates towards the base of the valley because of the topographically induced stresses. This condition leads to steeper valley flanks in the lower slope sectors.

221

Finally, non-equilibrium conditions between river erosion and debris production from slopes can cause further valley incision, in addition to slope stability/instability, adding more constraints on valley profile geometry and evolution as well as on DSGSDs, in terms of age, rate of displacement and acceleration, geometry, and presence of accumulation. These geometries are relatively common in Alpine areas that have been subjected to glaciation, and especially for large slopes affected by deepseated gravitational deformations (DSGD) (Harbor 1992a, b, 1995; Augustinus 1995a, b; Ambrosi & Crosta 2006). We adopted in any case a simplified geometry and it must be mentioned that steeper slopes, with irregular profiles, can be easily found in such terrains. Furthermore, Crosta et al. (2008a) showed that many of the DSGSDs mapped in the Alps occur in areas with total slope relief larger than 1500 m, which is often indicated as a threshold value typical of tectonically active areas (Montgomery & Brandon 2002). We simulated also the initial presence of a glacier filling the valley with a thickness variable between 2100 and 3000 m, comparable to that most frequently observed in Alpine areas (Florineth 1998; Kelly 2003; Kelly et al. 2004; Ambrosi & Crosta 2006) and its sequential melting and removal.

Model generation We discretized the model through a 3D grid, made up of prismatic grid cells (Fig. 6), and generated by means of a routine written in a built-in programming language. The final 3D shape is given by a number of i-zones in the x-direction, and j-zones in the y-direction. Each zone is defined as a column of

Fig. 5. Detailed description of the models used for numerical simulations. Three of the five different geometries are shown to show the changes in slope inclination, relief and the relative representative profiles (linear, with constant slope; and broken, with different slope sectors).

222

C. AMBROSI & G. B. CROSTA

Fig. 6. Example of the generated 3D grid adopted for the analyses.

blocks with constant vertical size. The vertices at the base of each zone have the same elevation. The procedure includes the following steps: (1) generation of a vector topographical map; (2) drawing of a series of orthogonal profiles; (3) identification of intersection points between profiles and contour lines; and (4) building of the 3D geometry. To limit computing time, a coarse grid cell with a size up of to 200 m was generated. The models include a total of more than 10 000 cells.

Model properties We adopted a Mohr–Coulomb yield rule associated with a ubiquitous joint model to simulate the presence of pervasive discontinuities controlling the failure mechanism. The rock mass behaviour is assumed to be elastic –perfectly plastic. Internal anisotropy allowed us to include a pervasive ‘structural’ feature with a specific dip and dip direction. Geomechanical properties for the slope material were set to average values characteristic of many

metamorphic rocks (e.g. phyllite, slate, paragneiss, gneiss: Kulhawy 1975), typical of an Alpine environment, and for which large slow rock slope deformations have been observed and studied (Zischinsky 1966; Agliardi et al. 2001; Ambrosi & Crosta 2006). We decided not to consider fatigue effects and degradation of physical and mechanical properties of the rock mass, even if this can happen during the long periods of time under which this type of phenomena take place. In fact, rock mass damage and shear localization (cataclastic zones) have been observed or suspected in very large rock slope instabilities at large depths (70 –200 m; Agliardi et al. 2001; Molnar 2004; Ambrosi & Crosta 2006; Seno & Thu¨ring 2006) and in the presence of pre-existing fault zones. These conditions could suggest the use of a strain-softening constitutive law or in some cases, visco-plastic behaviour. Values adopted in this study for the rock mass properties are summarized in Table 1 both for

Table 1. Mechanical properties adopted for the analyses

Rock mass properties Ubiquitous joint properties

K (MPa)

G (MPa)

t (kPa)

c (kPa)

fm (8)

r (kN m23)

2000

1000

200 100

1000 800

20 20

27

Abbreviations: K, bulk modulus; G, shear modulus; c, cohesion; t, tensile strength; fm, internal friction angle; r, bulk density.

VALLEY SHAPE AND SLOPE DEFORMATION

intact rock and for the pervasive set of weakness planes. As previously noted, we assumed that no weathering and mechanical degradation occurred along the valley flanks. However, they could have influenced the final slope geometry (Harbor 1995) of the natural slope just after the deglaciation or the occurrence of successive glaciations. Furthermore, rock masses with different initial physical–mechanical characteristics (strength, degree of fracturing, lithology) could undergo weathering at different rates, and length of glacial and interglacial periods could control the development of different valley shapes. For each slope geometry, 10 different discontinuity orientations (Fig. 7) have been introduced into the models. The 10 cases include high-angle (708) dip sloping and anti-dip sloping ubiquitous joints, with direction parallel to the trend of the main or tributary valley or at a 458 angle with respect to the trend of both valleys, and horizontal anisotropy planes. We generated a total of 30 different models. Glacier retreat has been simulated in all the models by sequentially deleting layers of ice initially filling the valley. The glacier mass was subdivided into four ice sheets to simulate a slow melting. Ice was modelled as an elastic material (bulk density ranging from 9.4 to 9.8 kN m23 with depth; bulk modulus, K, 5.6 GPa; and shear modulus, KG, 2.4 GPa). Water table conditions were imposed with a water table depth of 40 m below the topographical surface. This assumes larger groundwater availability during deglaciation.

Results Simple linear slope model The models show different degrees of deformation with the slope angle and attitude of pervasive (i.e. ubiquitous) planes, and for the adopted physical– mechanical properties. We do not observe significant displacements and failure for slopes below 288. So we present here the results for models characterized by slope angles of 288 and 358. The results are presented in terms of total displacements and shear strain increments for these 288 and 358 linear slope profile models because they verify the higher deformations. We observe that the amount, pattern and ‘localization’ of the deformation are strongly controlled by the orientation of the weakness planes for the same boundary conditions, as well as by the type of slope profile (continuous v. broken). We imposed zero horizontal displacement normal to the sides and fixed the base of the model.

223

Six out of the 10 models, obtained by combining ubiquitous planes attitude and slope morphology, present relevant slope deformation. These models are presented in Figure 7, whereas in Figure 8 we summarized the change in shape of the unstable mass for two profiles along the main and the tributary valley slopes. The most unstable conditions are those for models 2, 3 and 8 in Figure 7. These conditions are those for steeply dipping–vertical sloping anisotropy planes with direction parallel to the valley (main or tributary). Maximum slope displacements are computed for ubiquitous planes oriented parallel to the main valley, but the general pattern is quite similar for this model and the one with planes parallel to the tributary valley (model 10). For anisotropy planes dipping at 708 (models 1–4 in Fig. 7) the displacements affect the slope all the way up to the upper crest where the slope angle changes, but the toe of the moving mass does not affect the valley bottom. The basal shear zone of the moving mass is subparallel to the valley flank. Anti-dip slope planes, with directions parallel to the tributary valley, also induce instabilities along the main valley flank resulting in the worst case. For anti-dip sloping anisotropy planes with direction parallel to the main valley (model 2, Fig. 7), the failure zone climbs up the slope beyond the crest. This is the main difference with respect to the dip slope case. For vertical ubiquitous planes (models 8 and 10, Figs 7 & 8) the unstable sector propagates upwards beyond the slope ridge and part of the valley bottom is affected by the displacements. The base of the moving mass is more curved, even if not subcircular, and shear strain is distributed within the entire failing mass. The displacements are concentrated along the most unstable slopes and almost no movement is visible on the other slope sector, so there is no condition for which both slopes become unstable. Models that include horizontal ubiquitous planes show (model 7, Figs 7 & 8) displacements of the entire slope up to the main slope knick point. A welldefined subcircular shear zone daylights at the slope toe, along the main valley, but does not fully develop uphill. As for the other models, the unstable mass is limited to the main valley flank, but some minor displacements are visible along the tributary valley suggesting a dragging effect on the inner slope sector. The case with subvertical ubiquitous planes, oblique (transversal) with respect to both the main and the tributary valleys (model 9, Figs 7 & 8), is characterized by displacements that occur throughout the entire model up to the highest elevation. The valley bottom is not affected by the displacements.

224

C. AMBROSI & G. B. CROSTA

Fig. 7. Distribution of the total displacements computed for models characterized by linear slope profiles with two different inclinations (288 and 358). In the left-hand column the 10 simplified geometry and structural setting are shown.

VALLEY SHAPE AND SLOPE DEFORMATION

Fig. 8. Sketch showing the effect of different orientations of the anisotropy planes on the extent of the unstable mass. Results for some models are shown as isocontour lines of displacements along two profiles for the main and tributary valley slopes.

Finally, as shown in Figure 7, it is possible to observe that the shear zone development within the slope is controlled, in terms of both extension and thickness, by the slope geometry and the attitude of the weakness planes with respect to the slopes and the valleys. Increasing slope inclination results in a greater magnitude of displacements distributed in a similar way and at slightly higher depth, and a condition of dynamic equilibrium. In some cases the displacements become more spread (e.g. models 7 and 10) and in other cases less concentrated (e.g. models 2–4 and 6).

Broken slope profile models Valleys with complex geological histories (e.g. multiple glaciations, fluvio-glacial deepening) can undergo a different evolution under similar boundary conditions when compared to simple slope profiles. A series of models with a broken slope profile, with two main slope sectors inclined at 358 and 458, have been generated and simulations run with differently oriented ubiquitous planes. Results are shown in terms of total displacements in Figure 9 and can be directly compared with those obtained for linear slope profiles (at 358). The general effect of this new slope geometry, with respect to the linear profile slopes, is to intensify by increasing and localizing (see, e.g. models

225

1 –3 and 7 in Fig. 9) slope displacements along the main valley flank and, for ubiquitous planes, parallel to the tributary valley. At the same time the average and maximum depth of the slope movement increases with slope gradient. It is clear that our models in this case can represent a simplification of the real evolution because our simulations started with a broken slope profile and did not simulate the superposition of effects caused by the formation of the initial (linear) slope profile and its successive deepening. The general effect of the progressive steepening and of the presence of a broken slope profile is summarized in Figure 10 for a profile along the main valley slope. In this figure we observe the progressive increase of displacements with slope angle and the migration of the lower limit of the unstable mass. The orientation of the anisotropy planes strongly influences both the extent of the mass and its geometry and position along the slope. However, this influence is evident predominantly when a broken slope profile is adopted. In the case of a toppling or flexural failure mechanism (models 2 and 8) the unstable mass rises along the slope above its upper limit and becomes more elongated, with an average inclination of the basal zone comparable to that of the slope. As a consequence, this geometry controls the maximum depth of the failing mass. A different evolution is recognized for models 1 and 7. In particular, a chair-like instability, which extends only to the upper slope limit, is generated for horizontal anisotropy planes (model 7 in Figs 9 & 10). In this case, the maximum depth of the unstable mass grows considerably with respect to the other cases.

Effects of in situ stress regime Augustinus (1995b), Miller & Dunne (1996) and Molnar (2004) suggested that the stress regime can be relevant for development of large slope instabilities. Brown & Hoek (1978) and McGarr & Gay (1978) proposed that very high horizontal stresses are common in some structural settings. We simulated the effects of different stress regimes by combining gravity and tectonic stresses. The increase in tectonic stress was simulated by increasing the horizontal stresses in the model (Fig. 11). The ratio, K, between the horizontal and the vertical stresses has been used to describe the assumed tectonic stress regime (Brown & Hoek 1978). The K value for gravity-only loading can be assumed to be 0.3 and so we simulated three cases for K equal to 0.3 (e.g. in areas subjected to crustal extension or gravity-only loading), to 1 (e.g. weakly compressional and strike slip conditions) and 2 (e.g. convergence areas and thrust faulting). The maximum

226

C. AMBROSI & G. B. CROSTA

Fig. 9. Distribution of the total displacements computed for models characterized by linear 358 inclination and by broken slope profiles (358 and 458). In the left-hand column the simplified geometry and structural setting are shown.

VALLEY SHAPE AND SLOPE DEFORMATION

227

Total displacements (m) Slope 28°

10

10

50

30

65

0 0 150 10

50

10

25

15

250

100 50

600

300

20

25 20 15

15

20 0 100

15

10

20

10

Model 8

100

50

30

Model 2

10

10

50

15

Model 1

Model 7

Broken profile

Slope 35°

150

50

0

100

0

70

50

100

Fig. 10. Distribution of total displacements as a function of slope profile geometry at varying orientations of anisotropy planes. Results are shown for slopes of different geometry.

Fig. 11. Sketch of the applied in situ stresses. Sxx was incremented through the K ratio (horizontal to the vertical stress ratio) to study the change in behaviour under the action of tectonic stresses.

228

C. AMBROSI & G. B. CROSTA

Fig. 12. Distribution of the total displacements computed for models characterized by linear 358 inclinations under different in situ stress conditions. The K ratio (horizontal to vertical stress ratio) was imposed equal to 0.3, 1 and 2. In the left-hand column the simplified geometry and structural setting are shown.

horizontal stress was always imposed normal to the main valley (see Fig. 11). We observed a general increase in total displacement with increasing K values (Fig. 12). Furthermore, displacement becomes more developed along the main valley flank (normal to the increasing stress), with respect to the tributary valley flank, with an increase in K. A large part of the model, including both the main valley and the tributary valley slopes, is affected and shows a general displacement towards the main valley bottom. A particular pattern of deformation is recognized for subhorizontal planes of anisotropy. In this case, a stepped profile is generated at the toe of the valley flank with localized subhorizontal zones of larger displacements. The number of these localizations increases with the K ratio (see model 7, for K ¼ 1 and K ¼ 3, in Fig. 12). This behaviour also influences the general distribution of displacements with depth, with the basal failure zones moving progressively deeper and characterized by a general

low inclination (see both the model 1 and model 7 cross-sections in Fig. 12). The joint effect of an increasing stress and a different orientation of the anisotropy planes is summarized in Figure 13 for a 358 main valley slope. Results are shown for three different values of the K ratio (0.3, 1 and 2) and three different models (1, 2 and 7). As described earlier, we observed an increase in depth of the moving sector, with an almost bi-planar or chair-like sliding surface for models 1 and 2. Maximum displacements were computed for model 1, whereas the sliding surface becomes more subcircular from model 1 to model 7, with the lower depth for model 7. Furthermore, model 1 clearly shows the maximum uphill extent of the moving sector.

Discussion and conclusions The purpose of this study was to examine failure modes and to identify those that lead to major

VALLEY SHAPE AND SLOPE DEFORMATION

229

Fig. 13. Distribution of total displacements along the main valley slope profile for the 358 inclination model and different orientations of the anisotropy planes at varying values of the K ratio. Isocontour lines of displacements are shown for three models.

slope displacements in the presence of various or different geometrical and geological constraints. Valley geometry and the relative orientation of structural features can strongly control the stress and strain distribution within mountain slopes. As a consequence, in this study we analysed the relationship between some relatively simple valley geometries and structural features in terms of sensitivity to induce large slope deformations (Figs 3 & 7). Different valley transversal profiles have been introduced and modelled, together with homogeneously distributed anisotropies and tectonic stresses. We showed that even a simple 3D slope geometry, in combination with highly persistent discontinuities and glacial retreat, leads to the evolution of large-scale slope failures. Of the various possible failure mechanisms, toppling and dip-slip sliding are important. The pattern of deformation is strongly related to the constraining factors introduced in the analyses. As a consequence, we observed different shear zone geometries (shape, thickness, depth) and extension (below or beyond the slope crest) with different orientations of the ubiquitous planes (dip slope or anti-dip slope) with

respect to the two valleys (parallel, perpendicular, transversal). Orientation of the anisotropy planes plays a major role in the deformation. We observed that, in most of the cases but not all, the main valley (i.e. the one with highest relief) is mainly affected by slope instabilities and that the most unstable slopes are those containing vertical –subvertical ubiquitous planes (Figs 7–10 & 12). Linear slope profiles with constant inclination are less unstable than broken slope profiles (see Fig. 10), typical of valleys deepened by subsequent erosion phases. This erosion phase could be the result of rapid bedrock river erosion or the superimposed effects of successive glacial phases. In both cases, deepening is associated with an increase in the average slope angle and total relief, and, consequently, in stress intensity and concentration. As pointed out, we did not simulate the sequential formation of the broken slopes as obviously happens in nature. This can be at the origin of the particular or ‘anomalous’ geometries of slope instabilities observed in nature (e.g. suspended or more developed DSGSDs and associated features along the upper slope sectors of Alpine valley flanks).

230

C. AMBROSI & G. B. CROSTA

The existing tectonic regime can be another constraint to large slope deformations. We modelled the effects of different stress distributions by varying the ratio between horizontal and vertical stresses (Figs 12 & 13). Again, the increase in this ratio is associated with an increase in slope instability. This increase occurs both in terms of total displacement, strain localization and in the enlargement of the affected area (Fig. 13). In fact, both the main valley and the tributary valley slopes are affected in the case of applied higher stress ratios. We observe that passing from the linear to the broken slopes (Figs 7– 10) the depth of the affected zone increases slightly. According to model results, this becomes even more evident when in situ

horizontal stresses are increased. The DSGSD depth passes from about 50– 100 m up to more than 200 m with the increase in slope relief and the change in geometry of the slope profile. Much deeper masses are found when a chair-like shear zone develops because of particular anisotropies (e.g. model 7 for steep and broken profiles) and the in situ stresses (e.g. model 1). We have tried to compare the results of our modelling with examples of large DSGSDs in the Alps (Figs 1 & 2). From an inventory completed for most of the Alpine belt (Crosta et al. 2008a), and including about 900 DSGSDs and more than 1000 large landslides, we observed that many of these are characterized by the type of geometrical and structural conditions modelled in this study.

Fig. 14. Examples of complex DSGSDs (in the light grey colour) that are not directly modelled in this study but are frequently observed. (a) DSGSD corresponding to valley concavity, with its body narrowing down towards the valley bottom (confined-constricted toe); (b) DSGSD reactivating antecedent structural features and affecting only a part of the slopes included between two tributary valleys; (c) DSGSD on a dissected slope and subdivided by slope instabilities of different orders (other DSGSDs and large rotational landslides); (d) DSGSD affecting a long sector of an Alpine valley including more tributary valleys and some highly dissected slopes. DSGSDs involve both flanks of a ridge with different features and extent.

VALLEY SHAPE AND SLOPE DEFORMATION

Similar features have been simulated using physical models by Bachman et al. (2006) by running tests on 3D models, including main and tributary valleys or complex ridge and valley patterns. They demonstrated that morphology in the third dimension strongly controls the slope failure evolution. However, many other cases are located in more complex conditions (Fig. 14). In some cases, the DSGSDs can affect long sectors of the valley flanks, including more tributary valleys within themselves and sometimes very dissected slopes (Fig. 14b, c). In other cases, DSGSDs are found in valley concavities (Fig. 14a), like small glacial cirques, and their toes are confined and constricted by the main body narrowing down towards the main valley thalweg. We did not simulate this type of condition and our analyses considered only a part of the possible conditions (e.g. narrow and deeply cut main valleys, or partially infilled by sediments, abnormal groundwater conditions, main valley geometry concave or convex in the horizontal plane, very steep slopes). Other case studies are represented by long and continuous mountain ridges, sometimes symmetrical and in other cases asymmetrical (e.g. Fig. 14d). In these cases both ubiquitous structural features (foliation, schistosity, cleavage, jointing) and major structural– tectonic lineaments seem to play a major controlling action on DSGSDs that can develop along a single ridge flank or both flanks. Furthermore, at the scale of the affected mountain slopes and valleys, and of such slope deformations, these structural features can change in orientation thus controlling the evolution of the phenomena and causing instabilities within slopes with different orientation (Ambrosi & Crosta 2006). Beyond these, many other points need still to be analysed and discussed, and will be part of successive studies. A list of them can include the following. Are the observed slope deformations the result of the last glaciation cycle or do they originate by the successive action of different glacial cycles? What is the role (geometrical, physical, mechanical) of large weakness features in controlling the triggering and initiation of these type of instabilities? How do these features control the groundwater flow in association with the main structural anisotropies? What are the most reasonable assumptions about the groundwater flow and regime during a glacial cycle? What is the role of tectonic uplift and of the rate of valley erosion? What has been the history of slope activity since the triggering of the movement? What is the present state of activity of these phenomena? These points are of major interest both to assess potential hazards and also to help in understanding the general rock mass behaviour and properties. Some of the results show slopes in conditions of

231

dynamic equilibrium, then subjected to constant displacement rates except in the presence of changes in the boundary conditions (e.g. valley erosion, steepening and deepening, rising of the groundwater table). PS-InSAR (Interferometric Synthetic Aperture Radar using Permanent Scatterers) analyses (Crosta et al. 2008a) show that this seems a normal behaviour for many of the slopes mapped in the regional inventory covering the Alpine mountain belt (Crosta et al. 2008a, b). The average displacement rate ranges between 0 and 40 mm year21, with a maximum frequency in the 0–10 mm year21 interval. At the same time the average slope inclination of well-developed DSGSDs, in similar lithologies, ranges within a relatively narrow interval, suggesting a similar set of values for material properties and a possibly complete and well-developed failure plane. Strain localization in our models seems to suggest such a possibility, and the same is implied by slope displacement monitoring and in-depth field observations (Crosta et al. 2008a). As a consequence, many slopes with gradient lower than fresh deglaciated slopes, but larger than the typical slope for DSGSDs, may represent slopes in state of ‘dynamic equilibrium’. We thank an anonymous reviewer and also J. Hutchinson, for her careful review that greatly improved the quality of the manuscript.

