Submarine Landslides of the Upper East Australian Continental Margin

Home , Submarine landslide

Copyright and use of this thesis This thesis must be used in accordance with the provisions of the Copyright Act 1968. Reproduction of material protected by copyright may be an infringement of copyright and copyright owners may be entitled to take legal action against persons who infringe their copyright. Section 51 (2) of the Copyright Act permits an authorized officer of a university library or archives to provide a copy (by communication or otherwise) of an unpublished thesis kept in the library or archives, to a person who satisfies the authorized officer that he or she requires the reproduction for the purposes of research or study. The Copyright Act grants the creator of a work a number of moral rights, specifically the right of attribution, the right against false attribution and the right of integrity. You may infringe the author’s moral rights if you: - fail to acknowledge the author of this thesis if you quote sections from the work - attribute this thesis to another author - subject this thesis to derogatory treatment which may prejudice the author’s reputation For further information contact the University’s Director of Copyright Services sydney.edu.au/copyright Submarine landslides of the eastern Australian upper continental margin

Samantha Clarke

A thesis submitted for the degree of Doctor of Philosophy

at The University of Sydney in March 2014

Marine Geology Declaration by author

This thesis is composed of my original work, and contains no material previously published or written by another person except where due reference has been made in the text. I have clearly stated the contributions by others to jointly-authored works that I have included in my thesis.

I have clearly stated the contribution of others to my thesis as a whole, including data analysis, significant technical procedures, professional editorial advice, and any other original research work used or reported in my thesis. The content of my thesis is the result of work I have carried out since the commencement of my research higher degree candidature and does not include a substantial part of work that has been submitted to qualify for the award of any other degree or diploma in any university or other tertiary institution. I have clearly stated which parts of my thesis, if any, have been submitted to qualify for another award.

I acknowledge that an electronic copy of my thesis must be lodged with the University Library and, subject to the General Award Rules of The University of Sydney, immediately made available for research and study in accordance with the Copyright Act 1968.

I acknowledge that copyright of all material contained in my thesis resides with the copyright holder(s) of that material.

Supervisors

Associate Professor Thomas Hubble Professor David Airey Associate Professor Jody Webster

I Statement of Contributions to Jointly Au- thored Works Contained in the Thesis

Published

All works listed have been published in peer reviewed conference journals.

Clarke, S., T. Hubble, D. Airey, P. Yu, R. Boyd, J. Keene, N. Exon, J. Gardner, and S. Ward (2014), Morphology of Australia’s eastern continental slope and related tsunami hazard. Submarine Mass Movements and Their Consequences (6th), edited by S. Krastel, J.H Behrmann, D. Volker, M. Stripp, C. Berndt, R. Urgeles, J. Chaytor, K. Huhn, M. Strasser, and C.B. Harbitz, pp. 529-538, Springer Netherlands. • TH- assisted with data analysis, field work, and edited drafts of manuscript • DA - assisted with data analysis, and edited drafts of manuscript • PY - assisted with field work and data collection • RB - provided data, assisted with field work, and edited drafts of manuscript • JK - assisted with data analysis, field work, and edited drafts of manuscript • NE - assisted with field work and data collection • JG - assisted with field work and data collection • SW - assisted with data analysis, and edited drafts of manuscript • SC - completed all other work

Clarke, S., T. Hubble, D. Airey, P. Yu, R. Boyd, J. Keene, N. Exon, and J. Gardner (2012), Submarine Landslides on the Upper Southeast Australian Passive Continental Margin – Preliminary Findings. Submarine Mass Movements and Their Consequences (5th), edited by Y. Yamada, K. Kawamura, K. Ikehara, Y. Ogawa, R. Urgeles, D. Mosher, J. Chaytor and M. Strasser, pp. 55-66, Springer Nether- lands. • TH - assisted with data analysis, field work, and edited drafts of manuscript • DA - assisted with laboratory work, data analysis, and edited drafts of manuscript • PY - assisted with laboratory work and data collection • RB - provided data, and assisted with field work and data collection • JK - assisted with data analysis and data collection • NE - assisted with data analysis, and field work and data collection

II • JG - assisted with data analysis, and field work and data collection • SC - completed all other work

Clarke, S., D. Airey, P. Yu, and T. Hubble (2011), Submarine landslides on the south-eastern Austral- ian margin, in AGS Coastal and Marine Geotechnics - Foundations for Trade, edited by Australian Geomechanics Society. • DA - assisted with laboratory work, data analysis, and edited drafts of manuscript • PY - assisted with laboratory work and data collection • TH - assisted with field work, data analysis, and edited drafts of manuscript • SC - completed all other work

Statement of Contributions by Others to the Thesis as a Whole

Associate Professor Thomas Hubble - manuscript review Professor David Airey - manuscript review

Statement of Parts of the Thesis Submitted to Qualify for the Award of Another Degree

None.

III Published Works by the Author Incorporated into the Thesis

Published

Reproduced with permission of Springer Netherlands

• Incorporated in Chapter 3

Reproduced with permission of Australian Geomechanics Society

IV Published Articles Arising from this Thesis

Refereed

In Preparation

Clarke, S., T. Hubble, J.Webster, D. Airey, E. de Carli, C. Ferraz, P. Reimer, R. Boyd, and J. Keene (in prep) Sedimentology, Structure and Probable Age of Five Continental Slope Submarine Landslides, Nerrang Plateau, Eastern Australia.

Clarke, S., D. Airey, T. Hubble (in prep), Geotechnical properties of submarine sediments from submarine landslides on the eastern Australian continental margin and implications for slide initiation.

Clarke, S., T. Hubble, D. Airey, and S. Ward (in prep), Tsunami hazard assessment of Australia’s eastern continental slope.

Clarke, S., T. Hubble, D. Airey, P. Yu, R. Boyd, J. Keene, N. Exon, and J. Gardner (in prep) Subma- rine landslide geomorphology and quantitative slide characterization on the eastern Australian upper continental slope.

V Additional Published Works by the Author Relevant to the Thesis but not Forming Part of it

Fletcher, M., T. Hubble, S. Clarke, D. Airey, P. Yu (2014). Submarine Landslide Morphology of Box Slides Present on the Continental Slope Offshore Fraser Island, Queensland, Australia. Proceedings of 2014 Spring Meeting, EGU, Vienna, Austria. Fletcher, M., T. Hubble, S. Clarke, D. Airey, J. Webster, P. Yu, E. De Carli (2013). Submarine Landslide Morphology of Box Slides Present on the Continental Slope Offshore Fraser Island, Queensland, Australia. Proceedings of Geo- sciences 2013 Conference, Christchurch, NZ. Hubble T., Clarke S., Yu, P., Volker, D., (2013) Probable landslide detachment surfaces and post-slide deposits sampled in the Yamba Slide Complex, New South Wales, Australia. Proceedings of International Symposium on Submarine Mass Movements and Their Consequences, Kiel, Germany. Hubble, T., P. Yu, D. Airey, S. Clarke, R. Boyd, J. Keene, N. Exon, and J. Gardner (2012), Physical Properties and Age of Continental Slope Sediments Dredged from the Eastern Australian Continental Margin – Implications for Timing of Slope Failure. Submarine Mass Movements and Their Consequences, edited by Y. Yamada, K. Kawamura, K. Ikehara, Y. Ogawa, R. Urgeles, D. Mosher, J. Chaytor and M. Strasser, pp. 43-54, Springer Netherlands. Hubble T, Yu P, Clarke S, Airey D, et al. (2012) A conceptual model for the onset and occurrence of submarine landslides on the southeastern Australian continental margin, Proceedings of 34th International Geological Congress, Brisbane, Australia. Hubble T, Yu P, Clarke S, Airey D, et al. (2012) A conceptual model for the onset and occurrence of submarine landslides on the southeastern Australian continental margin, Proceedings of 2012 Ocean Science Meeting, Salt Lake City, Utah, USA. Hubble T, Yu P, Airey D, Clarke S, et al. (2010) Physical properties and age of mid-slope sediments dredged from the Eastern Australian Continental Margin and the implications for continental margin erosion processes, Proceedings of 2010 Fall Meeting AGU, San Francisco, USA. Rahman Talukder, A., R. Boyd, J. Keene, T. Hubble, S. Clarke, M. Kinsela, N. Exon, J. Gardner, and J. Felzenberg (2010), Submarine landslides and tsunami potential Off SE Australian Margin: results from the voyage SS2008/12, Pro- ceedings of EGU General Assembly 2010, Vienna, Austria. Volker, D., T. Hubble, S. Clarke, J. Webster, A. Puga-Bernabeu (2013) Mapping the South Queensland margin: collapses, depressions and general instability. Proceedings of International Symposium on Submarine Mass Movements and Their Consequences, Kiel, Germany. Yu, P., Chaproniere, G., Hubble, T., Airey, D., Clarke, S., Boyd, R., Keene, J., Exon, N., and Gardner, J. (2013) Timing of submarine canyon incision on the east Australian continental margin. Proceedings of International Symposium on Submarine Mass Movements and Their Consequences, Kiel, Germany.

VI Acknowledgements

The completion of this research would not have been possible without the assistance and support of a number of people to whom I am eternally grateful. First and foremost, I would like to thank my supervisors Tom Hubble, David Airey, and Jody Webster for their continued support, inspiration, and guidance throughout this process.

Associate Professor Thomas Hubble has provided unsurpassed supervision and support at every stage of this project. Tom’s ability to inspire, challenge, encourage, and entertain over the years has been phenomenal. Despite being run off his feet, Tom always finds the time to help me out, whether it be to give direction, help in the lab, or simply to tell a joke, chat, or get me to calm down during those, umm, rare moments that I started to stress out. I cannot thank you enough!

Professor David Airey (aka “the Engineering Guru”) has also been an enormous help to me throughout this project (especially with regards to all things engineering!) and I am hugely grateful for his expertise, guidance, patience, and insight, for which I owe him my sincere thanks.

Associate Professor Jody Webster has provided much appreciated advice and assistance (particularly in all things sedimentological!) during the project, especially when a fresh set of eyes was needed on many much mulled over problems.

Thanks to Dr Ron Boyd for being so very support during the whole process, and always offering great advice and assistance when required. Ron lead the Southern Surveyor cruise in 2008 that collected the data on which this project is based, so without him this work would not have been possible!

I am also especially grateful to Associate Professor Jock Keene for his guidance, support, and counsel throughout my candidature, all his positive words along the way, and the benefit of his many years of expertise on so many topics.

Thanks to Dave Mitchell, for his assistance and guidance in the collection of data onboard the Southern Surveyor, as well as for helping to solve many a problem that cropped up along the way. Dave’s expertise and enthusiasm is always invaluable and much appreciated. A special note also must go out to Dave’s wife Marie, who kept me and many others fed and watered so well during those trips

VII down on the Murray, especially during the writing up phase.

Thank you to Tom Savage, for not only his excellent assistance and instruction during sediment sample analysis, and also for his general amazing help is keeping everything ticking along in the sediment labs and always always offering a guiding hand when needed.

A huge thanks goes out to all those in the Engineering Soil Science labs – I don’t know what I would have done without you all! To Antonio Reyno, who helped me out in the early stages of my testing, for helping me with equipment, testing, and patient instruction on soil mechanics theory. To Guien Miao, my “engineering buddy”, who was always willing to help with my tests, or to simply make me laugh with our chats about Dr Who! To Shaiou Hue, Ken Chen, and Bhagaban Acharya for your assistance in running all those time consuming tests – I truly appreciate the help! And last but not least, to Ross Barker, who has provided me with an enormous amount of help over the years – Ross always generously allowed me to come and go in the soils lab as much as I needed, helping me with equipment, set-up, testing, and for providing so much ongoing patient instruction…he was never too busy to help me sort out a problem and I am especially grateful!

Thank you to Doug Bergersen (aka the Fledermaus go-to-guy) for his much-needed help in ne- gotiating the sometimes tricky but always much needed software, and for always making the time to help me make sense of confusing data!

Thank you to both Paula Reimer (Chrono Lab UK) and Quan Hua (ANSTO) for being so patient and generous with your time teaching me the ins and outs of radiocarbon dating and helping me get some many great dates for the project.

For all those who spent their time to assist me with lab work, I am eternally grateful. In particular I must mention Elyssa De Carli and Christina Ferraz for the amazing help in picking those pesky foram samples – I will not forget!! J

Thanks to Liz Abbey, Mike Kinsela and Dan Harris, who at various stages over the last few years have helped me to conquer the unconquerable – ArcGIS. Or at least they tried to…I’m not sure Arc is something that is ever conquered J

A special thanks to Rob Beaman for his knowledge and experience in helping me to sort through

VIII some very messy seismic data – your help was invaluable!

Thank you to my fellow postgrad peeps, both past and present: you have been a wonderful support network over the years and the source of many fun times. I will try and list you all but I apolo- gize if I forget anyone J: Matt de Paoli, Wai Wang, Liz Abbey, Kate Thornborough, Kellie Adlam, Mike Kinsela, Dan Harris, Marco Ferraz, Mark Daley, Tim Austen, Gus Gomez, Kayla Maloney, Elyssa De Carli, Rebecca Hamilton, Phyllis Yu, Melissa Fletcher, Bel Dechnik, Stephanie Duce, Anna Helfensdorfer, Sabin Zahirovic, Ana Gibbons, Grace Shepard, Kara Matthews, Luke Mondy, Logan Yeo, Aedon Talsma, Sarah Bembrick, Nathan Buterworth, Adrianna Rajkumar, Tim Chapman, Nick Barnett-Moore, Tegan Hall. You guys really did make these years much more enjoable and kept me sane (as much as humanly possible!). A special thanks goes to Elyssa De Carli and Rebecca Hamilton, who were an amazing help in many things, especially towards the end of this mammoth task J.

To my close friends Melanie de Leon, Hannah Power, Vicki Miller, and Allie Peery…thank you for always just being there for me, no matter what city, state, or country we happen to be living in J

My heartfelt thanks go to Amanda, whom I met at the beginning of my PhD and has been a source of much love and many laughs ever since. Amanda has always been there to keep me sane, make me laugh, allowed me to vent when it got hard, and also helped me to remember the big world out there J Thank you!

Lastly, to my family, thank you for all your love and support over the years. And a special massive huge thank you to my parents, Ian and Corralie, who have always been there for me, fed me, clothed me, housed me, and loved me. They introduced me to my love of learning from as young as I can remember and have always whole-heartedly supported me in all my pursuits and are a constant source of love and encouragement.

~Alea iacta est~ (Julius Caesar)

IX Abstract

Stable continental margins experience submarine landslides relatively frequently and some of the largest slides on record have been shed from these relatively passive terrains. Despite this fact and the obvious accompanying tsunami risk, slides from passive margins such as Australia are poorly understood when compared to other settings, such as the flanks of volcanic islands, active subduction-zone margins and submarine fans associated with large river deltas. This work presents an investigation into the submarine landslides occurring along eastern Australia’s upper continental margin, with a focus on investigating the causes, timing, and mechanisms responsible for these features. It has focused on analysing gravity core samples and interpreting of high-resolution multibeam and subbottom profiles. The age, morphology, composition, and origin of particular submarine landslides on the eastern Australian upper continental margin offshore New South Wales/Queensland has been described and, for the first time, the mechanical characteristics of sediments from the eastern Australian upper continental slope has been presented. The hazard posed by these submarine landslides has also been evaluated by investigating their potential to generate tsunamis along this margin.

Visually identified transition surfaces (boundary surfaces) are identified in five cores within, or adjacent to, five submarine landslides at depths of 0.8 to 2.2 meters below the present-day seafloor, which separate looser material from stiffer more compacted material. These boundary surfaces show distinct gaps in AMS 14C ages of at least 25 ka. Subbottom profiles across submarine landslides indicate that the youngest identifiable seismic reflectors located upslope of three slides terminate on and are truncated by slide rupture surfaces and that the studied slides are geologically recent. There is no obvious evidence in the subbottom profiles for a post-slide sediment layer draped over or otherwise burying slide ruptures or exposed slide detachment surfaces. This suggests that these submarine landslides are geologically recent, and that the boundary surfaces are either: a) erosional features that developed after the occurrence of the landslide in which case the boundary surface age provides a minimum age for landslide occurrence or b) detachment surfaces from which slabs of near-surface sediment were removed during landsliding in which case the age of the sediment above the boundary surface indicates approximately when landsliding occurred. While an earthquake triggering mechanism is favoured for the initiation of submarine landslides on the eastern Australian margin, this causal mechanism cannot be conclusively demonstrated.

Geomechanical test data are presented for 12 gravity cores taken at sites from the upper slope

X (<1200 m) of the east Australian continental margin in or adjacent to five submarine landslide features. The sediment from the study area presents as remarkably similar to sediment collected from other sections of the margin e.g. offshore Sydney, and the older compacted mud layers from lower in the stratigraphic section. The sediment is characterised by high shear strengths, low clay content and high void ratios, and brittle behaviour. It seems a major earthquake, or toe erosion could have contributed to initiate slides in this region. The identified boundary surfaces are interpreted to represent detachment surfaces or slide plane surfaces. The soil properties of the sediments show no evidence of weak clay layers, although they do contain significant but relatively small amounts of clay (<25%). Compres- sion testing indicates that the sediment above and below the boundary surface is apparently slightly overconsolidated. The friction angles of the sediments are in the range of 30o - 40o, so that conventional soil mechanics would suggest the slopes have high factors of safety. However, this is clearly not the case as slope failures are widespread. Slide surfaces have been identified by dating and it has been shown that these features are associated with relatively shallow slides (~5-10 m) and not the large slide features that are evident in the bathymetry. Triaxial tests have indicated a significant increase in the brittleness of the shear response with increasing vertical stress level (i.e. burial depth), and that below a depth of ~20 m the soil response will be compressive, leading to the build up of pore pressure when subjected to cyclic (earthquake) loading. This is thought to be significant in explaining why the slides have large thicknessess of 50 to 200 m. To date, no conclusive triggering mechanism has been identified for initiating submarine landslides in the region although the brittle nature of the sediments and grainsize distribution make the slopes susceptible to liquefaction during oscillatory shaking, favouring an initation mechanism related to seismic shaking.

Evidence of submarine landsliding is presented for the east Australian continental margin between Jarvis Bay and Fraser Island. Thirty-six submarine landslide along the eastern Australian continental margin are identified that had the potential to produce a tsunami flow depth >5 m at the coastline. Flow depths at the coast range from 3 to 38 m for blocks moving downslope with a landslide velocity of 20 ms-1. Thin (<100 m) and narrow (<5 km) landslides produce smaller tsunami, with coastal flow depths of <5 m, and thick (>100 m) and/or wide (>5 km) landslides generate coastal flow depths of 5-10+ m. The combination of both thick and wide landslides had the greatest potential to generate the largest coastal flow depths >10 m. Maximum inundation distances and run-up heights of 1.6 km and 22 m respectively are calculated for landslide velocities of 20 ms-1, but these values vary significantly depending on local coastal topography. The reoccurrence of submarine landslides with similar characteristics to those shed from the margin in the geologically recent past would therefore be expected to generate tsunami with maximum flow depths between five and twenty meters at the coast-

XI line, run-up of up to 20 m and inundation distances of up to 1.5 km. The number of tsunamogenic submarine landslides identified increases northward of Coffs Harbour, with the number and size of both past tsunamogenic submarine landslides being much greater offshore northern New South Wales and southern Queensland than in central and southern New South Wales. Tsunamogenic submarine landslide scars reported for central and southern New South Wales are not big enough to have shed a block able to generate a megatsunami similar to paleo-megatsunamis hypothesised for this section of margin.

The widespread occurrence of upper slope slides across the eastern Australian margin indicates that submarine sliding should be considered to be a common characteristic of this passive continental margin. Engineering properties (friction angle, cohesion, and unit weight) imply that the sediment forming the margin is reasonably strong and inherently stable and classical limit-equilibrium modeling indicates that submarine landslides should not be a common occurrence on the margin. This indicates that pre-conditioning trigger, or some other mechanism is required to destabilise the slope and enable failure. The most likely suspected processes include: 1) dramatic reduction of the shear strength of the sediments to extremely low values, possibly induced by creep or a build-up of pore-pressure; 2) long- term modification of the slope-geometry i.e., sedimentation on the head of the slope and/or erosion of the toe of the slope; and/or 3) seismic events large enough to trigger sediment liquefaction or a sudden increase of pore-fluid pressure.

Keywords

mass movement • southeastern Australia • multibeam • seafloor geomorphology • continental slope • passive margin • upper slope • sedimentation rates • tsunami hazard • geomechanics

Australian and New Zealand Standard Research Classifications (AN- ZSRC)

040305 – Marine Geoscience

XII Contents

1. Introduction 1.1 Scientific rational and background . 1-1 1.1.1 Definitions . 1-7 1.2 Aims and objectives ...... 1-8 1.3 Study Area...... 1-9 1.4 Regional setting...... 1-10 1.5 Continental slope morphology and submarine landslides ...... 1-11 1.5.1 Canyon morphology . 1-13 1.5.2 Plateau morphology...... 1-14 1.5.3 Submarine landslide style ...... 1-15 1.5.4 Submarine canyons. 1-15 1.6 Thesis outline ...... 1-17 2. Sedimentology, Structure and Age Estimate of Five Continental Slope Submarine Landslides, Eastern Australia 2.1 Introduction...... 2-3 2.2 Study Area Location . 2-4 2.3 Geologic Setting and Margin Structure ...... 2-9 2.3.1 Earthquakes and seismicity in Australia ...... 2-11 2.4 Data and Methods...... 2-13 2.4.1 Bathymetry ...... 2-13 2.4.2 Sub-bottom Profiles...... 2-14 2.4.3 Core Collection and Sediment Properties...... 2-14 2.4.4 14C Radiocarbon Dating...... 2-15 2.4.4.1 Planktonic foraminifera assemblage samples...... 2-16 2.5 Results . 2-17 2.5.1 Bathymetry and Sub-bottom Profiles. 2-17 2.5.1.1 Bribie Bowl Slide and GC1, GC2, GC3...... 2-18 2.5.1.2 Coolangatta-2 Slide and GC9; Coolangatta-1 Slide and GC8 . . . . 2-18 2.5.1.3 Cudgen Slide and GC5, GC6, GC7, and GC11...... 2-20 2.5.1.4 Byron Slide and GC12...... 2-24 2.5.2 Core Descriptions and Sedimentology (Physical Properties)...... 2-24 2.5.2.1 Lithological units...... 2-25 2.5.2.2 Carbonate Content (TIC) and Total Organic Carbon (TOC) . . . . 2-28 2.5.2.3 Grainsize...... 2-28

XIII 2.5.2.4 Density Change. 2-29 2.5.3 Radiocarbon Ages – Boundary Surfaces and Sedimentation Rates. . . . . 2-30 2.5.3.1 Age Discontinuities – Dating the Boundary Surface...... 2-30 2.5.3.2 Sedimentation Rate . 2-33 2.5.3.3 Comparison of bulk sample ages with picked foraminifera ages. 2-33 2.6 Discussion...... 2-34 2.6.1 Sedimentation Style and Rates...... 2-34 2.6.1.1 Limitations...... 2-36 2.6.2 Significance of Boundary Surfaces...... 2-36 2.6.3 Possible Triggers...... 2-38 2.7 Conclusions...... 2-43 3. Geotech 3.1 Introduction...... 3-2 3.1.1 Previous work...... 3-3 3.2 Determining Stress History ...... 3-4 3.2.1 Compression History ...... 3-4

3.2.2 Preconsolidation pressure (σpc’) . 3-5 3.2.3 Depth of sediment burial. 3-5

3.2.4 Rate of Consolidation (Cv) . 3-6 3.3 Study Area Location...... 3-6 3.4 Geologic Setting and Margin Structure ...... 3-10 3.5 Sampling and Testing Program . 3-11 3.6 Methods / Specimen Preparation. 3-11 3.6.1 Sediment Properties...... 3-11 3.6.2 Atterberg Limits...... 3-12 3.6.3 Oedometer Testing. 3-12 3.6.4 Triaxial Testing . 3-13 3.6.5 Slope Stability Modeling . 3-14 3.7 Results and Interpretation...... 3-15 3.6.1 Sediment Properties...... 3-15 3.6.1.1 Classification data - physical properties...... 3-15 3.6.2 Geotechnical Characterisation...... 3-17 3.6.2.1 Oedometer Tests...... 3-17 3.6.2.2 Triaxial Tests...... 3-22 3.6.3 Slope Stability Modeling . 3-27 3.8 Discussion...... 3-30

XIV 3.8.1 Uniformity of composition and geomechanical behaviour...... 3-30 3.8.2 The nature of the boundary surfaces – are they slide plane surfaces?. . . . 3-31 3.8.3 Landslide Initiation ...... 3-33 3.8.4 Static Liquefaction...... 3-35 3.8.5 “Missing” sediment ...... 3-36 3.8.6 Issues and limitations ...... 3-36 3.8.6.1 Core length . 3-36 3.8.6.2 Sample disturbance...... 3-37 3.9 Conclusions...... 3-37 4. Eastern Australia’s submarine landslides: implications for tsunami hazard between Jervis Bay and Fraser Island 4.1 Introduction and Aims . 4-3 4.2 Study Area...... 4-4 4.3 General Method Overview . 4-6 4.3.1 Identification and Selection of Potenially Tsunamogenic Submarine Landslides. 4-6 4.3.3 Bathymetry and Landslide Geometry - Calculation of Slide Characteristics. . 4-8 4.3.4 Coastal Topography (adjacent to slide sites)...... 4-9 4.3.5 Tsunami Calculations ...... 4-9 4.4 Results...... 4-11 4.4.1 Calculation of Required Slide Block Sizes...... 4-11 4.4.2 Morphometric Characteristics of Slide Regions ...... 4-19 4.4.2.1 Region 1...... 4-19 4.4.2.2 Region 2...... 4-20 4.4.2.3 Region 3...... 4-21 4.4.2.4 Region 4...... 4-22 4.4.3 Characteristics of Onshore Tsunami Surge Generated by Identified Slides. . 4-24 4.4.3.1 Region 1...... 4-25 4.4.3.2 Region 2...... 4-26 4.4.3.3 Region 3...... 4-26 4.4.3.4 Region 4...... 4-28 4.5 Discussion...... 4-35 4.5.1 Comparison to related prior work...... 4-35 4.5.2 Tsunamogenic submarine landslide frequency estimates...... 4-36 4.5.3 Tsunami record of eastern Australia and implications for the megatsunami hypothesis . 4-36 4.6 Conclusions...... 4-41

XV 5. Conclusions 5.1 Outline. 5-1 5.2 Specific aims and objectives...... 5-1 5.3 Summary of main conclusions...... 5-2 5.3.1 Important results – Chapter 2...... 5-3 5.3.2 Important results – Chapter 3...... 5-4 5.3.3 Important results – Chapter 4...... 5-5 5.4 Significance ...... 5-5 5.5 Future research . 5-6 6. Appendix 1 6.1 Paper 1. A1-2 6.2 Paper 2. A1-14 6.3 Paper 3. A1-24

XVI List of Figures

Figure 1.1 Types of submarine mass failures. Taken from: a) Stow and Mayall (2000); b) Lee et al. (2009); and c) Prior and Coleman (1984)...... 1-2 Figure 1.2 a) Classification of submarine slides and flows (from Prior and Coleman 1984) b) Two main geometric forms of submarine landslides: rotational slab slides and translation slab slides (from Highland and Johnson 2004) . 1-3 Figure 1.3 Features of submarine landslides (from Highland and Johnson 2004). 1-3 Figure 1.4 Geological settings of submarine landslide occurrence (from Morgan et al. 2009) . 1-4 Figure 1.5 a) Factors that may combine to initiate submarine sediment and slope instability (after Coleman and Prior 1998); b) factors that may contribute to the initiation of submarine landslides and example of where these types of submarine landslides have occurred (from Masson et al. 2006). 1-5 Figure 1.6 Summary of the main causes and consequences of submarine landslides (from Ca- merlenghi et al. 2007)...... 1-6 Figure 1.7 a) Google Earth image of the eastern Australian continental margin from the Bass Strait to the Great Barrier Reef; and b) Digital elevation model (DEM) of the northern NSW to southern QLD continental margin, showing the location of the study area, the five submarine landslide sites of interest, and the twelve gravity cores examined in the study...... 1-9 Figure 1.8 DEM of the east Australian continental margin looking landward showing the location of the study area, the five submarine landslide sites of interest, and the twelve gravity cores examined in the study . 1-12 Figure 1.9 DEM of the northern canyon region coloured by: a) water depth b) slope angle. . 1-13 Figure 1.10 DEM of the plateau region coloured by: a) water depth b) slope angle. . . . . 1-14 Figure 1.11 Example slide showing a distinct U-shaped trough in cross-section backed by an amphitheatre shaped crestal zone and the elongated longitudinal slide profile. . . . 1-16 Figure 2.1 Google Earth image of the eastern Australian continental margin from the Bass Strait to the Great Barrier Reef showing the location of the study area. . . . . 2-4 Figure 2.2 DEM of the northern NSW to southern QLD continental margin, showing the location of the study area, the five submarine landslide sites of interest, and the twelve gravity cores examined in the study.. 2-5 Figure 2.3 DEM of the east Australian continental margin looking landward showing the location of the study area, the five submarine landslide sites of interest, and the twelve gravity cores examined in the study.. 2-7 Figure 2.4 DEM of the five submarine landslides investigated in this work: a) Bribie Bowl Slide; b) Coolangatta-2 Slide and Coolangatta-1 Slide; c) Cudgen Slide; d) Byron Slide. Each slide is shown coloured by: i) water depth; and ii) slope angle. 2-8 Figure 2.5 Multi-channel seismic reflection line (GA206) located due east of Yamba showing the typical characteristics of the east Australian margin in NSW and southern QLD:

XVII a) uninterrupted and b) annotated. . 2-9 Figure 2.6 a) Locations and magnitudes of historic seismicity (M≥4) for Australia from 1841 to 2010, taken from Clark et al. (2012) ; b) Historic earthquakes in the Newcastle region and the epicenter location of the M5.6 1989 Newcastle earthquake. Modified from Clark et al. (2012). 2-12 Figure 2.7 a) Location of 6 subbottom profile lines taken across or adjacent to 3 submarine landslides. b) Sub-bottom profiling line L58a-WE across a slide feature, north of the Bribie Bowl Slide. c) Sub-bottom profiling line GC8-SN across the Coolangatta-1 Slide . . 2-19 Figure 2.8 a) Sub-bottom profiling line L79a-WE down the Cudgen Slide. b) Sub-bottom profiling line GC6-WE down the Cudgen Slide...... 2-21 Figure 2.9 a) Sub-bottom profiling line GC4 and 5-WE down the Cudgen Slide. b) Sub-bottom profiling line GC6and7-SN tie line across the Cudgen Slide...... 2-22 Figure 2.10 Core logs of canyon region gravity cores...... 2-26 Figure 2.11 Core logs of plateau region gravity cores...... 2-27 Figure 2.12 Close up images and interpretation of boundary surfaces.. 2-28 Figure 2.13 Grainsize results: a) Sand:Silt:Clay ratio diagram; b) Grainsize distribution plot.. . 2-29 Figure 2.14 Sedimentation rates determined from 14C ages.. 2-33 Figure 2.15 a) 14C age frequency distribution plot b) Eustatic sea level record from Waelbroeck et al. (2002)...... 2-39 Figure 2.16 The 500 year return period PGA (0.0 s RSA period) hazard map in a) 2D and b) 3D (from Burbridge et al. (2012))...... 2-42 Figure 3.1 Idealised void ratio-effective stress relations for compressible soil (after Mitchell 1993) ...... 3-4 Figure 3.2 Location map of the study area, showing the continental margin from Yamba (southern NSW) to Noosa Heads (southern QLD) . 3-7 Figure 3.3 DEM of the east Australian continental margin looking landward showing the location of the study area, the five submarine landslide sites of interest, and the twelve gravity cores examined in the study . 3-8 Figure 3.4 DEM of two slides showing typical slope profiles for a) canyon region slides and b) plateau region slides. 3-9 Figure 3.5 a) Typical core logs showing characterisation of sediment from plateau and canyon regions b) Core photos of two typical cores GC3 and GC11 c) Close up images of boundary surfaces. 3-16 Figure 3.6 Plasticity chart showing classification of sediment samples...... 3-17 Figure 3.7 Oedometer consolidation tests results showing void ratio (e) versus effective vertical stress (kPa) for representative samples: a) high vertical stress b) varying clay content (under high vertical stress) c) samples above and below boundary surfaces (low and high vertical stresses) d) remoulded samples ...... 3-19