References Agliardi, F., Crosta, G. & Zanchi, A. 2001. Structural constraints on deep-seated slope deformation kinematics. Engineering Geology, 59, 83–102. Agliardi, F., Zanchi, A. & Crosta, G. B. 2009. Tectonic vs. gravitational morphostructures in the central Eastern Alps (Italy): Constraints on the recent evolution of the mountain range. Tectonophysics, 474, 250– 270; doi: 10.1016/j.tecto.2009.02.019. Ambrosi, C. & Crosta, G. B. 2006. Large sackung along major tectonic features in the Central Italian Alps. Engineering Geology, 83, 183–200. Arsenault, A. M. & Meigs, A. J. 2005. Contribution of deep-seated bedrock landslides to erosion of a glaciated basin in southern Alaska. Earth Surface Processes and Landforms, 30, 1111– 1125. Augustinus, P. C. 1992. The influence of rock mass strength on glacial valley cross-profile morphometry: a case study from the Southern Alps, New Zealand. Earth Surface Processes and Landforms 17, 39– 51. Augustinus, P. C. 1995a. Rock mass strength and the stability of some glacial valley slopes. Zeitschrift fur Geomorphologie, 39, 55–68. Augustinus, P. C. 1995b. Glacial valley cross-profile development: the influence of in situ rock stress and rock mass strength, with examples from the southern Alps, New Zealand. Geomorphology, 14, 87– 97. Bachmann, D., Bouissou, S. & Chemenda, A. 2006. Influence of large scale topography on gravitational

232

C. AMBROSI & G. B. CROSTA

rock mass movements: New insights from physical modelling. Geophysical Research Letters, 33, L21406; doi: 10.1029/2006GL028028, 2006. Ballantyne, C. K. 2002. Paraglacial geomorphology. Quaternary Science Reviews, 21, 1935– 2017. Beck, A. C. 1967. Gravity faulting as a mechanism of topographic adjustment. New Zealand Journal of Geology and Geophysics, 11, 191– 199. Bovis, M. J. 1982. Uphill-facing (antislope) scarps in the Coast Mountains, southwest British Columbia. Geologicl Society of American Bulletin, 93, 804–812. Bovis, M. J. 1990. Rock-slope deformation at Affliction Creek, southern Coast Mountains, British Columbia. Canadian Journal of Earth Sciences, 27, 243– 254. Brown, E. T. & Hoek, E. 1978. Trends in relationships between measured in situ stresses and depth. International Journal of Rock Mechanics and Mining Sciences & Geomechanics Abstracts, 15, 211– 215. Chigira, M. 1992. Long-term gravitational deformation of rock by mass rock creep. Engineering Geolology, 32, 157– 184. Colesanti, C., Crosta, G. B., Ferretti, A. & Ambrosi, C. 2006. Monitoring and assessing the state of activity of slope instabilities by the Permanent Scatterers Technique. In: Evans, S. G., Scarascia Mugnozza, G., Strom, A. & Hermanns, R. L. (eds) Massive Rock Slope Failure: New Models for Hazard Assessment. NATO Advanced Research Workshop, Celano, Italy. Kluwer, Dordrecht, 1, 179– 212. Crosta, G. 1996. Landslide, spreading, deep-seated gravitational deformation: analysis, examples, problems and proposals. Geografia Fisica e Dinamica Quaternaria, 19, 297– 313. Crosta, G. B., Agliardi, F., Frattini, P. & Zanchi, A. 2008a. Alpine inventory of Deep-Seated Gravitational Slope Deformations. In: European Geosciences Union General Assembly 2008, Vienna, Austria, 13– 18 April 2008. European Geosciences Union, Munich, EGU2008-A-02709. Crosta, G. B., Allievi, J., Frattini, P., Giannico, C. & Lari, S. 2008b. Monitoring Deep-Seated Gravitational Slope Deformations in the Alps by PSInSARTM techniques. In: European Geosciences Union General Assembly 2008,Vienna, Austria, 13–18 April 2008. European Geosciences Union, Munich, EGU2008A-10662. Crosta, G. B. & Zanchi, A. 2000. Deep-seated slope deformations. Huge, extraordinary, enigmatic phenomena. In: Bromhead, E., Dixon, N. & Ibsen, M. (eds) Landslides in Research, Theory and Practice. Proceedings of the 8th International Symposium Landslides, Cardiff, June 2000. Thomas Telford, London, 351– 358. Cruden, D. M. & Hu, X. Q. 1993. Exhaustion and steady-state models for predicting landslide hazards in the Canadian Rocky Mountains. Geomorphology, 8, 279– 285. Drewry, D. 1986. Glacial Geologic Processes. Edward Arnold, London. Florineth, D. 1998. Geometry of the Last Glacial Maximum (LGM) in the Swiss Alps and its paleoclimatological significance. PhD thesis, University of Bern.

Harbor, J. M. 1992a. Numerical modelling of the development of U-shaped valleys by glacial erosion. Geological Society of American Bulletin, 104, 1364–1375. Harbor, J. M. 1992b. Application of a general sliding law to simulating flow in a glacier cross section. Journal of Glaciology, 38, 182–190. Harbor, J. M. 1995. Development of glacial-valley cross sections under conditions of spatially variable resistance to erosion. Geomorphology, 14, 99–107. Hasebe, N. & Wang, X. F. 2003. Irregular elastic halfplane gravity problem. International Journal of Rock Mechanics and Mining Sciences & Geomechanics Abstracts, 40, 863–875. Hirano, M. & Aniya, M. 1988. A rational explanation of cross-profile morphology for glacial valleys and of glacial valley development. Earth Surface Processes and Landforms, 13, 707 –716. Hirano, M. & Aniya, M. 1989. A rational explanation of cross-profile morphology for glacial valleys and of glacial valley development: a further note. Earth Surface Processes and Landforms, 14, 173–174. Holm, K., Bovis, M. & Jakob, M. 2004. The landslide response of alpine basins to post-Little Ice Age glacial thinning and retreat in southwestern. British Columbia. Geomorphology, 57, 201–216. Hutchinson, J. N. 1988. Morphological and geotechnical parameters of landslides in relation to geology and hydrogeology. In: Proceedings of the 5th International Symposium on Landslides, Lausanne, Switzerland, Volume 1. A. A. Balkema, Rotterdam, 3 –35. ITASCA. 2002. Fast Lagrangian Analysis of Continua in 3 Dimensions, User’s manual, Version 2.1. Itasca Consulting Group, Inc., Minneapolis, MN. Kelly, M. A. 2003. The late Wu¨rmian Age in the western Swiss Alps: last glacial maximum (LGM)ice-surface reconstruction and 10Be dating of late-glacial features: [Valais]. Dissertation Naturwissenschaften, Bern. Kelly, M. A., Buoncristiani, J. F. & Schluchter, C. 2004. A reconstruction of the last glacial maximum (LGM) ice-surface geometry in the western Swiss Alps and contiguous Alpine regions in Italy and France. Eclogae Geologicae Helvetiae, 97, 57–75. Kinakin, D. & Stead, D. 2005. Analysis of the distributions of stress in natural ridge forms: implications for the deformation mechanisms of rock slopes and the formation of a sackung. Geomorphology, 65, 85–100. Kulhawy, F. H. 1975. Stress deformation properties of rock and rock discontinuities. Engineering Geology, 9, 327–350. Liao, J. J., Savage, W. Z. & Amadei, B. 1992. Gravitational stresses in anisotropic ridges and valleys with small slopes. Journal of Geophysics Research, 97, 3325– 3336. Martel, S. J. 2000. Modeling elastic stresses in long ridges with the displacement discontinuity method. Pure and Applied Geophysics, 157, 1009–1038. Martel, S. J. & Muller, J. R. 2000. A two-dimensional boundary element method for calculating the elastic gravitational stresses in slopes. Pure and Applied Geophysics, 157, 989– 1007. McGarr, A. & Gay, N. C. 1978. State of stress in the earth’s crust. Annual Review of Earth and Planetary Sciences, 6, 405–436.

VALLEY SHAPE AND SLOPE DEFORMATION McTigue, D. F. & Mei, C. C. 1981. Gravity-induced stresses near topography of small slope. Journal of Geophysics Research, Solid Earth Series, 86, 9268–9278. McTigue, D. F. & Mei, C. C. 1987. Gravity-induced stresses near axisymmetric topography of small slope. International Journal for Numerical and Analytical Methods in Geomechanics, 11, 257–268. Miller, D. J. & Dunne, T. 1996. Topographic perturbations of regional stresses and consequent bedrock fracturing. Journal of Geophysical Research, 101, 25 523 –25 536. Molnar, P. 2004. Interactions among topographically induced elastic stress, static fatigue, and valley incision. Journal of Geophysical Research, 109, F02010; doi: 10.1029/2003/JF000097. Montgomery, D. R. & Brandon, M. T. 2002. Topographic controls on erosion rates in tectonically active mountain ranges. Earth and Planetary Science Letters 201, 481– 489. Nichol, S. L., Hungr, O. & Evans, S. G. 2002. Largescale brittle and ductile toppling of rock slopes. Canadian Geotechnical Journal, 39, 773– 788. Pan, E., Amadei, B. & Savage, W. 1994. Gravitational stresses in long symmetric ridges and valleys in anisotropic rock. International Journal of Rock Mechanics and Mining Sciences & Geomechanics Abstracts, 31, 293–312. Pan, E., Amadei, B. & Savage, W. 1995. Gravitational and tectonic stresses in anisotropic rock with irregular topography. International Journal of Rock Mechanics and Mining Sciences & Geomechanics Abstracts, 32, 201–214. Radbruch-Hall, D. 1978. Gravitational creep of rock masses on slopes. In: Voight, B. (ed.) Rockslides and Avalanches; Natural Phenomena. Elsevier, Amsterdam, 1, 607– 657.

233

Radbruch-Hall, D., Varnes, D. J. & Savage, W. Z. 1976. Gravitational spreading of steep-sided ridges (‘sackung’) in Western United States. Bulletin of Engineering Geology and the Environment, 13, 23–35. Savage, W. Z. 1994. Gravity induced stresses in finite slopes. International Journal of Rock Mechanics and Mining Sciences & Geomechanics Abstracts, 31, 471– 483. Savage, W. Z. & Varnes, D. J. 1987. Mechanics of gravitational spreading of steep-sides ridges (sackung). International Association for Engineering Geology Bulletin, 35, 31– 36. Savage, W. Z., Swolfs, H. S. & Powers, P. S. 1986. Gravitational stresses in long symmetric ridges and valleys. International Journal of Rock Mechanics and Mining Sciences & Geomechanics Abstracts, 22, 291– 302. Seno, S. 1992. Direzioni fluviali e tettonica recente in Canton Ticino (Svizzera). Atti Ticinesi di Scienze della Terra, 35, 213–230. Seno, S. & Thu¨ring, M. 2006. Large landslides in Ticino, Southern Switzerland: Geometry and kinematics. Engineering Geology, 83, 109–119. Sjoberg, J. 1999. Analysis of Large Scale Rock Slopes. PhD thesis, Lulea University of Technology. Varnes, D. J., Radbruch-Hall, D. & Savage, W. Z. 1989. Topographic and structural conditions in areas of gravitational spreading of ridges in the western United States. United States Geological Survey, Professional Paper, 1496, 1 –28. Whalley, W. B. 1983. The magnitude and frequency of large rockslides in Iceland in the postglacial. Geografiska Annaler, 65A, 99–110. Zischinsky, U. 1966. On the deformation of high slopes. In: Proceedings of the First International Conference of the International Society of Rock Mechanics, Lisbon, Volume 2. International Society of Rock Mechanics, Lisbon, 179–185.

The Celentino deep-seated gravitational slope deformation (DSGSD): structural and geomechanical analyses (Peio Valley, NE Italy) M. GHIROTTI1*, S. MARTIN2 & R. GENEVOIS2 1

Department of Earth and Geo-Environmental Sciences, University of Bologna, Via Zamboni, 67, 40127 Bologna, Italy

2

Department of Geosciences, University of Padova, Via Giotto 1, 35137 Padova, Italy *Corresponding author (e-mail: [email protected]) Abstract: Deep-seated gravitational slope deformations (DSGSDs) are fairly common phenomena in the Eastern Italian Alps. The Celentino DSGSDs (Trentino–Alto Adige Region, Peio Valley) extends over an area of approximately 5 km2 between 2400 and 1050 m a.s.l. (metres above sea level), involving metamorphic rocks. A set of parallel ridge-top trenches and antislope scarps are present in the upper part of the slope over a length of 2 km. The shear surface is inferred at 80–100 m in depth, and the involved volume is estimated to be 3.5  108 –4.0  108 m3. The displacement of the rock mass has diverted the Noce River for 200 –250 m. Structural, geomorphological and engineering geology surveys were conducted both within and outside the DSGSDs. The evolution of the whole ridge has been modelled using the two-dimensional (2D) finite element code Phase2, simulating the unloading of the glacier cover and assuming a progressive damage of the rock mass. The numerical analyses demonstrate that continuum modelling can be used to examine the evolution of stresses, strains and plastic yielding, providing results consistent with field observations. Structural control due to brittle geological structures seems to be extremely important in the location, shape and extent of the DSGSDs, and post-glacial debuttressing an important predisposing factor in its development and triggering. The DSGSDs are common phenomena in areas of high relief energy, associated with orogenic history and mountain unroofing where tectonics is still active.

Deep-seated gravitational slope deformations (DSGSDs) in rock slopes are a special category of mass movement, where small-scale movements prevail. The terms sackung or sagging have also been used (Hutchinson 1988; Dikau et al. 1996). Deformation is not restricted to a sliding or a narrow sliding zone but affects the whole moving rock mass. The DSGSD processes are characterized by an extremely slow displacement rate (0.4–5 mm year21) (Varnes et al. 1990); however, their slow, long-term deformations can damage surface structures (Agliardi et al. 2001; Mannucci et al. 2004; Ambrosi & Crosta 2006) Their evolution is rather complex and a catastrophic slope event such as rock avalanches or rockslides cannot be completely excluded (Evans & Couture 2002; Crosta & Agliardi 2003). Moreover, one of the main factors in DSGSD development is the effect of the stress distribution on natural ridges, particularly in formerly glaciated regions (Ballantyne 2002; Kinakin & Stead 2005). Links between stresses in the ridges and the formation of the typical features of DSGSDs have been postulated by several authors (Dramis & Sorriso Valvo 1994; Bovis & Evans

1996), the controlling factor for their formation being a combination of the rock mass discontinuities and the stress conditions. Continuum numerical modelling can be used to examine the evolution of stresses, strains and plastic yielding within the rock mass during the formation of large slope instabilities (Sitharam et al. 2001; Esposito et al. 2007). Generally, numerical analyses of these phenomena are carried out at the single slope scale (Forlati et al. 2001; Hu¨rlimann et al. 2006; Apuani et al. 2007) but very few papers analyse the whole ridge behaviour (Kinakin & Stead 2005). This last approach seems promising, even if the adoption of an equivalent continuum, incorporating both intact rock and discontinuity effects, is a simplified representation of the rock mass behaviour. DSGSDs in the Alps are widely distributed within in the Cevedale Massif area, in particular in the Campo-Ortler Nappe (Agliardi et al. 2001, 2009a; Martin 2008) because of its lithological composition and structural setting (Martin 2007). In this area the Peio Valley belongs to one of the most glaciated areas of the Italian Alps, where morphostructures and moraines have been attributed to Late Pleistocene–Holocene (Late Glacial Time)

From: Jaboyedoff, M. (ed.) Slope Tectonics. Geological Society, London, Special Publications, 351, 235– 251. DOI: 10.1144/SP351.13 0305-8719/11/$15.00 # The Geological Society of London 2011.

236

M. GHIROTTI ET AL.

(Gruppo Nazionale di Geografia Fisica e Geomorfologia CNR 1986; Dal Piaz et al. 2007). In this valley the slopes show evidence of post-glacial rebound morphostructures that may have facilitated the generation of DSGSDs (Martin 2007). In the Eastern Alps the DSGSDs predominantly depend on local extensional processes (Pazzaglia et al. 2007) that strongly influence the mountain landscape and stream drainage, in spite of the regional convergence caused by the Adria –Europe collision. However, DSGSDs also occur associated with other types of active tectonic structures as discussed by Pasquare´ (2001) and Ambrosi & Crosta (2006). In this paper we analyse the Celentino DSGSD, located on the northeastern side of the Peio Valley (Cevedale Massif, western Trentino), involving gneiss and micaschist rocks along an Alpine tectonic contact, presumably active during the Late Glacial Time (19 –11.5 ka BP) (Ravazzi et al. 2007). The main objectives of this paper are to investigate the relationship between tectonic structures and slope deformation for the Celentino area, and to model the ridge evolution considering a progressive glacier lowering. The approach used is based on structural, geomorphological and geomechanical surveys, carried out within and outside the DSGSD.

Geological setting At a regional scale, the study area is part of the Cevedale Massif, whose structural setting is due to a polyphased post-orogenic extension driven by Late Cretaceous and Neogene –Quaternary faults (Martin 2007). It corresponds to a km-scale NE-trending antiform (Martin et al. 1998). The exhumation of the Cevedale Massif area sensu lato started at the end of Cretaceous, facilitated by extensional shear zones such as the Peio Fault (Martin et al. 1991; Viola et al. 2003). Other regional faults and thrusts are present in this area; namely, the Tonale destral strike-slip fault to the south; the Cima Grande Thrust and the Lake Careser Fault to the north; and the Rumo Fault, the Cima Sternai Fault and the Giudicarie transpressive fault to the east (Fig. 1) (Martin et al. 2007). They were active at different times between the Late Cretaceous and Oligocene (Mu¨ller et al. 2001), and then reactivated during the Giudicarie contraction that has occurred since the Miocene (Castellarin et al. 1992; Vigano` et al. 2008). The Tonale Nappe basement has been involved in the pop-up Giudicarie tectonics that determined the formation of SE- and NW-verging tectonic brittle planes of movement (Viola et al. 2003). In the Cevedale Massif area several DSGSDs occur (Montresor & Martin 2007; Agliardi et al.

2009a, b), associated with extensional–transtensional faults of the Austroalpine Nappe that exhumed the Cevedale Massif. Among these DSGSDs the Celentino is located in the Peio Valley (between Celledizzo and Cusiano villages), excavated by the Noce River along a tectonic lineament into the Tonale Nappe gneisses. The Celentino area is characterized by local tectonic lineaments and morphostructures related to the NE-trending steeply dipping C.ma Vegaia Fault (parallel to the Giudicarie Fault), to some ENE-trending morphostructures and shear zones (parallel to the Tonale and Peio Faults), and to a NNW-trending fracture system parallel to the Noce lineament (Fig. 1b). The NE to NNE and the NNW and NW lineaments are considered still tectonically active on the basis of data yielded by the local seismic network of the Provincia Autonoma di Trento (http://www.protezionecivile.tn.it). The Tonale Nappe includes the Tonale Unit and the overlying Ulten Unit, both composed of highgrade gneisses (Fig. 1a). The former unit is characterized by sillimanite-bearing micaschists and quartzose gneisses rich in pegmatite, orthogneiss, marble, amphibolite and minor magnetite-bearing rocks; the latter, by monotonous kyanite –garnetbearing paragneisses with frequent quartz veins and lenses (Andreatta 1954). The Tonale and Ulten units are separated by a mylonitic horizon above a mineralized marble lens (Fig. 2a, b), not easily recognizable south of Celledizzo because of the Celentino DSGSD cover (Fig. 2c). This horizon has been observed in an abandoned magnetite mine at 1187 m a.s.l. Moraines, glacial debris, alluvial deposits and scree dated to the Upper Pleistocene are widespread along all of the Peio Valley slopes (Desio 1967; Martin 2007).

The Celentino DSGSD Geomorphological setting The DSGSD extends over an area of approximately 5 km2 between 2400 m and 1050 m a.s.l., including the village of Celentino (Fig. 3). It involves quartzrich paragneiss of the Ulten Unit, with a low-angle SE-dipping schistosity. The crown develops mainly along vertical fracture planes oriented NNW– SSE and parallel to the Peio Valley, reaching a maximum elevation of 2641 m a.s.l., and almost coinciding with the local watershed. Towards the NNW, the crown follows a SE-dipping fault plane. A set of almost parallel ridge top troughs and tensional antislope scarps characterizes the upper part of the slope, forming two main trenches developing

THE CELENTINO DSGSD (PEIO VALLEY, ITALY) 237

Fig. 1. (a) Geological sketch of the western Trentino (modified after Martin 2008). Geological units: 1, Presanella tonalite; 2, low-grade Val di Sole schists; 3, the Tonale Unit; gneiss and micaschists; 3a, the Ulten Unit; gneiss and migmatite; 4, the Campo-Ortler Nappe; 4a, Verdignana orthogneiss; 5, Quaternary deposits. (b) Lineaments and faults in the Peio Valley derived by airborne LiDAR (Light Detection And Ranging) photo-interpretation [digital elevation model (DEM) 10 m resolution].

238

M. GHIROTTI ET AL.

Fig. 2. (a–c) Geological sections of the Celentino DSGSD: the legend is in agreement with the 1:25 000 Peio geological sheet (Martin 2007); (d) geological section traces in the digital elevation model (DEM).

over a length of about 2 km with depressions and counterscarps. The NNW –SSE main scarp of the northern segment of the upper trench (50 –70 m wide) is covered by lichens, and only locally tectonically reactivated by steep (508– 608) west-dipping fracture planes with limited debris deposits. The upper trench, at 2150 m a.s.l., becomes narrower and deviates to the SSW; the troughs are better developed here and bordered by moraine-like debris ridges (Fig. 4b). Towards the south, the bottom of the trench becomes wider and more uneven with elongated (hundreds of metres long) moraine-like structures. At 2210 and 2310 m a.s.l. the skeletal moraines are composed of massive matrix-supported diamicton with angular blocks

(Fig. 4b). These moraine-like debris ridges could be related to the glacial reworking of oldest scree deposits. Other smaller and shorter trenches have been observed all over the upper part of the slope. Up to 1500 m a.s.l. the rock slope is covered by thick glacial undifferentiated deposits; at higher elevations coarse gravity deposits are present, covered by woods (Fig. 4a). Beginning at 1300 m a.s.l., the slope shows a pronounced bulging that influences the surface drainage (Fig. 3a). Based on surveys carried out inside the abandoned Celentino mine, the depth of the shear zone of the DSGSD at the same elevation may be inferred to be 80–100 m (Fig. 2a, b).

THE CELENTINO DSGSD (PEIO VALLEY, ITALY)

Fig. 3. Celentino DSGSD geological and geomorphological map.

239

240

M. GHIROTTI ET AL.

Fig. 4. (a) Frontal view of the Celentino DSGSD: main geomorphological and structural features; (b) the trench at 2210 m elevation with articulated glacial morphology and part of the lateral moraines. In the sketch a detail of the skeletal moraine is indicated.

Structural and geomechanical setting Major tectonic features border the DSGSD and strongly influence the DSGSD kinematics. Stuctural and geomechanical surveys were conducted within and outside the DSGSD (Fig. 5). The schistosity dips to the SE at a low angle (mean dip direction/mean dip: 1208/208). The fault plot shows very steep up to vertical NEtrending surfaces, and other NW-trending steep dipping planes. The NE direction coincides with that of the main local lineament (Fig. 1b), delimitating the Celentino DSGSD northwards, along which steep gullies of the DSGSD and small streams developed. The ENE-trending weakly SSE-dipping fracture cleavages (Nicolas 1987) reactivated the schistosity, in agreement with the Giudicarie brittle tectonics (Fig. 5, fractures and cleavages plot). In addition, very steep ENE-trending fracture planes occur at the base of the 2641 m peak north of Malga Campo. A NNW– SSE-trending trench system occurs parallel to the ridge itself and locally coincides with the Noce tectonic lineament (Fig. 1b). Discontinuity orientation measurements and their characteristics (ISRM 1978) were recorded at 14 field-based scanlines (locations and stereoplots shown in Fig. 6). The main attitude of the discontinuity sets, spacing, persistence, aperture and the joint roughness coefficient (JRC) (Barton 1976), are reported in Table 1. The geological strength index (GSI), developed by Hoek & Brown (1997) as a quantitative means of relating field observations to rock mass quality, is based on the structural conditions (number of joint sets, joint density, etc.) and on surface/weathering conditions of discontinuities observed in the rock mass. The GSI has been estimated inside and outside the DSGSD (Fig. 6).

Based on geomorphological observations and geomechanical data, three different units have been defined inside the DSGSD. Unit A corresponds to the crown and to the trenches area – sets J2 and J4 are closely related to the strike of the trench slope and antislope scarps – it is limited by a vertical lineament oriented NNW–SSE and is considered still tectonically active. Geomechanical data collected in this area support a scatter distribution around the main orientation sets and a disaggregated structure of the rock mass. Unit B corresponds to the NE-trending C.ma Vegaia Fault, which represents the lateral-release surface of the DSGSD: the attitude of the discontinuity set J3 provides the field expression of the main local fault. Unit C represents the main body of the DSGSD, widely covered by Quaternary deposits and vegetation; its limits were identified mainly by geomorphic evidence. The exterior (‘background’) DSGSD rock masses (units D and E in Fig. 6) are characterized by a medium alteration grade on the discontinuities surface and by a mean value of the uniaxial compressive strength (UCS), estimated using the Schmidt hammer test, of 70 –80 MPa. The average value of GSI in these areas is 55. Inside the Celentino DSGSD the characteristics of the rock masses are generally poorer. In Unit A (crown and trenches area) the surface-weathering conditions are very high, the UCS is 12–20 MPa and the average GSI is the lowest recorded (GSI ¼ 36). In Unit B the alteration grade is again very high, the UCS ranges between 20 and 30 Mpa, and the average GSI is 38. Unit C is characterized by an high grade of alteration, an elevated degree of fracturing of the rock mass with open discontinuities and a high variability of the schistosity planes owing to the rotation suffered by the mass. The mean value of the GSI estimated on the few available outcrops is about 40.

THE CELENTINO DSGSD (PEIO VALLEY, ITALY)

241

Fig. 5. Schmidt lower-hemisphere projection of the main structural [schistosity, faults, fractures and cleavages, joints and geomorphological linear features (trenches), and rose diagrams] frequency distributions of brittle tectonic structures. The number of collected data is given in brackets.

242 M. GHIROTTI ET AL.

Fig. 6. Geomorphological units accompanying the contour plot of discontinuity poles (Schmidt lower-hemisphere projection) and the mean GSI values.

THE CELENTINO DSGSD (PEIO VALLEY, ITALY)

243

Table 1. Summary of the geomechanical observations performed in the field (according to ISRM 1978). The location of the survey sites is reported in Figure 6 Discontinuity set J1 J2 J3 J4 SC

Inside Outside Inside Outside Inside Outside Inside Outside Inside Outside

Mean dip direction/dip

Spacing (m)

Persistence (m)

Aperture (mm)

JRC (Barton 1976)

1608/858 1608/858 358– 708/708 408/708 1408/758 Not observed 2408/508 2708/508 808–1508/158– 258 1208/158 – 208

0.4 0.5 0.4 0.2 0.3 – 0.3 0.3 – –

1.7 3.2 1.5 2.0 1.5 – 2.0 1.5 – –

2.0 –3.0 2.0 –3.0 2.0 –10.0 2.0 3.0 –10.0 – 3.0 –14.0 2.0 – –

8 10 9 12 9 – 10 12 11 12

The size of the tectonics lineaments that border the Celentino DSGSD, and the contrast in rock quality between the rock masses damaged by specific tectonic structures and the exterior ‘background’ rock masses, enabled us to map damage zones associated with the faults and shear zones. Zones of different rock mass damage were thus outlined in the headscarp and the sidescarp areas. Brittle tectonics structures have thus played an important role in reducing the rock mass quality and strength in the headscarp and in the lateralrelease surface of the Celentino DSGSD, thereby facilitating the development of a rear-release surface. When considering the undisturbed ‘background’ rock mass characteristics (GSI of about 55– 60) and those of the zones associated with tectonic damage or slow, long-term, deformations (GSI of 30–40), it follows that both the structure and surface conditions move downwards and to the right in the GSI table. Thus highlighting the progressive degradation of mechanical characteristics that accompanied the DSGSD evolution. This aspect can be explicitly taken into account in numerical modelling, using different decreasing GSI values to reflect the reduction in block size strength and the increase in the alteration of discontinuities.

Two-dimensional finite element modelling A range of complex factors, such as jointing, the in situ stress condition, and other geological and geomorphological features, affect the behaviour of a rock slope. Generally, the numerical modelling of a rock mass slope is carried out at the single slope scale, and does not consider the full ridge crosssection. However, the effects of the stress distribution on the natural topography in a ridge may have important implications for DSGSD development. Links between stresses in the ridges and the

formation of ridge-top trenches, antiscarps and tension cracks have been postulated by several authors (Dramis & Sorriso Valvo 1994; Bovis & Evans 1996). Kinakin & Stead (2005) investigated the spatial correlation between stress concentrations and sackung landforms in deglaciated regions of British Columbia and Colorado by means of numerical modelling. They adopted finite difference models in conjunction with an elasto-plastic constitutive criterion, based on a combined use of GSI and Hoek –Brown criterion (Marinos & Hoek 2000; Hoek et al. 2002; Marinos et al. 2005), obtaining specific field locations of trenches, antislope scarps and tension cracks for different ridge profiles (Fig. 7). In order to simulate the evolution of the Celentino DSGSD, a multistage stress– strain numerical modelling was carried out using the finite element continuum code Phase2 (Rocscience Inc. 2007). The discontinuous rock mass was then treated as an equivalent continuum in a large-scale analysis. The equivalent mechanical characteristics of the rock mass were derived using the Hoek –Brown criterion. The slope model assumes a constant dipping slope with a uniform lithology (orthogneiss and paragneiss) without a water table, and the modelling was performed simulating the lowering of the glacier, starting from its maximum elevation. The simulation examined the evolution of the whole Celentino ridge, considering that the rock mass deformation was related to the rock mass property degradation and to the consequent development of a deep shear surface. The mechanical behaviour of the rock masses involved was defined on the basis of in situ tests and GSI values mapped in the field. The simulation was carried out on the whole ridge subject to glacier loading and considering different schistosity dip angles. Successive unloadings steps, resulting from glacier retreat and

244

M. GHIROTTI ET AL.

Fig. 7. Mean ridge forms, stress distribution and slope deformation processes (modified from Kinakin & Stead 2005). The typical field locations of sackungen are compared to concentrations of stress produced in this study.

corresponding to moraine deposit elevations, were performed together with the progressive degradation of the rock mass properties, starting from a rock mass condition of GSI ¼ 60 down to GSI ¼ 40, a value corresponding to that of the rock mass inside the DSGSD.

Model geometry According to Cruden & Hu (1999), who distinguish different ridge form only on the basis of morphology, ignoring the geological setting, the Celentino ridge belongs to the Dogtooth ridge profile. In fact, few mountains affected by glacial erosion processes show symmetric profiles as a consequence of

differential erosion of smaller glaciers on one side of the ridge and of larger valley glaciers on the other. The Celentino ridge was affected, up to the elevation of about 2300 m a.s.l., by a thick glacial cover during the Late Glacial Time, and subsequently by post-glacial erosion processes and intensive debutressing. The geological section of the Celentino ridge (Fig. 2a) was chosen and simplified for the numerical analysis (Fig. 8a) with the 2D finite element code Phase2 (Rocscience Inc. 2007), adopting a sixnoded quadrilateral element mesh (Fig. 8b). The pre-event topography of the Celentino slope is based on the morphology of the surrounding areas and on a palinspastic reconstruction of the slope.

THE CELENTINO DSGSD (PEIO VALLEY, ITALY) Fig. 8. Numerical modelling: (a) simplified geological section used for the model; (b) finite element grid and stages of deglaciation; (c) vertical displacement contours (the marked line separates negative values from positive ones); (d) contour lines of the safety factor (for details see Fig. 9). 245

246 M. GHIROTTI ET AL.

Fig. 9. Final distribution of the safety factor for the Celentino slope. Tensile and shear failure zones are shown.

THE CELENTINO DSGSD (PEIO VALLEY, ITALY)

Topography is assumed constant because the effects of local morphology are less relevant compared to the total dimension of the model (length 8000 m and height 2300 m). Finally, the presence of the thrust (Fig. 2a) has not been considered because, due to its attitude, it involves not only a small sector of the DSGSD but also separates rock masses with similar geomechanical characteristics.

Boundary conditions The initial in situ stress ratio is an important but often overlooked input parameter, as it strongly controls the direction of the failure surface propagation. Data on the in situ stress state were obtained from hydraulic fracturing tests carried out 3 km NW of Celentino under an overburden of 250 m (ENEL 1994 pers. comm.). The measured horizontal to vertical stress ratio, K, is approximately 0.30. Stresses in the model were consequently initialized assuming this value. Gravity and tectonic loadings were considered in the numerical model, assuming a homogeneous, isotropic rock mass without the inclusion of both porewater pressures and discontinuities. The rock mass was considered to be completely dry as it was hypothesized that the glacier-related waters would have been rapidly drained, during each post-glacial phase, by the valley bottom river. The generalized Hoek –Brown failure criterion for jointed rock masses (Hoek et al. 2002) estimates static rock mass strength taking into account the specific stress conditions of slopes and the disturbance caused to rock masses owing to rock blasting or other factors (D factor). To apply the Hoek– Brown failure criteria and thus obtain the strength and deformability of jointed rock masses, three properties of the rock mass must be estimated (Cai et al. 2004): † the uniaxial compressive strength (UCS) of the intact rock: obtained from laboratory triaxial or uniaxial tests; † the Hoek –Brown constant, mi, for the intact rock: obtained from laboratory triaxial compressive tests; † the GSI of the rock mass: from field survey. When laboratory testing is not performed, mi and UCS may be obtained from published tables (Hoek & Brown 1997). For the Celentino modelling, the rock mass strength parameters have been evaluated by in situ tests and field surveys (Hoek et al. 2002). Intact Young’s modulus (Ei) was estimated using the equation Ei ¼ MR  UCS using the modulus ratio, MR, proposed by Deere (1968) and modified by Hoek & Diederichs (2006), and implemented in Roclab software (Rocscience Inc. 2002). In

247

addition, Hoek & Diederichs (2006) suggested assigning a disturbance factor D ¼ 1 in using the Hoek –Brown failure criterion to estimate rock mass properties relevant for slope stability. The adopted criterion requires values of the Hoek –Brown constant mb (related to the particle interlocking) and of parameter s (related to the degree of fracturing and representing the internal friction angle of intact rocks). The initial numerical model adopted rock mass properties derived from a GSI value of 60 (average value outside the DSGSD), UCS ¼ 80 MPa and mi ¼ 28 (typical value for gneiss), and for a minimum principle stress, s3, as proposed for slopes by Hoek et al. (2002).

Steps and material properties In order to investigate the ridge evolution during subsequent glacial unloadings, the initial model has considered the presence of a glacier reaching up to the top of the ridge (about 2300 m a.s.l.) and filling both of the valleys. After establishing equilibrium conditions in the model, successive stages have been simulated that consider a progressive glacier lowering, in three steps based on the moraine elevations observed in the field (Fig. 8b). The rock mass properties were then lowered in accordance with each considered glacial elevation. Initial material properties were scaled for each stage of modelling through the proposed approach of GSI reduction. The GSI decreased from 60, the undisturbed rock mass value, down to 30 –40, corresponding to the mean actual value inside the DSGSD. The rock mass deformation modulus (Erm) was not considered constant, but scaled as implemented in RocLab software (Rocscience Inc. 2002) (Table 2). Finally, we used a ubiquitous joint model (implemented in Phase2) to simulate the low-angle pervasive schistosity planes (with equivalent Mohr–Coulomb parameters w ¼ 278 and c ¼ 0.1 MPa).

Numerical modelling results Obtained results show that plastic deformations initiate near the crest of the ridge just after the deglaciation starts. At the end of the deglaciation process, the whole ridge profile shows that the maximum vertical downwards displacements (negative) are concentrated below the crest on both sides of the ridge as a consequence of the stress distribution, thus indicating the initiation of a sackungtype phenomenon. On the valley sides the vertical displacements are positive (upwards) due to the glacial rebound (Fig. 8c). If the present mechanical properties of the DSGSD rock mass were introduced into the model

248

M. GHIROTTI ET AL.

Table 2. Hoek–Brown material properties used for Phase2 modelling*

mb s Erm

Stage 0 GSI ¼ 60 D¼1

Stage 1 GSI ¼ 50 D¼1

1.608 0.0013 5166

0.787 0.0002 2801

Stage 2 GSI ¼ 40 D¼1

Stage 3 GSI ¼ 30 D¼1

0.385 4.5  1025 1677

0.189 8.6  1026 1185

*Properties were derived from field data and scaled values from RocLab. The rock density was assumed to be 2800 kg m23 (Bell 1987; Apuani et al. 2007).

at the beginning of the simulation, successive slope failures would be generated at each deglaciation step without resulting in a single deep-seated slope deformation. Numerical simulation results are very sensitive to the rock mass strength anisotropy: the distribution of the safety factor (Fig. 8d), obtained after considering different dip-angle schistosity, highlights a different response of the ridge slopes to the imposed conditions. Introducing the measured dip angle (i.e. Celentino slope), a deep and widespread slope movement may be observed on the main slope and only a shallow one on the secondary slope. Analysing in detail the Celentino DSGSD slope behaviour (Fig. 9), the rock mass reaches a tensile failure state only near the surface, from the crest down to an elevation of about 2100 m a.s.l. The entire tensile failure zone corresponds to the location of trenches as observed in the field (Fig. 3). Down-slope, and partially overlapping this area, a shear surface is evidenced, reaching a depth of about 100 m at the toe. The abandoned mine gallery crosses the shear surface at a depth similar to that observed in the field (approximately 100 m). The presence of a little bulging of the mass at the slope toe and the local shape of the shear surface might also lead to the presence of some rotational components of the general movement being considered.

Discussion and conclusion Sackung, frequently considered as relict, must be reconsidered from both a mechanical and a dynamic perspective, as they may suddenly accelerate and develop into large catastrophic rockslides and avalanches. In formerly glaciated rock slopes, the interaction of changing stress conditions associated with both glacial over-steepening and the relaxation of residual stresses following unloading and debuttressing, and rock mass strength controlled by lithology and jointing, may produce a complex time-dependent response. This complex condition

may lead to the development of rock slope failure during deglaciation, or delayed post-glacial rockslope deformation or failure (Ballantyne 2002). With the aim of investigating slope evolution during glacier unloading, a DSGSD, extending along the whole northeastern slope of the Peio Valley near the village of Celentino, was chosen for analysis because of its fairly simple geological and structural setting. Typical landforms produced by slow rock mass movements, including convex bulging slopes, ridge-top depressions, multiple rock ridges, uphill-facing scarps (antislope scarps) and a classical rearrangement of the screes as moraine-like structures, are present in the slope. Structural field surveys have pointed out close relationships between the observed morphostructures of the Celentino slope and the regional and local tectonics patterns: trench counterscarps and joint directions (Fig. 5) coincide with the direction of the Peio Valley and the related lineament (Fig. 1b). The NE-trending C.ma Vegaia Fault acted as a lateral right boundary to the DSGSD area, whereas the NNW-trending fault facilitated the main detachment just below the ridge crest. The structural setting also played a distinct role in conditioning the Celentino slope evolution as the rock mass strength was degraded by faulting, and so increasing the degree of fracturing (i.e. the number of degree of freedom in the rock mass kinematics) and the effects of alteration. A detailed in situ rock mass characterization has provided an assessment, by means of GSI estimate, of the intensity and extent of damage related both to brittle tectonic structures and to rock slope deformation. Field observations show that the DSGSD started before the Wu¨rmian glaciation: the initial phase of the DSGSD may, thus, be mainly related to the glacier debuttressing in the late interglacial age, which could have caused the propagation or the reactivation of pre-existing fractures. The results of the numerical analyses suggests that gravitational deformations occur on both slopes of the ridge, giving rise to a deep deformation

THE CELENTINO DSGSD (PEIO VALLEY, ITALY)

on the main slope and a shallow landslide on the secondary slope of the ridge, in agreement with the actual observed situation. These conditions have been reached by combining the degradation of the rock mass properties with the deglaciation stages. The relaxation of tensile stresses within a rock mass, exposed to decreasing loads as the glacier melts, causes a stress release within the rock mass, the magnitude of rebound being dependent on the amount of both the residual strain energy and the modulus of elasticity of the rock. Deglaciation simulation, together with a rock mass strength degradation, seems to be sufficient to explain the development of the DSGSD that is dominated by tensile fractures close to the crest, followed by a deep shear failure surface due to the rock mass degradation in association with a precise rock mass strength anisotropy. The orientation of planes of weakness (schistosity) strongly controls the shape and the localization of the shear surface. Different results, in terms of vertical downward displacements, location of tensile zones, displacements distribution, and depth and shape of the shear zone, have been obtained by varying the anisotropy direction. Introducing into the Celentino ridge model the measured schistosity, the numerical model simulations closely agree with the actual morphological features. These results have been obtained through the progressive lowering of rock mass properties derived from GSI values. If the present mechanical properties of the DSGSD rock mass are introduced into the model at the beginning of the simulation, successive rock mass failures are generated at each deglaciation step without resulting in a single deep-seated slope deformation. Glacial unloading alone seems to have been sufficient to generate the deep-seated deformation. The model results indicate that post-glacial debuttressing is the main trigger predisposing factor in the development of deformation in the slope, as concluded by other authors for the same area (e.g. Agliardi et al. 2001). In conclusion, the results demonstrate that continuum modelling can be used to examine the evolution of stresses, strains and plastic yielding at the ridge scale, as it provides results congruent with field observations. Discontinuum techniques, such as the distinct element method, will be also applied in future research both in two and three dimensions. Field surveys and desk analyses have increased our understanding of the kinematics of the Celentino DSGSD. The study of DSGSDs is still at an early stage and requires new theoretical developments to further our understandings of the operative processes. A better comprehension of these phenomena can be obtained by improving the modelling through

249

incorporating progressive failure concepts, investigating the influence of existing discontinuities systems and by introducing the concept of fatigue into the simulation.