XVIII Figure 3.8 Triaxial testing results: a) 1-D compression b) 1-D compression with two representative oedometer compression tests c) normalised deviator stress and pore pressure responses d) effective stress paths for saturated tests . 3-24 Figure 3.9 a) Slope stability models under static conditions of two representative slides showing Factory of Safety (FoS) values b) FoS values for two representative slides under seismic loading conditions showing the sensitivity of the FoS to the varying accelerations ...... 3-28 Figure 3.10 a) Zone of elevated pore pressures b) slope failure mechanism (after Puzrin et al. 2004) . 3-33 Figure 4.1 Location map of the study area, showing the continental margin from Jervis Bay (southern NSW) to Fraser Island (southern QLD). . 4-5 Figure 4.2 DEM of a section of the continental slope coloured by a) depth and b) slope angle showing five examples of investigated submarine landslide scars and examples of rejected submarine landslide scars ...... 4-7 Figure 4.3 a) DEM of Region 1 showing the location of investigated submarine landslides b) oblique perspective of Region 1 slope geometry and submarine landslides. . . . 4-13 Figure 4.4 a) DEM of Region 2 showing the location of investigated submarine landslides b) oblique perspective of Region 2 slope geometry and submarine landslides. . . . 4-14 Figure 4.5 a) DEM of Region 3 showing the location of investigated submarine landslides b) oblique perspective of Region 3 slope geometry and submarine landslides. . . . 4-15 Figure 4.6 a) DEM of Region 4 showing the location of investigated submarine landslides b) oblique perspective of Region 4 slope geometry and submarine landslides. . . . 4-16

Figure 4.7 Determining run-up (R(Xmax)) and inundation distance (Xmax) from coastal topography profiles using coastal elevation and distance inland from the coastline specific to each submainr landslide site ...... 4-18 Figure 4.8 a) Location map of a section of Region 4, showing the approximate location of the seismic line E-F and olistostromic block/slump mass reported by Hill (1992), offshore Double Island Point. b) Modified seismic line E-F reported by Hill (1992) c) Representative coastal topography profile adjacent to the olistostrome block and

maximum inundation distance (Xmax) / run-up (R(Xmax)) points . 4-29

Figure 4.9 a) The maximum expected flow depth at the coastlineF ( d(0)) for all submarine land- -1 slides investigated within the study area for maximum slide velocities (vs) 10-40 ms

b) The maximum inundation distance (Xmax) / run-up (R(Xmax)) points for all slides for slide velocities 20 ms-1 and 40 ms-1 ...... 4-30 Figure 4.10 a-f) Representative examples of coastal topography profiles for each region and

maximum inundation distance (Xmax) / run-up (R(Xmax)) points g) Example of the (sometimes) high local variation of coastal topography within a small area of the east

coast of Australia and how this can change Xmax and R(Xmax) values.. . . . 4-31, 4-32

XIX List of Tables

Table 2.1 Summary of gravity cores from the study area (location, depth, total recovery length, stratigraphy and target feature). 2-6 Table 2.2 Radiocarbon ages of 49 samples...... 2-31 - 2-32 Table 3.1 Summary of location, depth, total recovery length, stratigraphy, and target feature of the sediment cores retrieved from the study area...... 3-10 Table 3.2 Summary of the core sediment classification data ...... 3-15 Table 3.3 Plasticity index results showing classification of sediment samples...... 3-15 Table 3.4 Oedometer consolidation results . 3-18 Table 3.5 Triaxial results...... 3-23 Table 3.6 Numerical input parameters used for slope stability modeling with GEO-SLOPE/W. 3-27 Table 3.7 A summary of the FoS values from back analysis slope stability modeling from two representative slides (Cudgen Slide and Byron Slide) . 3-29 Table 3.8 Summary of available sediment classification data from the Sydney and Brisbane regions...... 3-30 Table 4.1 Summary of the four regions investigated...... 4-4 Table 4.2 Sensitivity analysis results of Equation 1 summarizing parameter combinations required to produce ~5 m flow depth at coastline in water depths 500-2500 m. . . 4-12 Table 4.3 Summary of the all the submarine landslides investigated within the study area including morphometric parameters for each submarine landslide listed...... 4-17

Table 4.4 Full list of the maximum expected flow depth at the coastline (Fd(0)) for all the submarine landslides investigated within the study area...... 4-33

Table 4.5 Full list of the inundation distance (Xmax) and run-up (R(Xmax)) values determined for each identified submarine landslide within the study area...... 4-34 Table 4.6 Summary of the minimum slide width values required for a range of water depths (500 – 2500 m) and slide thicknesses (50 m – 500 m) required for a submarine land-

slide to generate a maximum expected flow depth at the coastline (Fd(0)) of 20 m, 50 m, and 100 m at slide velocities of 20 ms-1 and 40 ms-1...... 4-39

XX List of Equations

Equation. 4.1 Maximum flow depth at the coastline . 4-10 Equation. 4.2 Propagation and beaching factor . 4-10 Equation. 4.3 Inundation depth over flat ground . 4-10 Equation. 4.4 Inundation distance allowing for arbitrary topography. 4-10

Function. 4.5 Run-up (R(Xmax)) and inundation distance (Xmax) ...... 4-10

XXI List of Symbols

Roman Symbols

Cc compression index

Cv coefficient of consolidation c cohesion z depth of burial q deviator stress r distance from the coastline to head of the landslide source

X distance inland from the coastline (X0 = 0 at coastline) g gravitational acceleration

A0 initially generated surface elevation

Xmax inundation distance

vs landslide mass velocity L length LL liquid limit n Manning’s coefficient

Fd(0) maximum flow depth at the coastline p’ mean effective stress PL plastic limit

Ip plasticity index P(r) propagation factor P at distance r from the source of the wave 14C radiocarbon age

Cr recompression index

Lref reference length (1 m)

R(Xmax) run-up

kh seismic lateral acceleration

kv seismic vertical acceleration

Gs specific gravity t thickness T(X) topographic elevation at location X

vt tsunami speed at the landslide e void ratio

h0 water depth at landslide centre of mass W width

XXII Greek Symbols

ɛa axial strain Ƴ’ effective unit weight

σ’ or σ’v effective vertical stress Ф’ friction angle

σ’pc preconsolidation pressure

Ƴsat saturated unit weight of the sample

Ƴwater unit weight of water

XXIII List of Abbreviations

AMS Accelerator Mass Spectrometry

ASTM American Society for Testing and Materials

CSL Critical State Line

DEM Digital Elevation Model

DEMs Digital Elevation Model - Smoothed

DSM Digital Surface Model

EAC East Australian Current

EA Eastern Australia

FoS Factor of Safety

GA Geoscience Australia

MTC Mass transport complexes

NEDF National Elevation Data Framework Portal

NSW New South Wales

OCR Overconsolidation Ratio

QLD Queensland

SRTM Shuttle Radar Topographic Mission

TIC Total Inorganic Carbon

TOC Total Organic Carbon

USCS Unified Soil Classification System

USGS United States Geological Survey

XXIV Clarke (2014) Chapter 1 Chapter 1

Introduction

This chapter provides an introduction to the thesis (scientific rationale, background information, thesis structure), and also introduces literature relevant to the following chapters.

1.1 Scientific rational and background

The commonality of turbidites in the sedimentary record demonstrates that submarine landslides have been common throughout geological history; particularly on seafloor slopes, areas where sedimentation rates are high, sediments are fine-grained, or seafloor rocks are weakened by fractures. Landsliding is an important process which controls the morphology of continental slopes and moves vast amounts of sediment away from the continental margin and onto the abyssal plain. The failures are thought to be triggered by a variety of stressors, including earthquakes loading, storms, waves, and sediments. Individual failures can involve the movement of volumes of material reported to be as large as 20,000 km3 over distances of more than 140km (Hampton et al. 1996).

Submarine mass failures encompasses a range of phenomena, including landslides, rock falls, debris flows, mudflows, soil and rock avalanches, deformational creep (Coleman and Prior 1988), and turbidity currents (see Fig. 1.1) (Gee et al. 2006; Masson et al. 2006). They mostly occur on the slopes of continental margins or volcanic islands where sediments are predominantly fine-grained (Masson et al. 2006). Submarine slides usually present in one of two geometric forms: slab slides (single or multiple), or rotational failures (Fig. 1.2). Slab slides are translational features with a basal shear plane oriented parallel to the slope surface upon which displacements of soft or partially lithified materials occur. Slab slides are commonly multiple event, retrogressive features where the displacement of an initial slab destabilises adjacent areas, causing the instability to migrate upslope. Rotational slides present a basal shear plane that has a curved or arcuate surface; this type of failure typically displaces discrete, intact blocks that tend to be progressively more deformed or dismembered towards the base of the feature (Coleman and Prior 1988). Figure 1.3 shows the features of a typical submarine landslide.

1-1 Clarke (2014) Chapter 1

Several major slides have been discovered and studied since the process of submarine mass failure was first recognized to be an important deep marine process in a variety of different geological settings (see Fig. 1.4). These include the Storegga Slide in Norway (Masson et al. 2006), the Brunei Slide in Borneo (Gee et al. 2007), the Goleta Slide in California (Greene et al. 2006), and slides in Angola (Gee et al. 2006), the Gulf of Mexico (Silva et al. 2004), the Hawaiian Is- lands (McMurtry et al. 2004) and Canary Islands (Masson et al. 2006). A large variety and combinations of causes, triggers, and controls have been proposed for the initiation of subma- Figure 1.1: Three representative diagrams showing types rine landslides (e.g. Masson et of submarine mass failures taken from: a) Stow and Mayall al. 2006; Locat and Lee 2002). (2000); b) Lee et al. (2009) using terminology recommended by Varnes (1958); and c) Prior and Coleman (1984) These include: earthquakes, storm wave loading, erosion and in particular slope over-steepening, rapid sedimentation leading to under-consolidation, the presence of weak layers, gas hydrate dissociation, sea-level changes, glaciations/isostatic uplift, volcanic activity, and mud-diapirism (see Fig. 1.5). It is also widely accepted that a combination of these factors is often required to initiate a landslide, especially where they occur on very gently inclined slopes (1° to 2°), often with the distinction being made between direct and indirect triggering mechanisms. That is, apparently stable slopes become unstable and fail due to a gradual change in state that builds to a critical condition triggered by a sudden or short-term event which releases the slide.

1-2 Clarke (2014) Chapter 1

Submarine mass failures generally require combinations of particular conditions for their occurrence and are the result of the interaction of a variety of factors such as geologic setting, mechanical properties, and environmental processes. As with subaerial failures, submarine failures usually occur due to the following changes: 1) loading of the head of the slope; 2) erosion of the toe of the slope; 3) steepening of the slope; and/or 4) variations in pore pressure conditions. In the case of a submarine setting, sedimentation supplies the load on the head Figure 1.2: a) Classification diagram for submarine slides and of the slope. It is possible that the flows (from Prior and Coleman 1984) b) Two main geometric forms of submarine landslides: rotational slab slides and trans- toe can be eroded by bottom cur- lation slab slides (from Highland and Johnson 2004)` rents, while steepening of the slope is achieved by regional uplift, either through active tectonics, basin subsidence, isostatic uplift due to deglaciation, or gradual subsidence of a rifted margin slope due to post-rifting cooling of the continental and oceanic lithosphere (c.f. Lutgens and Tarbuck 2012). However, these processes alone may not be enough to cause a failure. Further triggering mechanisms may be required to destabilise the system. The material properties of the sediments that constitute the slopes can be altered by a number of factors such that the slopes are more prone to failure. In Figure 1.3: Features of submarine landslides (from High- a fully saturated submarine situation, land and Johnson 2004) these triggering mechanisms are largely

1-3 Clarke (2014) Chapter 1 related to earthquake loading, pore water pressure changes, which acts as the opposing force to material cohesion, or sediment loading due to rapid deposition, as well as other processes which affect pore water pressure. The result of these effects is a loss of shear strength of a sediment layer or the liquefaction of a susceptible layer of material, subsequent destabilisation of the slope and failure.

Figure 1.4: Schematic showing the variety of geological settings where submarine landslides can occur (from Morgan et al. 2009)

Some of the interactions between environmental conditions and triggers are illustrated in the schematic diagram from Coleman and Prior 1988 (Fig. 1.5a). Some of the better-studied triggering mechanisms include seismic loading, sea level change, and toe erosion, biological processes (Coleman and Prior 1988), loading by waves, tides, glaciation and sedimentation, gas hydrate dissociation (Locat and Lee 2002; Greene et al. 2006), and volcanic activity (Masson et al. 2006). Sites where landslides have already occurred seem to be more likely to experience subsequent and further failures, with the failure scars acting as sediment traps on continental margin environments, and foci of volcanism for the slopes on volcanic islands (Masson et al. 2006).

In any case, failure occurs when the disturbing forces exceed the resisting forces. This is generally thought to require either a reduction in the strength of the slope sediments, an increase in the downslope stress, or a combination of both effects. A key parameter identified to be involved is pore water pressure. When pore water pressure is increased suddenly due to seismic loading, such as from dewatering as a result of tectonic compression, the slope stability suddenly and dramatically decreases (Hampton et al. 1996).

1-4 Clarke (2014) Chapter 1

Documented consequences of submarine landslides include damage to seabed infrastructure (communications cables and buried pipelines), subsidence of coastal land, and the generation of tsunamis (Masson et al. 2009). Figure 1.6 summarises and links the main causes of submarine landslides to their associated consequences. Ac- cepted examples of large submarine landslides that generated tsunamis include the famous 1929 earthquake triggered submarine slides of Grand Banks (Fine et al. 2005), the 1946 Scotch Cap Alaska slide (Fryer et al. 2004), the 1964 Seaward & Valdez Figure 1.5: Alaska slide (Haeussler et al. 2007), a) Factors that may combine to initiate submarine sediment and slope instability (after Coleman and the 1979 Nice Airport slide in France Prior 1998); b) factors that may contribute to the initiation (Assier-Rzadkiewiz et al. 2000), the of submarine landslides and example of where these types of submarine landslides have occurred (from Masson et al. 1998 Aitape Papua New Guinea 2006) slide (Tappin et al. 2001) and 2002 Stromboli volcanic island slide in Italy (Tinti 2005). Some of these submarine generated tsunami have resulted in significant casualties and property damage (e.g. Scotch Cap, Grand Banks, Aitape) but not at the same scale that large-earthquake generated tsunamis can produce, e.g. the devastating 2004 Sumatra Indian Ocean tsunami (Lay et al. 2004) and the March 2011 Japanese events. The causes and mechanisms of submarine landslide generated tsunami are not as well understood as the tsunami associated with large plate-boundary earthquakes. Consequently, submarine landslide induced tsunami present a significant but poorly quantified hazard to coastal and offshore development (Bardet et al. 2003; Watts 2004; Maretzki et al. 2007; Sue et al. 2011).

1-5 Clarke (2014) Chapter 1

Despite extensive literature on the nature and causes of submarine landslides, their dynamics and triggering processes are not well understood (Locat and Lee 2002; Bardet et al. 2003). One of the principal reasons for this is the limited data on the physical and mechanical properties of sediments from the slide plane, as these materials have not tradi- Figure 1.6: Summary of the main causes and consequences of submarine landslides (from Camerlenghi et al. 2007) tionally been collected in historical studies.

In the Australian context, only a few studies have investigated submarine landslides on the Aus- tralian continental slopes. The eastern Australian (EA) coast is potentially very vulnerable to tsunamis due to the high concentration of the population living (~85%) within 50 km of the coast and much of the critical infrastructure located close to the coast (Short and Woodroffe 2009), however, there has been little reason to suspect a local source for the generation of tsunami on the EA coastline. The identification of abundant, large, and apparently recent submarine landslide scars has changed this perception (Boyd et al. 2010; Clarke et al. 2012). While evidence of submarine landslides on the EA margin was first reported by Jenkins and Keene (1992), it was not until high resolution, multibeam bathymetric data became available (Glenn et al. 2008; Boyd et al. 2009) that the morphology of the slope has been recorded sufficient detail to determine the actual distribution of submarine slides on this continental margin and the extent and ubiquity of these slides could be established. These investigations of the EA continental slope morphology have established that a surprising number of submarine landslides are present on the continental slope such that submarine sliding should be considered to be a common and ongoing characteristic of this passive continental margin (Clarke et al. 2012; Hubble et al. 2012). Radiocarbon ages from the sediment indicate that they are also geologically young (< 25 ka) (Boyd et al. 2010; Clarke et al. 2012; Chapter 3). It is strongly suspected that several of these slide masses probably failed and moved down slope in the relatively recent geologic past.

1-6 Clarke (2014) Chapter 1

1.1.1 Definitions Submarine landslides are one of the main agents though which sediments are transferred across the continental slope to the deep ocean. They are ubiquitous features of submarine slope in all geological settings and at all water depths. Hazards related to such landslides range from destruction of offshore facilities to collapse of coastal facilities and the generation of tsunamis. Slope failure occurs when the downslope driving forces acting on the material composing the seafloor are greater that the forces acting to resist major deformations. Following slope failure, the failed mass moves downslope (mass movement) driven by gravity.

There is some complexity with regards to submarine landslide nomenclature and the frequent imprecise use of landslide terminology, especially in the submarine environment where information on landslide processes is often limited. Often the term ‘landslide’ is used as a generic term encompassing all forms of slope failure, irrespective of process (following terminology recommended by Masson et al. 2006):

Slide: movement of a coherent mass of sediment bounded by distinct failure planes. Debris flow: laminar, cohesive flow of clasts in a fine-grained matrix (e.g. wet concrete). Debris avalanche: rapid flow of cohesionless rock fragments with energy dissipation by grain contact. Turbidity current: gravity flow in which sediment grains are maintained in suspension by fluid turbulence.

When a submarine slope failure occurs, material is translated downslope above a basal shear surface which develops due to progressive shear failure. Once failure initiates, the event may progress by means of a number of mass movement processes. Although various subdivisions and classification schemes for these processes exist (e.g. Lee et al. 2009; Masson et al. 2006; Prior, D.B. 1984; Varnes 1958), each process represents part of a continuum, whereby one type may evolve into or trigger another. Many submarine slope failures are likely to have involved a number of processes, possibly active at different stages of failure. Therefore, it is common that the depositional units resulting from submarine mass movements are defined as ‘Mass-Transport Complexes (MTC)’.

A triggering mechanism, also called a short-term trigger, is an external stimulus that initiates the

1-7 Clarke (2014) Chapter 1 slope instability process instantaneously (direct mechanisms) (Sultan et al. 2004). Examples of triggering (or “external”) mechanisms include (from Locat and Lee 2002): oversteepening, seismic loading (earthquakes), storm-wave loading, rapid sediment accumulation and under-consolidation, gas charg- ing, gas hydrate dissociation, low tides, seepage, glacial loading, and volcanic island processes. Causal factors, also called long-term triggers, can contribute to instability but may not initiate failure and may take hundreds or thousands of years to influence slope instability (indirect mechanisms). These can include slope angle, mass-movement history, unloading. It is also important to note that more than one factor may contribute to a single landslide event (Masson et al. 2006).

1.2 Aims and objectives

Data collected on the RV Southern Surveyor research cruise (SS2008-V12) (Boyd et al. 2010) will be used to investigate a region of Australia’s eastern continental margin in order to gain a better understanding of the newly identified (Boyd et al. 2010) large and geologically-recent submarine landslides. This material presented will address fundamental questions about the causes, timing and mechanisms responsible for these features:

1. Why are submarine landslides occurring in this region? 2. When did submarine landslides occur in the past? 3. What caused submarine landslides to occur in the past and what could cause them to occur again in the future? 4. And lastly, if a submarine landslide was to occur now, similar to those which have occurred in the past, what impact could it have on the surrounding coastline?

Detailed characterization of the slope sediments provides essential information needed to explain the timing and occurrence of the large submarine landslides on the eastern Australian upper continental slope. The tsunamigenic potential of these submarine landslides is largely unknown, although it is recognised that a substantial number of the slides are volumetrically large enough and are located in water depths shallow enough (< 1500 m) to have generated substantial tsunamis that, if repeated in the future, could cause widespread damage on the SE Australian coast and threaten coastal communities.

Specifically, this study aims to: (1) define the age, morphology, composition and origin of par-

1-8 Clarke (2014) Chapter 1 ticular submarine landslides on the EA upper continental margin recently discovered offshore New South Wales/Queensland (NSW/QLD); (2) determine the geomechanical characteristics of sediments from the upper eastern Australian continental slope; (3) investigate when and why have slope failures occurred; and (4) improve the understanding of the hazard posed by these submarine landslides by investigating their potential to generate tsunamis along this margin.

1.3 Study Area

The study area is comprised of a segment of the eastern Australian continental margin located offshore Noosa Heads in southern Queensland (QLD) and Yamba in northern New South Wales (NSW), and is bounded by the 29°30’S and 26°20’S parallels of latitude and 153°40’E to 154°45’E meridians of longitude. It is situated approximately 30 km to 70 km seaward of the present coastline in water depths of 150-4500 m (Fig. 1.7).

Figure 1.7: a) Google Earth image of the eastern Australian continental margin from the Bass Strait to the Great Barrier Reef. Red box outlines the study area. Inset shows the location of the study area on a line map of Australia. b) Digital elevation model (DEM) of the northern NSW to southern QLD continental margin, showing the location of the study area. Insets show detailed views of the five submarine landslide sites of interest (black dashed lines) and the locations of the twelve collected gravity cores (GC) from the Southern Surveyor SS2008/12 voyage. Moving north to south: (i) Bribie Bowl Slide (ii) Coolangatta-2 Slide, (iii) Coolangatta-1 Slide (iv) Cudgen Slide, (v) Byron Slide

1-9 Clarke (2014) Chapter 1

Failure scars from five distinct submarine landslides identified by Boyd et al. (2010), along with twelve gravity cores from the upper continental slope (<1500 m) are the focus of this work. From north to south these slides are: 1) Bribie Bowl Slide (water depth 600 m); 2) Coolangatta-2 Slide (water depth 900 m); 3) Coolangatta-1 Slide (water depth 600 m); 4) Cudgen Slide (water depth 600 m); and 5) Byron Slide (water depth 800 m). These particular sites are the focus of this study due to the acquisition of twelve gravity cores within and adjacent to the five slides. At least one gravity core was recovered from each slide. These features are representative of the slope failures that occur in the two dominant slope morphologies present in the study area identified by Boyd et al. (2010) and Clarke et al. (2012, 2014) which are the: 1) relatively steep (3-7°) and canyon incised slope (Bribie Bowl Slide and Bryon Slide); and 2) relatively gentle slope (1-3°) of the Nerang Plateau (Coolangatta-1 & -2 slides and Cudgen Slide).

1.4 Regional setting

The EA continental margin stretches 1500 km north from Bass Strait to the Great Barrier Reef (Boyd et al. 2004). The margin, which is by world standards narrow, deep and sediment deficient, was formed by rifting in the Cretaceous period between 90M and 65M years ago (Gaina et al. 1998). Since then, margin subsidence is thought to have been relatively minor. In general, sedimentary deposits along the margin are no more than ~500m thick (Keene et al. 2008), due to a combination of factors such as low mainland sediment flux, limited accommodation space, reworking by the strong, northward acting longshore drift along the margin and southward moving sediment transport by the East Australian Current (EAC) (Keene et al., 2008). While longshore drift is generally assumed to have the greater effect on sediment transport of the shelf sediments (Short and Trenaman 1992; Keene et al. 2008), Boyd et al. 2004) has recently shown that the EAC has a more significant effect on upper and mid-slope margin sedimentation than previously thought.

After rifting ceased in ca. 65 Ma and Australia and Antartica separated at the end of the Eocene (Exon et al. 2004), the upper EA margin has experienced little subsidence and the post-rift sediment coverage is relatively thin (Boyd et al. 2004). The original shape of the basin, and the influence this has had on the tectonic formation of the margin, has largely controlled the modern margin physiog- raphy (Jones et al. 1975; Troedson and Davies 2001; Boyd et al. 2004), and when this is combined with the low mainland sediment input experienced during the previous low-stand, it helps explain the

1-10 Clarke (2014) Chapter 1 sediment-deficient nature of the region (Boyd et al., 2004).

The continental shelf ranges between 14-78 km wide and is essentially flat with a thin sediment cover. The continental margin sediment wedge reaches a peak thickness of about 500 m at the edge of the shelf, with the break in slope presented at depths ranging between 55 and 180 m. The continental slope is the region from the shelf edge to the Tasman Abyssal Plain where the water depth is around 4500 m. The width of the continental slope varies between 28-90 km wide and presents average slopes between 2.8° to 8.5°. The margin sediment wedge generally reduces from the shelf edge in thickness such that the lower slope is bare of sediment at about 3500 m below sea level and continental basement lithologies are exposed at the toe of the slope (e.g. Hubble et al. 1992; El Tanin Continental Margin Surveys).

The size, extent and apparent youth of the submarine landslides present on the EA continental slope is interesting when related to the tectonic setting and geomorphology of the margin. For this study many of the reported landslide triggers and/or contributing factors are apparently absent or min- imised relative to landslide locations elsewhere in the world where this phenomenon has been studied. In geological terms the EA margin is relatively inactive, and yet is yielding slides that are similar to those found elsewhere in the world despite the fact that many of the factors thought to be responsible for generating slides elsewhere in the world do not apply here. The EA continental slope presents relatively low seismicity, low sedimentation rates, and there is little or no evidence of gas-seepage and gas-hydrate decomposition. The Australian continent is considered to be relatively stable and there is no isostatic-rebound of the shelf and continental margin following ice-sheet retreat because there has been no ice-sheet loading. This is in stark contrast to the situations described for those slides on northern hemisphere continental margins that have been studied in detail (e.g. Atlantic margin; c.f. Twichell et al. 2009 and references therein). In all of these cases one or more of these behaviours (seismicity, rapid sedimentation, gas generation or isostatic rebound) is present and has been identified as causal or contributory to sliding.

1.5 Continental slope morphology and submarine landslides

Widespread erosional features occur on both the upper and lower continental slope across the entire section of the EA margin, with majority of the submarine mass failures occurring on the con-

1-11 Clarke (2014) Chapter 1 tinental slope. The extensive submarine landslides appear to have eaten away at the continental slope, giving the margin retrogressive character. Several large-scale sections of the slope appear to be incipient failures, suggesting landslides are still an active process of erosion the margin. Focusing on the upper slope (<1500 m), several distinct large sediment slides varying in volume from <0.5 km3 to 20 km3 can be identified. Within the study area, the continental slope alternates between regions of intense canyon incision (average slope ~6-7°, local maximums in excess of 35°), bisected by the relatively gentle sloping Nerang plateau (average slope ~2-3°) (see Fig. 1.8), with evidence of slope failure widespread in each morphological zone.

Figure 1.8: Digital elevation model (DEM) of the east Australian continental margin. Perspective looks west, landward onto the continental margin. Insets show details of five submarine landslides (black lines) and the locations of gravity cores within each slide. Moving north to south: a) Bribie Bowl Slide (northern canyon region): GC2 and GC3 are located in the slide; GC1 is located adjacent to the slide; b) Coolangatta-2 Slide and Coolangatta-1 Slide (plateau region): GC9 is located in the Coolangatta-2 Slide; GC8 is located in the Coolangatta-1 Slide; c) Cudgen Slide (plateau region): GC4, GC5, GC6, GC7, GC10, GC11 are located in the slide; GC7 is located adjacent to the slide; d) Byron Slide (southern canyon region): GC12 is located in the slide

1-12 Clarke (2014) Chapter 1

1.5.1 Canyon morphology The continental slope in the north and south of the study area is characteristic of the canyon region morphology experienced along the EA margin (Fig. 1.8). The continental slope in the canyon regions is steep and narrow, cut by several major steep-walled canyon features, and a large number of submarine slides and slumps (Fig. 1.9). On the whole, mass movement features that exhibit as a series of incipient, large-scale failures on both the upper and mid-lower slope, that that coalesce to form broad canyons. The slope appears to be progressively been “eaten back” towards the outer shelf by mass wasting processes. This is particularly evident when the surface of the slope is coloured by slope angle, which highlights the extent of submarine landslide scars (see Fig. 1.9b).

In the northern most section of the study area, the continental slope appears somewhat different to the majority of the canyon region. Here, the slope of the Noosa canyon system presents as a series of narrow, dendritic drainages incised on the upper and mid-lower slopes, separated by sharp ridges

Figure 1.9: Digital elevation model (DEM) of the northern canyon region coloured by: a) water depth; and b) slope angle.

1-13 Clarke (2014) Chapter 1 over a distance of 25 km, with an inter-ridge spacing of 500 – 1000 m (Fig. 1.9). This differs from the many large canyons and mass movement features that dominate the rest of the canyon region. Erosion and transport sandy sediment to the deep sea is thought to cause canyon ridges carved in this style (e.g. Boyd et al. 2004). 1.5.2 Plateau morphology The continental slope over the Nerang Plateau lacks the numerous steep-walled canyons seen in the northern and southern canyon regions, and is characterised by low slope gradients and a surface with the appearance of numerous and continued shedding of slope material (Fig. 1.10). The plateau has a wide (over 90 km wide in places), gently dipping slope (2-3°) to the southeast, with little apparent break in slope until you reach the abyssal plain. The surface is irregular with a general relief <20 m. While the plateau looks relatively smooth when compared to the canyon region, sliding is also widespread (see Fig. 1.10). The irregularity of the plateau surface is particularly evident when the slope is coloured by slope angle (see Fig. 1.10b).

Figure 1.10: Digital elevation model (DEM) of the plateau region coloured by: a) water depth; and b) slope angle. Pixelated segments of the image represent low resolution bathymetric coverage.

1-14 Clarke (2014) Chapter 1

The plateau may have once extended farther to the north and south into what is now the heavily canyoned region of the margin, but these areas have subsequently undergone massive failures and extensive canyon cutting. The style of erosion on the plateau is subtler when compared to the massive failures and extensive canyon cutting to the north and south. Despite this, extensive slide scarring does occur along its entire length. The slope is characterized by a “creepy” topography that appears to be caused by continuous incipient slumping downslope (Fig. 1.10b).

One explanation for the difference in erosion styles is that the underlying geology of the plateau may be different and much more resistant to erosion than the geology to the north and south. Alter- nately, it might simply be that the Nerang plateau section of the margin has yet to collapse.

1.5.3 Submarine landslide style The submarine landslides identified on the continental slope typically comprise a distinct U- shaped trough in cross-section (3-6 km wide and 20-250 m deep) backed by an amphitheatre shaped crestal zone (see Fig. 1.11). This slide morphology is similar to the classical circular failure profile described by Varnes (1978), with an elongated longitudinal profile. The sides and head walls of the scarps are relatively steep with slopes of up to 17°. Sub-bottom profiling data from multiple sections across the continental slope and in particular across the slides show the sediment is built up of well stratified beds, which have been suggested to be evidence of past slide events (see Chapter 2). In most locations, sediments derived from the slides cannot be detected on the slope and it appears that the slide material has been transported to the abyssal plain. However, in a small number of locations where the slopes are less steep (<2°), the slide debris flow deposits have remained on the slope and contain blocks up to 350 m wide and 50 m high. Based on current sediment accumulation rates between 0.017mka-1 and 0.2 mka-1, and with the estimation confirmed by radiocarbon dating (Clarke et al. 2012; Chapter 2), these failures are believed to be <25 ka.

1.5.4 Submarine canyons A large number of canyons also cut into the continental slope sediments (Fig. 1.7, 1.8). The head of the canyons vary in style from broad amphitheaters to relatively small, confined, canyon channels. The amphitheatre-style canyon head suggests a “gathering zone” of numerous upper-slope canyons that coalesce downslope into a broad canyon channel. The relatively confined upper-slope canyons are, for the most part, single canyon channels that lead downslope to narrow canyon channels. The

1-15 Clarke (2014) Chapter 1 canyon heads typically incise the upper-most slope but only rarely incise the outer-most shelf. Can- yon heads appear as large-scale failures of the tabular outcrops and reefs along the upper-most slope. Canyons have been categorized into large box canyons, and smaller narrow linear canyons (Boyd et al. 2010). The 46 large box canyons are on average 14 km wide, 20 km long and over 600 m deep, they stretch from the middle slope to the abyssal plain, and have slopes up to 34° on the walls, the steepest slopes found on this margin (Boyd et al., 2010). Narrow linear canyons occur in the upper slope sediments, most located in central NSW off major river systems such as the Tweed. Well-developed examples are 800-1900 m wide, 120-320 m deep and extend downslope for 14-22 km (see Chapter 2).

Figure 1.11: Example slide showing a distinct U-shaped trough in cross-section (3-6 km wide and 20-250 m deep) backed by an amphitheatre shaped crestal zone: a) slide morphology of the Byron Slide (black outline) and location of cross-section profiles though and across the slide (black) and reference profile adjacent to the slide (light blue); b) slope geometry and cross-section profiles across slide (black line S-N) and down slide (black line W-E); reference profile of the adjacent slope is shown in blue; c) DEM showing the slide morphology of the Byron Slide (black outline) and the elongated longitudinal profile of the slide

1-16 Clarke (2014) Chapter 1 1.6 Thesis outline

This thesis investigates submarine landslides along Australia’s eastern continental margin with a focus on the timing and mechanisms responsible for submarine landslides, and understanding of the related tsunami hazard. The thesis is built upon three discrete and self-contained, but closely related chapters. The chapters address the main aims and objectives of the study and build upon knowledge gained from each previous chapter. Each chapter is formatted as a scientific manuscript suitable for journal submission. As such, each chapter contains a small but unavoidable amount of repetition.

This chapter Chapter( 1) provides an introduction to the thesis (scientific rationale, background, thesis structure), and also introduces contextual information, a review of previous work and literature relevant to the following chapters. However, Chapters 2-5 each contains an introductory section with an additional discussion of the current literature.

Chapter 2 investigates the sediment characteristics and age of the submarine landslides. The role of this chapter was to examine the sedimentology and physical properties of sediment taken in or adjacent to five distinct submarine landslides, to obtain a detailed14 C age record of the sediments in order to investigate the age of the submarine landslides in the region, and to establish a sedimentation rate for this section of the margin. Chapter 3 investigates geotechnical characteristics of the submarine landslide sediments on the eastern Australian margin. The role of this chapter was to establish and quantify the geomechanical properties of the sediment from the margin and to model the currently stability of various sections of the continental slope based on slope morphology and geomechanical properties. Chapter 4 investigates the morphology of the Australian continental slope and examines the tsunami hazard related to submarine landslides for Australia’s eastern continental margin. It explores and estimates the coastal flow depth, run-up height, and inundation distance of tsunami waves generated by the submarine landslides that might occur along the margin.

Finally, Chapter 5 summarises the results from each chapter and places them in the context the thesis as a whole. The main conclusions are also highlighted and suggestions are given for future work.

1-17 Clarke (2014) Chapter 1 References

Assier-Rzadkiewicz, S., Heinrich, P., Sabatier, P. C., Savoye, B. & Bourillet, J. F., 2000. Numerical modelling of a landslide-generated tsunami: the 1979 Nice event. Pure and Applied Geophys- ics 157, 1717–1727. Bardet, J.P., Synolakis, C.E., Davies, H.L., Imamura, F., Okal, E.A., 2003. Landslide Tsunamis: Re- cent Findings and Research Directions Pure and Applied Geophysics 160, 1793–1809. Boyd, R., Ruming, K. and Roberts, J.J., 2004. Geomorphology and surficial sediments on the southeast Australian continental margin. Australian Journal of Earth Sciences 51, 743-764. Boyd, R., 2009. SS12/2008 Voyage Summary: Marine Geology and Geohazard Survey of the SE Australian Margin off Northern NSW and Southern Queensland. CSIRO. Boyd, R., Keene, J., Hubble, T., Gardner, J., Glenn, K., Ruming, K., Exon, N., 2010. Southeast Australia: A Cenozoic Continental Margin Dominated by Mass Transport, in: Mosher, D.C., Moscardelli, L., Baxter, C.D.P., Urgeles, R., Shipp, R.C., Chaytor, J.D., Lee, H.J. (Eds.), Sub- marine Mass Movements and Their Consequences. Springer Netherlands, pp. 491-502. Camerlenghi, A., Urgeles, R., Ercilla, G., Brückman, W., 2007. Scientific Ocean Drilling Behind the Assessment of Geo-Hazards from Submarine Slides, Scientific drilling, Hokkaido, Japan, pp. 45–47. Clarke, S., Hubble, T., Airey, D., Yu, P., Boyd, R., Keene, J., Exon, N., Gardner, J., 2012. Subma- rine Landslides on the Upper Southeast Australian Passive Continental Margin – Preliminary Findings, in: Submarine Mass Movements and Their Consequences. Yamada, Y., Kawamura, K., Ikehara, K., Ogawa, Y., Urgeles, R., Mosher, D., Chaytor, J., Strasser, M. (Eds.). Springer Netherlands, pp. 55-66. Coleman, J., Prior, D., 1988. Mass wasting on continental margins. Annual Review of Earth and Planetary Sciences 16, 101. Exon, N.F., Kennett, J.P., Malone, M.J., 2004. Leg 189 Synthesis: Cretaceous-Holocene History of the Tasmanian Gateway, in: Exon, N.F., Kennett, J.P., Malone, M.J. (Eds.), Proceedings of the Ocean Drilling Program, Scientific Results. Fine, I.V., Rabinovich, A. B., Bornhold, B. D., Thomson, R. E. & Kulikov, E. A. 2005. The Grand Banks landslide-generated tsunami of November 18, 1929: preliminary analysis and numerical modeling. Marine Geology 215, 45-57. Fryer, G.J., Watts, P., Pratson, L.F., 2004. Source of the great tsunami of 1 April 1946: a landslide in the upper Aleutian forearc. Marine Geology 203, 201-218. Gaina, C.M., Muller, R.D., Royer, J.-Y., Stock, J., Hardebeck, J., Symonds, P., 1998. The tectonic history of the Tasman Sea: a puzzle with 13 pieces. Journal of Geophysical Research 103, 12413–12433. Gee, M.J.R., Gawthorpe, R.L., Friedmann, S.J., 2006. Triggering and Evolution of a Giant Subma- rine Landslide, Offshore Angola, Revealed by 3D Seismic Stratigraphy and Geomoephology. Journal of Sedimentary Research 76, 9-19. Gee, M.J.R., Uy, H.S., Warren, J., Morley, C.K., Lambiase, J.J., 2007. The Brunei slide: a giant submarine landslide on the North West Borneo Margin revealed by 3D seismic data. Marine Geology 246, 9-23. Glenn, K., Post, A., Keene, J., Boyd, R., Fountain, L., Potter, A., Osuchowski, M., Dando, N., Party, S., 2008. NSW Continental Slope Survey – Post Cruise Report. Geoscience Australia, Can- berra. Greene, H.G., Murai, L.Y., Watts, P., Maher, N.A., Fisher, M.A., Paull, C.E., Eichhubl, P., 2006. Submarine landslides in the Santa Barbara Channel as potential tsunami sources Natural Haz- ards and Earth System Sciences 6, 63–88.

1-18 Clarke (2014) Chapter 1

Haeussler, P.J., Lee, H.J., Ryan, H.F., Labay, K., Kayen, R.E., Hampton, M.A., Suleimanip, E., 2007. Submarine Slope Failures Nears Seward, Alaska, During the M9.2 1964 Earthquake, in: Lykousis, V., Sakellariou, D., Locat, J. (Ed.), Submarine Mass Movements and Their Conse- quences, pp. 269–278. Hampton, M.A., Locat, J., Lee, H.J., 1996. Submarine landslides. Reviews of geophysics 34, 33–59. Highland, L., Johnson, M., 2004. Landslide Types and Processes. Department of the Interior: 4, USGS, USA. Hubble, T., 2013. Voyage Summary SS2013-V01: Marine Geology and Geohazard Survey of the SE Australian Margin off Northern NSW and Southern Queensland, CSIRO, Hobart. Hubble, T., Yu, P., Airery, D., Clarke, S., Boyd, R., Keene, J., Gardner, J., 2010. Physical properties and age of mid-slope sediments dredged from the Eastern Australian Continental Margin and the implications for continental margin erosion processes, 2010 Fall Meeting, AGU, San Francisco, Calif. Hubble, T., Yu, P., Airey, D., Clarke, S., Boyd, R., Keene, J., Exon, N., Gardner, J., 2012. Physi- cal Properties and Age of Continental Slope Sediments Dredged from the Eastern Australian Continental Margin – Implications for Timing of Slope Failure, in: Submarine Mass Move- ments and Their Consequences. Yamada, Y., Kawamura, K., Ikehara, K., Ogawa, Y., Urgeles, R., Mosher, D., Chaytor, J., Strasser, M. (Eds.). Springer Netherlands, pp. 43-54. Hubble, T.C.T., Packham, G.H., Hendry, D.A.F., McDougall, I., 1992. Granitic and monzonitic rocks dredged from the southeast Australian continental margin. Australian Journal of Earth Sciences 39, 619-630. Jenkins, C.J., Keene, J.B., 1992. Submarine slope failures on the southeast Australian continental slope. Deep Sea Research 39, 121-136. Jones, H., Davies, P., Marshall, J., 1975. Origin of the shelf break off southeast Australia. Journal of the Geological Society of Australia 22, 71-78. Keene, J., Baker, C., Tran, M., Potter, A., 2008. Geomorphology and Sedimentology of the East Marine Region of Australia. Geoscience Australia, Canberra. Lay, T., and others, 2005. The Great Sumatra-Andaman Earthquake of 26 December 2004. Science 308, 1127-1133. Lee, H.J., Locat, J., Desgagnés, P., Parsons, J.D., McAdoo, B.G., Orange, D.L., Puig, P., Wong, F.L., Dartnell, P., Boulanger, E., 2009. Submarine Mass Movements on Continental Margins, Con- tinental Margin Sedimentation. Blackwell Publishing Ltd., pp. 213-274. Locat, J., Lee, H.J., 2002. Submarine landslides: advances and challenges. Canadian Geotechnical Journal 39, 193-212. Lutgens, K., Tarbuck, E., 2012. Essentials of Geology, 11 ed. Pearson Education, USA. Maretzki, S., Grilli, S., Baxter, C.D.P., 2007. Probabilistic SMF tsunami hazard assessment for the upper east coast of the United States. Masson, D.G., Harbitz, C.B., Wynn, R.B., Pedersen, G., Lovholt, F., 2006. Submarine landslides: processes, triggers and hazard prediction. The Philosophical Transactions of the Royal Society A 364, 2009-2039. McMurtry, G.M., Watts, P., Fryer, G.J., Smith, J.R., Imamura, F., 2004. Giant landslides, megatsunamis, and paleo-sea level in the Hawaiian Islands. Marine Geology 203 219-233. Miller, D.J., 1960. The Alaska earthquake of July 10, 1958: Giant wave in Lituya Bay. Bulletin of the Seismological Society of America 50, 253-266. Prior, D.B., Coleman, J.M., 1984. Submarine slope instablilty, in: Brunsden, D., Prior, D.B. (Ed.), Slope Instability. John Wiley & Sons Ltd., Norwich, pp. 419-455. Short, A.D., Trenaman, N., 1992. Wave climate of the Sydney region, an energetic and highly variable ocean wave regime. Marine and Freshwater Research 43, 765-791. Short, A.D., Woodroffe, C.D., 2009. The Coast of Australia. Cambridge University Press, New York.

1-19 Clarke (2014) Chapter 1

Stow, D.A.V., Mayall, M., 2000. Deep-water sedimentary systems: New models for the 21st century. Marine and Petroleum Geology 17, 125-135. Sue, L.P., Nokes, R.I., Davidson, M.J., 2011. Tsunami generation by submarine landslides: comparison of physical and numerical models. Environmental Fluid Mechanics 11, 133-165. Tappin, D.R., Watts, P., McMurtry, G.M., Lafoy, Y., Matsumoto, T., 2001. The Sissano, Papua New Guinea tsunami of July 1998 - offshore evidence on the source mechanism. Marine Geology 175, 1-23. Tinti, S., 2005. The 30 December 2002 landslide-induced tsunamis in Stromboli: sequence of the events reconstructed from the eyewitness accounts. Natural Hazards and Earth System Sci- ences 5, 763-775. Troedson, A.L., Davies, P.J., 2001. Contrasting facies patterns in subtropical and temperature continental slope sediments: inferences from east Australian late Quaternary records. Marine Geology 172, 265-285. Twichell, D.C., Chaytor, J.D., ten Brink, U.S., Buczkowski, B., 2009. Morphology of late Quater- nary submarine landslides along the U.S. Atlantic continental margin. Marine Geology 264, 4-15. Varnes, D.J., 1978. Slope Movements and Types and Processes, Landslides: Analysis and Control, Special Report. Transportation Research Board, National Academy of Sciences, Washington, pp. 11-33. Watts, P., 2004. Probabilistic predictions of landslide tsunamis of Southern California. Marine Geol- ogy 203, 281-301.

1-20 Clarke (2014) Chapter 2 Chapter 2

Sedimentology, Structure and Age Estimate of Five Continental Slope Submarine Land- slides, Eastern Australia

Samantha Clarke1, Thomas Hubble1, Jody Webster1, David Airey2, Elyssa de Carli1, Cristina Ferraz3, Paula Reimer4, Ron Boyd5, 8, John Keene1, and Shipboard Party SS12/2008

1. Geocoastal Research Group, University of Sydney, Sydney, NSW, Australia

2. School of Civil Engineering, University of Sydney, Sydney, NSW, Australia.

3. Divisão de Geologia e Georecursos Marinhos, Instituto Português do Mar e da Atmosfera (IPMA), Lisbon, Portugal

4. Center for Climate, the Environment & Chronology (14CHRONO), Queen’s University, Belfast, UK

5. University of Newcastle, Newcastle, NSW, Australia.

6. ConocoPhillips, Houston, TX, United States.

Title of Team: Shipboard Party SS12/2008 [email protected]

Abstract

Sedimentological and AMS 14C age data are reported for calcareous hemipelagic mud samples cored from within, or adjacent to, five submarine landslides present on eastern Australia’s continental slope between Noosa Heads and Yamba. Sediments are mixtures of calcareous and terrigenous clay (10-20%), silt (50-65%) and sand (15-40%) and are generally uniform in appearance. Their carbonate contents vary between and 17% and 22% by weight, while organic carbon contents vary between 4% and 12% by weight. Dating of conformably deposited material indicates sediment accumulation rates between 0.017mka-1 and 0.2 mka-1, which is consistent with previous estimates reported for this area. Visually identified transition surfaces (boundary surfaces) are identified in five cores at depths of 0.8 to 2.2 meters below the present-day seafloor, which separate looser material from stiffer more compacted material. Boundary surfaces present a sharp colour-change across the surface; discern-

2-1 Clarke (2014) Chapter 2 able but small increases in sediment stiffness; a slight increase in sediment bulk density of 0.1 gcm-3; and distinct gaps in AMS 14C ages of at least 25 ka. Boundary surfaces are suspected to represent the removal of material, possibly a slide plane detachment surfaces. Examination of sub-bottom profiler records indicate that: 1) the youngest identifiable sediment reflectors upslope three submarine landslides terminate on and are truncated by slide rupture surfaces; 2) there is no obvious evidence for a post-slide sediment layer draped burying slide ruptures or exposed slide detachment surfaces; and 3) the boundary surfaces are unlikely to be a major slide surface, although they may represent a surface from which near-surface layers have been detached. This suggests that these submarine landslides are geologically recent, and that the boundary surfaces are either: a) erosional features that developed after the occurrence of the landslide in which case the boundary surface age provides a minimum age for landslide occurrence or b) detachment surfaces from which slabs of near-surface sediment were removed during landsliding in which case the age of the sediment above the boundary surface indicates approximately when landsliding occurred. While an earthquake triggering mechanism is favoured for the initiation of submarine landslides on the eastern Australian margin, this causal mechanism cannot be conclusively demonstrated.

Keywords: mass-failure • multibeam • seafloor geomorphology • continental margin • southeast Aus- tralia • continental slope • passive margin • sedimentation rates

2-2 Clarke (2014) Chapter 2 2.1 Introduction

Submarine landslides are common features along continental slopes and oceanic islands (Hamp- ton et al. 1996; Lee 2009). Their volumes range over 5 orders of magnitude from small shallow slides less than 0.1 km3 to more than 2400 km3, such as the Storegga slide off the Norwegian coast (Haflida- son et al. 2004; Masson et al. 2006). The larger slides are generally thought to be capable of generating damaging or catastrophic tsunami (Harbitz et al. 2013) and the suggested landslide triggers include earthquake loading (Fine et al. 2005), pore pressure effects (Locat and Lee 2002; Masson et al. 2006), gas generation (Maslin et al. 1998; Sultan et al. 2004), storm waves (Prior and Coleman 1982), and rapid sedimentation (Masson et al. 2006). While a number of slides have been identified and examined in detail, for example, the Storegga Slides in Norway (Haflidason et al. 2004), the Brunei Slide in Borneo (Gee et al. 2007), the Goleta Slide in California (Greene et al. 2006), slides in Angola (Gee et al. 2006), the Gulf of Mexico (Silva et al. 2004), the Hawaiian Islands (McMurtry et al. 2004), Canary Islands (Masson et al. 2006), and slides along the Hikurangi Margin in the Southwest Pacific Region (Lamarche et al. 2008), the physical processes that generate submarine landslides are not well- constrained or understood (Locat and Lee 2002; Bardet et al. 2003; Mosher et al. 2010; Urlaub et al., 2013). One of the principal reasons for the lack of definitive explanations for this phenomenon is the relative dearth of data on the physical and mechanical properties of the sediments, particularly sediments representative of the failure surface, as sediments of this type have only been recovered in a few cases (Urlaub et al. 2013).

Jenkins and Keene (1992) first identified submarine landslides on the eastern Australian (EA) continental margin south of Sydney in GLORIA swath maps produced in the late 1980’s. Subse- quent mapping of this margin with the next generation of high resolution, multibeam echo-sounding equipment between 2006 and 2013 has provided sufficient information and morphological data to establish that there are a surprisingly large number of submarine landslides evident on the continental slope (Glenn et al 2008; Boyd et al 2010; Hubble et al. 2013) given this margin’s well-recognized low sedimentation rates (Boyd et al. 2004) and passive margin tectonic setting. Instead of the relative qui- escence and stability that might be intuitively expected, recent studies have demonstrated that submarine sliding should be considered to be a common and ongoing characteristic of the continental slope offshore east Australia south of Fraser Island (Boyd et al. 2010; Clarke et al. 2012; Hubble et al. 2012).

2-3 Clarke (2014) Chapter 2

This chapter presents a study of upper continental slope sediments collected from the eastern Australian continental margin and is part of a larger body of work intended to determine the frequency and consequences of submarine landsliding in this part of the margin. This work aims to establish the geological and sedimentological characteristics of the materials in which geologically recent (<25 ka) submarine landsliding has occurred (Boyd et al. 2010; Clarke et al. 2012) and examines material sampled from ten gravity cores between two and five meters long that were collected onboard RV Southern Surveyor in November 2008 (SS2008-V12; Boyd et al. 2010). Five cores present boundary surfaces that are identified by a sharp, colour-change boundary; small increases in sediment stiffness; slight increases in sediment density; and distinct gaps in AMS 14C age of at least 25 ka. These boundary surfaces are interpreted to represent detachment surfaces or slide plane surfaces. It is suspected that the sediment below the boundary features has been buried to depths between 50-220 m (depending on the thickness of the submarine landslide scar) prior to the removal of the now-missing sediment. The sediment above the boundary features is believed to represent recent sediment drape. We report sedimentological data and AMS 14C isotopic dates for the cored sediments and then interpret this data in the context of the morphology of the landslides as evident in the multibeam bathymetry and constrained by sub-bottom profiler transects.

2.2 Study Area Location

The study area is located along the eastern Australian continental margin (Fig. 2.1), offshore Noosa Heads in southern Queensland (QLD) and Yamba in northern New South Wales (NSW). It is situated approximately 30 km to 70 km seaward of the present coastline in water depths of 150-4500 m (Fig. 2.2). Figure 2.1: Google Earth image of the eastern Australian continen- Ten gravity cores from the upper continental slope tal margin from the Bass Strait to (<1200 m) were examined in this work (Fig. 2.2). They are the Great Barrier Reef. Red box outlines the study area. Inset shows located at positions within or adjacent to five slides. There are: the location of the study area on a line map of Australia.

2-4 Clarke (2014) Chapter 2

1) Bribie Bowl Slide (water depth 600 m center of mass); 2) Coolangatta-2 Slide (water depth 900 m center of mass); 3) Coolangatta-1 Slide (water depth 600 m center of mass); 4) Cudgen Slide (water depth 600 m center of mass); and 5) Byron Slide (water depth 800 m center of mass) (see Fig. 2.2,2.3,2.4 and Table 2.1). These features are representative of the slope failures that occur in the two dominant slope morphologies present in the study area (see Fig. 2.3) identified by Boyd et al. (2010) and Clarke et al. (2012, 2014) which are: a) the relatively steep (3-7°) and canyon incised slope (Bribie Bowl Slide and Bryon Slide); and b) the relatively gentle slope (1-3°) of the Nerang Plateau (Cool- angatta-1 & -2 slides and Cudgen Slide). At least one gravity core was recovered from each of these

Figure 2.2: Digital elevation model (DEM) of the southern QLD to northern NSW continental margin, showing the location of the study area. Insets show detailed views of the five submarine landslide sites of interest (black line) and the locations of the twelve collected gravity cores (GC) from the Southern Surveyor SS2008/12 voyage. Moving north to south: (a) Bribie Bowl Slide (b) Coolangatta-2 Slide, (c) Coolangatta-1 Slide (d) Cudgen Slide, (e) Byron Slide

2-5 Clarke (2014) Chapter 2 submarine landslide scars.

The ten cores provide material for dating, sedimentology, and physical property testing. Seven cores (GC2, GC3, GC5, GC6, GC8, GC11, and GC12) were collected from within five submarine landslide scars (see Fig. 2.2,2.3 and Table 2.1). Three reference cores (GC1, GC4, and GC7) were recovered from adjacent sites in slopes that do not present obvious slide features or morphologies.

In addition to the ten cores dealt with in this work, two additional cores were collected for geotechnical testing (GC9 and GC10) and these cores are described and discussed in Chapter 3. Table 2.1: Summary of locations, depth, total recovery length, stratigraphy and target feature of the sediment cores retrieved from the study area.

Latitude Longitude Water depth Length Name Stratigraphy Target Feature °S °E (m) (m)

Single conformable Adjacent to Bribie Bowl GC1 -26°41.69’ 153°45.123’ 841 4.15 section Slide (reference core)

Single conformable GC2 -26°45.64’ 153°43.68’ 585 4.45 Bribie Bowl Slide section

Single conformable GC3 -26°45.22’ 153°43.85 759 4.35 Bribie Bowl Slide section

Surface sample, no GC4 -28°13.032’ 153°57.387’ 488 - Cudgen Slide (no return) stratigraphy

2 sections, one boundary Cudgen Slide – upper GC5 -28°14.102 153°57.871’ 677 2.26 feature slide block

2 sections, one boundary Cudgen Slide – base of GC6 -28°13.601’ 153°58.767’ 740 1.91 feature slide

4 sections, three Adjacent to Cudgen GC7 -28°12.71’ 153°59.69’ 748 5 boundary features Slide (reference core)

2 sections, one boundary Coolangatta-1 Slide – GC8 -28°07.263’ 154°02.752’ 929 2.52 feature lower slide region

Unsplit for geotechnical Coolangatta-2 Slide – GC9 -28°01.79’ 154°02.84’ 913 4+ testing upper slide region

Unsplit for geotechnical Cudgen Slide – slide GC10 -28°02.479’ 154°02.261’ 800 4+ testing hollow

2 sections, one boundary Cudgen Slide – base of GC11 -28°13.491’ 153°58.960’ 755 2.49 feature slide

2 sections, one boundary Bryon Slide – middle GC12 -28°37.96’ 153°58.09’ 1167 1.99 feature slide

2-6 Clarke (2014) Chapter 2

Figure 2.3: DEM of the east Australian continental margin. Perspective looks west, landward onto the continental margin. Insets show details of five submarine landslides (black lines) and gravity cores locations within each slide. Moving north to south: a) Bribie Bowl Slide (northern canyon region): GC2 and GC3 are located in the slide; GC1 is located adjacent to the slide; b) Coolangatta-2 Slide and Coolangatta-1 Slide (plateau region): GC9 is located in the Coolangatta-2 Slide; GC8 is located in the Coolangatta-1 Slide ; c) Cudgen Slide (plateau region): GC4, GC5, GC6, GC7, GC10, GC11 are located in the slide; GC7 is located adjacent to the slide; d) Byron Slide (southern canyon region): GC12 is located in the slide

2-7 Clarke (2014) Chapter 2

Figure 2.4: DEM of the five submarine landslides investigated: a) Bribie Bowl Slide; b) Coolangat- ta-2 and Coolangatta-1 Slide; c) Cudgen Slide; d) Byron Slide. Each slide is shown coloured by: i) water depth; and ii) slope angle. Black circles indicate gravity cores locations (see Fig. 2. 2 for core names).

2-8 Clarke (2014) Chapter 2 2.3 Geologic Setting and Margin Structure

The eastern Australian continental margin represents a distinct oceanographic, climatic and geological province (Boyd et al. 2004). It extends 1500km from the Great Barrier Reef to Bass Strait and forms the western continental physiographic boundary of the Tasman Sea. The margin is relatively narrow and typically presents a steep, thinly sedimented upper and middle slope with continental bedrock exposed on its lower slope (Fig. 2.5) (Marshall 1978, 1979; Keene et al. 2008; Boyd et al. 2010). The present day morphology of the slope is dominated by erosion and mass wasting (Glenn et al. 2008; Boyd et al. 2010; Clarke et al. 2012; and Hubble et al. 2012).

Figure 2.5: Multi-channel seismic reflection line (GA206) located due east of Yamba showing the typical characteristics of the east Australian margin in NSW and southern QLD: a) uninterrupted and b) annotated. Line drawing insert shows interpretation of the section. Note the thin sequence of post-rift sediment present on the upper half of the slope and exposed basement rock on the lower half of the slope (Symonds et al. 1973; Marshall et al. 1978, 1979; Cowell et al. 1987; Hubble et al. 1992).Vertical exaggeration is 15.5.

2-9 Clarke (2014) Chapter 2

The continental shelf is 20 km wide offshore Byron Bay in the south of the study area and gradu- ally increases to 70 km offshore Noosa Heads in southern Queensland in the north of the study area. The shelf break is located an average distance of 50 km from the shore at depths between 100 m and 150 m, which is typical of the eastern Australian margin’s structure south of Fraser Island. North of the study area, at the northern tip of Fraser Island, the shelf shallows considerably and at Breaksea Spit the shelf-break occurs in particularly shallow water, 20 m deep (Boyd et al. 2004). At this location modern-day sands are transported to, and then driven over, the shelf edge by the interacting northerly wave-driven littoral transport and easterly tidal flows.

The structure and morphology of the shelf in the study area can be divided into two distinct, inner and outer zones (Boyd et al. 2004). The inner shelf commonly presents regions of outcropping bedrock comprised of Palaeozoic fold belt rocks or Mesozoic sedimentary basin rocks while the outer continental shelf presents a variable but generally thin cover of sediment up to 300 metres thick above bedrock with some exposed bedrock pinnacles (Davies 1979; Marshall 1978 and 1979). A bedrock high may be present beneath the shelf break but more commonly the bedrock basement surface located beneath the outer shelf and upper slope presents as an inclined, irregular surface that drops away and joins the oceanic basement reflector at the base of the continental rise (Fig. 2.5).

The EA margin’s continental slope is relatively steep at the shelf break compared to typical (<5°) Atlantic style passive margins (McAdoo et al. 2000) and presents average slopes of 5-10° in water depths ranging from 150 m at the shelf break to 4500 m at the abyssal plain (Boyd et al. 2010). The relatively steep inclination of the slope is ascribed to the combination of asymmetric passive-margin rifting and insufficient sedimentation. The original continental rift that preceded sea-floor spreading in the Tasman Sea Basin between 74 Ma and 52 Ma before present (Gaina et al. 2003) is thought to have breached on its western flank (Falvey 1974; Shaw 1975; Lister et al 1986) resulting in a particularly steep continental slope. The lack of sediment on the lower slope is suggested to be a consequence of the absence of major river systems delivering material to the slope and submarine erosion during the Neogene (Keene et al 2008; Hubble et al. 2012). In any case, the slope is generally regarded to be sediment deficient, especially when compared to many other passive margins of similar age around the world (Boyd et al. 2004; Boyd et al. 2010).

Shelf sediments are generally comprised of siliciclastic sands on the inner shelf and carbonate-

2-10 Clarke (2014) Chapter 2 rich facies on the outer shelf, which is indicative of higher carbonate production at low stands on the outer shelf/continental slope and input from terrestrial sediment to the inner shelf (Troedson and Davies 2001; Boyd et al. 2004). The Eastern Australian Current, a strong (2 to 4 knots) southerly- directed oceanic boundary current whose influences the ocean column to depths greater than 500 m (Godfrey et al. 1980; Church 1987; Keene et al. 2008), is believed to be responsible for the absence of any significant accumulations of fine-grained sediments on the shelf in northern NSW and southern QLD region (Boyd et al. 2004). Sediment composition in this region is also controlled in part by water temperature, which influences the generation of carbonate component of the sediments and has resulted in a more sub-tropical to tropical assemblage of biogenic detritus (Troedson and Davies 2001; Roberts and Boyd 2004).

The upper continental slope sediments are generally deposited as a thin layer of material above continental basement rocks (see Fig. 2.5). These deposits are generally less than one kilometre thick and commonly less than 500 metres thick (Conolly 1969; Ringis 1972; Keene et al. 2008), which is a thin deposit in comparison to other passive margin deposits such as those of the North Atlantic (Heezen 1974) which are commonly an order of magnitude thicker. Previous studies are few, but they consistently indicate that these upper slope sediments are mixed siliciclastic-carbonate muds (Hubble and Jenkins 1984a, b; Troedsen and Davis 2001; Glenn et al. 2008; Hubble et al. 2012).

Only one published study on four cores of continental slope sediments from the Eastern Sea- board of Australia (depth range 350-2500 m) is currently available (Troedsen and Davis 2001), and only 40 cores, all obtained by gravity corer and <5 m in length, have been described for the ~1500 km length of continental slope (Hubble and Jenkins 1984a,b; Troedsen 1998; Troedsen and Davis 2001; Glenn et al. 2008). None are reported to have penetrated slide surfaces, although the cores collected by Glenn et al. (2008) were taken as part of a submarine landslide study. Sedimentation rates range from 0.01-0.02 mka-1 (long-term estimates based on current margin sediment thickness and the margin’s age) to 0.02-0.24 mka-1 (directly determined rates from the limited core samples; Troedson and Davies 2001). These rates are based on limited data and further investigation is needed to produce more robust short-term and long-term sedimentation rates for the margin.

2.3.1 Earthquakes and seismicity in Australia Earthquake loading (also known as cyclic loading and seismic loading) is the most commonly in-

2-11 Clarke (2014) Chapter 2 dicated submarine mass failure trigger (Masson et al. 2006). It is a short-term trigger, with slope failure occurring during the earthquake or shortly afterwards (e.g. Aitape submarine landslide; Tappin et al. 2008). During an earthquake, the ground accelerations may be of a sufficient magnitude to overtop static gravity forces (Bardet et al. 2003). Mechanically, there are three aspects to this phenomenon: the sediments exposed to seismic loading (where there can be an appli- cation of acceleration in both the horizontal and vertical directions) generally experience a decrease in stiffness, decrease in shear strength, and increase in pore pressure (Coleman and Prior 1988; Sultan et al. 2004). If the sediment is loosely-packed, compaction takes place at a rate controlled by permeabil- ity of the soil. Seismic loading usually takes place over a time period Figure 2.6: a) Locations and magnitudes of historic seismicity (M≥4) for that is too short to al- Australia from 1841 to 2010, major seismic zones, historic surface rup- low pore water drain- tures, and orientation of maximum horizontal stress (SHmax) overlying the simplified geological basement terrains of the Australian continent, age and the consequent taken from Clark et al. (2012) and references therein; b) Historic earth- pore-pressure increase quakes in the Newcastle region (error bars may be up to 50 km), fault scarps, and mapped basement geology (red stars, blue lines and red lines produces an equivalent respectively). Yellow star marks the epicenter of the M5.6 1989 New- decrease of the effective castle earthquake. Inset shows sparker profile after Huftile et al. (1999). Modified from Clark et al. (2012). confining stress. In some

2-12 Clarke (2014) Chapter 2 cases this process can reduce the vertical effective stress to zero with a corresponding loss of the soil’s shear strength which may produce a complete loss of shear strength in a non-cohesive granular soil (Sultan et al. 2004). All soil types (not over-consolidated or dense) expected on continental slopes are expected to experience pore pressure build up due to cyclic loading. Silty sands and, to a lesser extent, sands can be very sensitive to pore pressure build up and can collapse and loose strength due to this process (Mitchel and Soga 2005), which can lead to large ground deformations and possibly liquefaction (Coleman and Prior 1988).