References Agliardi, F., Crosta, G. B. & Zanchi, A. 2001. Structural constraints on deep-seated deformation kinematics. Engineering Geology, 59, 83– 102. Agliardi, F., Crosta, G. B., Zanchi, A. & Ravazzi, C. 2009a. Onset and timing of deep-seated gravitational slope deformations in the eastern Alps, Italy. Geomorphology, 103, 113–129. Agliardi, F., Zanchi, A. & Crosta, G. B. 2009b. Tectonic vs. gravitational morphostructures in the central Eastern Alps (Italy): Constraints on the recent evolution of the mountain range. Tectonophysics, 474, 250– 270; doi: 10.1016/j.tecto.2009.02.019. Ambrosi, C. & Crosta, G. B. 2006. Large sackung along major tectonic features in the Central Italian Alps. Engineering Geology, 83, 183–200. Andreatta, C. 1954. La Val di Peio e la Catena ViozCevedale. Studio geo-petrotettonico e minerario di una parte del massiccio dell’Ortles. Acta Geologica Alpina Contributi di Mineralogia Petrografia e Geologia, 5. Apuani, T., Masetti, M. & Rossi, M. 2007. Stress– strain–time numerical modelling of a deep-seated gravitational slope deformation: Preliminary results. Quaternary International, 171– 172, 80–89. Ballantyne, C. K. 2002. Paraglacial geomorphology. Quaternary Science Reviews, 21, 1935– 2017. Barton, N. 1976. The shear strength of rock and rock joints. International Journal of Rock Mechanics Mining Sciences and Geomechanical Abstract, 13, 1– 24. Bell, F. G. 1987. Ground Engineers Reference Book, Part 1 Properties of Intact Rock. Butterworth, London. Bovis, M. J. & Evans, S. G. 1996. Extensive deformation of rock slopes in southern Coast Mountains, southwest British Columbia, Canada. Engineering Geology, 44, 163– 182. Cai, M., Kaiser, P. K., Uno, H., Tasaka, Y. & Minami, M. 2004. Estimation of Rock Mass Deformation Modulus and Strength of Jointed Hard Rock Masses using the GSI system. International Journal of Rock Mechanics and Mining Sciences, 41, 3 –19. Castellarin, A., Cantelli, L. et al. 1992. Alpine compressional tectonics in the Southern Alps. Relationships with the N Apennines. Annales Tectonicae, 6, 62–94. Crosta, G. B. & Agliardi, F. 2003. Failure forecast for large rock slides by surface displacement measurements. Canadian Geotechnical Journal, 40, 176– 191. Cruden, D. M. & Hu, X. 1999. The shapes of some mountain peaks in the Canadian Rockies. Earth Surface Processes and Landforms, 24, 1229–1241. Dal Piaz, G. V., Castellarin, A. et al. 2007. Carta geologica e note Illustrative del Foglio geologico 042 – Male´ 1:50 000. Istituto Poligrafico e Zecca dello Stato, Rome.

250

M. GHIROTTI ET AL.

Deere, D. U. 1968. Chapter 1: geological considerations. In: Stagg, K. G. & Zienkiewicz, O. C. (eds) Rock Mechanics in Engineering Practice. Wiley, London, 1 –20. Desio, A. 1967. I ghiacciai del Gruppo Ortles-Cevedale. C.N.R., Comitato Glaciologico Italiano, Tamburini, Turin. Dikau, R., Brunsden, D., Schrott, L. & Ibsen, M.-L. (eds). 1996. Landslide Recognition. Wiley, Chichester. Dramis, F. & Sorriso Valvo, M. 1994. Deep seated slope deformations, related landslide and tectonics. Engineering Geology, 38, 231– 243. Esposito, C., Martino, S. & Scarascia Mugnozza, G. 2007. Mountain slope deformation along thrust fronts in jointed limestone: an equivalent continuum modelling approach. Geomorphology, 90, 55–72. Evans, S. G. & Couture, R. 2002. The 1965 Hope Slide, British Columbia; catastrophic failure of a sagging rock slope. Geological Society of America, Abstracts with Programs, 34, 16–26. Forlati, F., Gioda, G. & Scavia, C. 2001. Finite element analysis of a deepseated slope deformation. Rock Mechanics and Rock Engineering, 34, 135– 159. GRUPPO NAZIONALE GEOGRAFIA FISICA E GEOMORFOLOGIA CNR. 1986. Ricerche geomorfologiche nell’Alta Val di Peio (Gruppo del Cevedale). Geografia Fisica e Dinamica Quaternaria, 9, 137– 191. Hoek, E. & Brown, E. T. 1997. Practical estimates of rock mass strength. International Journal of Rock Mechanics and Mining Sciences, 34, 1165– 1186. Hoek, E., Carranza Torres, C. & Corkum, B. 2002. Hoek– Brown failure criterion – 2002 Edition. In: Hammah, R., Bawden, W., Curran, J. & Telesnicki, M. (eds) NARMS-TAC 2002: Mining and Tunnelling Innovation and Opportunity. University of Toronto Press, Toronto, 267–273. Hoek, E. & Diederichs, M. S. 2006. Empirical estimation of rock mass modulus. International Journal of Rock Mechanics and Mining Sciences, 43, 203– 215. Hu¨rlimann, M., Ledesma, A., Corominas, J. & Prat, P. C. 2006. The deep seated slope deformation at Encampadana, Andorra: representation of morphologic features by numerical modelling. Engineering Geology, 83, 343– 357. Hutchinson, J. N. 1988. General Report: Morphological and geotechnical parameters of landslides in relation to geology and hydrogeology. In: Bonnard, C. (ed.) Proceedings of the 5th International Symposium on Landslides, Lausanne, Volume 1. Balkema, Rotterdam, 3 –35. ISRM. 1978. Suggested methods for the quantitative description of discontinuities in rock masses. International Journal of Rock Mechanics and Mining Sciences & Geomechanics Abstract, 15, 319– 368. Kinakin, D. & Stead, D. 2005. Analysis of the distributions of stress in natural ridge forms: implications for the deformation mechanisms of rock slopes and the formation of sackung. Geomorphology, 65, 85–100. Mannucci, G., Ottenziali, L. & Martin, S. 2004. Instrumental problems connected with monitoring of deep gravitative deformations of slopes. International Geological Congress, Abstracts, 32, 678.

Marinos, P. G. & Hoek, E. 2000. GSI: A geological friendly tool for rock mass strength estimation. In: Proceedings of the International Conference on Geotechnical & Geological Engineering (GeoEng 2000), Technomic, Melbourne, 1422–1440. Marinos, V., Marinos, P. & Hoek, E. 2005. The Geological Strength Index: applications and limitations. Bulletin of Engineering Geology and the Environment, 64, 55–65. Martin, S. 2007. Carta geologica e note illustrative Tavola 042 IV – Peio, 1:25 000. Provincia Autonoma di Trento, Trento. Martin, S. 2008. Carta geologica e note illustrative del Foglio della Carta geologica d’Italia Rabbi n. 25, 1:50 000. Provincia Autonoma di Trento, Trento. Martin, S., Godard, G. & Ranalli, G. 1998. Evolution of the Deep Crust in the Central and Eastern Alps. Memorie di Scienze Geologiche, 50. Martin, S., Prosser, G. & Santini, L. 1991. Alpine deformation along the Periadriatic lineament in the Italian Eastern Alps. Annales Tectonicae, 5, 118–140. Martin, S., Vigano`, A., Godard, G., Laurenzi, M. & Fellin, M. G. 2007. Structural ‘old’ records in young faults: a case study from Eastern Italian Alps. Geophysical Research Abstracts, 9, 06782. Mu¨ller, W., Prosser, G., Mancktelow, N. S., Villa, I. M., Kelley, S. P., Viola, G. & Oberli, F. 2001. Geochronological constraints on the evolution of the Periadriatic Fault System (Alps). International Journal of Earth Sciences, 90, 623–653. Montresor, L. & Martin, S. 2007. Carta geologica e note illustrativa Tavola 025 III – Rabbi, 1:25 000. Provincia Autonoma di Trento, Trento. Nicolas, A. 1987. Principles of Rock Deformation. Reidel, Dordrecht. Pasquare´, G. (ed.) 2001. Tettonica recente e instabilita` di versante nelle Alpi Centrali. Fondazione Cariplo per la Ricerca Scientifica; CNR-Istituto per la Dinamica dei Processi Ambientali, Milano. Pazzaglia, F. J., Selverstone, J., Roy, M., Steffen, K., Newland-Pearce, S., Knipscher, W. & Pearce, J. 2007. Geomorphic expression of midcrustal extension in convergent orogens. Tectonics, 26, TC6010; doi: 10.1029/2006TC001961. Ravazzi, C., Peresani, M., Pini, R. & Vescovi, E. 2007. Il Tardoglaciale nelle Alpi italiane e in Pianura Padana. Evoluzione stratigrafica, storia della vegetazione e del popolamento antropico. Il Quaternario, Italian Journal of Quaternary Sciences, 20, 163–184. ROCSCIENCE INC. 2002. RocLab Version 1.0 – Rock Mass Strength Analysis using the Generalized Hoek-Brown failure criterion. www.rocscience.com, Toronto, Ontario. ROCSCIENCE INC. 2007. Phase2 Version. 6.0 – Finite Element Analysis for Excavations and Slopes. www. rocscience.com, Toronto, Ontario. Sitharam, T. G., Sridevi, J. & Shimizu, N. 2001. Practical equivalent continuum characterization of jointed rock masses. International Journal of Rock Mechanics and Mining Sciences, 38, 437–448. Varnes, D. J., Radbruch-Hall, D., Varnes, K. L., Smith, W. K. & Savage, W. Z. 1990. Measurement

THE CELENTINO DSGSD (PEIO VALLEY, ITALY) of ridge-spreading movements (sackungen) at Bald Eagle Mountain, Lake County, Colorado, 1975– 1989. United States Geological Survey Open-File Report, 90–543. Vigano`, A., Bressan, G., Ranalli, G. & Martin, S. 2008. Focal mechanism inversion in the Giudicarie – Lessini seismotectonic region (Southern Alps, Italy):

251

Insights on tectonic stress and strain. Tectonophysics, 460, 106–115; doi: 10.1016/j.tecto.2008.07.008. Viola, G., Mancktelow, N. S., Seward, D., Meier, A. & Martin, S. 2003. The Pejo fault system: An example of multiple tectonic activity in the Italian Eastern Alps. Geological Society of American Bulletin, 115, 515–532.

Megafans and outsize fans from catastrophic slope failures in Alpine glacial troughs: the Malser Haide and the Val Venosta cluster, Italy DAVID JARMAN1*, FEDERICO AGLIARDI2 & GIOVANNI B. CROSTA2 1

Mountain Landform Research, Scotland

2

Dipartimento di Scienze Geologiche e Geotecnologie, Universita` degli Studi di Milano-Bicocca, Piazza della Scienza 4, Milano 20126, Italy *Corresponding author (e-mail: [email protected]) Abstract: A cluster of exceptionally large sediment fans occurs in Val Venosta, a glacial trough in the east-central Alps, Italy. Its 59 tributary valleys generate 49 fans with volume:catchment area ratios varying across four orders of magnitude. Geomorphological and statistical analysis distinguish ‘allometric’ and ‘anomalous’ fans. Catastrophic massive slope failure origins are suggested for the anomalous cases. They comprise ‘outsize fans’ and ‘megafans’, the latter attaining 400 m cone height and 2700 m radius, and dominating the trough. Above most fans, evidence is found for source cavities of comparable volume. Reconstruction of the missing sides and heads of two tributary valleys reveals lost mountains 700 m deep. They are credible sources for the Malser Haide, a globally significant 11 km-long megafan with an estimated volume of 1650 Mm3, and the St Valentin outsize fans. Generally, anomalous fans occur where landslides are funnelled, comminuted and controlled through ‘debouchures’ high enough above the trough floor for conoidal deposition. Although sedimentological data are sparse, these fans may represent a new category of catastrophic slope failure outcome, mimicking conventional sediment fans of incremental origin. The Val Venosta cluster is the largest in the Alps, with concentrated glacial erosion in conducive geology among the possible factors explaining anomalous fan incidence.

Sediment fans in glacial troughs Fan-shaped deposits are common in formerly glaciated mountains, especially where steep tributary valleys join broad troughs; this being a classic locus where high volumes of mobilized sediment suffer a sudden loss of transport power (Harvey 2003). The term ‘alluvial fan’ now generally embraces both fluvial and debris-flow-dominated deposits (Blair & McPherson 1994; Sorriso-Valvo et al. 1998; Iverson 2003). However, Derbyshire & Owen (1990) considered the epithet ‘alluvial’ clearly inappropriate as the majority of fan deposits in alpine environments are not laid by flowing water, but are partly or wholly composed of debris flows. They propose ‘sediment fans’ as a processneutral term. Fan gradients vary widely, with fluvial fans tending to lower angles, but gradient is not a good discriminant given that deposits are often mixed and in confined topographical settings (SorrisoValvo et al. 1998; Crosta & Frattini 2004). Thus, Fischer (1965) found trough-wall fans in the Alps (his ‘murkegel’ type) in the range 7–21%. In the Karakoram, Derbyshire & Owen (1990) recorded considerable overlap between fluvial fans (,17%) and debris-flow fans (,27%) in gradient. They also suggested that extensive (,18%) fans ‘imply very wet, fluidized debris, and rapid flow and

sedimentation rates’ (p. 41). Fluidization can be achieved during transit by incorporation of ice, snow and wet sediments (Hewitt 1999). Legros (2006, p. 233) observed that ‘the difference between landslides and debris flows is wholly gradational and related to the water content’. It is generally assumed that cone-shaped fans in glacial troughs have grown incrementally, as evidenced by layered sections. Eisbacher & Clague (1984) described many incremental fan-building events in the Alps. Most of the contributions to Rachocki & Church (1990) and Harvey et al. (2005) follow this paradigm, although Lecce (1990) recognized a uniformitarian v. catastrophist issue in fan formation. The cases he cites for highmagnitude –low-frequency activity are, however, relatively modest, and he views their geomorphic role as still ‘controversial’. By contrast, Derbyshire & Owen (1990, pp. 28 and 50) observed that debrisflow landforms may ‘mimic closely the form of alluvial fans within high mountain valleys’; they concluded that a subtype of sediment fans can be identified that ‘derive essentially from a few large, perhaps catastrophic events’; for example, the 150 m-thick Batkor cone. Hewitt (2001) confirmed a rockwall detachment source for Batkor as a single large rock avalanche deposit, among other similar cases. Watanabe et al. (1998) found that the largest debris cones in Langtang Himal came from

From: Jaboyedoff, M. (ed.) Slope Tectonics. Geological Society, London, Special Publications, 351, 253– 277. DOI: 10.1144/SP351.14 0305-8719/11/$15.00 # The Geological Society of London 2011.

254

D. JARMAN ET AL.

slope failures, while Shang et al. (2003) described the deposit from a fresh 3000 Mm3 rock avalanche in Tibet as a sediment fan. In Tajikistan, Evans et al. (2009) documented a sediment fan emplaced by the 1949 Khait 75 Mm3 rockslide transformed in a rapid flow. Friele et al. (1999) identified a paraglacial collapse from the Mt Garibaldi volcano as the prime origin of the Cheekeye megafan, British Columbia, while conoidal deposits related to massive slope failure are documented at Huascaran, Peru (Plafker & Ericksen 1978). However, an overview of the possible association between large fans and catastrophic slope failures in glaciated mountain ranges remains lacking. An instructive precedent is the reinterpretation of anomalously large ‘terminal moraines’ in the Karakoram as huge rock avalanche deposits (Hewitt 1999). The classic model for sediment fans identifies an allometric (power-law) relationship between fan area and catchment area (Harvey 2003), albeit primarily for arid contexts. Crosta & Frattini (2004) reviewed its relevance for humid, deglaciated mountain environments. Their study of 209 ‘alluvial fans’ confirms the negative allometric relationship between fan and catchment areas identified in such environments (Allen & Hovius 1998). Larger catchments are inferred to generate relatively small fans because they are less likely to have completed paraglacial resedimentation, while small catchments generate disproportionately large fans because of their limited sediment storage capacities. But over and above this built-in tendency, Crosta & Frattini (2004) identified 12 ‘anomalous’ (non-allometric) fans, which they suggest may be associated with infrequent high-magnitude events (primarily large landslides). Exceptionally large fans occur widely across the European Alps (Fig. 1), but none have been interpreted as catastrophic except for Biasca where a 20 Mm3 outsize fan occurred as a single historical event in 1513, damming a temporary lake (Eisbacher & Clague 1984). Even though the Ponte-Chiuro megafan in Valtellina is located below a bold matching cavity, it was attributed by Crosta & Frattini (2004) to intense debris-flow activity rather than a single massive failure. The densest cluster of fans with areas greater than 3 km2 occurs in Val Venosta, including the worldranking Malser Haide at over 16 km2. Its gross mismatch with a tiny catchment area has been noticed (Fischer 1965; Eisbacher & Clague 1984; Crosta & Zanchi 2000; Agliardi et al. 2009a) but not analysed, providing initial inspiration for this study. The aims here are: † to catalogue the Val Venosta fans, identifying and characterizing anomalous cases;

† to reconstruct possible source configurations for the Malser Haide megafan and adjacent St Valentin outsize fans, where gross terrain anomalies suggest they could be the product of catastrophic landslides; † to discuss the conditions leading to the observed clustering of massive slope failure in the study area in relation to topography, lithology, structure and glacial –paraglacial history; † to consider, as an example of equifinality, how catastrophic slope failures can lead to the emplacement of conoidal deposits mimicking incremental fans, and thus to explore the possibility of anomalous fans as a new category of large slope failure outcome. Existing sedimentological and dating information is sparse; geotechnical investigations are essential to verify the interpretations advanced and to explore geohazard implications.

Geomorphological and geological setting Val Venosta (Vinschgau) is a major glacial trough ¨ tztal and Ortler ranges of the eastbetween the O central Italian Alps (Fig. 1). Its lower part trends west –east for 42 km, descending from 900 to 500 m asl with its east end above Meran closed by a large fan complex into the Adige (in German the Etsch) gorge. At its west end, the Upper Val Venosta ascends north into the lowest gap in the main Alpine divide west of Brenner, the Reschen Pass (1500 m). The infilled trough floors are 1– 2 km wide with a typical local relief of 1500 – 2000 m. The terrain character falls into three elevation ranges, reflecting Quaternary glacial impact on an orogenic landscape that had become generally ‘mature’ and well adjusted to late Tertiary base levels. Below 2400 m asl steep slopes fall to the Venosta glacial trough, without high rockwalls; tributary valleys are more V-shaped than U-shaped because of gravitational and fluvial overprinting. Between 2400 and 2900 m relief is subdued with open ridges and broad upland valleys, with remnants of preglacial relief suggesting rather limited glacial erosion. Only above 2900 m do narrow, periglacially shattered, alpine crests occur, with limited cirque development. This is probably a glacial trimline: at the Last Glacial Maximum, the upper ice limit was approximately 2500 m asl (Van Husen 1987), but may well have been higher in the midQuaternary stadials. The present main Alpine divide at the Reschen Pass is only 5 km from the Inn valley. The preglacial divide is inferred to have crossed the Upper Val Venosta at St Valentin (Fig. 1). South-carried erratics (J. Reitner pers. comm.) here suggest the cutting of a major glacial

MEGAFANS FROM CATASTROPHIC SLOPE FAILURES

255

Fig. 1. The Central and Eastern Alps, with the main Alpine divide and passes, the Val Venosta study area, and the most prominent megafans, outsize fans and rockslides identified in the literature and from satellite imagery. The co-ordinate grid is UTM 32.

breach, with post-glacial diversion of former Inn tributaries to the Adige (Etsch). The geology is predominantly Austro-Alpine metamorphic units (Fig. 2) (Agliardi et al. 2009b), comprising metapelites and metapsammites, with subordinate orthogneiss, metabasites and calcschists. Slices of sedimentary Upper Palaeozoic – Mesozoic cover occur around the Upper Val Venosta. Val Venosta exploits the Vinschgau Shear ¨ tztal Unit from the Zone, which separates the O underlying similar Campo Nappe. The entire stack is cut by north-, east-, NE- and SW-trending fractures, strongly constraining the drainage pattern and the strength of rock masses. Rock slope failure is so extensive here as to be nearly endemic. Susceptibility depends primarily on lithology, fracturing and anisotropy (Hoek & Brown 1997). Where stronger rock masses occur (e.g. orthogneisses in the Plawenn valley, carbonates east of St Valentin), rockfalls and rockslides dominate. Where less strong, anisotropic rock masses occur (e.g. metapelites, paragneiss), deepseated gravitational slope deformations (DSGSD) (Agliardi et al. 2001; Ambrosi & Crosta 2006) affect whole valley sides and interfluves (Agliardi et al. 2009b). In the Upper Val Venosta area, eastfacing slopes tend to slope deformation, while west and NW-facing slopes are more prone to rocksliding. This tendency is partly influenced by the

low-angle Schlinig Fault underlying the west side (Agliardi et al. 2009a).