The recurrence interval for large eastern Australian surface-breaking earthquakes (i.e. those that rupture on an individual fault source) is not well-constrained and estimates range over three orders of magnitude from one Mw>7.0 event every ten to twenty thousand years to one event every million years (i.e. a range in recurrence interval of 10 ka to 103 ka) (see Leonard 2008; Clark 2010; Clark et al. 2012; Burbidge 2012). The southeast Australian region is reported as being seismically active at a steady rate for over 100 yr, with hypocentral earthquakes depths ranging between very shallow (<4 km) and 17-km deep and aftershocks tending to be very shallow and numerous (Leonard 2008). Figure 2.6a shows the locations and magnitudes of historic seismicity (M≥4) for Australia from 1841 to 2010, major seismic zones, historic surface ruptures, and orientation of maximum horizontal stress (SHmax) overlying the simplified geological basement terrains of the Australian continent, taken from Clark et al. (2012) and references therein. Historically, a number of moderate earthquakes (M4.9-M5.6) have been recorded for the southeast Australian Newcastle region, including the 1989 M5.6 event (see Fig. 2.6b) that killed 13 people and caused 100s of millions of dollars in damages (Huftile et al. 1999; Sinadinovski et al. 2002 and references therein). The onshore 1989 Newcastle earthquake (depth to focus 11.5 km; McCue and Michael-Leiba 1993) (epicenter located on Fig. 2.6b) is suspected to have occurred on a fault system related to the Hunter-Mooki Thrust, that has been imaged offshore New- castle near the shelf edge (see Fig. 2.6b-insert; Huftile et al. 1999; Chaytor and Huftile 2000).

2.4 Data and Methods

Three types of data are presented and interpreted for this study. Firstly, remotely sensed high- resolution multibeam bathymetry data, sub-bottom profiling data and archived seismic reflection data contextualize both the location of core sites and the mechanisms of sliding taking place along the margin. Secondly, a detailed examination of the sediment recovered from the margin. Thirdly, radiocarbon

2-13 Clarke (2014) Chapter 2 dating provides an estimate on the timing of the slides features and sedimentation rates for the margin.

A particular focus of the sedimentology is the investigation of changes across visually identified transition surfaces (boundary surfaces), which separate looser material from stiffer more compacted material, suspected to represent slide planes.

2.4.1 Bathymetry Approximately 13,000 km2 of bathymetric data were acquired using a 30-kHz Kongsberg EM300 multibeam echosounder (Boyd et al. 2010). The multibeam data was processed to produce a 50 m gridded digital elevation model (DEM) covering the region investigated (Fig. 2.2). Using Fle- dermaus V7.3.3b software (http://www.qps.nl/, the DEM was used to examine the continental slope in the study area and the five individual slide sites where the 12 gravity cores were taken.

2.4.2 Sub-bottom Profiles Sub-bottom profile data were acquired in water depths less than ~1500 m using a Topas PS18 parametric sub-bottom profiler. The raw Topas data were processed and converted into SEG-Y format and displayed using SeiSee V2.16.1 software (http://www.dmng.ru/seisview/). Two-way travel time was converted to water depth and sediment thickness using a constant sound velocity of 1500 ms-1.

Sub-bottom profile penetration depended on the weather conditions and on the response of the seabed. Performance was best between 100 to 500 m water depth, but adequate performance was recorded in water depths up to 1500 m. Seismic penetration ranging from ~40 m and up to 100 m commonly in shallower water (<1200 m) and in deeper water (1200-2000 m) when the sea state and swell direction was favourable to the operation of the equipment.

2.4.3 Core Collection and Sediment Properties Coring was carried out using a gravity corer with a 1000 kg head-weight and 5 m long barrel containing plastic liners as the sediment casing. Core sampling locations were chosen after inspection of the multibeam bathymetric data and sub-bottom profiles in order to 1) identify submarine landslide scars (Boyd et al. 2010), and 2) examine the stratigraphy of the slide scars and their adjacent slopes. Despite limited seismic penetration (usually ≤50 m), various slide characteristics, morpho-structural features, layer geometry, and the physical removal of slabs of layered sediment can be inferred from the

2-14 Clarke (2014) Chapter 2 sub-bottom profiles. Figure 7a shows the location of the subbottom profiles.

Sediment cores were cut into one-meter lengths and then split, photographed, logged, and sub- sampled at regular intervals between 10-40 cm spacing (depending on test type), but additional samples were taken on either side and across the lithological boundaries, to test for variations with depth. Two cores were left unsplit for geotechnical testing.

Grainsize and mean grainsize distribution were determined using a laser particle sizer (Malvern Mastersizer 2000) using standard operating procedures (Malvern Instruments 1999), and statistically analyzed using GRADISTAT 8.0 (Blott and Pye, 2001). Distributions for each sample represent the average of the three sample runs. Grainsize subsamples were taken at approximately 30-40 cm intervals.

Carbonate and organic carbon content were determined using the sequential loss on ignition (LOI) method, which gives results to within a maximum error of 2% (Heiri et al. 2001). The average carbonate and organic carbon contents for each core were determined for each 1 m core section, based on three equi-spaced samples taken per metre. High-resolution changes were determined for three cores (GC8, GC11, GC12), where ten samples were taken from each 1 m core section.

Dry and bulk density, unit weight, void ratio, and water content were determined by testing sediment samples of a known volume using classical techniques (ASTM methods; Head 1982). Three samples were taken from each 1 m core section. For simplicity, only unit weight values are shown on included core logs (see Fig. 2.10 and 2.11), however these values directly correlate to the dry and bulk density, void ratio, and water content (see Head 1982). Full core logs are available upon request.

2.4.4 14C Radiocarbon Dating Radiocarbon dates were determined by accelerator mass spectrometry (AMS) on 49 samples taken from eight gravity cores (GC1, GC2, GC5, GC6, GC7, GC8, GC11 and GC12; see Table 2.2). Dates were obtained from both bulk sediment samples and planktonic foraminifera assemblages extracted from the hemipelagic sediments. Twenty five bulk and 24 picked radiocarbon ages were determined. Samples were run at three laboratories: 1) the CHRONO Centre, Queen’s University, Belfast UK (Lab ID: UBA); 2) Australian Nuclear Science and Technology Organisation (ANSTO),

2-15 Clarke (2014) Chapter 2

Lucas Heights, Australia (Lab ID: OZP); and 3) Radiocarbon Laboratory, University of Waikato, New Zealand (Lab ID: Wk).

Conventional 14C yrs BP were converted into calibrated calendar ages following Stuiver and Reimer (1993). Median calibrated ages (BP) were calculated with CALIB V6.1.1 (Stuiver et al. 2005) using marine calibration curve Marine09.14c data set (Reimer et al. 2009) with a reservoir correction (ΔR) value of 11 ± 13 yr (the average for eastern Australia; see http://calib.qub.ac.uk/marine/; Ulm et al. 2006) and reported here with 2σ errors. Some dates could not be calibrated as they fell outside the calibration range (0-50,000 years). Results are shown in Table 2.2.

Ages from GC1 and GC2 were specifically taken to determine a more robust sedimentation rate for the continental slope in this region, while ages from GC8, GC7, GC11, and GC12 were taken primarily to investigate boundary surfaces; however all ages can useful in constraining sedimentation rates.

2.4.4.1 Planktonic foraminifera assemblage samples Extraction of suitable planktonic foraminiferas was conducted by disaggregating each sample of ~1-4 cm of sediment in water and washing over a 63 µm wet sieve. The coarser fractions were dried at 40ºC, weighed and used for foraminiferal analysis. A minimum of ~6 mg of well preserved planktonic foraminifera shells was identified and hand-picked under a binocular microscope and included in each sample (e.g. Murray 1991). Pristine planktonic foraminifera were selected where possible. Foraminif- era shells that showed minimal or no indication of erosion or abrasion through transport were selected from the bulk sediment samples, thereby minimising disparity between death of skeletal organism and time of deposition (Woodroffe et al. 2007), and foraminiferas containing secondary cements or chamber infilling were rejected. For the determination of foraminifera we followed the methods of Loeblich and Tappan (1988) and Jones (1994). Planktonic foraminifera from the mixed surface layer of the water column were favoured for testing due to uncertainties in the deep-water reservoir correction associated with dating benthics.

A mixed assemblage of planktonic foraminifer species (maximum 2 to 3 species; polyspecific sample) were picked for each sample, consisting primarily of Globigerinoides ruber and Pulleniatina obliquiloculata. Where sufficient material was available, single specie (monospecific) samples were used, reducing the possibility of different species giving different ages. The species selection was -con

2-16 Clarke (2014) Chapter 2 sistent with foraminifera used in previous sedimentological studies from the east Australian margin (Troedson and Davies 2001, samples located approximately 100 km north of the study area; and Glenn et al. 2008, samples located approximately 700 km south of the study area) which used planktonic foraminifera species Neogloboquadrina, Pulleniatina, Globigerinoides rubber, Globigerinoides spp. and Globoquadrina spp. for radiocarbon dating.

2.5 Results

2.5.1 Bathymetry and Sub-bottom Profiles - Relationship of core locations to slide morphology Four cores are located within, or adjacent to, slide features from the steep canyon dissected slopes (3-7°) of this segment of the margin (Bribie Bowl Slide, north – 2 cores within the slide, 1 core adjacent to the slide; Byron Slide, south – 1 core). Eight cores are located within, or adjacent to, slide features on the southern end of the gentle dipping, dissected Nerang Plateau (1-3°) (Coolangatta-2 Slide – 1 core; Coolangatta-1 Slide – 1 core; and Cudgen Slide – 4 cores within the slide, 2 cores adjacent to the slide) (see Fig. 2.2,2.3 and Table 2.1). The multibeam bathymetry indicates that the slides sampled are dominantly translational except for the Cudgen Slide, which is rotational, but from which near-surface slabs have detached (see Fig. 2.7-2.9 and description below). The Bribie Bowl and Byron Slides within the canyon regions are significantly thicker (>100 m) in comparison to those developed on the adjacent Nerang plateau (~20-50 m thick).

The translational slides present as retrogressive features with headwall scarps presenting layers that are “unsupported” (i.e. truncated at the headscarp). In map view the crown/headwall scarps present a distinctive semicircular shape characteristic of circular failures (c.f. Varnes 1978). Detached slide slabs are essentially slope-parallel planar blocks. Sub-bottom profiles indicate that the surficial sediment layer reflectors upslope of the slide are truncated by the slide rupture surfaces (see Fig. 2.7- 2.9). Undeformed, parallel-bedded sediment reflectors are observed upslope of the headscarp, while contorted/distorted sediment reflectors are observed downslope. The undeformed units above the headscarp present a similar geometry and layer thickness to the distorted layers below the headscarp (see Fig. 2.8).

Sub-bottom profiles do not to present obvious evidence (such as continuous overlaying reflec-

2-17 Clarke (2014) Chapter 2 tors) of a post-slide sediment drape burying the slide ruptures or exposed slide detachment surfaces; i.e. any post-slide sediment layer is too thin to detect in the sub-bottom profile records (see Fig. 2.7-2.9). This indicates that these translational slides are either a) young, and/or b) the upper slope has experienced a continual erosive removal of material since the occurrence of the slide maintaining exposure of the detachment surface.

2.5.1.1 Bribie Bowl Slide and GC1, GC2, GC3 GC1 is a reference core, taken adjacent to the Bribie Bowl Slide, while GC2 and GC3 are located within the slide scar (Fig. 2.2a). The Bribie Bowl Slide is a translational slab failure located within the northern steeper canyon region (3-7°) and presents a thicker slide (>100 m) in comparison to those developed on the adjacent Nerang plateau. The Bribie Bowl Slide has an average slope of approximately 12° along the majority of the seafloor surface, increasing to 33° in the crestal amphitheater region at the head of the slide scar.

Approximately 8 km north of the Bribie Bowl Slide, subbottom profile line L58a (Fig. 2.7b) shows the subsurface sediment layering geometry of the upper slope between 300 m and 1100 m. It crosses a 100 m high escarpment at ~800 m water depth which is probably the crown headscarp of a slide. This escarpment is visible in the bathymetry (see Fig. 2.4a) and extends into the headscarp of the Bribie Bowl Slide. The upper slope of the profile presents two prominent reflectors (a and b), one at 15 m and the other at 30 m below surface slope (320 to 400 m), which onlap a third reflector (c) at the western end of the line. The a, b, and c (?) reflectors terminate at the slide scarp and a very similar geometric package of reflectors is evident in the near-surface, contorted sediment layers located immediately downslope of the slide scarp. Line L58a provides context for GC1 and demonstrates the relative geological youth of the prominent scarp, but the lack of downslope line length reduces its util- ity in confirming the slide type as a translational slab slide which is evident in the feature’s bathymetric expression, inter-level character and overlapping outline.

2.5.1.2 Coolangatta-2 Slide and GC9; Coolangatta-1 Slide and GC8 Cores GC9 and GC8 are located in the slide scar of the Coolangatta-2 and Coolangatta-1 Slides respectively on the Nerang Plateau (Fig. 2.2b). The Coolangatta-2 and Coolangatta-1 Slides are relatively shallow translation slab slides. The removed material is relatively thin (<50 m) and is representative of the numerous upper slope failures that occur ubiquitously on this very gently dipping (2-3°) plateau. The seafloor within the scars is hummocky, gently concave, with average slopes of approxi-

2-18 Clarke (2014) Chapter 2 a) Location of 6 subbottom profile lines (black) taken across or adjacent to 3 submarine landslides (gray): (i) Bribie Bowl Slide (ii) Cool - Bribie Bowl or adjacent to 3 submarine landslides (gray): (i) lines (black) taken across a) Location of 6 subbottom profile Figure 2.7: Figure angatta-1 Slide and Cudgen Slide. Gravity core locations are also shown; b) Sub-bottom profiling line L58a-WE moving west to east down the upper west to east down moving line L58a-WE profiling b) Sub-bottom also shown; locations are core Gravity Slide. and Cudgen angatta-1 Slide reflectors terminate at b, and c (?) The a, GC1, GC2, and GC3 (northern region). canyon Slide, Bowl north of the Bribie a slide feature, slope across line GC8- profiling c) Sub-bottom the slide scarp and distorted below; layers above undisturbed layers shows the slide scarp (see text for details). Insert Insert red. The location of GC8 is indicated in GC8 (southern plateau region). through south to north the Coolangatta-1 Slide, across SN moving distorted within the slide scar. sediment layers shows

2-19 Clarke (2014) Chapter 2 mately 3.5° within the failure plane, that increases up to 7.5° at the head scarp. Failures on the Nerang Plateau suggest a continuous shedding of material off the plateau slope, with failures retrogressively retreating towards the shelf break.

Sub-bottom profile GC8-SN runs parallel to the bathymetric contours across the Coolagatta-1 Slide and through the location of gravity core GC8 (Fig. 2.7c). The unfailed area to the south of the slide is characterised by gently undulating well-bedded material, whereas the sediments present as irregularly distorted with some obvious layering of the slide margin. The within slump’s surface is located at 1150 m water depth, and a slab of material approximately 50 m thick appears to have been removed. The youngest identifiable sediment layer reflectors upslope of these slides terminate on the slide margin surface. Penetration of about 50-60 m is achieved in the well-bedded sediments, while ~40-50 m of penetration is recorded in the irregular/distorted sediments down slope that are cut by the slump (Fig. 2.7c). The within-slide surface topography is irregular and presents numerous point diffractions upslope of the southern headwall, and distorted layering, which are also evident in the subsurface material. GC8 penetrated 2.52 metres, and at this location the strong continuous subbottom reflector was about 5 m below the surface (see Fig. 2.7c-insert).

2.5.1.3 Cudgen Slide and GC5, GC6, GC7, and GC11 Six cores are located within, or adjacent to, the Cudgen Slide from the southern end of the gently dipping (2-3°) Nerang Plateau (GC4, GC5, GC6, GC7, GC10, GC11) (see Fig. 2.2c). Two reference cores were retrieved adjacent to the Cudgen slide (GC4 and GC7), and four cores were retrieved through surfaces within the boundaries of the target slides (Table 2.1).

Sub-bottom profiles of the Cudgen Slide indicate it is a large area (50 km2), deep rotational slide from which near-surface slabs have been detached by translational sliding. The current slide depression shown in the bathymetric data is relatively thin (<50 m; see Fig. 2.3c,2.4c) and is representative of the numerous upper slope failures that occur on this very gently dipping plateau. It presents a hummocky surface terrain within the failure region, a gently concave slide shape, and average slopes of approximately 3.5° within the failure plane, up to 7.5° at the head scarp. These failures appear to be part of the continuous shedding of material off the plateau slope, with failures retrogressively moving back up the slope towards the shelf break.

Four sub-bottom profile transects (three west-east transects and one north traverse tie) are avail-

2-20 Clarke (2014) Chapter 2 a) Sub-bottom profiling line GC4 and 5-WE down the Cudgen Slide, through GC5 (southern plateau). The location of is indicated Slide, through Cudgen the down line GC4 and 5-WE profiling a) Sub-bottom Figure 2.8: Figure in red. Missing sediment slab location is shown in blue (see text for details). Line locations are indicated in Fig. 2.4a. Insert shows sub-parallel sedi - shows 2.4a. Insert indicated in Fig. in blue (see text for details). Line locations are sediment slab location is shown Missing in red. GC6, GC11, and the adjacent Slide, through Cudgen the line GC6and7-SN tie across profiling within the slide scar; b) Sub-bottom ment layers of modern upper layer the dark 2.7a. Note indicated in Fig. Line locations are indicated in red. locations are core GC7 (southern plateau). Gravity surface. the failure sediments above

2-21 Clarke (2014) Chapter 2 a) Sub-bottom profiling line L79a-WE down the Cudgen Slide, through GC11 (southern plateau). The location of is indicated in Slide, through Cudgen the down line L79a-WE profiling a) Sub-bottom Figure 2.9: Figure red. Missing sediment slab location is shown in blue (see text for details). Line locations are indicated in Fig. 2.7a. Insert shows distorted sediment shows 2.7a. Insert indicated in Fig. in blue (see text for details). Line locations are sediment slab location is shown Missing red. GC6 (southern plateau). The Slide, through Cudgen the west to east down moving line GC6-WE profiling within the slide scar; b) Sub-bottom layers indicated in in blue (see text for details). Line locations are shown missing sediment slabs are Locations of three location of GC6 is indicated in red. the slide scarp and distorted below. layers above undisturbed layers shows 2.7a. Insert Fig.

2-22 Clarke (2014) Chapter 2 able over the Cudgen Slide (Fig. 2.8 and 2.9). This provides the most detailed set of sub-bottom profiles for the slides available to this study. Penetration of ~50 m is achieved in the well-bedded sediments, while up to 40 m of penetration is recorded in the irregular/distorted sediments down slope that are displaced by this slump feature. The three downslope transects all contain detailed images of the landslide scarp and the two longer lines (transects L79a-WE and GC6-WE) present very irregular topography, dominated by diffractions, and highly disturbed material at the downslope terminations (Fig. 2.8), consistent with the hummocky terrain evident at the equivalent position in the bathymetric image. The apparent continuation of the upslope, above-scarp sediment layering geometry in the near-surface below-scarp landslide mass is consistent with and corroborates the interpretation of the Cudgen Slide’s bathymetric expression as a rotational failure.

West-East line L79a-WE down the Cudgen Slide presents a transition from well-bedded soft sediments above the main slide, to irregular distorted sediments with little obvious layering below a drop-off (especially > 5 m penetration), with the slumped surface dropping from ~875 to 950 m water depth (Fig. 2.8a). A 0.5 km long slab of material has obviously been removed from the upper slope just above the slide’s headscarp (~840 m water depth). This slab is ~20 m thick and there is no evidence for its incorporation as blocks deposited on the top of the rotated slump mass. This slab has apparently detached from its basal surface and either travelled beyond the extent of the both the sub-bottom profile and bathymetric image, or completely disaggregated during failure. GC11 is taken from the distorted within-slide sediment and has penetrated a distinct but non-continuous sub-bottom reflector surface at ~2 m below the surface, and at this location the strong continuous, but distorted, subbottom reflector was about 3.5 m below the surface (see Fig. 2.8a-insert).

Similarly, profile GC6-WE shows well-bedded soft sediments upslope of the slide’s headscarp and irregular distorted sediments with little obvious layering downslope of the scarp (especially > 5 m penetration), with the slumped surface dropping from ~850 to 950 m water depth (Fig. 2.8b). Three slabs of material have been removed from the upper slope above the slide (headscarps at ~690 m, 770 m, and 810 m water depth). The slabs are ~3-5 m, 5 m, 10 m thick, and ~700 m, 400 m, 450 m long respectively. They are not apparent as blocks on the slide surface in the lower hummocky terrain slumped material although they may be incorporated in this material. The slabs have also detached from strong sub-bottom reflector surfaces. GC6 is taken from the contorted layers located within the main body of the slide. There a strong non-continuous, sub-bottom reflector surface present about 3 m below the seafloor surface but the core probably does not sample this surface as the core length is

2-23 Clarke (2014) Chapter 2 only 1.91 m (Fig. 2.8b-insert).

Profile GC4 and 5-WE also presents well-bedded soft sediments above the slide’s headscarp, but does not show disturbed material as this line does not traverse the hummocky terrain evident in the bathymetry. As with the other two downslope transects, a slab of material is missing from the upper slope above the slide (headscarp at ~670 m water depth). This slab is ~10 m thick, ~900 m long, and is not obviously apparent in the lower slumped material. GC5 is taken from the within slide sediment and penetrates a distinct but non-continuous sub-bottom reflector surface at ~1.2 m below the surface, which correlates directly with the depth of the identified boundary surface at 1.16 m depth within the core (see Fig. 2.11, 2.12). GC5 terminates directly above a semi-continuous sub-bottom reflector surface at ~3 m below the surface (Fig. 2.9a-insert).

South-north tie line GC6 and 7-SN similarly shows a transition from well-bedded soft sediments in the sidewalls of the slide, to irregular distorted sediments with little obvious layering below a drop-off (especially > 5 m penetration), with the slumped surface dropping from ~900 to 1000 m water depth. Cores GC6 and GC11 are taken from the distorted sediment in the slide scar and penetrate a distinct but non-continuous sub-bottom reflector surfaces at ~1 m and 2 m below the surface respectively, correlating with the depth of the boundary surfaces at 1.83 m (GC6) and 2.11 m (GC11) for each core (see Fig. 2.11, 2.12). Both cores terminate directly above a strong continuous, but distorted, sub-bottom reflector surface at ~3-4 m below the surface (Fig. 2.9b-insert). Core GC7 is taken from parallel-bedded sediment adjacent to the slide, and penetrates at least three distinct continuous sub-bottom reflector surfaces at ~1.5 m, 2 m, and 4 m below the surface respectively (Fig. 2.9b-insert).

2.5.1.4 Byron Slide and GC12 One southern canyon core was collected from within the Byron Slide (core GC12; Fig. 2.2d). There are no sub-bottom profiles available for the Byron Slide but this feature’s bathymetric expression indicated it is a translational slab failure on the upper slope. It is located within the steeper canyon region (3-7°) and presents a significantly thicker slide (>200 m) in comparison to those developed on the adjacent Nerang plateau (~20-50 m thick). The seafloor within the Byron Slide has an average slope of approximately 12°, increasing to 33° in the crestal amphitheater region at the head of the slide scar. The Byron Slide displays both shallow slab slide failures towards the top that have failure planes parallel to the adjacent apparently unfailed slope segments, and deeper slump scars that seem to suggest rotational failure towards the bottom of the lower slope.

2-24 Clarke (2014) Chapter 2

2.5.2 Core Descriptions and Sedimentology (Physical Properties) Initial assessment of stratigraphic and lithologic characteristics of these upper slope (<1200 m water depth) sediments suggests that general sedimentological features remain fairly consistent across the margin.

2.5.2.1 Lithological units Two main lithological units that can be observed in the cores: Unit 1 and Unit 2. Unit 1 is the upper unit of younger, overlying sediment drape and generally presents as homogenous, bioturbated, hemipelagic clay bearing sandy silt to silty sand or clay-sand bearing silts. The sediment is dominantly biogenic, sandy carbonate mud, with some terrigenous silt and clay. The sand fraction is typically comprised of bioclastic material (mostly foraminifera). The colour of the unit sediments generally grades through olive grey at the top through dark olive grey at the base of the unit. The upper 5-20 mm presents as a thin oxidised layer of light yellowish-brown sediment. Burrows, laminae and mottling are present to varying degrees. Unit 2 is the lower unit of underlying sediment, occurring beneath a distinct break in material and presents as a homogenous, bioclastic, hemipelagic clay bearing sandy silts to silty sands, which appear in most cases (GC6, GC7, GC8, GC11, GC12) with discernable but small increases in sediment stiffness and a slight increase in sediment bulk density of 0.1 gcm-3 in comparison to the Unit 1 (GC5 experiences a large drop in stiffness directly below the boundary which then increases to normal levels down the core). Unit 2 generally has a uniform texture, with some faint laminations apparent. The sediment is comprised of foraminifera, shell fragments and other carbonate detritus. The colour of the material grades through light grey at the top of the boundary and into a dark grey right at the base. Faint to moderate mottling and occasional bioturbation is also present.

Cores GC1, GC2, GC3, GC9 and GC10 only present Unit 1; while cores GC5, GC6, GC7, GC8, GC11, and GC12 present both Units 1 and Unit 2 (Fig. 2.10-2.12). The difference between these two units is physical rather than compositional in that the second lower unit is stiffer and usually denser than the upper unit. Cores that sample both units tend to be shorter than those only pen- etrating Unit 1, with the exception of GC7, most likely due to the stiffer Unit 2 sediments. All cores show some degree of bioturbation, especially in the upper unit (Unit 1), with burrows, mottling, and laminations all recorded.

2-25 Clarke (2014) Chapter 2 ). The boundary surface is indicated with a dashed black line. Yellow boxes show close up location of sediment core photos (Fig. 2.9). photos (Fig. close up location of sediment core show boxes Yellow ). The boundary surface is indicated with a dashed black line. -3 Core logs of gravity cores (GC) analysed in this study collected from the canyon region of the study area: a) GC1, b) GC2, c) GC3, and of the study area: region the canyon (GC) analysed in this study collected from logs of gravity cores Core Figure 2.10: Figure d) GC12. Core logs show radiocarbon ages, core photos, and physical properties with depth. Physical properties include mean grain size ( μm ) and unit properties include mean grain size photos, and physical properties with depth. Physical radiocarbon ages, core logs show d) GC12. Core (kNm weight Legend is shown in Fig. 2.11. in Fig. Legend is shown

2-26 Clarke (2014) Chapter 2

Figure 2.11: Core logs of gravity cores (GC) analysed in this study collected from the plateau region of the study area: a) GC5, b) GC6, c) GC7, d) GC8, and e) GC11. Core logs show radiocarbon ages, core photos, and physical properties with depth. Physical properties include mean grain size (μm) and unit weight (kNm-3). The boundary surface is indicated with a dashed black line. Yellow boxes show close up location of sediment core photos (Fig. 2.12).

2-27 Clarke (2014) Chapter 2

Figure 2.12: Close up images and interpretation of the boundary surfaces analysed in this study: a) GC5, b) GC6, c) GC7, d) GC8, e) GC11 and f) GC12. See Fig. 2.1, 2.2 for core locations and Fig. 2.7, 2.8 for location of the photo down each core. The interpreted boundary surface is indicated with a dashed black line. Radiocarbon ages for each core are shown in yellow (ka = thousand years before present, RCD = radiocarbon dead).

Cores offshore southern QLD (GC1, GC2, GC3) generally appear to have a relatively thicker Unit 1 than cores obtained from offshore northern NSW (GC6, GC7, GC8, GC11, GC12). Unit 1 in the southern cores were typically around 1 m, substantially thinner than the corresponding Unit 1 from slide cores in the north, which accounted for the entirety of cores.

2.5.2.2 Carbonate Content (TIC) and Total Organic Carbon (TOC) Carbonate content values remain fairly constant throughout both regions at around 17-22% range (average 20%). Organic carbon content ranges between 4-12% in all cores (average 7%). While TOC may seem small, it could be considered high enough to have some impact on the strength of the sediment, with Bishop (2009) showing that TOC values as low as 3% can affect mechanical strength.

2-28 Clarke (2014) Chapter 2

2.5.2.3 Grainsize (Fig. 13) Sediments comprise mixtures of calcareous and terrigenous sand (15-40%), silt (50-65%) and clay (10-20%). While relative proportions of clay, silt, and sand vary to some degree between each core, the general uniformity of the sediments is evident across all cores, with the majority of the sediments classified as either clay/sand-bearing silts, clay-bearing sandy silts, or clay-bearing silty sands.

The northern region sediments are characterized by lower sand contents (<30%) compared to the southern region sediments (up to 60%). Clay contents remain low in all samples (<25%), and silt is the dominant grainsize. Sediments from above and below boundary surface generally display no significant change in grainsize characteristics.

Core sediments can be grouped into three main types: i. Clay/sand-bearing silts - northern cores ii. Clay-bearing sandy silts - southern cores

Figure 2.13: Grainsize analysis results for 142 sediment samples from 12 cores: a) Sand:Silt:Clay ratio diagram; b) Grainsize distribution plot (log scale). Coloured markers (a) or lines (b) represent “typical” responses for each main sediment type: Type 1: clay-sand bearing silts (green); Type 2: clay bearing sandy silts (blue and orange); and Type 3: clay bearing silty sands (red).

iii. Clay-bearing silty sands - southern cores Small variations in textural types and average grainsize are usually dependent upon the depth of the sample within the core, the spatial location of the core, and whether the sample was taken above a boundary surface (i.e. from the recent sediment drape) or below a boundary surface (i.e. from the underlying, older sediment). Gravel content was negligible or non-existent in all samples, with grain

2-29 Clarke (2014) Chapter 2 sizes ranging from fine clays to coarse sands. 2.5.2.4 Density Change Small but distinct changes in some sediment physical properties (bulk density, water content, unit weight) are recorded above and below the boundary surface. If no boundary surface is detected in the core and a uniform sequence of hemipelagic layers is presented, a gradual increase in bulk density and unit weight, and decrease in water content is shown down core. The variation and apparent disconnection of these properties above and below the boundary is probably related to different burial depths, with sediments below the boundary presenting densities consistent with compaction due to burial at 5-10 metres greater than their present depth below the sea floor (see Chapter 3).

The increase in density and stiffness of the sediment of the lower unit (Unit 2) when compared to the upper unit (Unit 1) was measured in four cores (GC7, GC8, GC11, GC12), with unit weight values increasing between 0.05-0.14 gcm-3 across the boundary (see unit weight in Fig. 2.10 and 2.11 core logs). There was insufficient sediment below the boundary surface to test this phenomenon in GC6, although its close proximity to GC11 makes it reasonable to infer that the trend would follow. In contrast, GC5 presents a slight decrease in unit weight directly below the boundary but the sediment density then increases down the core (unit weight decreases from 1.62 gcm-3 to 1.55 gcm-3 across the boundary and returns to 1.65 gcm-3 30 cm down core).

2.5.3 Radiocarbon Ages – Dating the Boundary Surface and Determining Sedi- mentation Rates

The14 C ages obtained from planktonic foraminifera samples and bulk samples (Table 2.2) span from 2459 cal yr B.P. onwards and are consistent with the previously established stratigraphy based on Clarke et al. (2012).

2.5.3.1 Age Discontinuities – Dating the Boundary Surface Multiple occurrences of obvious age discontinuities occur across identified boundary surface in five different cores (GC8, GC7, GC5, GC11, GC12) taken from four separate slides (Coolangatta-1, “GC7 Slide”, Cudgen, and Byron). Ages taken from either side of boundary surface indicate a significant time gap in deposition, suggesting distinct units. Sediment sampled from the basal layer of the overlying Unit 1 sediment return ages between 12.3 ka to 23.7 ka, while sediment sampled from the underlying unit directly below the boundary surface dates at 45.6 ka to >50 ka (see Table 2.2 and Fig.

2-30 Clarke (2014) Chapter 2 n o i t a m r o f n I l a n o i t i d d A No boundary surface No boundary surface No boundary surface No boundary surface No boundary surface No boundary surface No boundary surface No boundary surface No boundary surface No boundary surface Above boundary Below surface boundary surface Above boundary surface Above boundary surface Above of picked sample Replicate boundary Below surface boundary Below surface boundary Below surface boundary Below surface boundary Below surface boundary Below surface boundary Below surface boundary Below surface boundary Below surface boundary Below surface d o h C C C C C C C C C C C C C C C C C C C C C C C C C t 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 e M g n AMS AMS AMS AMS AMS AMS AMS AMS AMS AMS AMS AMS AMS AMS AMS AMS AMS AMS AMS AMS AMS AMS AMS AMS AMS i t a D

l a i r e t a e p m y e t Bulk Bulk Bulk Bulk Bulk Bulk Bulk Bulk Bulk Bulk Bulk Bulk Bulk Bulk l Foram Foram Foram Foram Foram Foram Foram Foram Foram Foram Foram p m a S r o r r σ ) ------e

2 ±77 ±62 ±90 ±92 ±94 ±94 ±65 ( ±107 ±109 ±105 ±154 ±188 ±661 C ±2858 ±1884 ±1154 ±1560 ±1240 ±1280 4 1

d e t ) a P r b B i l ( a e g c 8268 4443 6035 6364 2442 n 45599 11989 16429 13910 19144 22145 41025 41572 47857 47650 48333 42378 11658 12268 A a >50000* >50000* >50000* >50000* >50000* >50000* i C d 4 1 e M

% 4 . 5 9 ) ( 8351 4536 6163 6471 2601 48609 12202 16817 14056 19403 22416 42476 42445 49678 49337 49896 42911 11905 12398 e P g B n ( a ) r y e t i ------g l i a b d a e b t o a r r p b i l a 8173 4342 5919 6272 2334 43272 11732 15860 13764 18894 21717 38985 40553 46155 46145 46726 41857 11383 12061 σ c C yrs BP were converted into calibrated calendar ages following Stuiver and Reimer and Reimer Stuiver converted into calibrated calendar ages following C yrs BP were 2 14 r o r r σ ) e

1 ±38 ±54 ±55 ±31 ±45 ±53 ±77 ±94 ±46 ±47 ±47 ±33 ( ±577 ±942 ±780 ±620 ±770 ±770 ±640 ±330 ±930 C ±1429 ±3600 ±1800 ±1500 4 1

C 4 1

l ) a P n B o i ( t e n 7809 4323 5644 5965 2750 g e 10665 13738 12479 16394 18945 42559 37002 10511 36505 10820 51800 45090 44910 48360 48250 45550 53100 38200 51700 50510 v A n o C h ) t p m e 43 21 11 91 20 c 41* 41* 120 220 320 295 391 102 113 191 376 430 98.5 ( D D I e l p m a S D I b a L OZP951 GC7/1E/10-12cm OZP952OZP953 GC7/1E/40-41cm OZP954 GC7/1E/53.5-55cmOZP955 GC7/1E/90-92cm OZP956 GC7/2D/148-152cm 54.25 OZP957 GC7/2D/189-192cmOZP958 GC7/3C/234-237cm 150 OZP959 GC7/3C/280-283cm 190 OZP960 235.5 GC7/4B/334-337cmOZP961 281.5 GC7/4B/374-378cm 335.5 GC7/5A/428-432cm Wk27865 GC6/2A/191cm Wk27866 GC7/1E/41cm UBA22880UBA22881 GC1/3C/120cm UBA22882 GC1/4B/220cm UBA22883 GC1/5A/320cm UBA22884 GC2/1E/42-44cm UBA22885 GC2/2D/97-100cm UBA22886 GC2/3C/197-200cmUBA22887 GC2/4B/294-296cm UBA22888 198.5 GC2/5A/390-392cm UBA22889 GC5/2B/102cm GC5/2B/113cm UBA22890 GC7/1E/21cm UBA22879 GC1/2D/20cm Radiocarbon ages of 49 core samples. Conventional samples. Conventional 2.2: Radiocarbon ages of 49 core Table (1993). Median calibrated ages (BP) were calculated with CALIB V6.1.1 (Stuiver et al. 2005) using marine calibration curvedata set Marine09.14c V6.1.1 (Stuiver calculated with CALIB calibrated ages (BP) were (1993). Median (see text for details). with 2 σ errors here of 11 ± 13 yr and reported (ΔR) value et al. 2009) with a reservoir correction (Reimer

2-31 Clarke (2014) Chapter 2 n o i t a m r o f n I l a n o i t i d d A Below boundary Below surface boundary surface No boundary surface Above boundary surface No boundary surface Above boundary surface No boundary surface Above boundary surface No of picked sample Replicate boundary surface No boundary surface Above boundary surface No of picked sample Replicate boundary surface No boundary surface Above boundary surface No boundary Below surface boundary surface No boundary surface Above boundary surface Above boundary Below surface boundary Below surface boundary Below surface boundary surface Above boundary surface Above boundary surface Above boundary surface Above boundary surface Above boundary Below surface of picked sample Replicate boundary Below surface boundary Below surface boundary Below surface boundary Below surface boundary surface Above boundary Below surface boundary surface Above boundary Below surface of picked sample Replicate boundary Below surface boundary surface Above boundary Below surface boundary surface Above boundary Below surface boundary Below surface boundary Below surface boundary Below surface boundary Below surface boundary Below surface d o h C C C C C C C C C C C C C C C C C C C C C C C C C C C C C C C C C C C C C C C C C C C C C C C C C t 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 14 e M g n AMS AMS AMS AMS AMS AMS AMS AMS AMS AMS AMS AMS AMS AMS AMS AMS AMS AMS AMS AMS AMS AMS AMS AMS AMS AMS AMS AMS AMS AMS AMS AMS AMS AMS AMS AMS AMS AMS AMS AMS AMS AMS AMS AMS AMS AMS AMS AMS AMS i t a D

l a i r e t a e p m y e t Bulk Bulk Bulk Bulk Bulk Bulk Bulk Bulk Bulk Bulk Bulk Bulk Bulk Bulk Bulk Bulk Bulk Bulk Bulk Bulk Bulk Bulk Bulk Bulk Bulk Bulk l Foram Foram Foram Foram Foram Foram Foram Foram Foram Foram Foram Foram Foram Foram Foram Foram Foram Foram Foram Foram Foram Foram Foram p m a S r o r r σ ) ------e

2 ±94 ±71 ±96 ±90 ±92 ±84 ±98 ±80 ±65 ±77 ±62 ±90 ±92 ±94 ±94 ( ±192 ±202 ±162 ±105 ±109 ±150 ±184 ±115 ±156 ±107 ±109 ±105 ±154 ±188 ±661 C ±2534 ±3164 ±2410 ±1000 ±2858 ±1560 ±1240 ±1280 ±1154 ±1884 4 1

d e t ) a P r b B i l ( a e g c 4217 5192 3685 3003 7447 8809 2442 8268 4443 6035 6364 n 46807 10214 12396 20625 23744 48078 10186 12780 21116 47170 40199 20193 20195 12700 15756 45599 11989 16429 11658 13910 19144 22145 12268 47857 47650 48333 42378 41572 41025 A a >50000* >50000* >50000* >50000* >50000* >50000* >50000* >50000* >50000* i C d 4 1 e M

% 4 . 5 9 ) ( 4375 5290 3821 3168 7541 8975 2601 8351 4536 6163 6471 49455 10352 12572 21133 24142 50000 10279 12935 21436 49632 41244 20400 20438 12876 16406 48609 12202 16817 11905 14056 19403 22416 12398 49678 49337 49896 42911 42445 42476 e P g B n ( a ) r y e t i ------g l i a b d a e b t o a r r p b i l a 4076 5048 3558 2844 7369 8623 2334 8173 4342 5919 6272 44753 10125 12203 20259 23377 45869 10096 12635 20772 45166 38944 19883 19856 12588 15172 43272 11732 15860 11383 13764 18894 21717 12061 46155 46145 46726 41857 40553 38985 σ c 2 r o r r σ ) e

1 ±47 ±35 ±48 ±45 ±96 ±46 ±42 ±49 ±81 ±53 ±40 ±55 ±75 ±92 ±57 ±78 ±45 ±33 ±38 ±54 ±55 ±31 ±46 ±53 ±77 ±94 ±47 ±47 ( ±101 ±980 ±500 ±780 ±620 ±770 ±770 ±640 ±330 ±930 ±577 ±942 C ±3966 ±1267 ±1582 ±1205 ±3600 ±1600 ±1429 ±1800 ±1500 4 1

C 4 1

l ) a P n B o i ( t e n 4157 4887 9382 3763 9349 3207 6948 8275 5520 5644 2750 7809 4323 5965 g e 10886 17732 20298 11330 18095 51825 50882 43888 35535 17410 45672 17417 11240 13463 44288 51800 10665 13738 10511 12479 16394 18945 42559 10820 45090 44910 48360 48250 45550 53100 38200 51700 50510 37002 36505 v A n o C h ) t p m 6 3 5 e 13 77 92 79 75 81 88 20 43 11 21 91 c 46* 46* 85* 26* 26* 41* 41* 141 200 221 122 206 120 220 320 295 391 102 113 376 430 191 98.5 ( 215* 215* 86.5* D D I e l p m a S GC8/1C/6cm D I b a L OZP963 GC11/3A/215cm OZP962 GC7/5A/469-472cm 470.5 OZP951 GC7/1E/10-12cm OZP952OZP953 GC7/1E/40-41cm OZP954 GC7/1E/53.5-55cmOZP955 GC7/1E/90-92cm OZP956 GC7/2D/148-152cm 54.25 OZP957 GC7/2D/189-192cmOZP958 GC7/3C/234-237cm 150 OZP959 GC7/3C/280-283cm 190 OZP960 235.5 GC7/4B/334-337cmOZP961 281.5 GC7/4B/374-378cm 335.5 GC7/5A/428-432cm Wk27868 Wk27867 GC8/2B/85cm Wk27869 GC11/1C/3cm Wk27870 GC11/3A/206cm Wk27871 GC12/1B/5cm Wk27872Wk27873 GC12/1B/81cm GC12/1B/88cm Wk27865 GC6/2A/191cm Wk27866 GC7/1E/41cm UBA18917UBA18918 GC8/1C/13cm UBA22891 GC8/1C/46cm UBA18919 GC8/1C/46cmR UBA18920 GC8/2B/77cm UBA18921 GC8/2B/85-88cm UBA22892 GC8/2B/92cm UBA22893 GC11/2B/79cm UBA18926 GC11/2B/141cm GC11/3A/200cm UBA18927 GC11/3A/215cm UBA18928 GC11/3A/221cm UBA18930UBA22894 GC12/1B/26cm UBA18932 GC12/1B/26cmR GC12/1B/75cm UBA19514 GC12/2A/122cm UBA22879UBA22880 GC1/2D/20cm UBA22881 GC1/3C/120cm UBA22882 GC1/4B/220cm UBA22883 GC1/5A/320cm UBA22884 GC2/1E/42-44cm UBA22885 GC2/2D/97-100cm UBA22886 GC2/3C/197-200cmUBA22887 GC2/4B/294-296cm UBA22888 198.5 GC2/5A/390-392cm UBA22889 GC5/2B/102cm GC5/2B/113cm UBA22890 GC7/1E/21cm

2-32 Clarke (2014) Chapter 2

2.10-12). Ages close to or at 50 ka are at the limit of the 14C technique and could be radiocarbon dead, while those greater than 50 ka are considered radiocarbon dead. We also suspect that bioturbation has introduced a small amount of modern sediment into the older sediment below the boundary, which could potentially result in a younger age being detected.

Small differences among ages above the boundary surface between gravity cores from the same slide (i.e. multiple cores with boundary surface taken from the Cudgen Slide, GC5 and GC11) can be ascribed to a consequence of differing sample types (e.g. GC5 – bulk; GC11 foram picked). 2.5.3.2 Sedimentation Rate 14C ages were used to establish sedimentation rates for the northern NSW/southern QLD continental slope using material from six gravity cores: five cores from within submarine landslides (GC2, GC7, GC8, GC11 and GC12) and one adjacent reference core (GC1) (Fig. 2.14; see Fig. 2.2 for core locations).

Assuming continuous and constant sedimentation, the upper slope offshore Bribie Island within the northern canyon region (southern QLD) was determined to have a sedimentation rate between 0.21 mka-1 (GC1; 4 samples) and 0.24 mka-1 (GC2; 5 samples) (Fig. 2.14). The upper slope offshore Coolangatta/Cudgen on the plateau region (northern NSW) returned sedimentation rates of 0.04 m/ka (GC8; 6 samples), 0.02 mka-1 (GC7; 2 samples), and 0.12 mka-1 (GC11; 5 samples). The upper slope offshore Byron within the southern canyon region (northern NSW) returned a sedimentation rate of 0.06 mka-1 (GC12; 4 samples). These rates apply over the last ~20 ka (ages used to determine the sedimentation rate are all <23 ka and all lo- 14 Figure 2.14: Sedimentation rates determined from C ages (see Table cated above the boundary features). 2.2) within 6 gravity cores (GC1, GC2, GC7, GC8, GC11, and GC12). Sedimentation rate values range between 0.21-0.24 mka-1 2.5.3.3 Comparison of Bulk for the northern cores (GC1, GC2), and 0.02-0.12 mka-1 for the Sample Dates with Picked Fo- southern cores (GC7, GC8, GC11, and GC12). These rates apply raminifera Dates over the last ~20 ka (ages used to determine the sedimentation rate are all <23 ka and all located above the boundary surfaces – see text Age correspondence between for details).

2-33 Clarke (2014) Chapter 2 bulk sediment and foram picked samples was investigated using five sample pairs (see Table 2.2) to test the difference/error between the two different sampling techniques. Bulk ages are usually deemed acceptable or not depending on the depositional environment of the sediment and the type of material tested (Harney et al. 2000). If sediment re-working and/or turbidite deposition is characteristic of the site then error associated with bulk samples would be much greater. There is no obvious physical reworking of the sediment evident. Neither is graded bedding characteristic of turbidite deposition evident in any of the cores from this region, apart from GC7, which does present a reworked unit (Unit 2, 28-48 cm). There is good correspondence between the ages of foram picked samples and bulk samples for three out of the five sample pairs tested (GC8, GC11, GC12), with ages differences ranging between <1 ka to 2.18 ka for those three pairs. One sample pair (GC8) showed moderate correspondence, with an age difference of 3.2 ka. The fifth sample pair (GC7) showed little to no correspondence with a recorded age difference of approximately 8.4 ka. This sample pair was taken from GC7 Unit 2 (41 cm down core) within a reworked unit and is therefore expected to have a greater error range when using a bulk sample. The bulk sample ages should be considered acceptable and representative of the age of the sediments, so long as the samples are not taken from within reworked units.

2.6 Discussion

2.6.1 Sedimentation Style and Rates There is no major difference in the texture of the deposited material either down individual cores or between the 10 different sites located in this slope. Grainsize, carbonate and organic carbon content present no significant change between units across the boundary features. While displaying an obvious change in colour, stiffness, density and water content between the upper and lower units, grainsize analysis reveals that the sediment composition remains generally uniform with all materials in the cores presenting as poorly sorted, medium to coarse sandy silts, with relatively low clay and relatively high silt and sand percentages. Carbonate and organic carbon content also remains fairly uniform, with carbonate content varying between 17% and 22% by weight and organic carbon content less than 10% by weight. These trends in the physical and mechanical properties are consistent across all 10 split gravity cores.

Given the age of the eastern Australian margin since subsidence ended (>50 million years) and sediment thickness of between 500 m to 1000 m, the average sedimentation rate for the margin is

2-34 Clarke (2014) Chapter 2 between about 10 and 20 metres per million years (0.01-0.02 mka-1). This low rate has been ascribed to a combination of factors such as low mainland sediment flux, limited accommodation space, and reworking by the strong, southward moving sediment transport by the East Australian Current (Keene et al. 2008). Since rifting ceased ca. 52 Ma, the southeast section of the east Australian margin has experienced minimal subsidence (Boyd et al. 2004), which is expressed by the convergence and apparent combination (or “pinching-out”) of post-rift seismic reflectors on the shelf and uppermost slope (Marshall et al. 1979; Keene et al. 2008; Fig. 5) which indicates that the present day inner-shelf has been maintained at an elevation near sea-level (within 200 m) during the tertiary.

In contrast to this low, approximate long-term determination of the rate of sediment accumulation on the upper slope Troedson and Davies (2001) have directly determined sedimentation rates for Late Quaternary to Recent upper slope sediment deposits offshore Noosa Heads (southern QLD) and Sydney (central NSW) from three gravity core samples (<4.5 m). Offshore Noosa Heads, a sedimentation rate of 80 m per million years (0.08 mka-1) was recorded from the last glacial lowstand, while the last post-glacial transgression produced more rapid sedimentation rates of 150 - 240 m per million years (0.15-0.24 mka-1). Offshore Sydney, mean sedimentation rates are reported to be between 20 - 50 m per million years (0.02-0.05 mka-1) over the last 71 thousand years, with mean glacial/intersta- dial rates higher than Holocene rates by a factor of 1.36 (Troedson and Davies, 2001). Sedimentation rates around 0.02-0.11 mka-1 have been calculated from radiocarbon ages from gravity cores described by Glenn et al. (2008) from the central NSW continental slope. It is of interest to note that the directly determined sediment accumulation rates are generally an order of magnitude higher than the whole- of-Tertiary rate determined above, and are also quite low in comparison to reported rates of sediment accumulation worldwide (Urlaub et al. 2013); e.g. 0.2 mka-1 for the Atlantic margins of North and South America, 0.17-36.0 mka-1 for Europe and 0.12-0.25 mka-1 for Western Africa.

The late Pleistocene and Holocene sedimentation rates reported in this work are comparable with previous directly determined sedimentation rates reported for the margin and are of the same order of magnitude (Fig. 2.14). Offshore central and northern NSW have a slightly lower sedimentation rate (~0.02-0.12 mka-1), with sediments offshore southern QLD being deposited at a slightly higher rate (~0.15-0.24 mka-1). In comparison to approximate long-term determination of the rate of sediment accumulation on the upper slope of 0.01-0.02 mka-1, the directly determined rates for the last 50,000 years from southern QLD (~0.15-0.24 mka-1) are an order of magnitude higher.

2-35 Clarke (2014) Chapter 2

As the rate of sedimentation is expected to have been higher in the geologically recent past due to lower sea levels, the consistent rates of sedimentation suggest that either these are in some way anomalously high, or that removal of sediment accumulated on the slope is a normal process. If the rates are representative or similar to the long-term Neogene average surface accumulation then sediment removal has been a consistently occurring event since the formation of the margin 60 million years ago. Given that the sediment wedge deposit is generally less than 500 m thick and there is abundant evidence for submarine landsliding (Boyd et al. 2010; Clarke et al. 2012; Hubble et al. 2012; Hubble 2013; and this work) then it follows that the mechanism for sediment removal on this section of the margin is the current and geologically recent mass wasting. Assuming constant sedimentation rates of between 0.02-0.20 mka-1 on the margin since its formation, a sediment deposit between ~1-3 km is “missing” from the margin, suggesting that much of the sediment is moved off the shelf and slope and down to the abyssal plain in the geologically recent past. If these rates are indicative of average conditions during the Pleistocene, a full glacial cycle (~100 ka) is represented by ~20 ms ± 10 ms of subbottom profiler record, or somewhere between 8 to 24 m of sediment (e.g. Fig. 2.8a).

2.6.1.1 Limitations All sedimentation rate calculations have been on cores that are <5 m in length, which is far from ideal. Only a small part of the overall record can be investigated, making it necessary to extrapolate back over a long period of geological time, thereby creating an overreliance on the assumption that rates have remained relatively constant over the entire period. Additionally, these calculations do not take into account the possibility of sediment removal from the slope, resulting in an underestimation of the actual sedimentation rate. Given the prevalence of mass wasting along this margin, submarine landslide events would more than likely affect the calculated sedimentation rate due to the removal of near-surface sediment layers.

2.6.2 Significance of Boundary Surfaces One slide-adjacent core, and four cores from within the Coolangatta-1, Cudgen, and Byron Slide scars in the south of the study area all present a boundary surface which are located at depths of 0.8 to 2.2 meters below the present-day seafloor and are all identified by a sharp, colour-change boundary; discernable by small increases in sediment stiffness; as well as a slight increase in sediment bulk density of 0.1 gcm-3. Distinct gaps in AMS 14C age of at least 20 ka are recorded across the boundary surface in most cases.

2-36 Clarke (2014) Chapter 2

In all five cases these boundaries are paraconformities or obvious disconformity surfaces representing age hiatuses recorded by the 14C age gaps. Coupled with the small but distinct down-core increases in sediment density, this suggests that the surfaces probably represent the removal of material.

The subbottom profiles of the Coolangatta-1 (Fig. 2.7c) and the Cudgen Slides (Fig. 2.8, 2.9), demonstrate that the boundary surface is unlikely to be a major slide surface, although they may represent a surface from which near-surface layers have been detached, either as slabs or possibly removed due to liquefaction of the near surface sediments (Steedman and Sharp 2001; Ozener et al. 2009) or scour erosion. The boundary surfaces therefore do not enable a direct estimation of the time of slide occurrence. They do, however, enable dating of a relatively recent disturbance event, and the termina- tion of the youngest reflectors in the subbottom profiles across the Cudgen Slide scar suggest that these slope failure events are geologically recent enough to be the event which removed significant amounts of material downslope. We suggest that the boundary surfaces can be used to constrain the timing of sliding events with some confidence and that the boundary surfaces themselves probably date either the disturbance associated with slide motion or a penecontemporaneous post-slide erosion event.

If a disturbance/hiatus relationship or a post-disturbance erosion event is accepted, it can be inferred that the consistent dates for the boundary surfaces at difference sites (~12 ka (GC7), ~15 ka (GC12), and ~20-23 ka (GC5, GC8, and GC11)) suggests a regional erosional or disturbance event, with the sliding age to be constrained as the maximum ages for slide occurrence. The truncation of the material shown in subbottom profiles shows that the slides are young (Fig. 2.7-2.9). Slide headscarps are shown to rupture the modern bedding and this rupturing of the near surface sediment layers indicates that the oldest age the thinner (20-50 m) translational slides can be is ~100 ka (this is based on the calculated recent surface sedimentation rates, a deposit ~20 m thick has a maximum age of ~100 ka).

The truncation of near surface sediment packages and lack of recent sedimentation above slide features for both the Coolandatta-1 and Cudgen submarine landslides suggests they are also geologically recent. If this is the case then the surfaces identified in the Coolangatta-1 and Cudgen within- slide cores are either: a) erosional features that developed after the occurrence of the landslide in which case the hiatus surface age provides a minimum age for landslide occurrence or b) detachment surfaces from which slabs of near-surface sediment were removed during landsliding in which case the post-

2-37 Clarke (2014) Chapter 2 hiatus sediment dates indicates approximately when landsliding occurred. In either case, it is reasonable to suggest that these two spatially adjacent slides occurred penecontemporaneously approximately 20 ka years ago.

A number of thin planar translational slab slide blocks have been removed from the seafloor upslope of the major slide scarps of the Cudgen Slide (Fig. 2.8, 2.9). These translational slides are significant because they indicate that subsequent to the large rotational Cudgen Slide failure there has been a continued retrogressive dislocation and removal of young sediment slabs which were present prior to the formation of the Cudgen rotational slide surface. The translational slides present as retrogressive features with headwall scarps presenting layers that are “unsupported” (i.e. truncated at the headscarp) and could possibly slide in the future with further retrogressive failures.

2.6.3 Possible Triggers A variety of causes for the initiation of submarine landslides have been suggested (e.g. Locat and Lee 2002; Masson et al. 2006). These include: earthquakes, storm wave loading, erosion and in particular slope over-steepening, rapid sedimentation leading to under-consolidation, the presence of weak layers, gas dissociation, sediment creep, sea-level changes, glaciations/isostatic uplift, volcanic activity, and diapirs. It is also widely accepted that a combination of these factors is required to initiate a landslide, especially where these occur on very shallow slopes. There are data indicating that several large landslides have coincided with earthquakes, (e.g. Tappin et al., 2001; Synolakis et al., 2002; Bar- det et al., 2003; Masson et al., 2006), and the role of weak layers, oriented parallel to the sedimentary bedding, has long been used to explain the scale of some large slides, similarly the importance of weak layers in controlling sliding at all scales has been suggested by several works (Masson et al. 2009).

Observations of the widespread occurrence of submarine slides suggests that weak clay layers are probably not be a major causal component in the case of the eastern Australian margin slides and tends to favour earthquakes as the triggering mechanism. The sediments of this area are ubiquitously uniform and the similarity of the hemipelagic muds from a relatively large geographic area has been noted by the works that describe them (40 cores - Hubble and Jenkins 1984a,b; Troedsen 1998; Tro- edsen and Davis 2001; Glenn et al. 2008).

The wide occurrence of upper slope slides across the east Australian margin indicates that sub-

2-38 Clarke (2014) Chapter 2 marine sliding should be considered to be a common characteristic of this passive continental margin. This indicates that one or more of the potential triggering mechanisms can operate in passive margin settings to destabilise the slope. The processes suspected to be most likely include: 1) reduction of the shear strength of the upper-slope sediments to very low values, possibly induced by creep or a build-up of pore-pressure; 2) long-term modification of the slope-geometry i.e., sedimentation on the head of the slope and/or erosion of the toe of the slope; and/or 3) seismic events large enough to trigger sediment liquefaction or a sudden increase of pore-fluid pressure.

The major contender – earthquakes leading to a loss of sediment strength - for triggering submarine landslides can be investigated with respects to the east Australian margin as a set of the following three questions:

1. Is submarine landsliding along the east Australian margin related to sea-level change?

The effects of climate change over hundreds of years, particularly with regards to sea-level fluctuations, have previously been identified as a possible long-term triggering mechanisms for submarine landslides and slumping, contributing to slope instability (Sultan et al. 2004; Masson Figure 2.15: a) 14C age frequency distribution plot: a total et al. 2006). With this in mind, it is of 49 samples are plotted (31 samples above the boundary surface, 18 samples below) into 10-year bins. Samples located important to determine when major above the boundary surfaces cluster in <30 ka age ranges, sea-level fluctuations have occurred while samples taken below the boundary surfaces fall only in the 41-50+ ka age range; b) Eustatic sea level record from in the past within the study area and Waelbroeck et al. (2002) (foraminifera isotope records using whether the timing of these events oxygen18/16 ratios) over the last 150 ka. Stars highlight the location of approximate 14C age groupings from the gravity correspond to the dates of the slid- cores samples: 12 ka (blue), 15 ka (orange), 20 ka (yellow), and 50 ka (green). See Table 2 for the full list of 14C ages.

2-39 Clarke (2014) Chapter 2 ing events.

Lee (2009) has suggested that landslides were more frequent during and just after the last glaciation maximum than they are today. One of the suggested reasons for this is that glacial maxima coincide with periods of low relative sea level. The relative sea level curve for Australia for the last 0.5 million years (Waelbroeck et al. 2002; see Fig. 2.15b) indicates that sea-levels have been over 100 m below present on several occasions, and the times are associated with glacial maxima. The lowered sea-level can increase the likelihood of sliding because it results in the shoreline migrating closer to the shelf edge, leading to increased erosion and higher rates of sedimentation offshore, which occurs directly onto the slope. The lower water pressures (and possibly change of temperatures) can lead to release of gas from gas hydrates increasing pore pressures and reducing strength, although there is little evidence of gas hydrates in this section of the east Australian margin, and related changes to stress levels in the crust can increase seismic activity (King et al. 1994; Clark 2010). Additionally, increased groundwater flows from underlying rocks can occur and contribute to reduced strength. It has also been suggested that changes to deep ocean currents are associated with glaciations and erosion from these currents can contribute to toe erosion and slope steepening as suggested by Hubble et al. (2012).

AMS 14C ages determined for sediment sampled directly above the boundary surfaces do not appear to correlate sliding with any one particular sea-level event (Fig. 2.15b). The oldest above- boundary surface age (~20 ka) correlates to around the time of the most recent lowstand; the next above-boundary surface age (~15 ka) correlates to the beginning of the rising limb of sea-level rise; and the third above-boundary surface age (~12 ka) corresponds to mid-way up the rising limb of sea- level rise, moving towards more stable sea levels (Fig. 2.15). These data are not conclusive, especially given the limit in both the datable age of the sediment (50 ka) and maximum depth for the sediment cores (5 m), but are not consistent with lowered sea-level as a primary factor inducing sliding. This is consistent with Urlaub et al’s (2013) recent conclusion that sea-level change does not appear to correlate with submarine landslides occurrence and that sea-level changes are a contributing factor only if sediments cannot dissipate pore pressure. However, the age database for submarine landslides is still extremely limited.

2-40 Clarke (2014) Chapter 2

2. Do gas hydrates play a role in triggering submarine landsliding along the east Aus- tralian margin?

Gases formed by the dissociation of methane hydrate can dissipate though porous permeable sediments that drain relatively freely (such as granular silts). Gas hydrates are usually located much deeper beneath the seafloor (i.e. >500 m), and while gas build-up is recognized as a contributing causal factor for some submarine landslides (e.g. Storegga Slide or Ruatoria; Collot et al. 2001) they generally relate to deep, very large, slope failures (Collot et al., 2001).However, provided that there is a source of gas, gas hydrates are able to precipitate in a sediment at a more shallow depth (< 500 m) if the water in the sediment occurs at a depth and temperature range which is sufficient to take it below the solid gas hydrate stability zone. The process of gas hydrate dissociation is not considered to be likely in the case of the upper eastern Australian slides due to the more shallow presentation of the slides combined with the low clay content (<20%), relatively loose packing (low density), and high porosity of the materials considered by this study.

Our bathymetric data show no obvious evidence of gas hydrate dissociation for this section of the continental slope. There have been circular depressions, referred to as pockmarks, reported for the continental margin offshore Newcastle (approximately 400 km south of the study area), which are believed to be associated with gas leakage from the underlying Permian coal measures (Glenn et al. 2008); however, similar features are not observed in the study area.

Biogenic (shallow) gas formation could also impact the stability of the slopes, causing near- surface, superficial failure of the slopes (Canals et al. 2004). Biogenic gas is usually identifiable in high-resolution seismic profiles by bright spots; however to date, no such evidence has been observed.

3. Do the grain-size distributions of the sediments from the east Australian margin make the slopes prone to earthquake-induced strength loss, triggering submarine landsliding?

The non-cohesive, granular soils recovered from on the eastern Australian margin described in this work (i.e. silty sands/ sandy silts) present grainsize distributions similar to those materials demonstrated to be prone to earthquake induced strength loss that results in liquefaction of the upper sediment layers (Steedman and Sharp 2001; Ozener et al. 2009; Wiemer et al. 2012). This, combined with the ubiquity of submarine landslides on the east Australian margin and an increased incidence of earthquakes during the Pliocene and Pleistocene due to the ongoing collision of the N-W Australia

2-41 Clarke (2014) Chapter 2 with SE Australian (c.f. Müller et al. 2012) as suggested by Hubble et al. (2012), favours an earthquake triggering mechanism for the initiation of submarine landslides on this passive continental margin.