Anomalous fans and source cavities in Val Venosta There are 59 tributary valleys in the 65 km length of Val Venosta plus Val Mustair; 49 have discernible fans at their mouths (Fig. 3, anomalous sites numbered and identified below as #). This indicates that fan production is here the norm across all catchment sizes and types. The equally diverse ‘no-fan’ group includes some large upper Venosta tributaries where the Malser Haide has either buried their fans or displaced their effective valley mouths down to its own toe, and a sequence at the east end of Val Venosta possibly attributable to damming by the Partschins fans (#12 and #13) and subsequent alluvial burial.

Preliminary fan assessment by proxy volume The fans vary enormously in prominence, their scale often bearing little relationship to their catchment area. Adjacent pairs of similar side valleys commonly juxtapose tiny fans with outsize cones. To sort this extreme variation, a simple initial classification was carried out based on estimated volume

256

D. JARMAN ET AL.

Fig. 2. Geological sketch map of the study area in the context of the eastern Central Alps. EW, Engadine Window; EL, Engadine Line; GL, Giudicarie Line; IL, Insubric Line; JA, Jaggl Permo-Mesozoic; JL, Jaufen Line; MA, Matsch Unit; SB, Schneeberg Unit; PA, Passiria Line; PJ, Pejo Line; PL, Pustertal Line; SF, Schlinig Fault; TW, Tauern Window; VSZ, Vinschgau Shear Zone; ZL, Zebru` Line.

following the method of Fischer (1965). Clearly, actual fan volumes cannot be determined without geotechnical investigation, but area and height can be combined via conoidal geometry to give ‘proxy volumes’. Although these proxy volumes imagine fan deposition on flat floors abutting vertical trough walls, they offer a reasonable basis for preliminary comparison where trough form is consistent. Proxy volume should normally exceed true volume, but at St Valentin (#2 and #3) it is less than half the estimate obtained from boreholeconstrained cross-sectioning, illustrating an ‘iceberg effect’ where fans coalesce or have buried toes. Proxy volume:catchment area ratios vary across four orders of magnitude, and allow five types of fan to be distinguished in ascending order of visual prominence within the trough (Table 1). These five types fall readily into two broad categories. † Allometric fans are the majority and represent ‘normal’ fan production, not standing out as discordant in the landscape. They are broadly proportionate to their catchment areas, and are inferred to have built incrementally from varying proportions of fluvial and debris-flow input;

hence their wide range of gradients. They subdivide naturally into groups with smaller and larger volume:catchment area ratios, which may well correspond with the Type I and Type II fans of Blair & McPherson (1994), the latter having a greater debris-flow input attested by steeper angles. † Anomalous fans stand out conspicuously, whether as ‘outsize fans’ emanating from microcatchments (,3.7 km2) or as ‘megafans’ dominating the troughs with huge volumes exceeding 250 Mm3 (Table 2). In the megafan group, the Malser Haide is exceptional in both scale and character. The excess volume in these anomalous fans, over and above normal production by conventional incremental processes, can be inferred to derive from one or more highmagnitude events (Crosta & Frattini 2004). Finally, the four very extensive but very lowangle alluvial fans (‘schwemmkegel’ of Fischer 1965) (Fig. 3) are from the five largest catchments. Although they are allometric in respect of fan area, they have extremely low proxy volume:catchment ratios, suggesting that they fall into a separate process domain.

MEGAFANS FROM CATASTROPHIC SLOPE FAILURES

257

Fig. 3. Sediment fans and related catchment areas in Val Venosta and the Mustair valley. Fans are classified according to the procedure described. Outsize fans and megafans are numbered (Malser Haide marked MH). The co-ordinate grid is UTM 32.

Regression analysis by fan area Building on this preliminary assessment of fan typology, a statistical analysis was carried out. Detailed measurement of key parameters from highresolution LiDAR (Light Detection And Ranging) topography (2.5 m cell size) enabled regression analysis to be performed using fan area as the standard size descriptor. While most studies using this parameter deal with pediment fans in arid environments (e.g. Allen & Hovius 1998), it was also used in a study of the neighbouring Valtellina troughs (Crosta & Frattini 2004). Fan area is a more objective measure than proxy volume, although both encounter problems with variable underlying trough topography, and where fans are constrained by neighbours or are partly concealed by lakes or alluviated flood plains; fan area cannot allow for varying gradient. The distorting effect of confinement by neighbouring troughs was identified by Sorriso-Valvo et al. (1998) and, to identify a truer relationship, Crosta & Frattini (2004) excluded such constrained cases from their analysis. Larger fans

are, of course, more likely to be ‘confined’, both distally and laterally, with the smaller ones best representing ‘normal’ or base-load fan production. Iterative least squares, non-linear regression analysis showed that for the whole of the Venosta fan population (N ¼ 49) the classic power-law (i.e. ‘allometric’) relationship is weak, with correlation coefficient R ¼ 0.36. By excluding 15 ‘distally confined’ fans and 12 ‘laterally confined’ fans, a subset of the population comprising the unconfined fans (N ¼ 22) gives a much better regression of the fan area –catchment area relationship, with R ¼ 0.83. Applying 99.7% confidence bands rigorously identifies ‘outlier’ fans with respect to allometric assumptions (i.e. fans whose morphometric character has a negligible chance of being explained by the allometric model). The results then compare well with other studies (Table 3), although the dataset includes large low-angle alluvial fans that were excluded by Crosta & Frattini (2004). The log–log plot of fan area v. catchment area (Fig. 4) clearly distinguishes most of the apparently ‘anomalous’ fans as lying well outside the

258

Table 1. Val Venosta fan types and size:catchment ratios Fan type

33 4

(m)

Fan proxy volume (V ) (Mm3)

20 –70 (45)

1550 – 2500 (1840)

4 – 6 (4)

18 – 92 (56)

0.02– 0.05 (0.03) 0.2– 0.5 (0.4)

25 –150 (70) 75 –250 (150)

280 – 900 (550) 400 – 1650 (830)

8 – 27 (15) 8 – 27 (18)

0.4– 12 (5.3) 4 – 55 (27)

0.01– 0.09 (0.05) 0.1– 1.8 (0.8) 0.12– 0.45 (0.27) 4 – 16 (9)

13 – 23 (17) 11 – 21 (14) 11.5

35 – 100 (69) 268 – 916 (513) 1650

0.24– 1.14 (0.63) 0.42– 0.64 (0.56) 1.83

13 – 57 (35) 33 – 67 (47) 200

–

–

–

–

Fan area (Af )

Fan height (H )

Fan radius (R)

(km2)

(m)

95–167 (133)

2.2 –8.6 (4.4) 0.1 –0.7 (0.33) 0.2 –1.3 (0.64)

19 2.6–49.6 (10.9) 10 0.8–10.4 (3.2)

Anomalous fans 16 Outsize 10 1.4–3.7 (2.2) Megafans 5 5.3–18.7 (11.0) Malser Haide (MH) 1 9 Total fans

49

No fan

10 2.0–236 (44)

Total tributary valleys

59

0.8 –1.6 (1.3) 3.4 –10.6 (6.2) 16.5 –

Af/Ac ratio

Fan gradient (Gf ) (%)

Catchment area (Ac) (km2)

130 –350 (195) 800 – 1500 (1050) 230 –400 (300) 1600 – 2700 (2100) – – –

–

V/Ac ratio (Mm3 km2)

Notes: 1. Data given as ranges with averages (in parentheses). Figures in bold highlight the marked distinctions between categories. 2. ‘Proxy volume’ data are obtained by conoidal geometry for comparative purposes only, see the text. 3. One unclassifiable fan is included in the outsize fans tally but is excluded from the data tabulation (see Table 2). 4. Malser Haide is shown separately from the more typical megafans. Its volume is a best estimate from cross-sectioning, not comparable with proxy volumes (data in italics).

D. JARMAN ET AL.

Allometric fans Large low-angle alluvial Smaller Larger

N

Table 2. Val Venosta outsize fans and megafans # ref in Fig. 3

Fan

Outsize fans Reschen St Valentin North St Valentin South Fischerhauser Taufers NE St Martin/Glurns Mareinweisen Tschars Algund/Meran Avignatal

7 9 11 12 13 MH

Megafans Allitz/Laas Tarsch/Latsch Tabland Partschins West Partschins East Malser Haide

1 1 1

2 2 3 6 4 5 5

Catchment area (Ac) (km2)

Fan area (Af ) (km2)

Fan gradient (Gf ) (%)

1.7 1.4 2.3 1.9 1.9 2.1 3.7 1.8 2.7 21.0*

1.1 1.2 2.0 1.4 1.4 1.5 0.9 1.1 0.8 2.2

22 15 16 14 23 17 13 23 18 14

18.6 9.8 11.4 10.0* 5.3 8.3

10.6 6.9 6.2 4.2 3.4 16.5

13 16 11 12 17 16 . 11 . 7

Fan height (H ) (m)

Debouchure width (W ) (m)

350 200 150 140 200 150 160 280 130 150

125 375 450 460 175 150 250 230 – –

300 300 230 240 400 (200)

500 600 500 450 – (650)

Fan proxy volume (V ) (Mm3) 70 80 100 47 74 78 54 97 35 90 916 567 385 332 268* 1650*

Volume Fischer (Mm3) 67 72

Af/Ac ratio

V/Ac ratio (Mm3 km2)

– – 80 – – – –

0.65 0.86 0.87 0.74 0.74 0.71 0.24 0.61 0.30 0.10*

1350 630 420 280 240 1550

0.57 0.61 0.54 0.42* 0.64 1.99

30 57 43 25 39 37 15 54 13 4.3* 49 58 34 33* 51* 200*

Notes: ‘Volume Fischer’ is from Fischer (1966), obtained by conoidal geometry (basis of calculation for #MH unclear). *Data not comparable, see the numbered notes. 1. Toes in reservoir: proxy volumes taken to former lake shore; true volumes estimated from valley cross-sections with allowance for prior infill (see text) are: (#1) 65 Mm3; (#2) 200 Mm3; and (#3) 230 Mm3. 2. These cases have significantly lower area:volume ratios to catchment, and are on the regression confidence limit (Fig. 4). They are transitional with large allometric fans. 3. This unclassifiable fan has a volume much greater than any allometric fan and a conoidal shape, but debouches into Mu¨nstertal from a large catchment giving it very low ratios. It is likely that the deposit emanates from a small part of the outer basin, but a suspected catastrophic source has yet to be verified. Compare #12. 4. The fan emanates from the large Zielbach catchment (32 km2) but is evidently sourced from a side bay just inside its mouth; the arbitrary catchment area given here is for the outer basin, offering best comparablility, but the specific catchment is only 4 km2.. 5. These fans are not conoidal and proxy volumes cannot be obtained. #13 is estimated assuming overlap onto #12 and infill of a V-valley to 100 m average thickness; for #MH see text. 6. Subsequent field work and valley section calculations indicate that this site has a larger true fan volume and lost mountain source than Malser Haide. This fan is the Gadriamure of Fischer (1966).

MEGAFANS FROM CATASTROPHIC SLOPE FAILURES

1 2 3 4 5 6 8 10 14 –

Note

259

260

D. JARMAN ET AL.

Table 3. Values of coefficient c and exponent k in the power-law relationship (Af ¼ cAkb) between fan area and catchment area in humid mountain ranges Location Present study Valtellina area (Crosta & Frattini 2004) Banff, Alberta*

All fans Unconfined fans All fans Unconfined fans Fluvial fans Debris-flow fans

Roan Mt, North Carolina* Dellwood, North Carolina*

Sample

c

k

R

49 22 209 64

0.93 0.25 0.29 0.15 0.48 0.17 0.38 0.23

0.33 0.52 0.33 0.65 0.32 0.48 0.76 0.53

0.36 0.83 0.35 0.81

*North American data from Crosta & Frattini (2004, table 1). R, coefficient of correlation.

regression confidence limit, with only a few borderline cases. They are grossly disproportionate to catchment area, with the two distinct clusters of ‘outsize fans’ and ‘megafans’ almost an order of magnitude apart, suggesting responses to different scales of terrain in the study area. Both point to catastrophic events for which sources can be sought.

Anomalous sources for anomalous fans Recognizing anomalous cavities as possible sources for problematic deposits (e.g. Hewitt 2001), let alone reconstructing them to determine comparability of volume (e.g. Jarman 2002), is not yet common practice in geomorphology. Thus, in the Central

Fig. 4. Log–log plot of fan area v. catchment area for the 49 sediment fans mapped in Val Venosta. Fans are classified according to confinement. The solid line represents the allometric best fit for the unconfined (UC) data subset (N ¼ 22). The dashed lines represent 99.7% confidence bands. Different groups of fans are identified (see Table 1); anomalous fans are numbered as per Table 2.

MEGAFANS FROM CATASTROPHIC SLOPE FAILURES

Alps, Fischer (1966) simply regarded the whole catchment valley space (‘hohlraum’) as the anomalous fan source: for 12 cases he found fan:hohlraum ratios averaging 1:1.8 (excluding Tabland (#11) as exceptional at 1:6). However, in Val Venosta most of the 15 megafans and outsize fans have specific source cavities or broader anomalous landform configurations identifiable from topographical maps, LiDAR topography and satellite imagery. These candidate sources often appear to be sharply bounded landslide scars, and usually occupy more than a quarter of the catchment area (Fig. 5). These anomalous cavities are generally not just simple voids in the open trough wall, but are linked to the trough floor by somewhat narrower ‘debouchures’ of measurable width (Table 2); their significance is discussed below. Evidence for the most striking cases is now examined: † #1–Reschen (Fig. 5a): a remarkably steep cone debouches from a broad chute below an embayment only 1.7 km2 in catchment area. The cirque bowl beneath the 2981 m summit appears well adjusted. Below 2500 m, a subarcuate cavity bites roughly 150 m into the trough wall, faceted by scars controlled by NW –SEand NE–SW-trending fractures, and yielding a volume broadly equivalent to the fan. Intense post-collapse debris flows may have steepened the cone. The boldest wedge facet cuts into a broad paragneiss ridge on the SE flank where incipient retrogressive scarps indicate deepseated deformation. † #2–#4, #MH. see detailed reconstructions in the following sections on ‘The Malser Haide megafan’ and ‘The St Valentin outsize fans’. † #5–Taufers NE: at the foot of the Mustair valley, a sequence of contiguous fans of different types suggests intermeshing of catastrophic and incremental processes. The boldest outsize cone debouches from an open chute beneath an evident cavity that sharpens the deforming orthogneiss crest. Two adjacent bays have even wider debouchures above bajada-like debris banks; the smaller retains bulky slip-masses, the larger appears evacuated. † #6–St Martin, Glurns (Fig. 5b): a steep cone emanates from an open chute below a sharply defined wedge cavity biting into the end of a broad split-crested ridge; residual slipped material lies in the apex and an arrested rockslide occurs on the left side just upslope from the debouchure. The scar in mylonitic orthogneiss and mica-schist is controlled by north –south and NE– SW fractures. Extrapolating a precavity hill form yields a volume of 96 Mm3, equal to the cone if it extends below the toe of

†

†

†

†

261

the Malser Haide. This is a particularly good candidate for a single catastrophic event, given that the adjacent bay of similar catchment area produces only a tiny fan (Lichtenburg). In this extreme case, virtually the entire catchment is the inferred source cavity. #7 –Allitz/Laas (Fig. 5c): one of the largest symmetrical fans in the Alps deflects the Adige (Etsch) river to the opposite slope foot and dominates Val Venosta as an extraordinary halfbarrier. The upper catchment has an open valley floor and north side well adjusted to the c. 2500 m asl trough rim topography. It hangs above a more rectilinear lower V-basin, cut into mylonitic paragneiss and mica-schist, and sharply delimited by NE –SW structural controls. This inferred slope failure cavity has rendered the upper basin highly asymmetric, and beheaded the adjacent valley (cf. the Vivana valley, Fig. 5d discussed below. Reconstructing these anomalies identifies a lost slice between the upper and lower basins of some 4 km2 and averaging 200 m in depth, which yields a volume comparable to the c. 900 Mm3 proxy volume calculated for the fan (but see Table 2, note 6). #9 –Tarsch/Latsch: although resembling #7 in its remarkable cone height, gradient and ratios to catchment, a single source cavity is less evident here. Several mid-slope gully bays carved in orthogneiss and metapelites would yield a volume comparable to the fan if reinstated by an average depth of only 75 m. This could point to multiple event origins for this cone. The debouchure is unusually wide (600 m). The adjacent valley to the west, which is of similar catchment area and topographical character, has no significant fan. The broad summit ridge from here eastwards is lower relief (c. 2500 m) and much deformed. #11–Tabland: this is similar to #9 except that a single large cavity could be reconstructed more readily across the lower half of the valley side. Tree logs near the toe have only been buried to a depth of 7 m since 2600 BP (Staffler & Nicolussi 2004). None of the six similar conventionally gullied catchments to the east display fans, although they may have been concealed by alluvium. #12 and #13–Partschins West/East: these fans overlap, with the west one (Zielbach) being conoidal and rather dissected, and the east one (To¨ll) confined by and funnelled down the Adige (Etsch) gorge. The trough is cut into metagranitoids and paragneiss, with a thick cover of slope and glacial debris. Widespread east –west- and NE–SW-trending gravitational deformation morphostructures behead a series of glaciated upland basins at the 2500–3000 m rim.

262

D. JARMAN ET AL.

Fig. 5. Anomalous cavity– fan systems in Val Venosta. Thick and thin solid white lines represent fan and cavity boundaries, respectively. (a) Reschen outsize fan (#1 in Fig. 3). (b) St Martin outsize fan (#6) and the adjacent Lichtenburg Valley with minimal fan. (c) Allitz/Laas megafan (#7). (d) St Valentin outsize fans (#2 and #3). Note the widespread slope deformation (DSGSD) indications. Basemaps: LiDAR topography at 2.5 m resolution.

Immediately below this rim, sharp breaks may define the sources of the two failed masses that have generated the cones. † #14– Algund: the catchment is within the same paragneiss slope affected by extensive deepseated deformation, suggesting a source of fractured material. It emanates from a dog-leg ravine that may be a debouchure closed in by subsequent rock slope failure. Despite intense gullying of the high scarp, there has only been 5 m of debris overlay to the main gravelly deposit since 5200 BP (Spindler 1994). Key diagnostic landforms for anomalous source cavities thus include: interruption of the continuity of topographical features; beheaded valleys; and bold planar/wedge/subarcuate forms with unusually clean-cut boundaries. These are often associated with regional structural orientations, split

ridges, and other evidence of deep-seated deformation and slope failure. Distinct or probable source configurations are discernible in 11 cases, with Tarsch, Tabland and Partschins requiring closer investigation. All except Partschins have distinct debouchures ranging from broad gullies and wider recesses to hanging bays and short tributaries. The secondary Venosta valleys generally lack significant fans, except at Melag in Langtaufers (similar to #6 St Martin), and one in Schlinigertal beneath a distinct source bay. At these locations (Fig. 3) the secondary valleys are broad troughs; elsewhere their narrowness, V-shape and gradient are inimical to fan development. There is little evidence for fan ages. They presumably post-date final deglaciation, with the destabilized mountain sources either collapsing soon after withdrawal of glacier buttressing or delayed until progressive failure and/or trigger events

MEGAFANS FROM CATASTROPHIC SLOPE FAILURES

263

Fig. 6. The Malser Haide megafan and source area (Plawenntal), within the upper Val Venosta. Contour lines of the Malser Haide are at 50 m intervals. Traces of the long section of Figure 8 and cross-sections of Figure 9 are shown. Note the splitting crests of main ridges indicating slope deformation. Basemap: LiDAR topography at 2.5 m resolution.