Large onshore and offshore earthquakes have the potential to trigger a submarine landslide on the east Australian continental slope given that the narrowness of the continental margin (30-70 km; Boyd et al. 2004). In the context of this study, ground shaking from all sources (i.e. not only those able to rupture the ground surface) have the potential to destabilize or liquefy continental slope sediments and induce a submarine landslide event. The Australian fault catalogue is not considered to be complete, this is especially so for the continental shelf and slope (Clark 2009) and there is no existing evidence for sediment liquefaction from the catalogue of onshore earthquakes. We are not aware of evidence for liquefaction occurring in unconsolidated sediment deposits on the Australian east coast due to earthquake shaking although there are many such Quaternary sediment deposits present on the eastern seaboard that contain materials that would liquefy if subjected to a large enough seismic event (e.g. around Botany Bay; c.f. Sydney Geological Map Sheet notes; McNally 1998). This makes it difficult to determine with confidence the number of earthquakes that have occurred on this margin that have had the potential to induce sediment liquefaction or a seismically induced loss of shear strength which could cause submarine landslide events. Despite the uncertainties in the earthquake record and the dearth of knowledge about the location and distribution of the eastern Australian margin’s offshore

Figure 2.16: The 500 year return period PGA (0.0 s RSA period) hazard map calculated using the maximum hazard value of the combined Regional and Background hazard map (Regional_robust) and the Hotspot hazard map in (a) 2D and (b) 3D. Taken from Burbridge et al. (2012).

2-42 Clarke (2014) Chapter 2 faults, it is reasonable to expect that there are sufficient of these faults and that they have generated enough earthquakes to explain the large number of submarine landslides reported along the east Aus- tralian continental slope.

Peak Ground Acceleration (PGA) threshold for the generation of slope failure for the Hikurangi Margin in the South Pacific region ranges from 0.08 to 0.1 g (Patton et al. 2013; Pouderoux et al. 2014 and reference therein). PGA values for eastern Australia are lower than this and range between 0.02 and 0.05 g (see Fig. 2.16) at the locations of the slope failures along the east Australian margin.

2.7 Conclusions

Specific conclusions are as follows:

1. Calcareous hemipelagic muds, comprised of mixtures of calcareous and terrigenous clay (10-20%), silt (50-65%) and sand (15-40%) and generally uniform in appearance, have been sampled in cores taken from within, or adjacent to five submarine landslides on eastern Australia’s continental slope in 450 to 1150 m of water depth.

2. Dating of conformably deposited material indicates a range of sediment accumulation rates between 0.017mka-1 and 0.2 mka-1 which are consistent with previous estimates reported for this area.

3. Subbottom profiler records of transects through three within-slide core-sites and their nearby landslide scarps (Coolangatta-1 and Cudgen slides) indicate that the youngest identifiable seismic reflectors located upslope of these slides terminate on and are truncated by slide rupture surfaces and that the studied slides are geologically recent. There is no obvious evidence in the subbottom profiles for a post-slide sediment layer draped over or otherwise burying slide ruptures or exposed slide detachment surfaces.

4. Boundary surfaces were sampled in one slide-adjacent core, and four within-landslide cores at depths of 0.8 to 2.2 metres below the present-day seafloor. These boundary surfaces are identified by a sharp, colour-change boundary; discernable but small increases in sediment stiffness; and slight increases in sediment bulk density of 0.1 gcm-3. Distinct gaps in AMS 14C age of at least 25 ka are recorded across the boundary surfaces.

5. The boundary surfaces are either: a) erosional features that developed after the occur-

2-43 Clarke (2014) Chapter 2

rence of the landslide in which case the hiatal surface age provides a minimum age for landslide occurrence or b) detachment surfaces from which slabs of near-surface sediment were removed during landsliding in which case the post-hiatus sediment dates indicate approximately when landsliding occurred. In either of these scenarios, it is reasonable to suggest that these two spatially adjacent slides occurred penecontemporaneously approximately 20 ka years ago.

6. While an earthquake triggering mechanism is favoured for the initiation of submarine landslides on the eastern Australian margin and is consistent with available data, this causal mechanism cannot be conclusively demonstrated.

Acknowledgments

We would like to acknowledge: 1) the P&O crew and scientific crews of the SS2008-V12 voyage, and 2) Daniel Clark from Geoscience Australia for his valuable advice on the Australian earthquake climate. Funding for this voyage was provided by ARC Australia and ConocoPhillips Company. Funding for radiocarbon dating was provided by: 1) Professor Ron Boyd and Newcastle University, Australia; and 2) Australian Nuclear Science and Technology Organisation (ANSTO), Lucas Heights, Australia. This paper benefitted from reviews by Dr Geoffroy Lamarche and an anonymous reviewer.

2-44 Clarke (2014) Chapter 2 References

Bardet, J.P., Synolakis, C.E., Davies, H.L., Imamura, F., Okal, E.A., 2003. Landslide Tsunamis: Recent Findings and Re- search Directions Pure and Applied Geophysics 160, 1793–1809. Blott, S.J., Pye, K., 2001. Gradistat: a grain size distribution and statistics package for the analysis of unconsolidated sediments. Earth Surface Processes and Landforms 26, 1237-1248. Boyd, R., Ruming, K. and Roberts, J.J., 2004. Geomorphology and surficial sediments on the southeast Australian continental margin. Australian Journal of Earth Sciences 51, 743-764. Boyd, R., Keene, J., Hubble, T., Gardner, J., Glenn, K., Ruming, K., Exon, N., 2010. Southeast Australia: A Cenozoic Continental Margin Dominated by Mass Transport, in: Mosher, D.C., Moscardelli, L., Baxter, C.D.P., Urgeles, R., Shipp, R.C., Chaytor, J.D., Lee, H.J. (Eds.), Submarine Mass Movements and Their Consequences. Springer Netherlands, pp. 491-502. Burbidge, D.R., 2012. The 2012 Australian Earthquake Hazard Map. Record 2012/71, in: Burbidge, D.R. (Ed.). Geosci- ence Australia: Canberra. Canals, M., Lastras, G., Urgeles, R., Casamor, J.L., Mienert, J., Cattaneo, A., De Batist, M., Haflidason, H., Imbo, Y., Laberg, J.S., Locat, J., Long, D., Longva, O., Masson, D.G., Sultan, N., Trincardi, F., Bryn, P., 2004. Slope failure dynamics and impacts from seafloor and shallow sub-seafloor geophysical data: case studies from the COSTA project. Marine Geology 213, 9-72. Chaytor, J.D., J., H.G., 2000. Faulting in the Newcastle area and its relationship to the 1989 M5.6 Newcastle earthquake, Dams, fault scarps and earthquakes. Australian Earthquake Engineering Society, Australia. Clark, D., 2009. Potential geologic sources of seismic hazard in the Sydney Basin, in: Clark, D. (Ed.). Geoscience Australia Record 2009/11, p. 115. Clark, D., McPherson, A., Van Dissen, R., 2012. Long-term behaviour of Australian stable continental region (SCR) faults. Tectonophysics 566–567, 1-30. Clark, D.J., 2010. Large Earthquake Recurrance in New South Wales: Implications for Earthquake Hazard, Seismic Engi- neering - design for management for geohazards, Australian Geomechanics Society, Sydney. Clarke, S., Hubble, T., Airey, D., Yu, P., Boyd, R., Keene, J., Exon, N., Gardner, J., 2012. Submarine Landslides on the Upper Southeast Australian Passive Continental Margin – Preliminary Findings. Submarine Mass Movements and Their Consequences, in: Yamada, Y., Kawamura, K., Ikehara, K., Ogawa, Y., Urgeles, R., Mosher, D., Chaytor, J., Strasser, M. (Eds.). Springer Netherlands, pp. 55-66. Clarke, S., Hubble, T., Airey, D., Yu, P., Boyd, R., Keene, J., Exon, N., Gardner, J., Ward, S., 2014. Morphology of Aus- tralia’s Eastern Continental Slope and Related Tsunami Hazard, Submarine Mass Movements and Their Conse- quences. Springer, pp. 529-538. Coleman, J., Prior, D., 1988. Mass wasting on continental margins. Annual Review of Earth and Planetary Sciences 16, 101. Collot, J.Y., Lewis, K., Lamarche, G., Lallemand, S., 2001. The giant Ruatoria debris avalanche on the northern Hikuran- gi margin, New Zealand: Result of oblique seamount subduction. Journal of Geophysical Research: Solid Earth (1978–2012) 106, 19271-19297. Conolly, J.R., 1969. Western Tasman Sea Floor. New Zealand Journal of Geology and Geophysics 12, 310-343. Colwell, J.B., Coffin, M., 1987. Rig Seismic research cruise 13 : structure and stratigraphy of the Northeast Gippsland Basin and southern New South Wales margin, initial report. Australian Govt. Pub. Service, Canberra. Davies, P.J., 1979. Marine geology of the continental shelf off southeastern Australia. BMR Bulletin 195, 51pp. Falvey, D.A., 1974. The development of continental margins in plate tectonic theory. APEA J 14, 95-106. Fine, I.V., Rabinovich, A. B., Bornhold, B. D., Thomson, R. E. & Kulikov, E. A. (2005). The Grand Banks landslide-generated tsunami of November 18, 1929: preliminary analysis and numerical modeling. Marine Geology 215, 45-57. Gaina, C., Muller, R. D., Brown, B., Ishihara, T., 2003. Microcontinent formation around Australia in The Evolution and Dynamics of the Australian Plate, in: Hillis, R.R., Muller, R. D. (Ed.), Spec. Pap. Geol. Soc. Am., pp. 405-416. Gee, M.J.R., Gawthorpe, R.L., Friedmann, S.J., 2006. Triggering and Evolution of a Giant Submarine Landslide, Offshore

2-45 Clarke (2014) Chapter 2

Angola, Revealed by 3D Seismic Stratigraphy and Geomoephology. Journal of Sedimentary Research 76, 9-19. Gee, M.J.R., Uy, H.S., Warren, J., Morley, C.K., Lambiase, J.J., 2007. The Brunei slide: a giant submarine landslide on the North West Borneo Margin revealed by 3D seismic data. Marine Geology 246, 9-23. Glenn, K., Post, A., Keene, J., Boyd, R., Fountain, L., Potter, A., Osuchowski, M., Dando, N., Party, S., 2008. NSW Continental Slope Survey – Post Cruise Report in: Australia, G. (Ed.). Greene, H.G., Murai, L.Y., Watts, P., Maher, N.A., Fisher, M.A., Paull, C.E., Eichhubl, P., 2006. Submarine landslides in the Santa Barbara Channel as potential tsunami sources Natural Hazards and Earth System Sciences 6, 63–88. Haflidason, H., Sejrup, H.P., Nygard, A., Richter, Mienert, J., Bryn, P., Lien, R., Fosberg, C.F., Berg, K. & Masson, D.G., 2004. The Storegga Slide: Architecture, Geometry and Slide Development. Marine Geology 213, 201-234. Hampton, M.A., Locat, J., Lee, H.J., 1996. Submarine landslides. Reviews of Geophysics 34, 33–59. Harney, J., Grossman, E., Richmond, B., Fletcher Iii, C., 2000. Age and composition of carbonate shoreface sediments, Kailua Bay, Oahu, Hawaii. Coral Reefs 19, 141-154. Head, K.H., 1982. Manual of Soil Laboratory Testing. Pentech Press Limited, Devon. Heezen, B.C., 1974. Atlantic-type continental margins. The geology of continental margins: New York, Springer-Verlag, 13-24. Heiri, O., Lotter, A.F., Lemcke, G., 2001. Loss on ignition as a method for estimating organic and carbonate content in sediments: reproducibility and comparability of results. Journal of Paleolimnology 25, 101-110. Hubble, T.C.T., Packham, G.H., Hendry, D.A.F., McDougall, I., 1992. Granitic and monzonitic rocks dredged from the southeast Australian continental margin. Australian Journal of Earth Sciences 39, 619-630. Hubble, T., 2013. Voyage Summary SS2013-V01: Marine Geology and Geohazard Survey of the SE Australian Margin off Northern NSW and Southern Queensland, CSIRO, Hobart. Hubble, T., Jenkins, C.J., 1984a. Sediment Samples and Cores from the Southern New South Wales Upper Continental Slope 36’00’S-37’15’S (HMAS Kimbla November 1982), Ocean Sciences Institute Report No.8. The University of Sydney, Sydney. Hubble, T., Jenkins, C.J., 1984b. Sediment Cores from the N.S.W Continental Slope of Port Stephens, Port Macquarie and Coffs Harbour (31’-33’), Ocean Sciences Institute Report No.6. The University of Sydney, Sydney. Hubble, T., Yu, P., Airey, D., Clarke, S., Boyd, R., Keene, J., Exon, N., Gardner, J., 2012. Physical Properties and Age of Continental Slope Sediments Dredged from the Eastern Australian Continental Margin – Implications for Timing of Slope Failure. Submarine Mass Movements and Their Consequences, in: Yamada, Y., Kawamura, K., Ikehara, K., Ogawa, Y., Urgeles, R., Mosher, D., Chaytor, J., Strasser, M. (Eds.). Springer Netherlands, pp. 43-54. Huftile, G., Van Arsdale, R., Boyd, R., 1999. Pleistocene faulting offshore Newcastle, New South Wales. Earthquake En- gineering Society, Sydney. Jenkins, C.J., Keene, J.B., 1992. Submarine slope failures on the southeast Australian continental slope. Deep Sea Research 39, 121-136. Jones, R.W., 1994. The Challenger foraminifera. Oxford University Press, Oxford Keene, J., Baker, C., Tran, M., Potter, A., 2008. Geomorphology and Sedimentology of the East Marine Region of Aus- tralia, in: Australia, G. (Ed.), Canberra. King, G.C.P., Stein, R.S., Lin, J., 1994. Static stress changes and the triggering of earthquakes. Bulletin of the Seismological Society of America 84, 935-953. Lamarche, G., Joanne, C., Collot, J.-Y., 2008. Successive, large mass-transport deposits in the south Kermadec fore-arc basin, New Zealand: The Matakaoa Submarine Instability Complex. Geochemistry, Geophysics, Geosystems 9, Q04001. Lee, H.J., 2009. Timing of occurrence of large submarine landslides on the Atlantic Ocean margin. Marine Geology 264, 53-64. Leonard, M., 2008. One hundred years of earthquake recording in Australia. Bulletin of the Seismological Society of America 98, 1458-1470. Lister, G.S., Etheridge, M.A., Symonds, P.A., 1986. Detachment faulting and the evolution of passive continental margins. Geology 14, 246-250. Locat, J., Lee, H.J., 2002. Submarine landslides: advances and challenges. Canadian Geotechnical Journal 39, 193-212.

2-46 Clarke (2014) Chapter 2

Loeblich, A., Tappan, H., 1988. Foraminiferal genera and their classification, vol. 2Van Nostrand Reinhold. New York 847. Marshall, J.F., 1978. Morphology and shallow structure of the continental shelf of southern Queensland and northern New South Wales. BMR Record 1978/100, 25p. Marshall, J.F., 1979. The development of the continental shelf of northern New South Wales. BMR Journal of Australian Geology and Geophysics 4, 281‐288. Maslin, M., Mikkelsen, N., Vilela, C., Haq, B., 1998. Sea-level -and gas-hydrate-controlled catastrophic sediment failures of the Amazon Fan. Geology 26, 1107-1110. Masson, D.G., Harbitz, C.B., Wynn, R.B., Pedersen, G., Lovholt, F., 2006. Submarine landslides: processes, triggers and hazard prediction. The Philosophical Transactions of the Royal Society A (Phil. Trans. R. Soc. A) 364, 2009-2039. McAdoo, B.G., Pratson, L.F., Orange, D.L., 2000. Submarine landslide geomorphology, US continental slope. Marine Geology 169, 103-136. McCue, K., Michael-Leiba, M., 1993. Australia’s Deepest Known Earthquake. Seismological Research Letters 64, 201- 206. McMurtry, G.M., Fryer, G.J., Tappin, D.R., Wilkinson, I.P., Williams, M., Fietzke, J., Garbe-Schoenberg, D., Watts, P., 2004. Megatsunami deposits on Kohala volcano, Hawaii, from flank collapse of Mauna Loa. Geology 32, 741-744. McNally, G.H., Branagan, D.F., 1998. An overview of the engineering geology of the Botany Basin, in: McNally, G.H., Jankowski, J. (Eds.), Environmental Geology of the Botany Basin. Conference Publications, Sydney. Mitchell, J.K., Soga, K., 2005. Fundamentals of soil behavior. John Wiley & Sons Ltd, Chichester, UK. MS2000 Operators Manual (1999). Malvern Instruments, Worcestershire, England. Mosher, D.C., Moscardelli, L., Baxter, C.D.P., Urgeles, R., Shipp, R.C., Chaytor, J.D., Lee, H.J., 2010. Submarine Mass Movements and Their Consequences, Submarine Mass Movements and Their Consequences. Springer Nether- lands, pp. 1-8. Müller, R.D., Dyksterhuis, S., Rey, P., 2012. Australian paleo-stress fields and tectonic reactivation over the past 100 Ma. Australian Journal of Earth Sciences 59, 13-28. Murray, J.W., 1991. Ecology and paleoecology of benthic foraminifera. Longman, Harlow. Özener, P., Özaydın, K., Berilgen, M., 2009. Investigation of liquefaction and pore water pressure development in layered sands. Bulletin of Earthquake Engineering 7, 199-219. Patton, J., Goldfinger, C., Morey, A., Romsos, C., Black, B., Djadjadihardja, Y., Meltzner, A., 2013. Seismoturbidite record as preserved at core sites at the Cascadia and Sumatra-Andaman subduction zones. Natural Hazards & Earth System Sciences 13. Pouderoux, H., Proust, J.-N., Lamarche, G., 2014. Submarine paleoseismology of the northern Hikurangi subduction margin of New Zealand as deduced from Turbidite record since 16 ka. Quaternary Science Reviews 84, 116-131. Prior, D.B., Coleman, J.M., 1984. Submarine slope instablilty, in: Brunsden, D., Prior, D.B. (Ed.), Slope Instability. John Wiley & Sons Ltd., Norwich, pp. 419-455. Reimer, P.J., Baillie, M. G. L., Bard, E., Bayliss, A., Beck, J. W., Weyhenmeyer, C. E., 2009. Intcal09 and Marine09 radiocarbon age calibration curves, 0-50,000years cal BP. Radiocarbon 51, 1111-1150. Ringis, J., 1972. The Structure and History of the Tasman Sea and Southwest Australian Margin. University of NSW, p. 338. Roberts, J.J., Boyd, R., 2004. Late Quaternary core stratigraphy of the northern New South Wales continental shelf. Aus- tralian Journal of Earth Sciences 51, 141-156. Shaw, R.D., 1978. Sea-floor spreading in the Tasman Sea: a Lord Howe Rise – eastern Australia reconstruction. Australian Society of Exploration Geophysics 9, 75-81. Silva, A.J., Baxter, C.D.P., LaRosa, P.T., Bryant, W.R., 2004. Investigation of mass wasting on the continental slope and rise. Marine Geology 203 355-366. Steedman, R.S., Sharp, M., 2001. Liquefaction of deep saturated sands under high effective confining stress, Proceedings of 4th International Conference Recent Advances in Geotech Earthquake Engineering and Soil Dynamics. University of Missouri-Rolla, San Diego, California, pp. 441-446. Stuiver, M., Reimer, P., Reimer, R., 2005. CALIB 6.1.1 radiocarbon calibration. Execute Version 6.1.1 http://calib.qub.

2-47 Clarke (2014) Chapter 2

ac.uk/calib/. Stuiver, M., Reimer, P.J., 1993. Extended (super 14) C database and revised CALIB 3.0 (super 14) C age calibration program. Radiocarbon 35, 215-230. Sultan, N., Cochonat, P., Foucher, J.P., Mienert, J., & Sejrup, H.P., 2004. Effect of gas hydrates dissociation on seafloor slope stability. Marine Geology 213, 379-401. Synolakis, C.E., others, 2002. The slump origin of the 1998 Papua New Guinea Tsunami. Proceedings of the Royal Society of London A 458, 763-789. Symonds, P.A., 1973. The structure of the north Tasman Sea. Bureau of Mineral Resources Australia. Record 1973/167. Tappin, D.R., Watts, P., Grilli, S.T., 2008. The Papua New Guinea tsunami of 17 July 1998: anatomy of a catastrophic event. Natural Hazards and Earth System Sciences 8, 243-266. Tappin, D.R., Watts, P., McMurtry, G.M., Lafoy, Y., Matsumoto, T., 2001. The Sissano, Papua New Guinea tsunami of July 1998 - offshore evidence on the source mechanism. Marine Geology 175, 1-23. Troedson, A.L., 1998. Late Quaternary sedimentation on the East Australian continental slope: responses to palaeoenvi- ronmental change, Dept. Geology and Geophysics. University of Sydney, Sydney, p. 305. Troedson, A.L., Davies, P.J., 2001. Contrasting facies patterns in subtropical and temperature continental slope sediments: inferences from east Australian late Quaternary records. Marine Geology 172, 265-285. Ulm, S., Petchey, F., Ross, A., 2009. Marine reservoir corrections for Moreton Bay, Australia. Archaeology in Oceania 44, 160-166. Urlaub, M., Talling, P.J., Masson, D.G., 2013. Timing and frequency of large submarine landslides: implications for understanding triggers and future geohazard. Quaternary Science Reviews 72, 63-82. Varnes, D.J., 1978. Slope Movements and Types and Processes, Landslides: Analysis and Control, Special Report. Trans- portation Research Board, National Academy of Sciences, Washington, pp. 11-33. Waelbroeck, C., Labeyrie, L., Michel, E., Duplessy, J.C., McManus, J.F., Lambeck, K., Balbon, E., Labracherie, M., 2002. Sea-level and deep water temperature changes derived from benthic foraminifera isotopic records. Quaternary Sci- ence Reviews 21, 295-305. Wiemer, G., Reusch, A., Strasser, M., Kreiter, S., Otto, D., Morz, T., Kopf, A., 2012. Static and Cyclic Shear Strength of Cohesive and Non-cohesive Sediments. Woodroffe, C.D., Samosorn, B., Hua, Q., Hart, D.E., 2007. Incremental accretion of a sandy reef island over the past 3000 years indicated by component-specific radiocarbon dating. Geophys. Res. Lett. 34, L03602.

2-48 Clarke (2014) Chapter 3 Chapter 3

Geotechnical properties of submarine sediments from submarine landslides on the eastern Australian continental margin and implications for slide initiation

Samantha Clarke1, David Airey2 ,Thomas Hubble1,

1. Geocoastal Research Group, University of Sydney, Sydney, NSW, Australia

2. School of Civil Engineering, University of Sydney, Sydney, NSW, Australia. [email protected]

Abstract

Geomechanical test data are presented for 12 gravity cores, up to 5 m long, taken at sites from the upper slope (<1200 m) of the east Australian continental margin in or adjacent to five submarine landslide features. Sediments uniformly consist of olive grey to grey sandy silts (MH-ML), with clay content ranging from 7-23%. Total unit weight varies between 14.1 to 17.4 kNm-3, bulk density 715- 2065 kgm-3, water content 43-90+%, and specific gravity 2.5-2.74. Sediments present low plasticity, liquid limits 43-63%, and plasticity indices of 8.7-34%. Measured strength values, friction angle (Ф’) and apparent cohesion (c’), vary between 30-40°, and 0-10 kPa respectively. One slide-adjacent core, and four within-landslide cores present boundary surfaces located at depths of 0.8 to 2.2 meters below the present-day seafloor that are identified by a sharp, colour-change boundary; small increases in sediment stiffness; slight increases in sediment bulk density of 0.1 gcm-3; and distinct gaps in AMS 14C age of at least 25 ka. Compression testing indicates that the sediment above and below the boundary surface is slightly overconsolidated. Triaxial tests indicate a significant increase in the brittleness of the shear response of the sediment with increasing vertical stress, which would cause a progressive increase of pore pressure if the sediment was subjected to cyclic (earthquake) loading. The boundary surfaces are interpreted to represent detachment surfaces or slide plane surfaces. Slope stability models based on classical soil mechanics and measured sediment shear-strengths indicate that the upper slope sediments

3-1 Clarke (2014) Chapter 3 should be stable. However, multibeam bathymetry data reveal that many upper slope landslides occur across the margin and that submarine landsliding is a common process. We infer from these results that: a) the margin experiences seismic events that act to destabilise the slope sediments, and/or b) an unidentified mechanism regularly acts to reduce the shear resistance of these sediments to the very low values required to enable slope failure.

Keywords: southeast Australia • submarine landslides • slide plane • upper slope • continental slope • geomechanical • oedometer • triaxial • compression test

3.1 Introduction The geological record contains many examples of submarine landslides, which can vary in size from small, thin shallow slides <0.1 km3 in volume to very large, deep seated slides, such as the Storegga slides off the Norwegian coast which have a total volume of 2400-3000 km3 (Haflidason et al. 2004). Statistics on known landslides on the eastern continental slope of North America, which has geological similarities to Australia’s southeastern margin (c.f. Hubble et al. 2012; Boyd et al. 2010), have been published by Masson et al. (2006). These show that between 30°N and 45°N there are 152 large landslides (1 km3 to 392 km3) affecting an area of nearly 40,000 km2. Most failures occur on slopes of between 1° and 7°, with the greatest number of failures (~35%) occurring on slopes of 3° to 4°. The area affected by failures decreases and the volume of the failures increases with increasing slope. Water depths at which slide scars present ranges from 250 to 2500m, and the greatest number of failures (~40%) occur at water depths of around 1000 m. Despite extensive investigation on the nature and causes of these North American submarine landslides (c.f. Masson et al. 2006), their dynamics and triggering processes are not well understood (Locat and Lee 2002; Bardet et al. 2003). One of the principal reasons for lack of consensus is the limited data on the physical and mechanical properties of the slide plane sediments, as these materials are rarely sampled with confidence. The geotechnical characterisation of sediments representative of slide surfaces (physical behaviour and frictional stability) is required to assess the sediments response to consolidation, earthquake induced acceleration and vibration, as well as pore pressure response to oscillatory loading (Glenn et al., 2008; Minning et al., 2006), as these are the characteristics which will enable identification of probable triggering mechanisms.

Jenkins and Keene (1992) were the first to describe submarine landslides on the southeast (SE)

3-2 Clarke (2014) Chapter 3

Australian margin using GLORIA data collected onboard HMAS Cook; but it was later voyages on the RV Surveyor using high resolution, multibeam bathymetric data (Glenn et al 2008; Boyd et al. 2010; Hubble et al. 2013) that provided bathymetric data with the detail required to demonstrate the widespread occurrence of large and small mass failures on the eastern Australian continental margin. These investigations have established that submarine landsliding should be considered to be a common and ongoing characteristic of this passive continental margin (Clarke et al., 2012; Hubble et al, 2012).

This study presents results from physical property and geotechnical tests on 12 gravity cores taken from the east Australian continental slope from sites located within or adjacent to five submarine landslides. The cores were collected during the RV Southern Surveyor (SS12/2008) examination of the east Australian continental margin (Boyd et al., 2010) and used to: 1) characterize both the physical and geotechnical properties of marine sediments collected from the east Australian continental slope; 2) analyse the stress history of the sediment found on the margin; 3) test whether boundary surfaces identified within five sediment cores represent slide plane surfaces; and 4) determine input parameters (i.e. cohesion and friction angle) for slope stability modelling to assess the frictional stability of the submarine slopes. These data were then used to consider the stability of the continental slope sediments and potential submarine landslide trigger mechanisms.

3.1.1 Previous work The investigation of the mechanical behaviour of the eastern Australian slope materials conducted prior to this study consists of four triaxial and four shear box tests performed by Glenn et al. (2008). These samples were obtained from the vicinity of large (20 km3) slides off southern NSW. The tests indicated a friction angle of 40°, whereas the average slope of these slides was 5°. Conventional soil mechanics analysis would suggest a failure angle of 40° for an infinite submerged slope. The data from the tests on the specimens from the central Australian continental slope (Glenn et al. 2008) showed that following 1-D compression all four specimens were only able to sustain a small increase in load when subjected to undrained loading, and their resistance progressively decreased as they approached an ultimate frictional state. This susceptibility to collapse suggests that static liquefaction may be an important part of the mechanism for the slope failures (c.f. Hight and Leroueil 2003). However, only four specimens have been tested and to what extent they are representative of the sediment from the greater depths at which failure occurred is not certain. Data on tests from other marine sediments show that they often have considerable structure, due to the presence of microfossils and diatoms, as well

3-3 Clarke (2014) Chapter 3 as variable silt and clay contents. Leroueil and Hight (2003) show that increasing silt and clay fines contribute to an open structure, which can increase the brittleness and compressibility of the soil, and can be critical to the collapse potential.

3.2 Determining Stress History

3.2.1 Compression History The stress history undergone by a sample can be determined from the shape of the e-logσ’ plot after consolidation, where e is void ratio and σ’ is effective vertical stress. A normally consolidated sample will have a near linear e-logσ’ relationship which is called the virgin compression line. The slope of the virgin compression line is defined as the compression index (Cc). If a sample is overconsolidated, its state will be represented by a point on the expansion or recompression path of the e-logσ’ plot. The slope of the recompression line is defined as the recompression index (Cr). The compression index

(Cc) indicates the compressibility of a normally-consolidated soil, while the recompression index (Cr) indicates the compressibility of an over-consolidated soil. Line AB on Figure 3.1 shows the plot of a normally consolidated sample, and line ABCD shows the plot of an overconsolidated sample, with an initial compression AB, followed by expansion BC and recompression CD. The recompression curve will ultimately re- join the virgin compression line AB, with further compression occurring along the virgin line. During compression, changes in soil structure take place continually and the sample does not revert to the original structure during expansion. An overconsolidated sample will be much less compressible than a normally consolidated one. The compressibility of the sample is Figure 3.1: Idealised void ratio-effective stress relations represented by a dimensionless coefficient for compressible soil (after Mitchell 1993). Normally called the compression index (Cc), the consolidated sediments follow line AB, while overconsolidated sediments follow ABCD (AB = Initial or slope of the linear portion of the e-logσ’ virgin compression; BC = Rebound or swelling; CD = plot. Disturbance of the sample during Recompression; Cc = Compression index; Cs = Swelling index)

3-4 Clarke (2014) Chapter 3 testing can slightly decreases the slope of the virgin compression line and can, in some cases, result in a minor underestimation of the preconsolidation pressure and thus decrease the calculated depth of burial slightly (Craig 2004).

3.2.2 Preconsolidation pressure (σ’pc) Using the Casagrande method (Craig 2004), it is possible to estimate, from the e-logσ’ curve for an overconsolidated sample, the maximum effective vertical stress that has acted on the sample in the past. This value is referred to as the preconsolidation pressure σ( ’ pc). At some stage in the sediment history after the maximum stress has been applied, the vertical pressure is removed, causing the sediment to then undergo expansion (see Fig. 3.1). Examples of processes that would cause an expansion of this type is removal of overlying sediments because of submarine sliding, or bottom current erosion of the surface sediments.

For a normally consolidated sample, the preconsolidation pressure will be the same as the vertical overburden stress (due to the weight of the overlying sediment) existing at the depth from which the sample was taken within the core. An overconsolidated sample will exhibit a preconsolidation pressure that is much larger than the overburden stress at the level from which the sample was taken. This is described numerically as the overconsolidation ratio (OCR). OCR is defined as the ratio of the preconsolidation pressure to the current (in situ) effective vertical stress: OCR =σ ’ pc/ σ’ . For normally consolidated sediment, OCR is equal to 1. If no submarine landslide has occurred (i.e. no overlying sediment has been removed from above the sampled sediment) and the sampled core sediment is simply a result of continuous hemipelagic sedimentation, the OCR should equal 1. If sediment has been removed from above the sampled sediment, the sample should be overconsolidated and have an OCR greater than 1.

Although this is “standard” soil mechanics, it is also understood that yield stress (apparent σ’ pc) does not always coincide with the past maximum stress. This is due to sediment creep, sediment aging, or post-depositional changes (e.g. cementation, bio-activity) (Leroueil and Vaughan 1996).

3.2.3 Depth of sediment burial From the preconsolidation pressure (i.e. the maximum vertical effective stress), the maximum depth of burial experienced by the sediment can be estimated from the following relationship (see

3-5 Clarke (2014) Chapter 3

Craig 2004):

σ’pc = Ƴ’z = (Ƴsat – Ƴwater)z

where σ’pc is the preconsolidation pressure, Ƴ’ is the effective unit weight, z is the depth of burial in meters, Ƴsat is the saturated unit weight of the sample and Ƴwater is the unit weight of water (10.1 kNm-3 for sea water).

3.2.4 Rate of Consolidation The rate of consolidation is the time period over which compression takes place under a vertical load. This can be determined from compression tests and is expressed in terms of the coefficient of consolidation (Cv). Once a vertical load is applied, the consolidation of the soil is measured over time.

Cv is determined once 90% of primary consolidation is complete using Taylor’s square root of time method (Craig 2004) by plotting void ratio (e) verses square root of time, determining the time taken for 90% consolidation, and then applying the following equation:

2 Cv = (0.848d )/t90

where Cv is the coefficient of consolidation, d is half the height of the sample at t90 (two-way drainage assumed), and t90 is time at which 90% of consolidation is complete.

3.3 Study Area Location

The study area is comprised of a segment of the eastern Australian continental margin located offshore Noosa Heads in southern Queensland (QLD) and Yamba in northern New South Wales (NSW). It is situated approximately between 30 km to 70 km seaward of the present coastline in water depths of 150-4500 m (Fig. 3.2). Twelve gravity cores from the upper continental slope (<1200 m) were examined in this work (Fig. 3.2, 3.3) and are located within or adjacent to the 1) Bribie Bowl Slide (water depth 600 m); 2) Coolangatta-2 Slide (water depth 900 m); 3) Coolangatta-1 Slide (water depth 600 m); 4) Cudgen Slide (water depth 600 m); and 5) Byron Slide (water depth 800 m) (see Fig. 3.2, 3.3 and Table 3.1). These features are representative of the slope failures that occur in the two dominant slope morphologies present in the study area identified by Boyd et al. (2010) and Clarke et al. (2012, 2014) which are the: 1) relatively steep (3-7°) and canyon incised slopes (Bribie Bowl Slide and Bryon Slide); and 2) relatively gentle slopes (1-3°) of the Nerang Plateau (Coolangatta-1 & -2 slides and Cudgen Slide). At least one gravity core was recovered from each of these submarine landslide scars.

3-6 Clarke (2014) Chapter 3

The twelve cores provide material for physical and geotechnical testing. Dating, sedimentology, and physical property properties of the slope sediments are described in Chapter 2 and summarised in

Figure 3.2: Digital elevation model (DEM) of the southern QLD to northern NSW continental margin, showing the location of the study area. Insets show detailed views of the five submarine landslide sites of interest (black line) and the locations of the twelve collected gravity cores (GC) from the Southern Surveyor SS2008/12 voyage. Moving north to south: (a) Bribie Bowl Slide (b) Coolangatta-2 Slide, (c) Coolangatta-1 Slide (d) Cudgen Slide, (e) Byron Slide

3-7 Clarke (2014) Chapter 3 section 3.7.1.1 below. Seven cores (GC2, GC3, GC5, GC6, GC8, GC11, and GC12) were collected from within five submarine landslide scars between 50 to 220 m thick; that is, 50 to 220 m of sediment is missing from within the slides where the cores have been collected (see Fig. 3.2, 3.3 and Table 3.1). Three reference cores (GC1, GC4, and GC7) were recovered from adjacent sites in slopes that do not present obvious slide features or morphologies. Figure 3.5a shows representative core logs from the plateau region and canyon region.

Chapter 2 demonstrated sediments are mixtures of calcareous and terrigenous clay (10-20%),

Figure 3.3: Digital elevation model (DEM) of the east Australian continental margin. Perspective looks west, landward onto the continental margin. Insets show details of five submarine landslides (black lines) and the locations of gravity cores within each slide. Moving north to south: a) Bribie Bowl Slide (northern canyon region): GC2 and GC3 are located in the slide; GC1 is located adjacent to the slide; b) Coolangatta-2 Slide and Coolangatta-1 Slide (plateau region): GC9 is located in the Coolangatta-2 Slide; GC8 is located in the Coolangatta-1 Slide ; c) Cudgen Slide (plateau region): GC4, GC5, GC6, GC7, GC10, GC11 are located in the slide; GC7 is located adjacent to the slide; d) Byron Slide (southern canyon region): GC12 is located in the slide

3-8 Clarke (2014) Chapter 3

Figure 3.4: DEM of two slides showing typical slope profiles for a) canyon region slides and b) plateau region slides. Slope geometry and cross-section profiles across slide (black line S-N) and down slide (black line W-E) as shown. Thickness of the material missing from the slide scar is indicated on each profile. A reference profile of the adjacent unfailed slope is shown (light blue line). For depth scale, see Fig. 3.2 silt (50-65%) and sand (15-40%) and are generally uniform in appearance. Their carbonate contents vary between and 17% and 22% by weight, while organic carbon contents vary between 4% and 12% by weight. Dating of conformably deposited material identified in ten of the twelve cores indicates a range of sediment accumulation rates between 0.017mka-1 and 0.2 mka-1 which are consistent with previous estimates reported for this area. One slide-adjacent core, and four within-landslide cores present boundary surfaces which are located at depths of 0.8 to 2.2 meters below the present-day seafloor that are identified by a sharp, colour-change boundary; small increases in sediment stiffness; slight increases in sediment bulk density of 0.1 gcm-3. Distinct gaps in AMS 14C age across the boundary surface show a depositional hiatus of at least 25 ka. These boundary surfaces are interpreted to represent detachment surfaces or slide plane surfaces. It is suspected that the sediment below the boundary features has been buried to depths between 50-220 m (depending on the thickness of the submarine landslide scar; see Fig. 3.4) prior to the removal of the now-missing sediment. The sediment above the boundary surface is believed to represent recent sediment drape. If this is the case, the geomechanical properties of the sediment above and below the boundary surface should present significantly different maximum burial depths. The purpose of this chapter is to use geomechanical data to further constrain dates and impacts associated with the boundary surfaces.

3-9 Clarke (2014) Chapter 3 3.4 Geologic Setting and Margin Structure

The eastern Australian (EA) continental margin stretches 1500 km north from Bass Strait to the Great Barrier Reef (Boyd et al., 2004). The margin, which is by world standards narrow, steep and sediment deficient, was formed by rifting in the Cretaceous period between 90 Ma and 65 Ma (Gaina et al. 1998). Since then, margin subsidence has been relatively minor. The continental shelf ranges between 14 and 78 km wide and is relatively flat with a thin sediment cover. The sediment reaches a peak thickness of about 500 m at the edge of the shelf, which occurs at water depths ranging between 55

Table 3.1: Summary of location, depth, total recovery length, stratigraphy, and target feature of the sediment cores retrieved from the study area

Latitude Longitude Water depth Length Name Stratigraphy Target Feature °S °E (m) (m)

Adjacent to Bribie Bowl GC1 -26°41.69’ 153°45.123’ 841 4.15 Single conformable section Slide (reference core)

GC2 -26°45.64’ 153°43.68’ 585 4.45 Single conformable section Bribie Bowl Slide

GC3 -26°45.22’ 153°43.85 759 4.35 Single conformable section Bribie Bowl Slide

Surface sample, no GC4 -28°13.032’ 153°57.387’ 488 - Cudgen Slide (no return) stratigraphy

2 sections, one boundary Cudgen Slide – upper slide GC5 -28°14.102 153°57.871’ 677 2.26 feature block

2 sections, one boundary GC6 -28°13.601’ 153°58.767’ 740 1.91 Cudgen Slide – base of slide feature

4 sections, three boundary Adjacent to Cudgen Slide GC7 -28°12.71’ 153°59.69’ 748 5 features (reference core)

2 sections, one boundary Coolangatta-1 Slide – lower GC8 -28°07.263’ 154°02.752’ 929 2.52 feature slide region

Unsplit for geotechnical Coolangatta-2 Slide – upper GC9 -28°01.79’ 154°02.84’ 913 4+ testing slide region

Unsplit for geotechnical GC10 -28°02.479’ 154°02.261’ 800 4+ Cudgen Slide – slide hollow testing

2 sections, one boundary GC11 -28°13.491’ 153°58.960’ 755 2.49 Cudgen Slide – base of slide feature

2 sections, one boundary GC12 -28°37.96’ 153°58.09’ 1167 1.99 Bryon Slide – middle slide feature

and 180 m. The continental slope is the region from the shelf edge to the Tasman Abyssal Plain where the water depth is around 4500 m. The continental slope ranges from 28-90 km wide and has average slopes in the range from 2.8° to 8.5°. The sediment cover generally reduces from the shelf edge to the Abyssal plain (Boyd et al. 2004).

3-10 Clarke (2014) Chapter 3

Figure 3.2 shows the study area where the sediments have been recovered. The survey consisted of both sub-bottom profiles and high-resolution multibeam bathymetric mapping to provide a detailed picture of the seafloor and reveal the underlying geology. An overview showing the bathymetry looking landwards is given in Figure 3.3. At this scale it is possible to see that a number of canyons cut into the slope sediments and most of these are off the major rivers. Further details of the five-cored submarine landslides are shown in Table 3.1.

3.5 Sampling and Testing Program

The SS2008-V12 cruise obtained 12 1.91-4.45 m long undisturbed sediment cores (90 mm in diameter) in or adjacent to five submarine landslide scars from the eastern Australian upper continental slope (Table 3.1). Slide scar thickness ranges from ~50 to 250 m (see Fig. 3.4). Coring was carried out using a gravity corer with a 1000 kg head-weight and 5 m long barrel containing plastic liners as the sediment casing. Sampling sites were chosen after inspection of the multibeam bathymetry to identify slide sites (Boyd et al. 2010). For most of the gravity cores the soil has been logged and basic properties, particle size distribution, carbonate and organic carbon content, void ratio, and densities have been obtained (results presented in Chapter 2). The basic classification tests have been supplemented by a series of one-dimensional (1-D) compression tests at a range of stress paths, followed by undrained shearing to failure (static triaxial), and oedometer consolidation tests to investigate the mechanical behaviour of the sediments by imposing conditions onto a specimen that mimic seafloor loading.

In order to gain a thorough understanding of the sediment behaviour, it is usual for a large number of tests to be completed. Triaxial tests are used to generate shear-strength parameters while consolidation tests (oedometers) are used to identify geotechnical parameters including preconsolidation pressure, settlement, compression index, and coefficient of consolidation.

3.6 Methods / Specimen Preparation

3.6.1 Sediment Properties Sediment core was cut into one-meter lengths and ten were split, photographed and logged. Two cores were left unsplit for geotechnical testing. Core stratigraphy was completed by manual logging of visual observations of the core. Observations were made methodically from bottom to top of each core

3-11 Clarke (2014) Chapter 3 and sub-samples taken at regular intervals. In addition, where logging identified a distinct change in lithological features, further samples were taken both either side and also across the boundary to test for variations with depth. Sedimentology (stratigraphy, grain-size, and carbonate and organic carbon content), radiocarbon ages, and physical properties (bulk density, water content, and unit weight) of the slope sediments have been reported in Chapter 2 and are summarised in section 3.7.1.1 below.

Specific gravity (Gs) of the core sediment was also measured using the American Society for Testing and Materials (ASTM) D 854 standard method. Specific gravity is a ratio of the density of the core sediment to the density (mass of the same unit volume) of water.

3.6.2 Atterberg Limits Atterberg Limits, plastic limit (PL) and liquid limit (LL), were determined on 13 representative samples from six separate cores using standard testing procedures (Head 1982). From these values, plasticity index (Ip) was determined. Seawater was substituted as the mixing agent instead of distilled water in order to simulate in-situ conditions. Soil classification is based on the Unified Soil Classifica- tion System (USCS) using the ASTM D 2487 standard method. Plastic limit (PL) is defined as the minimum water content at which the soil will deform plastically, while liquid limit (LL) is the minimum water content at which the soil will flow under a small disturbing force. The plasticity index (Ip) is simply the difference betweenLL and PL, and is the range of moisture content over which the soil is plastic or malleable. The Atterberg Limits and relationships derived from them are simple measures of the water absorbing ability of soils containing clays. For example, if a clay has a very high LL it is capable of absorbing large amounts of water. More importantly, the LL and PL are also related to the soil strength.

3.6.3 Oedometer Testing The characteristics of a soil during one-dimensional consolidation or swelling can be determined by conducting an oedometer test. Oedometer consolidation tests were performed on 18 representative sediment samples (~110 cm3) from six separate cores. Representative sediment types were tested based on percentage clay content, ranging from low (~7-10%; GC5), mid-range (~10-20%; GC7), and high (~23%; GC2), as well as sediment from above and below the boundary feature surfaces (GC5, GC7, GC8, GC11, and GC12), and remoulded sediment (GC2-R, GC5-R, and GC7-R). The remoulded test samples were reconstituted to a moisture content of ~45-60% using seawater as the mixing agent, and consolidated in a cylinder to ~25 kPa (similar to pressures experienced in situ, ~5 m overburden),

3-12 Clarke (2014) Chapter 3 and progressively loaded following standard methodology (Head, 1982; Craig 2004). Samples were tested under both high stress (maximum load ~6.6 MPa; approximately 800 m burial depth) and low stress (maximum load ~288 kPa) conditions, following standard methodology (Head 1982; Craig 2004). Average measured grain density of 2.64 gm-3 was used in oedometer calculations, which is a typical value for sandy silts (Mitchell and Soga 2005). Results are presented by plotting void ratio (e) against the corresponding effective stress after each stress increment has reached equilibrium.

Oedometer tests provide the compression index (Cc) and coefficient of consolidationC ( v) for the samples. The results were used to 1) validate void ratio and density results; 2) simulate overburden pressures greater than those believed to have been experienced in situ based on inspection of bathymetry data (see Fig. 3.4) calculate the depth to burial of the sediments based on their preconsolidation pressures; and 4) determine the coefficient of consolidation Cv (i.e. the rate of consolidation of the sediment).

3.6.4 Triaxial Testing Monotonic triaxial tests were performed on eight unsplit core sections ~126 mm-long and 52 mm-diameter taken from gravity core GC9 (see Fig. 3.2/Table 3.1) following standard testing procedures (Head, 1982; Craig 2004). The triaxial test equipment setup comprises of a conventional triaxial cell with an internal load cell for accurate measurement of the deviator state. The equipment is fully computer controlled and measurements are made of cell pressure, pore pressure, deviator load, axial displacement and volume change. Care was taken when setting up triaxial tests to prevent moisture loss, volume change and vibration interference, to help prevent degradation on the soil skeleton structure.

The specimens were mounted in the triaxial cell, with porous disks and filter paper placed at each end to facilitate drainage. The specimen was enclosed in a rubber membrane that was sealed at the top and bottom platens by O-rings. Drainage could take place through the porous disks on the top and bottom faces of the specimen. After placement in the triaxial apparatus, an isotropic stress of approximately 25 kPa was applied (similar in situ load at time of coring). The specimens were saturated by raising cell pressure and back pressure together until the back pressure reached 200 kPa. The specimens were then left for at least 12 hours to allow the removal of any remaining air bubbles.

3-13 Clarke (2014) Chapter 3

A 1-D stress state was estimated before commencing shearing (k0 consolidation). This was achieved either by specifying a stress path to follow from the start of loading, or by loading isotropi- cally to the required cell pressure. During testing, specimens were compressed so that deviator stress approximately equalled mean effective stress (q = p’) in order to approximately follow 1-D loading, before subjecting the specimens to a shear test under undrained conditions. Following 1-D consolidation theory (see Craig 2004), the rate of stress increase was kept low ensure the excess pore pressures did not exceed 30 kPa. Rate of loading was determined by the coefficient of consolidationC v. For saturated tests any volume changes of the specimen are calculated from the quantity of water flowing in or out from the specimen. Accurate measurement of volume change is important in calculating volume strains, changes in void ratio and in calculating the deviator stress. Deviator stress is calculated by di- viding the measured force by the cross-sectional area of the specimen. Because of the large strains that occur in soil tests the change in area must be used to calculate the stresses correctly, and this requires knowledge of axial and volume strains. Specimens were sampled at 2:1 height to diameter ratio so that volume and axial strains estimated from measurements extend to the triaxial cell. On the completion of the test the sample was unloaded and removed from the apparatus. The moisture content values were determined by measuring the wet and dry masses of the specimen in order to check the consist- ency of void ratio calculations (see Craig 2004).

3.6.5 Slope Stability Modeling Geomechanical modelling of the landslides was undertaken using the slope stability software GEO-SLOPE/W (2007) (http://www.geo-slope.com/) to examine the influence of several parameters on the stability of the slope profiles (cohesion, friction angle, and slope geometry). Bishop’s simplified method of slices was used for slope stability calculations (Craig 2004). The slope profile used to perform the stability modelling was determined from high resolution multibeam bathymetry data acquired during the SS2008-V12 cruise (see Fig. 3.2). The profiles of adjacent unfailed sections of the landslide region were used to indicate the pre-failure geometry of the failed slope (following the pro- tocols of McAdoo et al. 2000 and Hubble and Rutherford 2010). For each slope stability calculation, a dimensionless parameter, the Factor of Safety (FoS), is determined. It is defined as the ratio of the restoring forces to the disturbing forces (stable slope, FoS>1; unstable slope, FoS<1; critically stable slope, FoS=1).

3-14 Clarke (2014) Chapter 3

Table 3.2: Summary of the gravity core sediment classification data Final Void Ratio Initial Depth Clay (<2 Silt (2-600 Sand (600- Dry Bulk (after Unit Weight Moisture below um) um) 2000 um) Density Sample ID oedometer Content seafloor tests) (cm) (%) (%) (%) (e) (kg/m 3) (kN/m 3) (%)

GC2/4B/296-314-R 296-314 12.48 87.52 0.00 1.5 901.7 15.5 75.2 GC2/5A/401 401 11.14 88.85 0.02 1.6 978.5 15.9 65.9 GC2/5A/433 433 11.88 88.12 0.00 1.5 1001.6 16.0 63.5 GC5/2B/48 48 1.92 98.07 0.01 1.4 1082.0 16.5 56.0 GC5/2B/63-76-R 63-76 5.54 94.46 0.00 1.4 1089.4 16.6 55.3 GC5/2B/90 90 3.84 96.02 0.14 1.4 1079.1 16.2 53.1 GC5/2B/113^ 113 2.95 96.70 0.35 ---- GC7/2D/185^ 185 2.55 97.45 0.00 1.3 1145.9 17.0 51.2 GC8/2B/68.5 68.5 10.28 88.64 1.08 1.8 1601.0 15.7 65.5 GC8/2B/136.5^ 136.5 3.12 95.53 1.34 1.5 1699.7 16.7 54.6 GC11/3A/196.5 196.5 2.94 95.91 1.15 1.5 1022.2 16.2 62.0 GC12/1B/35 35 6.08 93.92 0.00 1.4 814.2 14.9 86.4 GC12/2A/135^ 135 2.92 96.20 0.87 1.6 1091.4 16.6 55.5 Parameters prior to testing ^ Below the boundary feature

3.7 Results and Interpretation

3.7.1 Sediment Properties 3.7.1.1 Classification data - Physical properties A summary of classification data for this study is provided in Table 3.2. This shows that the continental slope sediments are predominantly comprised of silt-sized material, with about 15% clay, variable amounts of Table 3.3: Plasticity index results showing classification of 13 sediment sand sized particles, samples. Liquid limit, plastic limit, plasticity index, and soil classification are and significant or- shown. Soil classification is based on USCS standards (see ASTM D 2487).

Depth within Liquid Limit Plastic Limit Plasticity Index Soil ganic content. The Core number core (cm) (LL) (PL) (Ip) Classification sediments contain a GC9/4B-1 307-321 48.6 39.9 8.7 ML significant amount GC9/4B-2 279-293 46.5 37.7 8.8 ML GC9/4B-3 265-279 49.8 37.4 12.4 ML of carbonate grains GC2/3C/166-175cm 166-175 54.3 38.1 16.2 MH derived from the re- GC5/2B/50-63cm 50-63 55.0 38.2 16.8 MH mains of living or- GC5/3A/162-176cm 162-176 55.2 43.7 11.5 MH 136-156 MH ganisms, and also GC7/2C/136-156cm 50.1 28 22.1 GC8/2B/60-77 cm 60-77 54 33 21 MH significant amounts GC8/2B/120-132 cm 120-132 54.5 31.1 23.4 MH of terrigeneous fine GC11/2B/165-178 cm 165-178 60.3 31.1 29.2 MH grained silt, believed GC11/2B/220-234 cm 220-234 45.8 29.2 16.6 ML GC12/1B/52-72 cm 52-72 70.2 38.5 31.7 MH to be transported as GC12/2A/120-134 cm 120-134 62.8 35 27.8 MH

3-15 Clarke (2014) Chapter 3

Figure 3.5: a) Typical core logs showing characterisation of sediment two cores: GC3 (plateau region) and GC11 (canyon region); b) Core photos of GC3 and GC11 c) Close up images and interpretation of boundary surfaces identified in all six cores analysed in this study: i) GC5, ii) GC6, iii) GC7, iv) GC8, v) GC11 and vi) GC12. See Fig. 3.2 for gravity core locations. The inferred boundary feature is indicated with a dashed black line. Radiocarbon ages for each core are also shown in yellow (ka = thousand years before present, RCD = radiocarbon dead). dust by the westerly winds from the interior of the adjacent Australian continent (Hubble pers. comm. 2014). Although there is some variability in the composition from core to core there is a broad similarity in the particle distributions all along the study area.

The majority of sediments are sandy silts (MH-ML) with a small clay component, comprised of mixtures of calcareous and terrigenous clay (7-23%), silt (50-65%) and sand (15-40%) and generally uniform in appearance. No major difference in the texture of the deposited material is observed either down individual cores or between the 12 different sites located in this slope (see Fig. 3.5). They are also very similar in particle composition and grainsize distribution to the underlying compacted Pliocene, Miocene, and Oligocene muds recovered in dredge hauls from older layers of the continental slope sequence described by Yu et al. (2010) and Hubble et al. (2012). Grainsize, carbonate and organic carbon content present no significant change between units across the boundary surfaces. While

3-16 Clarke (2014) Chapter 3

Figure 3.6: Plasticity chart showing sediment classification of 13 representative samples from six separate cores based on Atterberg limit testing. Samples are classified as either sandy silts (MH) or sandy elastic silts (ML). displaying an obvious change in colour, stiffness, density and water content between the upper and lower units, grainsize analysis reveals that the sediment composition is consistently uniform with all materials in the cores presenting as poorly sorted, medium to coarse sandy silts, with relatively low clay content but relatively high silt and sand contents. Carbonate and organic carbon content also remains fairly uniform, with carbonate content varying between 17% and 22% by weight and organic carbon content varying between 4% and 12% by weight. These trends in the physical and mechanical properties are consistent across all 10 split gravity cores. Sediments are of moderately low plasticity with a liquid limit of 46-70% and plasticity index of 9-32%. Atterberg Limit tests indicate similar plastic and liquid limit values for all samples (see Fig. 3.6 and Table 3.3). The water content ranges from 45-85+%, the bulk density ranges 815-1700 kgm-3, and the specific gravity of the solids ranges from 2.5-2.74.

3.7.2 Geotechnical Characterisation 3.7.2.1 Oedometer Tests The results show typical consolidation plots with specimens showing reasonably high compressibility with compression index Cc values ranging from 0.31 to 0.78, and recompression index Cr values ranging from 0.04 and 0.10 (Table 3.4). This is similar to what would be expected from their

3-17 Clarke (2014) Chapter 3

n o i t

a e d ------p i l y t o s n o C

a n i n o r i o u t ) a d b m e d f s i ( l ------a s o o s e s b e t n r h a t t o s c m p i e e t r s d p E R ------C O

) c i

p t ’ f a σ d o i ) e l t a ( o a s ------s P s 30 6.4 4 Overconsolidated 32 4.6 5 Overconsolidated k n e m ( i r o t t c s s e E r n p o

s s )

t a r l t a n P s c e k i r - - e - t r ( 4.7 v r 6.9 ) i u ’ e t v c C σ e ( f f e

) f 1 n - o o r i t y t n 2 a e i d i m c l i ( f o ) f s v e n o C o ( c C o i s s x e e ) r r d p n C ( m I o n c e R

n o i s x ) s c e e r d C p n ( I m o C

) d f i o e ( V l o i a t n a i r F

o i

l t ) a a i 0 t r i e d ( n i I o V ) t n e Wf t ) n o % ( C ) e r W u )( t ( s i o 43 30 1.37 0.78 0.57 0.06 4.3 9.54 ( Wi Initial Final M - ) 0 0 m ) 6 u ( % 0 d ( 0 n 0 a 2 S ) - m 2 ) ( u t % l 0 ( i 0 S 6

2 < ) ) ( m y % ( 3.12 95.53 1.34 a u l C

r h o ) w t o o p l m l f e c e a ( b e D s D I e l p m a S GC5/2B/90 90 3.84 96.02 0.14 54 27 1.47 0.72 0.47 0.05 3.5 6.04 GC5/2B/48 48 1.92 98.07 0.01 56 19 1.34 0.5 0.41 0.04 5.6 3.40 GC12/1B/35 35 6.08 93.92 0.00 77 35 2.02 0.93 0.78 0.10 2.4 1.92 GC2/5A/401GC2/5A/433 401 433 11.14 11.88 88.85 88.12 0.02 0.00 62 61 29 28 1.63 1.74 0.77 0.75 0.61 0.63 0.10 0.10 0.7 0.8 25.26 26.18 GC8/2B/68.5 68.5 10.28 88.64 1.08 52 28 1.53 0.73 0.48 0.05 3.9 4.48 GC5/2B/113^ 113 2.95 96.70 0.35 57 20 1.24 0.52 0.41 0.04 5.0 8.36 GC7/2D/185^ 185 2.55 97.45 0.00 48 29 1.28 0.76 0.53 0.07 3.6 13.44 GC12/2A/135^ 135 2.92 96.20 0.87 53 31 1.45 0.83 0.65 0.10 3.4 9.13 GC8/2B/136.5^ 136.5 GC11/3A/196.5 196.5 2.94 95.91 1.15 55 30 1.52 0.79 0.54 0.07 3.1 12.92 GC8/2B/77-LW 77 11.18 88.82 0.00 59 47 1.72 1.26 0.34 0.04 2.3 GC5/2B/63-76-R 63-76 5.54 94.46 0.00 45 25 1.24 0.65 0.33 0.03 3.6 GC12/1B/81-LW 81 12.96 87.04 0.00 72 47 1.67 1.21 0.34 0.04 1.4 5.03 39.0 7.8 6 Overconsolidated GC12/1B/91-LW^ 91.5 11.27 88.12 0.61 54 48 1.68 1.21 0.33 0.04 2.4 5.66 39.0 6.9 6 Overconsolidated GC8/2B/100-LW^ 100 13.21 86.79 0.00 59 50 1.84 1.36 0.35 0.04 2.5 GC2/4B/296-314-R 296-314 12.48 87.52 0.00 53 25 1.29 0.65 0.41 0.06 3.1 GC7/3C/227-249-R^ 227-249 10.47 88.79 0.75 57 24 1.07 0.63 0.34 0.05 4.7 R = remoulded LW = low weight ^ Below the boundary feature Oedometer consolidation results from 6 gravity cores tested (GC2, GC5, GC7, GC8, GC11, GC12). 6 gravity cores from consolidation results 3.4: Oedometer Table

3-18 Clarke (2014) Chapter 3 - repre for (kPa) vertical stress effective ratio (e) versus void showing consolidation tests results Oedometer Figure 3.7: Figure sentative samples tested. Effective vertical stress is plotted on log10 scale. Solid lines are samples taken above the bound - samples taken above Solid lines are is plotted on log10 scale. vertical stress samples tested. Effective sentative ary the boundary samples taken below surface surface:b) and dashed lines are a) samples under high vertical stress; boundary and below surfaces under samples with varying c) samples taken above clay content (under high vertical stress); boundary and below surfaces for above plots from vertical stress high effective and four equivalent vertical stresses low - lines repre tests and natural samples included for comparison. Dash-dot-dot-dash comparison; d) samples of remoulded samples. sent remoulded

3-19 Clarke (2014) Chapter 3

index test results (Table 3.4), as the correlation Cc = IpGs/2 would suggest Cc values of 0.33 to 0.58 for plasticity indices of 11% to 34%. It can also be noted that the moisture contents in the upper 5 m are significantly higher than expected from their liquid limits. The oedometer specimen size was small (~100 cm3) due to the apparatus used when testing to high vertical stresses. Small sample size can lead to higher sample disturbance, which is taken into consideration when interpreting the results below.

Figure 3.7 shows representative compression plots from 1-D oedometer compression tests on undisturbed specimens (Fig. 3.7a-c); as well as for remoulded specimens (Fig. 3.7d). Representative results of 18 specimens from six cores (GC2, GC5, GC7, GC8, GC11, and GC12; see Fig 3.1, 3.2 for core locations) are shown. Although there is some variability, the similarity of the response of the specimens is remarkably consistent.

Figure 3.7a shows 11 tests completed to high effective vertical stresses. Consolidation tests run to a high effective vertical stress (6.6 MPa) are used to investigate whether the sediment experienced burial to depths significantly greater (i.e. >20+ m) than their current in situ burial depths (<5 m). Yield stresses of 2-3 MPa (representative of burial depths of ~200-250 m) would be indicative of deep failure and suggest that there has been the removal of significant sediment overburden at some point in the sediment’s compression history. However, there is no evidence of yield at this high stress and it is reasonable to say that none of the sediment tested has been buried deeply. Cc values are consistent with values expected for normally consolidated silts and suggest that there has been no significant (i.e. >300+ kPa, representative of burial depths of ~50 m) preconsolidation stress within the sediment, i.e. none of the sediment tested have been buried deeply and have remained within 50 m of the seafloor.

Figure 3.7b shows representative high stress compression plots for samples with varying clay content: i) low - 7-10% (GC5; blue squares); ii) mid-range - 12-15% (GC8; green diamonds); and iii) high – 20-23% (GC2; red circles). The compressibility of specimens on the continental slope is similar despite varying clay content (<7% to 23%), both above the boundary surfaces (solid lines) and below the boundary surfaces (dashed lines). The higher clay content samples show higher Cc values when compared to the more silty samples due to greater compressibility, and also a higher initial void ratio and moisture content (see Table 3.4).

Figure 3.7c shows four representative samples taken above and below boundary surfaces subjected to lower vertical effective stresses to obtain more accurate estimation of preconsolidation pres-

3-20 Clarke (2014) Chapter 3 sures (GC8-LW and GC12-LW). Four typical equivalent high effective vertical stress plots above and below boundary surfaces from the same two cores are also included for comparison. Specimens above and below identified boundary surfaces are almost identical under lower stresses. Some small variability is seen in the sediments above and below the boundary surface under high stress, however they too are also remarkably similar, despite suspected differences in maximum burial depths. The response of the sediment at low vertical stresses shows a small varibility from the high vertical stress results, with

Ccvalues slightly lower (0.3 compared to 0.5-0.8; see Table 3.4), but other data show that Cc tends to increase until the vertical stress applied is greater than 0.5 MPa, above the stress levels applied to the lower stress tests.

The samples taken below the boundary surface (dashed lines Fig. 3.7) were initally suspected to be significantly overconsolidated as they were taken from below a suspected slide plane surface that the geometry of the landslide scar suggested was located at a depth below 50-220 m below the sea-floor (see Fig. 3.4). Following the Casagrande method, preconsolidation pressure estimates (i.e. yield stress) between 30-39 kPa were obtained for the four samples run to low vertical stresses (Table 3.4). This equates to ~5-6 m burial depth, not the 50-220 m suggested from the bathymetry of the slides. Ad- ditionally, there is no significant difference in the preconsolidation pressure estimates for samples taken above and below the boundary surface (see Table 3.4), and no useful indication of slide plane can be obtained. If these boundary surfaces do represent the slide plane of a submarine landslide, <10 m of material has been removed during the slide. However, based on 14C dating of the sediment above and below the boundary surface and determined sedimentation rates (see Chapter 2), a minimum of 0.5 m to 5 m of sediment is missing from above the boundary surface. This would represent preconsolidation pressures between ~3 to 30 kPa, similar to what is determined for this sediment (Table 3.4), and this is not inconsistent with a slide plane interpretation. This suggests that the tested material is a younger sediment fill of the slide feature which has itself failed more recently (see discussion below)

Based on the preconsolidation pressure estimates above, the over-consolidation ratio (OCR) can be determined. The OCR for the low vertical stress samples is ~6.5 for all cores (Table 3.4), indicating slightly overconsolidated sediments. Estimates of the preconsolidation pressure indicates ~6 m burial for the samples whereas in-situ burial depths within the cores were all less than 1 m. For the samples taken above the boundary surface (solid lines, Fig. 3.7) there is no evidence suggesting that these sediments are anything but normally consolidated; i.e. that the sediment has only experienced burial equal to its in-situ depths within the core (<2 m). However, the slightly overconsolidated presentation of the

3-21 Clarke (2014) Chapter 3 sediments seems to contradict this. A possible explanation for this unusual result is biological activity (bioturbation and biocementation) that is introducing structure into the sediment and these data indicate that the slope sediments are structured, with some sensitivity due to the similarity of compression results (i.e. similar Cc values, see Table 3.4). Sensitivity could be affected by the relatively high organic content of up to 8%, which is known to be a factor in sensitivity in other soils (Mitchell and Soga 2005; Watts and Dexter 1998).