264

D. JARMAN ET AL.

exceeded critical thresholds (Prager et al. 2008). At Allitz/Laas (#7) the fan toe overlies Adige (Etsch) valley-floor sediments, with logs dated to around 7300 BP (Fischer 1990). Most fans have undergone remarkably little fluvial dissection of their surfaces or erosion of their toes, possibly suggesting recency or continuing aggradation, but alternatively it could be that they are so large and massive that Holocene processes have had little impact on them.

The Malser Haide megafan The Malser Haide has long been celebrated as the largest ‘alluvial fan’ in the Alps (Penck & Bru¨ckner 1909; Taylor 1940; Eisbacher & Clague 1984). With an area exceeding 16 km2 (Table 3), it is one of only three fans with Allitz/Laas (#7) and Illgraben (Fig. 1) noted by Fischer (1966) that exceeds 10 km2. It is remarkable for its exceptional overall height (900 m) and length (11 km), extending far beyond the main cone both headwards and down-valley. The Malser Haide occupies much of upper Val Venosta (Fig. 6). It emanates from the small Plawenn side valley, which has a catchment area of 8.3 km2 and unexceptional available relief (summit: Mittereck 2908 m; Fig. 7a). At its oblique point of entry into the upper Venosta trough

(1600 m) it fans out in a conventional smooth conoid of ‘distally confined’ form (Sorriso-Valvo et al. 1998) to dam the small Haidersee (1450 m). It then continues on down the 1–2 km-wide main trough, with a progressively more planar surface to the confluence with Val Mustair at Glurns (910 m). The head of the cone retrogresses more steeply 3 km into the Plawenn source valley. The deposit dramatically overfills the trough (Fig. 7b), displacing the Adige (Etsch) and Puni rivers against their respective valley sides. Its notably smooth overall surface (Fig. 6) may partly reflect extensive boulder clearance and cultivation. It being named as a heath and its irrigation attest to its porosity. There are numerous shallow dry channels, with floods and superficial debris flows recorded in recent centuries (Eisbacher & Clague 1984). The exceptional length of the Malser Haide is attributable to its topographical context. Most megafans in the Alps debouch into broad, lowgradient glacial troughs (eg. Rhoˆne, Valtellina). The upper Venosta trough is a glacial breach and, therefore, descends relatively steeply (450 m in 8 km, or 5.6%) into the main low-gradient trough (400 m in 42 km, or 1%). The Plawenn valley enters the upper Venosta trough at a 17% gradient at an oblique angle that is conducive to onward travel: if it debouched into a typical trough a more

Fig. 7. The Malser Haide megafan. (a) Cone debouching from Plawenntal into the upper Val Venosta; the inferred lost mountain ‘Plawennspitz’ location is above the head. View NE from Mt Watles ski lift base station.

MEGAFANS FROM CATASTROPHIC SLOPE FAILURES

conventional fan form would have resulted. Nevertheless, the Malser Haide has a 250 m cross-trough cone height and a gradient profile (16% at the head, 10– 12% on the cone, declining to 7% at the toe: Fig. 8), which are within the typical range of Alpine megafans (Fischer 1965). There is no information on the age of the Malser Haide, but it clearly post-dates final local deglaciation.

Quantifying the Malser Haide There is no available borehole or geophysical evidence on the thickness of the Malser Haide fan. Although mathematical procedures exist to obtain glacial trough bedrock profiles (Schrott et al. 2003; Jaboyedoff & Derron 2005), the possible form of the atypical upper Venosta breach trough is more naturally obtained by graphically extrapolating cross-sections at c. 1 km intervals (Figs 6 & 9). The profiles assume a gently parabolic form that can be projected and reasonably constrained from the extant valley side slopes, which vary between 35 and 70%. A relatively continuous long profile of the bedrock valley floor is obtained (Fig. 8), although its varying width may suggest a more stepped long profile. The marked widening at the Planeil valley mouth may be illusory, if the

265

rock ‘island’ near Alsack (Fig. 6) is bedrock and indicates the trough margin proper. In estimating a volume for the Malser Haide fan, the key assumption is the depth of prior infill by glacial, glacifluvial, debris-flow and alluvial deposits. Major Alpine glacial troughs can have kilometric infill depths, but this is implausible here. A conservative prior infill depth of 100– 125 m has, therefore, been adopted for the core 7 km (Plawenn–Mals). The long section (Fig. 8) shows total sediment thickness as fairly constant, but has the prior infill thinning back to zero in the Plawenn valley, and likewise the Malser Haide debris layer thinning out to zero at Glurns. Applying the cross-sectional areas net of prior infill to the relevant approximately 1 km lengths gives a best estimate megafan volume of 1650 Mm3. Simple robustness checks on this volume address three variables: † Bedrock floor shape –this is well constrained because the width of the Malser Haide surface is fixed, and the considerable taper in the crosssection even above the prior infill reduces the effect of any deviation. Deepening the floor by 100 m increases the fan volume by 80 Mm3 over a typical 1 km length or 600 Mm3 over

Fig. 7. (Continued) (b) Sediment infilling upper Val Venosta and tapering into the main Val Venosta trough head. Asymmetrical cross-profile is real– compare Figure 9, sections 6 and 7. View north from St Martin above Glurns.

266

D. JARMAN ET AL.

Fig. 8. Longitudinal section of the Malser Haide megafan (profile trace in Fig. 6), showing the inferred depths of the catastrophic failure deposit and prior infill. Asymmetry of Plawenntal is demonstrated by profiles of its NW and SE crests. The cross-section locations (Fig. 9) are shown.

the core 7 km. The precise shape of the Plawenn valley itself is less material because it contributes only about 100 Mm3 of the volume. † Depth of prior infill–as this is in the tapering floor and its width is only about half that of the Malser Haide surface, varying it by +50 m alters volume over the core length by 200 Mm3 per 50 m depth, maximally +500 Mm3. † Toe concealment by alluvial deposits–debris layer extension by 2 km at 100 m thickness increases the volume by 200 Mm3. Until geophysical depth profiles are available, a broad range for the estimated fan volume is appropriate. These robustness checks suggest a volume of 1150–2450 Mm3, assuming that a deeper floor is unlikely to have shallower prior infill.

Problems with reinterpreting the Malser Haide origin as catastrophic Eisbacher & Clague (1984, Appendix A86) observed that the Malser Haide ‘drops off from an astonishingly small catchment basin, forcing the Etsch river against the west side of the valley for almost 10 km’. Viewed from any vantage point (Fig. 7a), it seems improbable that the small Plawenn valley could yield sufficient debris to build such a megafan by episodic debris flows and floods. However, unlike the other Venosta cases reviewed earlier, there is no obvious cavity of sufficient size within the Plawenn valley. Yet, the area around the valley head displays many largescale terrain anomalies, especially the obviously beheaded Vivana valley, and extensive evidence of slope failure (Fig. 10).

The possibility of a catastrophic mass movement origin for the Malser Haide has been noted previously (Crosta & Zanchi 2000). Fischer (1966) recognized its exceptional scale but, as with all Venosta fans, attributed it to numerous fluvial events, including glacial melting, ice-dammed lake bursts and ‘catastrophic’ rainstorms, with ‘piles of rubble driven by water like lavaflows’. He treated the entire Plawenn valley space or ‘hohlraum’ as the source cavity, obtaining his closest fan volume: hohlraum ratio (1:1.1). Despite this close coincidence, it is an implausible origin. By reference to valley spacing in the surrounding area, a short side valley would naturally have pre-existed here; the two dendritic gully complexes on the SE side of the valley are inconsistent with a rockslide scar, by comparison with the smoother valley-head gullies; the terrain form achieved by extrapolating perimeter ridge contours across the present valley is incongruous; and the anomalous terrain around the valley head cannot be reconciled with it. Nonetheless, a catastrophic interpretation must address three principal difficulties. † Can a plausible source for such a large deposit be identified, given that it must be substantially above the rims of the present Plawenn valley if the hohlraum-source explanation is rejected? At the Ko¨fels catastrophic rockslide (Fig. 1), the failure plane daylighted well behind and sheared off the crest of the Fundus ridge, which flanks ¨ tztal (Heuberger 1994), the west side of O leaving no measurable cavity. The lost ridge is inferred to have been some 300–400 m higher (Sørensen & Bauer 2003; Prager et al. 2009b). A similar ‘lost mountain ridge’ is proposed here.

MEGAFANS FROM CATASTROPHIC SLOPE FAILURES

267

Fig. 9. Schematic cross-sections of the Malser Haide megafan (profile traces in Fig. 6), showing their parabolic extrapolation from existing valley side slopes. Section 8 appears flat floored because it extends into the mouth of Planeiltal; it overstates the deposit if the main trough is narrower at Alsack island (‘A’ in Fig. 6). Sections 1– 3 include the ‘Plawennspitz’ lost mountain and the Plawenntal lost NW ridge (Fig. 13).

268

D. JARMAN ET AL.

Fig. 10. Geomorphological sketch of the source areas for the Malser Haide megafan and the St Valentin outsize fans. Note the well-adjusted terrain character of adjacent valleys used for ‘clone-stamping’. Intact or near-intact terrain elements define perimeters of inferred catastrophic slope failure events at ‘Plawennspitz’ (P) and ‘Valentinkopf’ (V). Basemap: LiDAR topography.

† How can catastrophic failure be asserted when it is not yet known whether the Malser Haide is a homogenous diamict of disintegrated bedrock, and not a layered structure of incremental flows

and outwashes? In the absence of quarries or deep river cuts, this must await geotechnical investigations. However, at St Valentin (#2 and #3), eight closely-spaced boreholes near the fan

MEGAFANS FROM CATASTROPHIC SLOPE FAILURES

margins up to 100 m deep show chaotically intermixed sediment grades with no consistent layering (Fischer 1966), while at Allitz/Laas (#7) the 40 m-thick megafan toe is a massive unstratified diamict with plurimetric blocks (Fischer 1990). † If a lost mountain comparable with the Fundus ridge at Ko¨fels is identifiable here, how could its catastrophic collapse and discharge have been mobilized so as to settle out into the smooth form of a fan? Does the smooth surface indicate a high degree of comminution of mobilized debris or an overlay by subsequent debris flows and alluvial deposition? If the latter, how could such relatively low-energy processes extend smooth deposition for 11 km? This is addressed in the Discussion below.

Reconstructing a lost mountain source for the Malser Haide A lost mountain can be reconstructed above the present head of the Plawenn valley of sufficient

269

scale to support a catastrophic origin for the Malser Haide. Its former existence is apparent from extensive anomalous upland terrain (Fig. 10). Reconstruction of a geomorphologically authentic pre-failure landscape is possible by an iterative process for which published exemplars appear lacking. This process recognizes the typically sharp boundaries between pre- and post-failure terrain, and rebuilds a reasonably constrained proto-relief by reference to intact landform ‘templates’ in the vicinity (‘clone-stamping’). The methodology comprises: † preparing a geomorphological map based on terrain character (Fig. 10); † identifying slope facets and ridge lengths that appear little modified by post-glacial mass movements; † identifying landscape elements that may have been truncated or displaced, for example, valley heads, rock glaciers (Figs 11 & 12); † identifying large landslips and slope deformations, and reinstating them to former elevations; † defining the boundary of the pre-catastrophe terrain form around the inferred cavity;

Fig. 11. Source areas for the St Valentin outsize fans and the Malser Haide megafan, with the dismembered Vivana valley. The Plawenn valley head displays bold wedge structures fresher than adjacent gully complexes, inferred to have assisted detachment of the failed mass. Oblique view from SW (3D rendering of LiDAR topography).

270

D. JARMAN ET AL.

Fig. 12. Anomalous morphological features indicative of the lost mountains ‘Plawennspitz’ and ‘Valentinkopf’. (a) The Vivana valley remnant west rim, looking north from the foot of Grosshorn: a beheaded rock glacier drapes back over the headwall of the St Valentin south fan cavity. Note the anomalous ‘crater’ at the far end of the ridge, source of north fan. (b) The Vivana valley with beheaded rock glacier complex, looking west from the Habicherkopf ridge, the St Valentin breach shown beyond: the far side of the valley is the inferred site of the lost mountain ‘Valentinkopf’. (c) View from Kofelboden north across the head of the Plawenn valley into the dismembered head of the Vivana valley, through the location of the inferred lost mountain ‘Plawennspitz’. (d) The Vivana valley grossly asymmetrical cross-section and missing head, looking south along the decapitated west-side ridge from Pleiskopfl. (e) and (f) Hectometric gravitational morphostructures indicating structural propensity for large-scale slope instability. (e) Rim of the Plawenn valley at the head of the Vivana valley; note the remnant NE flank of Grosshorn from which the lost mountain contours and summit trend were extrapolated; (f) the Vivana valley west-side ridge viewed from the NE.

† extrapolating contours from this boundary across the cavity so far as their initial trend is clear; † taking the form beyond the limits of contour extrapolation by ‘clone-stamping’ suitable slope angles and curvatures from non-failed

topography in the vicinity. These templates take into account local rock-type variations, structural grain, valley asymmetry, etc.; † governing the lost mountain form by iterating the process, converging from its perimeters, until a well-integrated proto-landscape is achieved.

MEGAFANS FROM CATASTROPHIC SLOPE FAILURES

Implementing this procedure reveals that most of the pre-failure boundary can readily be fixed and extrapolated from Figure 10; only the SW boundary within the Plawenn valley is problematic. The position and steepness of a pre-catastrophe headwall across the valley between Grosshorn and Kofelboden has to be fixed. While the position is somewhat arbitrary, its form can be clone-stamped from the mature gully complexes in the SE Plawenn valley, interpreted as outside the failure. It may also have been supported by a higher NW Plawenn ridge, the present one being anomalously low and deformed. The lost mountain (here named ‘Plawennspitz’) reconstructed by this procedure has a summit above 3100 m (Fig. 13). This is about 200 m higher than the remnant Mittereck, and is consistent with other summit elevations along the inferred preglacial main watershed (Fig. 1). Plawennspitz emerges as the natural culminating peak at the convergence of several ridges and valley heads. Its upper slopes and summit ridge have been kept relatively broad and the encroaching cirques gentle, consistent with the subdued character of this watershed section. The failed mass is located predominantly above the head of the Plawenn valley, attaining a maximum thickness of 700 m (Fig. 13c). It almost completely beheads the Vivana valley (Fig. 12c), where incipient gravitational scarps 150 m behind the rim indicate the propensity for failure (Fig. 12e). It slightly beheads the Planeil and Rosell valleys, both of which display broad splitting and sagging crests. The volume of the failed mass obtained by digital elevation model (DEM) calculation is 1500 Mm3. This excludes a small sector east of the present Mittereck south ridge, where the present broad bay is inferred to be the cavity of a sliding failure into the Planeil valley, which is widely affected by slope deformation and rockslides (Fig. 10). Because it is not known whether this Planeil valley failure pre- or post-dated the main event, an arbitrary vertical separation is placed between them, attaining 200 m in height; the difference to the volume is minor. It is assumed that all of the failed mass within the present Plawenn catchment contributed to the Malser Haide, except for c. 50 Mm3 exported down the adjacent valley to the west to build the contiguous and, also outsized, Fischerhauser fan (#4). Simple robustness checks on this result indicate: † retreating the arbitrarily-positioned pre-failure Plawenn valley headwall by 100 m shrinks Plawennspitz by 65 m in height and 130 Mm3 in volume. A maximal retreating of 250 m reduces volume by 300 Mm3;

271

† lowering the summit by 100 m cuts the volume by 100 Mm3; † conversely, steepening the headwall and building a sharper summit, consistent with others above 3000 m along the divide, increases volume by 300 Mm3. An elevation of 3200 m is feasible for a summit here; † excluding the Plawenn valley, NW ridge reconstruction cuts the total by 100 Mm3. The lost mountain volume could thus be in the range of 1000– 1800 Mm3. Applying a bulking-up factor to the deposit, conventionally 25–33% (Rapp 1960; Crosta et al. 2007) but possibly less (say 15%) with comminuted and well-consolidated debris, yields a range of 1150–2400 Mm3 for the evacuated debris.

Relating source and deposit At this preliminary level of quantification, the source and deposit volumes match well (Table 4). Note that the reconstruction of ‘Plawennspitz’ was made purely on geomorphological criteria, prior to calculating the volume of the Malser Haide and with no attempt to equal it. The two anomalous landscapes require further investigation to determine whether they constitute a catastrophic event. Nonetheless, to deliver such a deposit volume from the 8.3 km2 catchment of the Plawenn valley by episodic processes would require erosion of the whole catchment by an average of 200 m over the Holocene, a scale for which there are no precedents in deglaciated mountains (Hinderer 2001). In the Alps, some of the highest post-glacial erosion rates occur in the upper Rhine tributary catchment areas in debrisflow- and landslide-prone weak Bu¨ndner schist and flysch: 4 mm per annum or c. 40 m over the Holocene (Korup & Schlunegger 2009). If verified as catastrophic slope failure, Plawennspitz –Malser Haide would be among the five largest landslides in the Alps, after Flims, and similar to Sierre, Engelberg, and Ko¨fels (von Poschinger 2002). It would be exceptional (with Ko¨fels) in occurring in crystalline rather than carbonate rocks. It would be in the top 60 recorded globally (Korup et al. 2007).

The St Valentin outsize fans The reconstruction also identifies a second lost mountain (‘Valentinkopf’) on the west side of the Vivana valley (Fig. 13). This valley is grossly asymmetrical (Figs 11 & 12) and, with its lost crest parallel to the upper Venosta trough, it closely resembles the Fundus Valley above Ko¨fels. Clear evidence of catastrophic collapse is

272

D. JARMAN ET AL.

Fig. 13. Reconstruction of the lost mountains ‘Plawennspitz’ (P) and ‘Valentinkopf’ (V) by contour extrapolation and the ‘clone-stamping’ procedure. (a) Present-day topography (LiDAR), with cavities of areas lost by catastrophic failure outlined and contoured. (b) Reconstructed topography (DEM at 10 m resolution), with 100 m contours (solid) superimposed on present-day ones (dashed). Plawennspitz has a pyramidal form typical of an apex location; Valentinkopf has a more rounded form, comparable to adjacent mid-level terrain (Fig. 7). (c) The thickness of the Plawennspitz failed mass by 100 m contours.

MEGAFANS FROM CATASTROPHIC SLOPE FAILURES

273

Table 4. Estimates of the Malser Haide source and deposit volumes Volume (Mm3) Source Deposit

First estimate from reconstruction applying robustness checks with robustness checks and bulking-up First estimate from cross-sections plus Fischerhauser (#4) fan applying robustness checks Fischer (1966) ‘murkegel’ deposit

1500 1000 – 1800 1150 – 2400 1650 1700 1100 – 2500 1550

1000 Mm3 ¼ 109 m3 ¼ 1 km3.

provided by the extensive rock glaciers that could not have developed beneath the present low to negligible slopes (Figs 10 & 12b); the head of one rock glacier actually drapes back over the remnant crest (Fig. 12a). This crest also displays splits indicating extensive structural weaknesses (Fig. 12f). ‘Valentinkopf’ is the inferred source for the St Valentin outsize fans (#2 and #3: Figs 5d & 11). A maximal reconstruction, restoring a continuous Venosta trough wall, yields a broad ridge at over 2650 m, some 2–300 m higher than the present Vivana valley rim, although still 250 m short of full symmetry. This is consistent with the chain of mid-level hills preserving rounded preglacial forms along this rim of the trough, reflecting its incision as a major breach (Fig. 10). Maximum thickness of the lost mountain here is 550 m. However, a gross cavity volume at 1100 Mm3 vastly exceeds the volume of the St Valentin fans, estimated using the same procedure as the Malser Haide at 430 Mm3 and constrained by borehole evidence. Even with deep gullying and a cliffed summit factored in, as elsewhere along this trough wall, the cavity volume remains excessive at 680 Mm3. The south fan appears to be overlapped by the north, and its source bay is more maturely dissected by gullying, suggesting that ‘Valentinkopf ’ may have collapsed in several phases, with earlier ones possibly removed by the last glacier.

Discussion In proposing a new category of catastrophic slope failure outcome, many vexed questions are raised, including source rock-mass character and flow dynamics. The key issue discussed here is why similar mountain relief should yield such contrasting catastrophic collapse forms as hilly rock avalanche deposits and smooth fan-mimicking cones. The character of catastrophic slope failure deposits is conditioned by: the source geometry; the failure mechanism; the degree of fragmentation (Crosta et al. 2007); the wetness of the failed mass and any materials incorporated in transit, for example

colluvium, glacial debris, ice (Crosta et al. 2009); the trajectory resulting from the interaction of moving debris and topography; and whether the emplacement zone permits unconstrained run-out or induces deflection (Hewitt et al. 2008). Here, the source ! trajectory ! run-out configuration appears to be especially significant.

Topographical controls on fan-mimicking deposits– the ‘debouchure’ Most large catastrophic failures in the Alps are on open trough walls and have produced hilly deposits with substantial preservation of source-area fabric, for example Sierre (Burri 1997), Ko¨fels (Sørensen & Bauer 2003), Flims (Ivy-Ochs et al. 2009), Fernpass (Prager et al. 2009a), Tschirgant, and Disentis (Fig. 1). A key difference between these hilly deposits and the Venosta anomalous fans is that the former have laterally unconstrained descent trajectories, whereas the latter have been funnelled through constricted ‘debouchures’ between the source and the slope foot (Fig. 5). We thus propose that conoidal deposits mimicking incremental sedimentary fans may originate from catastrophic slope failure if: † the exposure of the source area above a main glacial trough is high and steep enough to allow rapid movements to originate; † an open mouth or ‘debouchure’ exists on the trough wall, greater in width than the typical deeply incised ravine or hanging V-shaped tributary, but narrow enough to funnel the sliding mass through a chute. † the debris is originally, or becomes, sufficiently comminuted to discharge through the constriction as a freely flowing mass or stream, rather than remaining so blocky as to form a hilly deposit or be arrested as a choke; † the debouchure is sufficiently elevated above a wide, flat trough floor for a cone to form. We suggest that the conoidal form is a product of rapid flow-energy dissipation achieved: (a) by

274

D. JARMAN ET AL.

constriction as the debris mass passes through the debouchure; (b) by fragmentation during transit, an energy-consuming process (Hungr 2006); and (c) by sudden reduction of gradient as the debris mass enters the proximal run-out zone. The momentum-starved flow decelerates abruptly and is deposited in an efficient conoidal form. The absence of opposite slope run-up in all of the Venosta cases, despite several being distally confined, supports this interpretation. The absence of prominent leve´es (except possibly at the Malser Haide debouchure, see later) is more surprising. Debouchure origination is clearly significant. Some may pre-exist as small side valleys, but others might result from initial lower-slope failure of more conventional scale, either propagating rapidly upslope or setting in train progressive weakening above. The main Venosta trough displays classic alpine mid-slope benches, often with signs of slope deformation: failure biting into them would magnify the effective relief exposure of the upper slopes. There are 18 such debouchures in Val Venosta. They occur above all except two of the anomalous fans; the five not above such fans may indicate evacuated or latent large events. They range between 150 and 600 m in width (Table 2). Significantly, they are proportionate to cone size, averaging 275 and 540 m in width above outsize fans and megafans, respectively, implying that broader openings destabilize larger source areas. Fans often retrogress up the chutes, obscuring their profiles, but these bold trough-wall openings are anomalous landforms in their own right.

Single or multiple events– the unified Malser Haide A single catastrophic event issuing from a debouchure will experimentally produce a smooth conoidal form if the material is uniformly granular. The greater the extent to which a source area is ‘prefailed’ or ‘pre-disintegrated’, clearly the greater is the potential for comminution to a rapidly flowing fine debris stream. In reality, heterogenous debris might be expected to generate more irregular forms. Equifinality allows that fans of Venosta type may be attained via various combinations of circumstances. One scenario emplaces a subconoidal core deposit, whether by one or a few events, which is either rapidly or gradually smoothed by debris-flow and fluvial processes as the unstable source area re-equilibrates. Subsequent overlays were observed by Blair (1987), Derbyshire & Owen (1990) and Watanabe et al. (1998). In the Alps they are have been located at St Barthele´my and Corberier-Yvorne (Eisbacher & Clague 1984), and notably at Illgraben (Fig. 1), where the megafan is concentrically layered, most recently

by a 5.4 Mm3 event in 1961 (Oppikofer et al. 2006). These examples suggest that slope instability propagates headwards to trigger successive catastrophic (but relatively small-scale) events. However, the remarkable smoothness of the entire length of the Malser Haide (Figs 6 & 7) can hardly be accounted for by multiple-event and incremental-overlay models. Multiple large events would tend to increase the cone thickness at the debouchure, not extend it; overlays from conventional debris flows could not travel as thin veneers over such a distance. Even so, clear evidence for a single event is lacking, although Fischer (1965) described anomalous ‘kegelimse’ terraces sloping outwards from 1750 m asl above down to 1550 m on either side of the Plawenn valley mouth about 100 m above the fan. They bear morainic deposits from sources north of St Valentin, and may resemble the leve´es often found parallel to catastrophic rock avalanches (Abele 1974). The steeper gradient within the Plawenn valley (Fig. 8) suggests that subsequent debris-flow activity has retrogressed up it, while the concave gradient in the Upper Val Venosta suggests that alluvial and debris-flow overlay deposition has been limited to the proximal part of the fan. This is consistent with the historical record: debris flows have swept across the conoidal part of the fan in recent centuries, destroying the village of Plawenn three times (Eisbacher & Clague 1984). Applying Occam’s razor suggests that the simplest explanation for the remarkably unified form of the Malser Haide is that it was emplaced by a single catastrophic event of fluidized character funnelled through a debouchure, with relatively minor smoothing by subsequent processes including human activity. This is consistent with inferences of rapid fluidized flow elsewhere (Derbyshire & Owen 1990). The Flims mega-failure also has an exceptionally long run-out: dating confirms that it was most probably a single event, although multiple events within a short period are not ruled out (Ivy-Ochs et al. 2009).

Clustered incidence of anomalous fans The propensity for large slope instabilities to cluster has been noted in Scotland (Jarman 2006), and in the Central Alps (Agliardi et al. 2009b), where it applies equally to the spatial incidence of anomalous fans (Fig. 1). The occurrence of the largest fan cluster in Val Venosta has no immediately obvious explanation, and may reflect a complex interaction of bedrock type, structure, active tectonic processes and paraglacial slope evolution. But this area has a varied geology, prone to both slope deformation and rockslides, and intensely dissected by tectonic lineaments (Agliardi et al. 2009a, b); while tectonic

MEGAFANS FROM CATASTROPHIC SLOPE FAILURES

damage is an important predisposing factor to slope failure (Brideau et al. 2009), it is no more prevalent here than in many other parts of the Alps. In the Upper Val Venosta all of the anomalous fans emanate from east-side sources carved into varied resistant rock types (e.g. carbonates, orthogneiss) with extensive fracturing of brittle fabric. Even though the west side is affected by widespread slope deformation, it is not a fan source; unlike Val Mustair where the fan sequence is below a sagging ridge. The physical asymmetry of the Upper Val Venosta, with much of its east side a cliff wall, may be a factor. By contrast, anomalous fans occur on both sides of the main Venosta trough, as does extensive slope deformation. All of the megafans, except for the Malser Haide, are in this main trough, a not unreasonable scale effect. Throughout the area, arrested rockslides are fairly small while large rock avalanches are absent, suggesting that here they develop into fans. The main Venosta trough is unusually large and overdeepened, possibly exploiting the Vinschgau shear zone (Fig. 2). The glacier was augmented by transfluent input of ice from the Engadine ice centre (Florineth & Schlu¨chter 1998) through several glacial breaches of the main Alpine divide, among which Reschen Pass– St Valentin is exceptionally large (Fig. 1). An association between large slope failures and breaches has also been identified in Scotland (Jarman 2006). Glacial debuttressing, post-glacial unloading, topographical stress rotation, regional seismicity, permafrost, meltwaters, climatic variation and progressive weakening conventionally explain slope failure in the Alps. However, these are endemic factors operating along all glacial troughs, and cannot explain local clusters such as the Val Venosta fans (cf. Jarman 2006). Geology and structure are clearly influential, but only enabling factors. Proximity of these fans to the St Valentin breach, which entailed the removal of more than 1000 m of bedrock (Fig. 10), and to the overdeepened main Venosta trough could suggest that concentrated erosion has induced high rebound stresses akin to quarry-floor bursting (Hutchinson 1988; Ustaszewski et al. 2008). Such rebound breakouts could also have provoked higher magnitude seismicity around deglaciation (Persaud & Pfiffner 2004). Topographical amplification of seismic shaking may favour large failures at or near summits (Murphy 2006), possibly contributing to the catastrophic ‘Plawennspitz’ and ‘Valentinkopf’ collapses.

Conclusions This study has explored for the first time the possibility that oversized sediment cones in glacial

275

troughs are not conventional alluvial or debris-flow fans but have primarily catastrophic origins. It identifies a cluster of anomalous fans in Val Venosta that are grossly disproportionate in scale relative to catchment area. These ‘outsize fans’ and ‘megafans’ emanate from a quarter of the side valleys, with similar valleys yielding allometric or no fans. This suggests that they are the product of one or more high-magnitude low-frequency events. If nonstratified constitution can be demonstrated by sedimentological and geotechnical investigation then these ‘catastrophic fans’ would represent a new category of slope failure deposit. They appear to be generated where the topography funnels a fastflowing comminuted debris mass through a relatively confined ‘debouchure’ well above a wide trough floor, with rapid deceleration in the proximal run-out zone producing an efficient conoidal form. The exceptionally large Malser Haide would, if verified, be one of the largest landslide deposits in the Alps, estimated at 1650 Mm3. For it to have been emplaced incrementally by floods and debris flows would have required erosion of the entire source valley catchment by an average of 200 m during the Holocene, an unprecedented scale and process rate. Instead, its unified form sustained over 11 km suggests a single catastrophic event, for which reconstruction of anomalous terrain at the source valley head yields a lost mountain of comparable volume. Anomalous fans of catastrophic provenance are identified widely across the Alps, with the Venosta cluster being the most numerous. Their incidence here cannot be explained by generic factors, such as glacial unloading, and while favourable geology is clearly important, locally concentrated glacial trough erosion and rebound stresses may be involved. This new category also embraces certain fans reported in the Himalaya (prehistoric), British Columbia (Cheekye) and the Andes (Huascaran, Peru in 1970), where active tectonics, vulcanicity and seismicity are more relevant. This geomorphological study addresses the catastrophism v. uniformitarianism controversy via a specific dataset that provides evidence for both. As anomalous fans mimic allometric fans on all criteria other than scale, equifinality is once again a caution against monogenetic interpretations. We are grateful to Michel Jaboyedoff for valuable inputs; Ian Evans (Durham) and an anonymous referee for detailed and perceptive reviews; John Clague, Ken Hewitt and Monique Fort for their encouragement and examples; Sven Lukas and Richard Scothorne for discussion of the Vivana valley landscape anomalies; Hans-Petter Staffler (Natural History Museum, Bolzano) for Fischer’s papers on Val Venosta; Provincia Autonoma di Bolzano for providing LiDAR topography; and the Forest Service for permitting vehicular access to the Vivana valley. The

276

D. JARMAN ET AL.

first author thanks the third for laying down the original challenge to prove feasible a lost mountain source for the Malser Haide, and for resourcing this study.

References Abele, G. 1974. Bergstu¨rze in den Alpen, ihre Verbreitung, Morphologie, und Folgeerscheinungen. Wissenschaftliche Alpenvereinshefte, 25. Agliardi, F., Crosta, G. & Zanchi, A. 2001. Structural constraints on deep-seated slope deformation kinematics. Engineering Geology, 59, 83–102. Agliardi, F., Crosta, G., Zanchi, A. & Ravazzi, C. 2009a. Onset and timing of deep-seated gravitational slope deformations in the eastern Alps, Italy. Geomorphology, 103, 113–129. Agliardi, F., Zanchi, A. & Crosta, G. B. 2009b. Tectonic vs. gravitational morphostructures in the central Eastern Alps (Italy): constraints on the recent evolution of the mountain range. Tectonophysics, 474, 250–270. Allen, P. A. & Hovius, N. 1998. Sediment supply from landslide-dominated catchments: implications for basin-margin fans. Basin Research, 10, 19–35. Ambrosi, C. & Crosta, G. 2006. Large sackung along major tectonic features in the Central Italian Alps. Engineering Geology, 83, 183– 200. Blair, T. C. 1987. Sedimentary processes, vertical stratification sequences, and geomorphology of the Roaring River alluvial fan. Journal of Sedimentary Petrology, 57, 1 –18. Blair, T. C. & Mcpherson, J. G. 1994. Alluvial fan processes and forms. In: Abrahams, A. D. & Parson, A. J. (eds) Geomorphology of Desert Environments. Chapman & Hall, London, 354–402. Brideau, M.-A., Yan, M. & Stead, D. 2009. The role of tectonic damage and brittle rock fracture in the development of large rock slope failures. Geomorphology, 103, 30–49. Burri, M. 1997. Ge´ologie re´cente de Finges et de ses environs (VS). Bulletin Murithienne, 155, 5– 27. Crosta, G. B. & Frattini, P. 2004. Controls on modern alluvial fan processes in the Central Alps, Northern Italy. Earth Surface Processes and Landforms, 29, 267– 293. Crosta, G. B., Frattini, P. & Fusi, N. 2007. Fragmentation in the Val Pola rock avalanche, Italian Alps. Journal of Geophysical Research, 112, F01006; doi: 10.1029/2005JF000455. Crosta, G. B., Imposimato, S. & Roddeman, D. 2009. Numerical modelling of entrainment/deposition in rock and debris-avalanches. Engineering Geology, 29, 135–145. Crosta, G. B. & Zanchi, A. 2000. Deep seated slope deformations –huge, extraordinary, enigmatic phenomena. In: Bromhead, E., Dixon, N. & Ibsen, M. (eds) Landslides in Research, Theory and Practice. Proceedings of the 8th International Symposium on Landslides, Cardiff, June 2000. Thomas Telford, London, 351– 358. Derbyshire, E. & Owen, L. A. 1990. Quaternary alluvial fans in the Karakoram Mountains. In: Rachocki, A. H. & Church, M. (eds) Alluvial Fans: A Field Approach. Wiley, Chichester.

Eisbacher, G. H. & Clague, J. J. 1984. Destructive mass movements in high mountains: hazard and management. Geological Survey of Canada, Professional Paper, 84-16. Evans, S. G., Roberts, N. J., Ischuk, A., Delaney, K. B., Morozova, G. S. & Tutubalina, O. 2009. Landslides triggered by the 1949 Khait Earthquake, Tajikistan, and associated loss of life. Engineering Geology, 109, 195 –212; doi: 10.1016/j.enggeo.2009.08.007. Fischer, K. 1965. Murkegel, Schwemmkegel und Kegelimse in den Alpenta¨lern. Mitteilungen der Geographischer Gesellschaft in Mu¨nchen, 56, 127– 159. Fischer, K. 1966. Die Murkegel des Vinschgaues. Der Schlern, 40, 24– 34. Fischer, K. 1990. Entwicklunggsgeschichte der Murkegel im Vinschgau. Der Schlern, 64, 93– 96. Florineth, D. & Schlu¨chter, C. 1998. Reconstructing the Last Glacial Maximum (LGM) ice surface geometry and flowlines in the Central Swiss Alps. Eclogae Geologicae Helvetiae, 91, 391– 407. Friele, P. A., Ekes, C. & Hickin, E. J. 1999. Evolution of Cheekeye fan, Squamish, British Columbia: Holocene sedimentation and implications for hazard assessment. Canadian Journal of Earth Sciences, 36, 2023– 2031. Harvey, A. M. 2003. Alluvial fan. In: Goudie, A. S. (ed.) Encyclopedia of Geomorphology. Routledge, London. Harvey, A. M., Mather, A. E. & Stokes, M. (eds) 2005. Alluvial Fans: Geomorphology, Sedimentology, Dynamics. Geological Society, London, Special Publications, 251. Heuberger, H. 1994. The giant landslide of Ko¨fels, ¨ tztal, Tyrol. Mountain Research and Development, O 14, 290– 294. Hewitt, K. 1999. Quaternary moraines v. catastrophic rock avalanches in the Karakoram Himalaya, Northern Pakistan. Quaternary Research, 51, 220–237. Hewitt, K. 2001. Catastrophic rockslides and the geomorphology of the Hunza and Gilgit River valleys, Karakoram Himalaya. Erdkunde, 55, 72–93. Hewitt, K., Clague, J. J. & Orwin, J. F. 2008. Legacies of catastrophic rock slope failures in mountain landscapes. Earth-Science Reviews, 87, 1– 38. Hinderer, M. 2001. Late Quaternary denudation of the Alps, valley and lake fillings and modern river loads. Geodinamica Acta, 14, 231–263. Hoek, E. & Brown, E. T. 1997. Practical estimates or rock mass strength. International Journal of Rock Mechanics and Mining Sciences and Geomechanics Abstracts, 34, 1165– 1186. Hungr, O. 2006. Rock avalanche occurrence, process and modelling. In: Evans, S. G., Scarascia Mugnozza, G. S., Strom, A. & Hermanns, R. L. (eds) Landslides from Massive Slope Failure. Springer, Dordrecht, 243–266. Hutchinson, J. N. 1988. General report: morphological and geotechnical parameters of landslides in relation to geology and hydrogeology. In: Proceedings of the 5th International Symposium on Landslides, Lausanne, Volume 1. Balkema, Rotterdam, 3– 35. Iverson, R. 2003. Debris flow. In: Goudie, A. S. (ed.) Encyclopedia of Geomorphology. Routledge, London. Ivy-Ochs, S., Von Poschinger, A., Synal, H.-A. & Maisch, M. 2009. Surface exposure dating of the

MEGAFANS FROM CATASTROPHIC SLOPE FAILURES Flims Landslide, Graubu¨nden, Switzerland. Geomorphology, 103, 104– 112. Jaboyedoff, M. & Derron, M.-H. 2005. A new method to estimate the infilling of alluvial sediment of glacial valleys using a sloping local base level. Geografia Fisica e Dinamica Quaternaris, 28, 37– 46. Jarman, D. 2002. Rock slope failure and landscape evolution in the Caledonian Mountains, as exemplified in the Abisko area, northern Sweden. Geografiska Annaler, 84A, 213–224. Jarman, D. 2006. Large rock slope failures in the Highlands of Scotland: characterisation, causes and spatial distribution. Engineering Geology, 83, 161 –182. Korup, O., Clague, J. J., Hermanns, R. L., Hewitt, K., Strom, A. L. & Weidinger, J. T. 2007. Giant landslides, topography, and erosion. Earth and Planetary Science Letters, 261, 578–589. Korup, O. & Schlunegger, F. 2009. Rock-type control on erosion-induced uplift, eastern Swiss Alps. Earth and Planetary Science Letters, 278, 278–285. Lecce, S. 1990. The alluvial fan problem. In: Rachocki, A. H. & Church, M. (eds) Alluvial Fans: A Field Approach. Wiley, Chichester, 3 –24. Legros, F. 2006. Landslide mobility and the role of water. In: Evans, S. G., Scarascia Mugnozza, G. S., Strom, A. & Hermanns, R. L. (eds) Landslides from Massive Slope Failure. Springer, Dordrecht, 233–242. Murphy, W. 2006. The role of topographic amplification on the initiation of Rock Slope Failures during earthquakes. In: Evans, S. G., Scarascia Mugnozza, G. S., Strom, A. & Hermanns, R. L. (eds) Landslides from Massive Slope Failure. Springer, Dordrecht, 139–154. Oppikofer, T., Bardou, E & Jaboyedoff, M. 2006. DTM analysis of the 1961 rock avalanche, Illgraben creek (Switzerland). Geophysical Research Abstracts, 8, SRef-ID: 1607-7962/gra/EGU06-A-02291. Penck, A. & Bru¨ckner, E. 1909. Die Alpen im Eiszeitalter. Tauchnitz, Leipzig. Persaud, M. & Pfiffner, O. A. 2004. Active deformation in the eastern Swiss Alps: Post-glacial faults, seismicity and surface uplift. Tectonophysics, 385, 59–84. Plafker, G. & Ericksen, G. E. 1978. Nevados Huascaran avalanches, Peru. In: Voight, B. (ed.) Rockslides and Avalanches Natural Phenomena. Elsevier, Amsterdam, 277 –314. Prager, C., Zangerl, C., Patzelt, G. & Brandner, R. 2008. Age distribution of fossil landslides in the Tyrol (Austria) and its surrounding areas. Natural Hazards and Earth System Sciences, 8, 377–407. Prager, C., Ivy-Ochs, S., Ostermann, M., Synal, H.-A. & Patzelt, G. 2009a. Geology and radiometric 14C,

36

277

Cl and Th/U-dating of the Fernpass rockslide (Tyrol, Austria). Geomorphology, 103, 93–103. Prager, C., Zangerl, C. & Nagler, T. 2009b. Geological controls on slope deformations in the Ko¨fels rockslide area (Tyrol, Austria). Austrian Journal of Earth Sciences, 102, 4– 19. Rachocki, A. H. & Church, M. (eds) 1990. Alluvial Fans: A Field Approach. Wiley, Chichester. Rapp, A. 1960. Recent development of mountain slopes in Ka¨rkevagge and surroundings, northern Scandinavia. Geografiska Annaler, 42, 71–200. Schrott, L., Hufschmidt, G., Hankammer, M., Hoffmann, T. & Dikau, R. 2003. Spatial distribution of sediment storage types and quantification of valley fill deposits in an alpine basin, Reintal, Bavarian Alps, Germany. Geomorphology, 55, 45– 63. Shang, Y., Yang, Z., Lihui, L., Liu, D., Liao, Q. & Wang, Y. 2003. A super-large landslide in Tibet in 2000: background, occurrence, disaster, and origin. Geomorphology, 54, 225 –243. Sørensen, S. A. & Bauer, B. 2003. On the dynamics of the Ko¨fels sturzstrom. Geomorphology, 54, 11–19. Sorriso-Valvo, M., Antronico, L. & La Pera, E. 1998. Controls on fan morphology in Calabria, southern Italy. Geomorphology, 124, 169–187. Spindler, K. 1994. The Man in the Ice. Weidenfeld & Nicolson, London. Staffler, H. & Nicolussi, K. 2004. Fund subfossiler Holzsta¨mme in Naturns. Der Schlern, 78, 4 –9. Taylor, G. 1940. Trento to the Reschen pass: a cultural traverse of the Adige corridor. Geographical Review, 30, 215 –237. Ustaszewski, M. E., Hampel, A. & Pfiffner, O. 2008. Composite faults in the Swiss Alps formed by the interplay of tectonics, gravitation and postglacial rebound: an integrated field and modelling study. Swiss Journal of Geosciences, 101, 223– 235. Van Husen, D. 1987. Die Ostalpen in den Eiszeiten. Geo¨ sterreichs, Wien. logische Bundesamstalt O Von Poschinger, A. 2002. Large rockslides in the Alps: a commentary on the contribution of G. Abele (1937– 1994) and a review of some recent developments. In: Evans, S. G. & de Graaff, J. V. (eds) Catastrophic Landslides: Effects, Occurrence and Mechanisms. Geological Society of America, Reviews in Engineering Geology, 15, 237 –255. Watanabe, T., Dali, L. & Shiraiwa, T. 1998. Slope denudation and the supply of debris to cones in Langtang Himal, central Nepal Himalaya. Geomorphology, 26, 185 –197.

Index Page numbers in italic refer to Figures. Page numbers in bold refer to Tables. active movement 99, 102, 104, 109 faulting 187, 196, 197 landslide 218, 219 rockslide 74, 147, 188 ˚ knes rockslide, structural control 74, 147– 160 A geology and morphology 148–151 potential movement 91– 93 structures 37, 151–154 ˚ knes–Tafjord project 147 A allometric fan 256, 258 alluvial fan 253, 264 alluvium, buried 4 anhydrite 121, 122, 125, 128 Aosta Valley, slope deformation case study 113 –129 Apennines, geology and geomorphology 201–206 Arbino landslide 217, 218–219 Ayasse Torrent 116, 117, 118 back-crack structure 62, 65, 68– 71 Romsdalen 45, 46, 50, 58 Storfjord 44, 48, 49, 59 basal sliding plane 81–83, 99, 107, 109 ˚ knes rockslide 151, 153 A Gamanjunni slide 191, 193 Biasca fan 254 block size 177, 178 block-field, Gamanjunni slide 190, 191, 193, 194 Boaco see Santa Lucia Bosco del Conte, slope deformation 218 breccia 82 Breuil– Cervinia, slope deformation 115–116, 122–123 brittle deformation 27, 107 brittle fracture 241, 243 brittle structures and slope stability 41–70, 71 building damage 121 –124, 125 bulging 238, 248 caldera, Santa Lucia Depression 136– 142 Campo di Giove rock avalanche 204, 205, 206, 213 Canada see Turtle Mountain Caramanico Basin 207– 208 Caramanico Breccias 204, 205 Caramanico Fault System 202 –204, 207, 208, 212 Caramanico Valley, slope evolution 201–213 analysis 208–212 geology and geomorphology 202– 206 landforms 206–208 palaeo-landslides 204 results 212– 213 catastrophic slope failure fans 253– 276 source reconstruction 268-273 catastrophist v uniformitarian 253, 276 catchment area, fans 256, 257, 258, 260, 261 cavities (anomalous) and fan cones 261, 262, 272 Celentino, gravitational slope deformation 235 –249 geology 236, 237, 238 geomorphology 236, 238, 239 numerical modelling 243– 248 Cervinia, slope deformation 115– 116, 122– 123

Cevedale Massif 236 Champorcher Valley 116, 117 climate 109, 127 cohesion values 28, 73, 74 Colle Longue, slope deformation 14 columnar failure 62, 104 convex slope break and instability 107, 109 CROP II seismic profile 204 Curon Venosta, slope deformation 217, 218, 219 dam, risk to 4, 102 data acquisition 114 debris flow deposits 253, 275, 276 deep-seated gravitational slope deformation 4, 97 controlling factors classification 125– 129 historical data 113– 114 landforms 12– 14, 206–208 Tine´e Valley 11–22 see for example Aosta, Celentino, Cervinia, Hoˆne, Livigno, Motto d’Arbino, Quart, Sogn, Villeneuve deglaciation simulation 249 dip-slip sliding 229 dissolution and slope deformation 119– 122, 124, 125, 126, 127, 128 dissolution, limestone 166, 173, 177 dogtooth ridge profile 244 DSGSD see deep-seated gravitational slope deformation ductile deformation, Aosta 124 ductile fabric 27 ductile structures and slope stability 41–58, 71, 73 early warning system 147 earthquake 139 engineering model, slope development 208–212 erosion 1 –2 rate of 216, 229, 273, 276 evaporite 121, 124, 203 see also anhydrite exfoliation 58, 59, 63, 64, 67, 89 fabric, type 1 failure surface 45, 49, 82 fans, catastrophic slope failure 253–276 anomalous 256, 258, 274 area of 257– 260, 264 geomorphology 254– 255, 269, 272 in glacial troughs 253 –254 Malser Haide 264 –273 St Valentin 273–274 Val Venosta 255 –264 fatalities due to slope failure 97, 98 fault evolution 196 fault gouge 15 see also gouge fault reactivation 139 fault rock breccia 82 fault scarp 192, 193, 195–197 fault zones and gravitational deformation 123– 125

280 faults and slope instability 58–68 faults, active 187, 196, 197 ˚ knes rockslide 153 faults, A Ferret Valley 217, 218 field investigations 31–37, 69 Alpine valley shape 218– 219, 231 La Clapie`re 12, 14–19 Sogn and Fjordane 104, 107, 109 Tine´e Valley 12– 19 Turtle Mountain 165– 166 finite difference method code (FLAC) 212, 219 fjord, structural ranking study 79–93 Fjordane see Sogn fjords, slope instability 27–75 Fla˚m rockslide 32 Flatmark valley, collapse 46 fluidization 253 flysch, geomechanical parameters 211 fold geometry 71 ˚ knes rockslide 151– 152, 155, 156, 158 fold hinges, A fold-related fracturing, Turtle Mountain 165, 166–170, 174– 179, 181 folds and slope instability 43–50 foliation ˚ knes rockslide 151, 152, 154–156, 157–158 A Sogn and Fjordane 99, 100– 101, 102, 103, 107 Storfjorden 80, 81–83, 85 potential failures 90 foliation, slope instability 41, 43, 61– 63, 65, 66, 71 field observation 31–37, 69 fracture plot, Celentino 241 fracture transfer 103 fractures 56, 60, 63, 64, 66, 82, 83 open 58, 64, 66, 67, 69, 109 site investigation 100 –101 fractures and slope failure, Turtle Mountain 163–181 analytical methods 165– 166 fold form 165, 166– 170 geology and geomorphology 164– 165 lineaments 167, 168, 175 model 179– 181 rock conditions 170–174 slope instabilities 175–179 fractures, back-bounding 104, 108 fractures, slope instability 31– 37, 173– 179 France (SW Alps), slope stability 11– 23 Frank Slide 163, 181 friction angle 28, 73, 74 Turtle Mountain 176, 179, 180 frost wedging 28 Gamanjunni slide 186, 188–198 geology 188, 189 geomorphology 191 –194 InSAR 188–191 Geiranger Fjord 62 Geirangerfjorden 81 geochronology 10 Be dating 14, 19 14 C dating 137, 138–139, 201 geology and slope failure ˚ knes rockslide 148–151 A Aosta Valley 114, 115 Celentino 236, 237, 238

INDEX Turtle Mountain 164–165 Sogn and Fjordane 97, 100–101, 106–107 geomechanical parameters, Caramanico 209, 210–211, 222 geomechanical survey, Celentino 240– 243, 249 geomorphical features, Italian Alps 215, 218 geomorphology 21 ˚ knes rockslide 148– 151 A Apennines 201–206 Celentino 236, 238, 239, 240, 242 Gamanjunni slide 191– 194, 198 megafans 254– 273 numerical models 220–222, 224–230 Turtle Mountain 164–165 Villeneuve 122 geotechnical parameters (Caramanico) 208–209 Gikling rockslide 37, 51, 54, 55, 56, 71 glacial cirque, slope deformation 231 glacial erosion 218 rate of 216 glacial trimline 254 glacial trough fans 253– 254, 255, 276 glacial unloading 236, 247 –249, 275 glaciated regions and slope deformation 235, 243, 245 glacier fill, modelling 221, 223, 229, 247 glacier retreat and slope stability 11–14, 28 gneiss 28, 31– 37, 70, 71 Aosta 116–117, 119 Storfjorden 81, 93 gouge 15, 43, 55, 69, 158 recent 89, 90, 92, 93 graben ˚ knes rockslide 152, 154 A Tussen rockslide 102 granitic rocks 12, 20 gravitational deformation, south-west Alps 11– 22 chronology 14, 19 field investigation 12–19 gravitational duplex 81, 82 gravitational mass movement see deep-seated . . . gravity and tectonics 2 –3 groundwater 51, 52, 173 and rockslide 156, 158, 160 slope deformation 216, 217, 223, 231 Heggursaksla, potential rockslide 91, 92, 93 ˚ knes 149, 157 historical rockslide, A historical slope failure 28–41, 98, 109 data 113 –114 Hoek-Brown failure criterion 247, 248 Hoˆne, gravitational slope deformation 114– 119 hummocky topography 122, 142 hydraulic pressure 51, 52, 58, 74 Hyllestad unstable area 104– 105 ice-river 43 inherited structure 28, 31– 37, 39, 43 InSAR (interferometric synthetic aperature radar) study of rock slope failure 185– 198 discussion 194–196 methodology 185–188 results 188– 194 synthesis 196– 197 intermontane basin 201– 203

INDEX Italian Alps see Aosta Valley and Celentino Italy, megafans 253–276 Italy, Plio-Quaternary slope evolution 201– 213 Italy, valley shape and deformation 215–231 Ivasanen rockslide 34, 56–58 joint roughness coefficient 240, 241, 243, 247 joints 27, 41, 46, 248 ˚ knes rockslide 152–153, 158 A Arbino 219 Storfjorden 85 Turtle Mountain 174, 176– 179, 181 karst 206, 213 Turtle Mountain 170, 176, 177, 181 kink fold 2 knick-point 100–101, 107, 108, 109 La Clapie`re, field investigation 12, 14– 19 landforms and anomalous cavities 262, 263 landforms/geomorphology, Caramanico 206–208 landslide active 218, 219 and erosion 1– 2 Aosta Valley 116 Caramanico 204, 205, 206, 207, 208, 213 inventory 217, 230 mechanism 4–5, 15, 20 very large 253, 273, 276 landslide and structural control 147– 160 analytical methods 151 faults 153 folds 151–152, 155, 156, 158 foliation 151, 152, 154 –156, 157–158 joints 152–153, 158 movement 154 scars and back-scarps 153– 154, 155 LiDAR (Light Detection And Ranging) data analysis 3, 81, 84 Turtle Mountain 164, 169 limestone, geomechanical parameters 210 lineament mapping 83, 84, 87–88, 92 lineaments 240 Gamanjunni 191 Sogn and Fjordane 99, 107 Turtle Mountain 167, 168, 175 listric fault 18–19, 46, 48, 49 lithology, slope instability 31–37, 41 Little Ice Age 122 Livigno, gravitational slope deformation 217, 218, 219 lost mountain 270, 271, 272, 273– 274, 276 Maiella Massif 202, 203, 204, 206, 207, 208, 212 Malser Haide megafan 254, 261, 263, 264– 273, 274 Mannen rockslide 36, 49, 50, 58, 60 mass movement see deep-seated . . . mechanical processes and slope failure 20, 21 megafan map, Alps 255 metamorphic rocks 28 mine, observations in 238, 239, 245, 246 model, extension/compression 3 modelling, gravitational slope deformation ˚ knes 147, 150 A Italian Alps 113, 219– 223, 230

281

La Clapie`re 15–19, 21–23 slope geometry 6, 20–21, 219– 223 Mohr–Coulomb parameters 209, 222, 247 monitoring mass movement 93 ˚ knes rockslide 147, 160 A Turtle Mountain 163 Møre og Romsdal County 28, 41, 71, 109 Morrone Ridge 202, 206, 207, 208, 212 Motto d’Arbino, slope deformation 217, 218– 219 movement rate, slope deformation 122, 125, 231 ˚ knes rockslide 150, 154, 159 A Celentino 235 Gamanjunni slide 185, 187, 188, 189, 194– 198 movement, active 99, 102, 104, 109 movement, potential 100–101 Mt Arezzo landslide 201 Mt Macellaro Fault Zone 208, 212 nappes, pre-Alpine 1 neotectonics 1, 81, 89, 90–92 Aosta 124, 127 Nicaragua, palaeostress analysis 133– 144 Norddal village, rockfall 45 Norddalsfjorden 81 Norway rockslide study 79–93 slope instability 27–75 structure and geomorphology 97– 109 see also Gamanjunni slide numerical modelling 216, 235 2D finite element 243 –248 3D FLAC 219–223 back-analysis, Caramanico Valley 208–212 results, Caramanico Valley 212–213 olistostrome 5 opaline fill 139 open cracks 67, 118, 152, 159 orthophotography and InSAR 187, 188, 190, 193, 194 palaeo-landslides 204 palaeostress analysis, Holocene rockslide 133– 144 history 136, 138–139 methods 135–136 stress field 139–142, 143 tectonic features 139 paraglacial deformation, SW Alps 11– 23 La Clapie`re 14–20 Tine´e Valley 12–19 paragneiss 12 Peio Valley, LiDAR 237 periglacial see paraglacial permafrost 109 planar failure 173 planning issues 194 Plawenn Valley 270, 271, 273 Plawenn village, destructive debris flow 275 Plawennspitz ‘lost mountain’ 270, 271, 272, 273, 276 Plio-Quaternary slope evolution 201–213 power plant (hydroelectric), damage to 121–124, 125 proxy volume of fans 255– 257, 258 Quart, gravitational slope deformation 115 –116, 123– 125, 126, 127–129

282

INDEX

ranking matrix, rockslide development 86 rebound 109, 236, 249, 276 ridge evolution, Celentino 235, 236, 238, 243– 247 rock avalanche 28, 41, 163 rock avalanche deposits 254 Maiella 203, 208 rock bridge 5, 105, 158 rock mass characteristics 170– 173 rock mass properties 222 rock mass strength 248 rock slope failure, InSAR study 185– 198 rock slope instability, Norway 27– 75 geology 28– 41 field observations 31–37 reactivated structures 27, 41–70 rockfall 20, 27, 28, 41–45, 47, 60 field observations 31–37 rockslide 51, 59, 64 active 74, 147, 188 development 68–70, 71, 73 potential 91, 92, 187–188 structural ranking 79– 93 rockslide, La Clapie`re 14, 16, 17, 19 Romsdalen Valley 28, 39, 43, 60, 107, 109 back-crack 45, 46, 50 sackung 12, 92, 215, 235, 243, 248 sag 235, 271, 275 Santa Lucia volcanic complex geology 133, 134–135 palaeostress analysis 135– 144 scars and back-scarps 54 ˚ knes rockslide 153–154, 155, 156, 157 A Gamanjunni slide 190, 191– 196 Schmidt hammer test 240 sediment fans in glacial troughs 253–254 shear failure 246, 248 shear moduli 211, 212, 223 shear plane 47, 49, 65, 128 shear zone, Celentino 238 shear zones and slope stability 114 –119, 124, 125, 127 sinkhole 121, 191, 193, 194 slickenside record 139 slickenside, Turtle Mountain 175 sliding block 193 sliding plane 19, 50, 57, 60, 66, 73, 74 breccia 89–90 field observation 31–37 Storfjord 45, 47– 49, 61 slope deformation see deep-seated . . . slope evolution chronology 19 slope failure, historical 28– 41, 51, 68, 73, 98, 109 slope geometry modelling 2D, La Clapie`re 15–19, 21– 23 3D 6, 219– 223 4D 20– 21 slope instability, Turtle Mountain 175– 179 slope profile and instability 107 numerical model 220– 222, 223, 224–229 slope tectonics definitions 1– 2 mechanism 5 –6 structures 3 –5 summary 6– 7

Sogn and Fjordane County, slope failure 28, 38, 41, 51, 68, 73 analysis 97–109 database 99 sole detachment, Caledonian structures 50–58 source reconstruction for megafan 260– 264, 268–274, 276 spring 51, 52, 54, 58 St Valentin fans 254, 256, 261, 262, 273– 274 Storfjord 28, 40, 41–50, 58– 68, 70–72 field observations 32– 37 Storfjorden, rockslide and structural ranking study 79– 93 Strandanipa unstable area 105, 106 strength index 166 Celentino 240, 242, 243, 247, 248, 249 strength of rock mass, Turtle Mountain 176 –179, 181 strength reduction and weathering 20 stress and slope instability 215– 217, 225, 227, 228–230 stress field 139, 140– 141, 143 stress ratio 247 stress, Plio-Quaternary 208, 209, 212 stress– strain analysis 202, 213 structural control and rockslide 79–93 development 68–70, 82– 89 ranking 89– 90, 92 structural criterion and rock slides 86, 88 Sunndalen Valley 28, 39, 51, 56–58, 107 Sunnylyvsfjorden, potential rockslide 81, 90–91, 93 surficial deformation, Aosta Valley 125–129 Switzerland, morphostructures 217, 218–219 Tafjorden, potential rockslide 81, 93 tensile failure 244, 246 terminal moraine, reinterpretation 254 thrust fault, modelling 19 time-dependent factors in slope deformation 114, 127, 129 time-dependent model 20, 21 Tine´e Valley, field investigation 13, 12–19 toe zone compression 196, 217 toe zone movement 159, 160 topographical analysis 30, 165 topography and active faults 196, 197 topography and slope instability 217 toppling 3, 4, 6, 14, 31, 73, 104 Alps 218, 219, 225, 229 Turtle Mountain 173, 174, 176, 177, 179, 180 Trafjord rockslide 90– 91 trench 122, 135, 206, 213 trench, ridge-top Celentino 235, 236, 238, 240, 248 Troms County, rockslides 186, 187, 188– 198 troughs and faults 15, 19 tsunami 97, 105, 147 tunnel damage 119, 123, 124 Turtle Mountain, fractures and slope failure 163– 181 geology and morphology 163–165 structures 166–174 Tussen instability 51, 53, 99, 102, 107 ultramafics 31, 41, 42, 70 uniformitarian v catastrophist 253, 276 uplift and instability 1, 81

INDEX uplift, Apennines 201– 202, 204, 212, 213 rate of 207, 208 uplift, fjords 30 uplift, post-glacial 109 U-shaped valley 28, 39 Val Venosta fans 254, 255– 264, 265, 274, 276 Valentinkopf ‘lost mountain’ 271, 272, 273, 274, 276 valley shape and deformation 215–231 numerical models 220–222, 224–229 Vallunga Curon, slope deformation 219 Viddalen instability 32, 51, 52, 74, 102– 104, 107 Vik rockslide 31 Villeneuve, slope deformation 115 –116, 119–122 Vivana Valley 270, 271, 272 volcanic terrain, slope failure 133 Vollan 31, 56– 58 Vollein necropolis 125

volume of fans 254–256, 261 Malser Haide 265, 268, 273, 274, 275, 276 volume of instabilities and slides 70, 72, 73 field observation 31– 37, 104 Gamanjunni slide 194 Santa Lucia volcanic complex 133, 136 Sogn and Fjordane 99, 105 Storfjord 66, 71 Storfjorden 87, 90–93 Turtle Mountain 173 volume of rock avalanche 204, 206 Arbino 219 water pressure 51, 74 weathering 41, 70 and strength reduction 20 weathering, limestone 166 Western Gneiss Region 28, 79, 81

283