Figure 3.7d shows 3 representative samples of remoulded tests. Three natural samples from the same cores are included for comparison. Remoulded samples show a lower Cc value than natural samples, indicating that there is some loss of structure during the remoulding process and therefore some sensivity in the sediments. Remoulded samples also show similar compressibility to natural samples, however all remoulded samples showed lower initial void ratios (see Table 3.4) showing that the sediments preserve a loose structure created during deposition and this is most pronounced for the specimens with higher clay content. This trend is similar to other offshore sediments (Kuo and Boulton and references therein; Chaney et al. 1988).

The oedometer tests were also used to calculate the coefficient of consolidation (Cv) using Tay- lor’s method (see Craig 2004; section 3.2.4). Cv values range between 0.8-5.6 m2yr-1 at a vertical effective stress of 500 kPa (Table 3.4) and are low for silty materials (Mitchell and Soga 2005), especially given the low clay content of the specimens (<25%). Silty samples should present with higher Cv values than clay-rich samples (Mitchell and Soga 2005), and when compared to published values for clay- rich specimens that generally range from <0.1 to 5 m2yr-1 (Mitchell and Soga 2005), the data suggest that the EA margin specimens should contain more clays than they actually do. Cv is important for controlling the rate of shear for triaxial testing.

3.7.2.2 Triaxial Tests In addition to the oedometer tests, eight triaxial tests were carried out on samples from one unsplit core (GC9) at different loading amplitudes (Fig. 3.8 and Table 3.5). The sediment in GC9 is a sandy silt (ML) with clay content of <9%, and is a hemipelagic sediment presenting characteristics typical of the other cores collected. The samples were compressed to ~25 kPa, which is similar to what they currently experience in situ (i.e. 3-4 m of overburden), and are then consolidated to higher pressures, greater than what the sediment is currently experiencing, in order to observe the response of the sediment under elevated stresses (see Table 3.5).

3-22 Clarke (2014) Chapter 3 ◦ ) ( Ф ’ 32.29 43.81 39.17 39.17 39.17 M 1.3 1.8 1.6 1.6 1.6

d d e ) c e h z v t i n l g i a a σ ’ n r e 0.9 m / 1.08 1.37 r d u r t s n o s ( u N ) a

P d k e ( u n i s a , r h d 58.5 94.2 t n g n U e r t s

f ) e a o v P t i t r k c a ( t e c v f s f t 54 69 132 108.5 0.82 250145 225.8 120.8 0.83 E σ ’ a l , s r a s i a e x e r t h A s S

s g s s s n e i

e r y r c t l i t s p p s g r p o n r o i 25 22 25 25 25 t a t e n o a i r i s f I v o f n e e o d b c ) e ( o i t a 1.51 1.51 1.51 1.42 1.52 R d i o V ) l f a n W i 46 48 ( t 46.4 44.3 51.7 F n e t n ) o % C ( ) e r u W t ( s i l o ) a i i t M i W n 50.5 51.5 52.7 52.8 ( 48.02 I # D I l a i x a i r T

) w m o l c e ( r b o h t o l p f e a e D s D I e l p m a GC9-4BGC9-4BGC9-4B 307-321 279-293 265-279 T3/1 T1/2 T3/2 GC9-5AGC9-5A 364-380 395-410 T1/3 T3/3 S Triaxial sample sediment classification data and triaxial results sample sediment classification data and triaxial 3.5: Triaxial Table

3-23 Clarke (2014) Chapter 3 - - Triaxial testing results: a) response of core specimens to 1-D compression. Plots show void ratio (e) versus effec ratio (e) versus void show Plots specimens to 1-D compression. of core a) response testing results: Triaxial Figure 3.8: Figure responses from pressure and pore is plotted on log10 scale; c) normalised deviator stress vertical stress Effective (kPa). axial strain; d) effective versus (kPa) normalised deviator stress show specimens (5 samples). Plots 1-D compressed triaxial for five (kPa) mean stress effective versus (kPa) deviator stress show paths for saturated tests. Plots stress (black dashed lines). also shown Line 1 (M = 1.6) and 2 1.3) are State GC9. Critical specimens from core tive vertical stress (kPa) for 8 saturated tests on core GC9. Effective vertical stress is plotted on log10 scale; b) response of 8 is plotted on log10 scale; b) vertical stress GC9. Effective for 8 saturated tests on core (kPa) vertical stress tive tests included for compari oedometer compression with two representative specimens during 1-D compression triaxial core (GC8/2B/136.5; black dashed line) and one sample compressed stress to high vertical effective son: one sample compressed vertical stress versus effective void ratio (e) Plots show black solid line). (GC12/1B/81.5-LW; stress vertical effective to low

3-24 Clarke (2014) Chapter 3

Figure 3.8a shows typical compression plots from 1-D compression tests on undisturbed triaxial specimens (the relation between void ratio and mean effective stress). Results of eight specimens from GC9 (see Fig 3.2 for core location) are shown. The compressibility response of the eight samples is similar despite the different initial void ratios. While GC9 is dominantly comprised of sandy silts, some variability is present between the specimens, with GC9-5A/T3-3 presenting greater compressible due to a lower sand and corresponding higher clay content.

Figure 3.8b shows the same triaxial 1-D compression results as Figure 3.8a but includes two oedometer compression tests for comparison: one sample compressed under high vertical effective stress (GC8/2B/136.5; black dashed line) and one sample compressed under low vertical effective stress (GC12/1B/81.5-LW; black solid line). The triaxial specimens show compressibility similar to those seen in oedometer tests (see Fig. 3.8b), with Cc values ranging from 0.3 to 0.7. Triaxial specimens have slightly lower Cc values (0.3-0.4 compared to 0.5-0.8) in comparison to Cc values determined in oedometer tests (expect for a more compressable specimen: GC9-5A/T3-3 - 0.7). The results are similar to the Cc results for the low stress oedometer samples, due to the vertical stresses not yet reaching ~0.5 MPa. Cc tends to increase until vertical stresses are greater than 0.5 MPa. Yield stresses range from 80-200 kPa for these triaxial test materials, which are significantly higher than the yield stresses measured from oedometer testing (~30-40 kPa). Given that triaxial test yield stresses are the result of K0 compression tests, this suggests that the oedometer results have been affected by disturbance, which is consistent with the small size of the sample (see section 3.5.3).

The responses of five specimens to undrained shearing in triaxial compression are shown in Fig- ure 3.8c, and the associated effective stress paths are shown in Figure 3.8d. Only five of the eight test results are presented, as there were difficulties with the equipment in the other tests. Figure 3.8c shows the deviator stress (q) versus axial strain (ɛa) responses. Specimens were loaded to a range of mean effective stress values (p = 200-600 kPa) before shearing undrained to failure. To enable comparison of the tests, the deviator stress and excess pore pressures have been normalised by the vertical effective stress at the start of shearing.

Figure 3.8c shows that very different resistances are mobilised depending on the state of the soil at the start of shearing. It should also be noted that the initial void ratio is not the void ratio controlling the soil behaviour during shearing. From Figure 3.8a it can be seen that the void ratio of the speci-

3-25 Clarke (2014) Chapter 3 mens only reduces slightly for the more silty material (e.g. e = 1.43 – 1.42, GC9/4B/T4-1) whereas the void ratio of more clay rich samples decreases substantially (e.g. e = 1.45 – 1.2, GC9/4B/T3-3). In undrained loading it is the void ratio presented by the material at the start of shearing that controls the ultimate resistance.

Figure 3.8d shows the effective stress paths followed in all the tests as a plot of deviator stress (q) versus mean effective stress. All five specimens approach one of two ultimate failure loci, referred to as the Critical State Line (CSL) on the plot (all specimens approach a similar ultimate frictional resistance). All specimens end up on or very close to one of these lines, which is conventionally described by q = Mp` (1). M is the slope of the line given here by either M = 1.6 or M = 1.3. These values are equivalent to effective friction angles of 39.2° and 32.3° respectively. Clay rich sediments have a lower effective friction angle of ~30 (GC9/4B/T3-3) when compared to more silty samples (e.g. GC9/4B/ T1-2). Although all the tests shown in Figure 3.8d reach one of two ultimate failure loci, they reach it at very different points because they have different void ratios, and possible different ultimate e-logp’ loci.

Two specimens (GC9/4B/T3-1 and GC9/5A/T1-3) were run to low vertical stresses (σ’v= 69 kPa and 59 kPa respectively; see Table 3.5). They had a dilative response to shearing and did not reach a maximum effective stress until relatively large strain (Fig. 3.8c), effectively elevating the positions of the failure loci on the graph(see Fig. 3.8d). From the pore pressure responses and the effective stress paths it can be seen that this difference is a consequence of a transition from dilative to compressive behaviour as the stress level increases. This dilative response represents increased stability at shallow depths (<15 m) due to the low vertical stresses.

Two specimens (GC9/4B/T1-2 and GC9/4B/T3-2) were run to mid to high vertical stresses (σ’v = 250 kPa and 145 kPa respectively; see Table 3.5). These specimens show a more brittle response with the peak deviator stress attained at a very small strain, after which the resistance rapidly decreases to its ultimate value (Fig. 3.8c). These specimens have a similar high frictional resistance (39.2°), lowering the positions of the failure loci on the graph once the maximum effective stress was achieved (Fig. 3.8d) indicating pore pressure build up. The increasing brittleness of the sediment behaviour indicates decreasing stability at burial depths greater than 15 m due to higher vertical stresses.

One specimen is clearly different to those discussed above, GC9/5A/T3-3. This specimen was

3-26 Clarke (2014) Chapter 3

also tested at moderate vertical stresses (σ’v = 132 kPa) and also shows a more brittle response, with the peak deviator stress attained at a very small strain, after which the resistance drops away continuously, very rapidly, until its ultimate value. This specimen has a lower frictional resistance (32.3°), lowering the position of the failure loci after the maximum effective stress was achieved (Fig. 3.8d). This is a typical response of a loose soil. The shear response of GC9/5A/T3-3 is more compressible and shows the tendency for increasing brittleness with increasing stress level, however, this result is not considered reliable due to non-uniform deformation during shearing. The combination of increased brittleness and higher compressibility is significant as it is associated with a sudden loss of strength if the sediment is subjected to earthquake acceleration, which causes a rapid increase in pore pressure.

The sediment presents an increasingly brittle behaviour as stress levels increase and there is a transition from dilative to compressive behaviour, which is typical of normally consolidated sediments. The sediments also present rapid pore pressure build up as stress levels increase. Therefore, increasing stress levels (depth of burial) change the way the sediment responds. This pattern of reducing dilation and increasing brittleness with stress level might explain why deeper failure surfaces develop. There is no evidence of a boundary feature within GC9 and as such, no reason to suspect that the sediment is anything other than normally consolidated. However, from the consolidation tests of these sediments (see Fig. 3.8a), it is only the clay-rich specimen GC9/4B/T3-3 that clearly presents a normally consolidated response. Specimens GC9/4B/T1-2 and GC9/4B/T3-2 are possibly normally consolidated but additional testing at higher vertical stresses is required to confirm this.

3.7.3 Slope Stability Modeling

Geomechanical modeling of the submarine landslides has been undertaken using the slope stability program GEO-SLOPE/W (2007) to examine the influence of cohesion, friction angle, and slope geometry on the stability. Static modeling Table 3.6: Numerical input parameters used for slope of the slopes associated with each of the stability modeling with GEO-SLOPE/W. The friction angle (Ф’) represents the friction component of the soil five landslides indicate that they are all strength and the apparent cohesion (c’) represents the inherently very stable. Using parameters cohesive component of the soil strength. close to measured values (friction angle Parameter Unit Input value range ~30-40o and slopes from 3-6o; Table 3.6) Unit weight (γ) kN/m³ 15 - 17 Apparent cohesion (c’) kPa 0 - 22 static analyses predict very high factors of Friction angle (Ф) ° 0 - 40

3-27 Clarke (2014) Chapter 3 a) slope stability models under static conditions of two representative slides showing Factory of Safety (FoS) values using measured strength strength using measured values (FoS) of Safety Factory slides showing a) slope stability models under static conditions of two representative Figure 3.9: Figure = 40°) as input parameters: i) Cudgen Slide and ii) Byron Slide. The pre-failure profile was generated from a was generated profile The pre-failure Slide. and ii) Byron Slide Ф’ = 40°) as input parameters: i) Cudgen for the materials (c = 0 kPa, values - for two representa values b) FoS region slopes within the failure to simulate pre-failure site in order section of unfailed slope adjacent to the failure The Slide. and ii) Byron Slide to the varying accelerations for i) Cudgen the sensitivity of FoS slides under seismic loading conditions showing tive to the varying (kh) and vertical (kv) seismic acceleration coefficients). accelerations (horizontal the sensitivity of FoS graph shows

3-28 Clarke (2014) Chapter 3

Table 3.7: A summary of the FoS values from back analysis slope stability modeling from two representative slides (Cudgen Slide and Byron Slide) from reducing c and Ф’ as shown. Critical FoS values are underlined. Scenario Friction angle Site Cohesion (kPa) FoS (lowest) description (°) 40 >10 Residual 30 >10 cohesion, 0 decreasing 15 4.8 Cudgen Slide friction angle 7.5 2.4 (plateau 3.75 1.2 morphology) Peak friction 11 >10 angle, 5.5 >10 40 decreasing 2.75 >10 cohesion 1.375 >10 40 6.19 Residual 30 4.26 cohesion, 0 15 1.98 decreasing Byron Slide friction angle 7.5 0.97 (canyon 3.75 0.48 morphology) Peak friction 11 8.8 angle, 5.5 7.8 40 decreasing 2.75 7.28 cohesion 1.375 6.98 safety (>5). Table 3.7 summaries the resultant FoS values for a typical plateau morphology scenario (Cudgen Slide; Fig 3.9a) and a typical canyon morphology scenario (Byron Slide; Fig 3.9b). Back analysis modelling determined parameter combinations required for failure to occur. The range of tested parameters are summarised in Table 3.7. For failure to occur, results show the friction angle value must drop to values less than 8°, with cohesions of zero, values well below the measured strength for the materials (c = 0 kPa, Ф’= 30-40°). Analyses have also been conducted to investigate the effects of earthquake loading by including a factor for seismic accelerations in the standard pseudo-static limit- equilibrium calculations. Results indicate that lateral and vertical accelerations of 0.3 g (kh = 0.3 g, kv = -0.3 g), would be sufficient to destabilise the slopes of the seafloor in the present study by giving FoS equal to 1 with the measured values of Ф’ (see Fig 3.9c). While this approach has been widely used to assess the stability of submarine slides to earthquake events (e.g. Puga-Bernabeu et al. 2013; Urgeles et al. 2006) and is a useful if very crude guide to the size of seismic event that might be required; the assumption that all the soil in the slide mass has the same accelerate at the same time is a very simplistic approximation. Several works have questioned its applicability to the large volumes of soil involved in submarine landslides. The approach is indicative, but not conclusively indicative, in identifying the mechanisms leading to submarine slope failure (Li et al. 2009; Kramer 1996).

3-29 Clarke (2014) Chapter 3

3.8 Discussion

Mechanical tests have been performed to provide evidence to support our model of slope processes, and in particular the role of large submarine landslides. As noted, slide features are widely distributed on the continental slope and so not appear to be associated with differences in soil type or soil gradient. Slides up to 50m deep are present on the shallow slopes of the Nerang Plateau and seem unrelated to progressive/regressive failures. Slides >100 m are present on the steeper slopes in the canyon regions, which seem to evolve landwards through retrogressive processes with a number of slides appearing to have progressively extended upslope as a series of retrogressive failures. There is some observable steepening at the toe of the present upper slope, which is possibly a sign of small failures that may continue to migrate upslope.

The identified boundary surfaces are either: a) erosional hiatal features that developed after the occurrence of the landslide in which case the boundary surface age provides a minimum age for landslide occurrence; b) deep-failing sediment detachment surfaces where the boundary surfaces represent the main submarine landsliding event in which case the post-boundary surface sediment dates indicate approximately when the main landsliding occurred; or 3) near-surface sediment detachment surfaces from which thinner slabs were removed subsequent to the main submarine landslide event (slabs <50 m thick) in which case the post-boundary surface sediment dates indicate approximately when smaller, ongoing landsliding occurred.

Table 3.8: Summary of available sediment classification data from the Sydney and Brisbane regions

Moisture Clay Silt Sand Carbonate Organic content Region LL Ip (%) (%) (%) (%) (%) (%)

Brisbane 10-20 50-65 15-40 20 8 46.5 9 50-100 Sydney 8-25 30-80 10-60 35 - 44 18 55-85

3.8.1 Uniformity of composition and geomechanical behaviour The sediment properties investigated in this study and geotechnical tests results are relatively uniform and consistent all cores. Preconsolidation pressures indicate slightly overconsolidated sediments, with reasonably high compressibility, and low sensitivity. The only other study reporting geomechanical data for sediments of the east Australian continental slope is available for sediments taken from

3-30 Clarke (2014) Chapter 3 submarine landslides offshore Sydney (Glenn et al. 2008). They also completed one-dimensionally consolidated undrained triaxial tests, with some additional shear box tests to evaluate residual strength properties. Although there is some variability in the responses, the similarity of the response of the specimen from Sydney and SE Queensland is remarkable (see Table 3.8). 3.8.2 The nature of the boundary surfaces – are they slide plane surfaces? Several of the gravity cores (GC2, GC3, GC5, GC6, GC8, GC9, GC10, GC11, GC12) were obtained from within detected landslide features in an attempt to penetrate through the base of the slides to constrain the age of sliding and characterise the materials of the slide surface. In most locations this was unsuccessful as the recent sediment drape overlying the slide surface was thicker than the gravity corer could penetrate. However, five of the cores described here present a distinct boundary surface in four within-landslide cores and one slide-adjacent core at depths of 0.8 to 2.2 meters below the present-day seafloor that have been interpreted as possible slide plane boundaries. These boundary surfaces are all identified by a sharp, colour-change boundary; discernable but small increases in sediment stiffness; and slight increases in sediment bulk density of 0.1 gcm-3. Most importantly, distinct age breaks of at least 25 ka are recorded across these boundary surfaces (AMS 14C dates; see Chapter Two). The sedimentalogical and geotechnical properties of the sediments, and their variation with depth, from one of these cores (GC11) are shown in Figure 3.5. It can be seen from the figure that there is a distinct change in density and moisture content, as well as appearance of the material at a depth of 220 cm. It is also noticeable that this change in density is not associated with any significant change in the grading, carbonate content or organic content of the material.

Based on the location of gravity cores within submarine landslide scars (see Fig. 3.3), changes in the physical properties of the sediments across the boundary surfaces (e.g. increased in sediment stiffness), and distinct gaps in sediment age, these boundary surfaces have previously been interpreted to represent the basal surfaces of large submarine landslides (e.g. Boyd et al. 2010). This interpretation is based on distinct down-core changes in bulk density, unit weight, and moisture content across the inferred slide-plane boundary (see Chapter 2). The depth reconstructed at the slide scar sites by replacing the material apparently missing from the U-shaped trough, i.e. by maintaining the continuity and shape of the adjacent slope and projecting it above the failure site, is approximately 50-250 m thick (see Fig. 3.4).

The geotechnical analyses do not clearly differentiate an upper post-landslide unit and a lower,

3-31 Clarke (2014) Chapter 3 more deeply buried unit with a boundary surface that represents the surface of a major submarine landsliding event. Compression testing indicates that the sediment both above and below the boundary features is slightly over consolidated and that the removal of significant thickness of sediment overburden (>50 m) from above these particular boundary surfaces has not occurred. Therefore these boundary surfaces do not represent the deep-seated failure surfaces required by the bathymetric expression of the slide features. The sediment below the boundary surfaces is older than the sediment above of the boundary surfaces (16-23 ka vs. 48-50+ ka), and sediment below the boundary surfaces has been buried to a greater depth (0.8-2.2 m vs. 5-10 m). The burial depths are not great enough to suggest the sampled boundary surface represents the slide plane of the slide features evident in the bathymetric head scarps (e.g. 220 m for the Byron Slide, 50 m for the Cudgen Slide, and 50 m for the Coolangat- ta-1 Slide). If this were the case, the sediment below the boundary surfaces would present as obviously overconsolidated, with past burial depths well in excess of 50 m ((i.e. preconsolidation stress estimates ~300+ kPa). The sampled boundary surfaces, therefore are interpreted to represent subsequent slide events either: a) sliding of sediment which has back-filled the original slide scar (e.g. the Byron Slide; Fig. 3.4a); or b) the detachment of near surface slabs, contemporaneous with the main sliding event (e.g. the Cudgen Slide; Fig. 3.4b).

Another possible explanation for the sediment below the boundary surface showing no great geomechanical effects of deeper burial could be due to the influence of bioturbation on the sediment, which is known to alter sediment compression signatures such that the sediment presents less consolidation than measured (i.e. actually experienced) (see Kuo and Bolton 2013).

The bulk density determined for the sediment sampled below the identified boundary surfaces is consistent with burial and compaction by column of material at least 5-10 m thick. In contrast, the sediment sampled just above the boundary feature presents a density consistent with its present-day depth (i.e. approximately 1 m) below the sea floor. The down-core changes in unit weight and water content are also consistent with 5-10 m of burial. The values of Cc reported, and change in moisture content at the inferred slide plane can be interpreted as representing a slide depth of anywhere between 5 m and 10m for each of the slides. The depth reconstructed at the five sites by replacing the material apparently missing from the U-shaped trough and head-wall scarp height, i.e. by maintaining the continuity and shape of the adjacent slope and projecting it above the failure site, is between 50-250 m. Thus while it is possible to date the material above and below the boundary features, there is insufficient information to determine whether this is the date of the main slide at this location, and further

3-32 Clarke (2014) Chapter 3 mechanical and dating studies are in progress to further constrain the result.

3.8.3 Landslide Initiation Static modeling of the slopes associated with each of the five landslides indicate that they are all inherently very stable (FoS values >5). This apparently inherent stability is not consistent with the ubiquitous occurrence, and is contradicted by, the widespread presence of slope failures in the study area and the occurrence of slide features in the “slide adjacent zone”. We infer from these data that that the landslides are the result of external triggers, either short term or long term, which reduce the shear strength of the sediments, and/or change the geometry of the slope such that the margin is frequently destabilised. Back analysis modeling further indicates a dramatic reduction of sediment shear strength is required for slope failure to occur. Values well below the measured strength for the materials (c = 0 kPa, Ф’ = 30-40°) are required. For failure to occur, the friction angle value must drop to values less than 8°, with cohesions of zero. These are unrealistically low strengths for the sedimentary materials that accumulate on these slopes and some other process must be responsible for these failures. There- fore there must be some mechanism that leads to pore pressure build-up to reduce the vertical effective stress.

While it is possible to demonstrate that a ground shake associated with a magnitude 7 or larger earthquake could destabilize these slopes, this approach does not consider a number of other potential mechanisms that could cause these sediments to undergo a catastrophic loss of shear strength. The pseudo-static analysis treats the landslide mass as a rigid block on a plane and does not consider the substantial effects of soil liquefaction, which is likely given their saturated, loose to moderately compacted nature of the sediments (c.f. Ozkan 1998).

Puzrin et al (2004) have argued that it is unlikely that a failure can develop over distances of

Figure 3.10: a) Zone of elevated pore pressures; b) Slope failure mechanism (after Puzrin et al. 2004)

3-33 Clarke (2014) Chapter 3 several kilometres instantaneously, and that a progressive failure mechanism must be considered. On land progressive failures are often observed to result from oversteepening of the toe of a slope, where failure at the toe leads to a retrogressive failure that migrates upslope. This mechanism is also considered to be responsible for the large Storrega submarine slide. Puzrin et al (2004) have suggested an alternative progressive failure mechanism that involves a weakened zone propagating down slope. The basis of the analysis can be explained by considering Figure 3.10. The starting point is that a zone of elevated pore pressures develop (Fig. 3.10a), possibly owing to an earthquake, where the soil reaches a state of failure for which the mobilisable soil resistance is lower than the stresses at equilibrium from the weight of the overlying soil (tr < tg Fig. 3.10b). If the length of this failed zone is sufficient a global and catastrophic failure will occur. However, if the zone of failure is more limited the soil will tend to move downslope into the currently unfailed region. If the failure plane can propagate because the energy released is greater than that needed to progress the failure then the shear plane can grow, and if conditions are unfavourable it may continue to advance until it reaches a length where global failure results. For significant energy release to occur the ultimate resistance of the soil needs to be lower than that required to resist the gravitational stresses, and the soil needs to respond in a brittle manner. The triaxial test data presented here above display the type of brittle behaviour that can potentially lead to this type of mechanism.

The analysis of Puzrin et al (2004) suggests that failure begins upslope, but depending on the soil type and behaviour progressive failure may be limited or not occur and it is possible that the resulting length of the failure surface may be less than the critical value required for a catastrophic failure. There is some evidence for this from a number of head scarps present on the east Australian margin where the soil mass below the headscarp has not moved significantly downslope (e.g. Cudgen Slide).

The size of the area where soil reaches a state of failure will depend on earthquake magnitude, distance from its source, slide thickness, and properties of the soil column. Triaxial data suggest that below a depth of ~20 m the soil response will be compressive, and this will lead to the build up of pore pressure when subjected to cyclic (earthquake) loading. More testing is required to investigate the cyclic response of the sediments, but the observed failures are not inconsistent with earthquakes being the trigger (see Chapter 2).

3-34 Clarke (2014) Chapter 3

3.8.4 Static Liquefaction A potential mechanism for submarine landslides has been suggested by Hight and Leroueil (2003) that can explain how failures can be initiated on slopes at about one third of the frictional resistance. They quote an example from Bangladesh where failure occurred at a mobilised friction angle of 11o for a material with a friction angle of 29o . The mechanism is based on the phenomenon of static liquefaction. It has been demonstrated for sandy soils that there is a collapse surface (Sladen et al. 1985), or instability line (Lade, 1993), that is reached before the soil mobilises its full frictional resistance. When the collapse surface is reached large compressive strains start to develop, or in undrained tests excess pore pressures build up to the point that a soil loses all resistance, the latter being called static liquefaction. Sladen et al, (1985) showed that the peaks in the normalised stress paths for loose sand all lie on a unique curve in q, p’, e space, and similar results have been reported (e.g. Lade and Yamamuro, 1998) for other sands and silty sands. Leroueil and Hight (2003) suggest that many soil types show this type of behaviour, particularly in simple shear and triaxial extension stress paths that are relevant to slide failures.

The static liquefaction phenomena, caused by a tendency for materials to be contractile on shearing which leads to pore pressure generation, has been widely reported in triaxial tests for a variety of soil types. However, demonstrating that this effect can trigger large slides in practice has been difficult. Ideally field data are required for verification, but this is unrealistic for large submarine landslides because of the scale, remote access and high cost involved, and also the rarity of these events. Both numerical analysis and experiments have been carried out and demonstrated that it is possible to reproduce the slide behaviour, but the value of these tools is limited by lack of knowledge of the mechanical properties of the sediments and by limited site characterisation. One approach which has been explored by some researchers (e.g. Phillips and Byrne, 1994, Coulter and Phillips, 2003, Take et al. 2004) is to use model scale centrifuge tests. Phillips and Byrne (1994) reported that static liquefaction triggered at the toe of the slope in a model centrifuge test was responsible for the collapse of a normally consolidated silty clay slope with a slope angle of only 12 degrees, a similar mechanism to that suggested by Hight and Leroueil (2003) for the Bangladesh failures mentioned above. However, although the concept has been demonstrated there appears to be little published data from centrifuge tests on the susceptibility of different soil types, and Take et al. (2004) have questioned whether the static liquefaction mechanism is correct.

3-35 Clarke (2014) Chapter 3

3.8.5 “Missing” sediment While no conclusive triggering mechanism has been determined, it can be reasonably concluded from the widespread presence of slides on the margin that some process is acting on the margin sediments to destabilise the slope and transport large amounts of sediment to the abyssal plain. Dating of conformably deposited material indicates a range of sediment accumulation rates between 0.017 mka-1 and 0.2 mka-1 (see Chapter 2), these rates suggest that removal of sediment accumulated on the east Australian continental slope is a normal process. If the rates are representative or similar to the long- term Neogene average surface accumulation then sediment removal has been a consistently occurring event since the formation of the margin 60 million years ago (Gaina et al., 2003). Given that the sediment wedge deposit is generally less than 500 m thick and there is abundant evidence for submarine landsliding (Boyd et al. 2010; Hubble et al. 2012) then it follows that the mechanism for sediment removal on this section of the margin is the current and geologically recent mass wasting. Assuming constant sedimentation rates of between 0.02-0.20 mka-1 on the margin since its formation, a sediment deposit between ~1-3 km is “missing” from the margin, suggesting that much of the sediment is moved off the shelf and slope and down to the abyssal plain in the geologically recent past (<15 Ma).

3.8.6 Issues and limitations 3.8.6.1 Core length Accurate sediment property data is essential to properly assess the stability of these deep sea slopes and to identify areas of increased hazard worth further study. The use of short gravity cores (<5 m) for investigating the submarine landslides presented here is a useful but also somewhat limited method when compared to coring techniques such are piston cores and vibro-cores with core recover- ies between 5-15 m, or deep sea drill cores, which are able to sample >50 m below the sea bed. Only a small part of the overall record can be investigated, making it necessary to extrapolate back over a long period of geological time, thereby creating an overreliance on the assumption that sedimentation style and rates have remained relatively constant over the entire period. Additionally, the sedimentation rate calculations do not take into account the possibility of sediment removal from the slope, resulting in an underestimation of the actual sedimentation rate. Given the prevalence of sliding along this margin, submarine landslide events would more than likely affect the calculated sedimentation rate due to the removal of near-surface sediment layers. Investigations into the deeper sediment are needed for analysis of these failures.

3-36 Clarke (2014) Chapter 3

3.8.6.2 Sample disturbance The majority of sediment specimens tested in this work represent “natural samples”. Specimens were sampled directly from the recovered in situ gravity cores and placed straight in the various testing apparatuses, with no resculpting/remoulding occurring (unless stated otherwise). Whist this is considered to be sampling in the natural state, the sediments should still be considered somewhat disturbed due to the nature of gravity core sampling and subsequent transportation to storage facilities. Distur- bance can lead to a reduction of σ’pc and greater compressibility of the sediments.

3.9 Conclusions

This chapter has presented an analysis of the physical and geotechnical properties of submarine landslide sediment taken from east Australian upper continental slope (<1200 m). The geotechnical properties of the sediment from the east Australian upper continental slope are described and then discussed in the context of slope stability in order to identify or at least improve our understanding of the origin of the observed slides. To date, no conclusive triggering mechanism has been identified for initiating submarine landslides in the region although the brittle nature of the sediments and grainsize distribution make the slopes susceptible to liquefaction during oscillatory shaking, favouring an initiation mechanism related to seismic shaking. Studying the geotechnical response of the sediment helps to evaluate the slope stability. Using a variety of geotechnical tests, a baseline of sediment properties has been determined for failure regions within the study area. The sediment from the study area is remarkably similar to sediment collected from other sections of the margin e.g. offshore Sydney, and the older compacted mud layers from lower in the stratigraphic section. The sediment is characterised by high shear strengths, low clay content and high void ratios, and brittle behaviour. It seems a major earthquake, or toe erosion could have contributed to initiate slides in this region.

The chapter has shown evidence of large-scale mass wasting phenomena on the EA continental slope, including the existence of many landslide features. Slides are evident all along the margin from Shoalhaven to the Sunshine Coast, and on slopes ranging from 1o to 9o. The soil properties of the upper sediments are similar along the margin and show no evidence of weak clay layers, although they do contain significant but relatively small amounts of clay (<25%). The friction angles of the sediments are in the range of 30o - 40o, so that conventional soil mechanics would suggest the slopes have high factors of safety. However, this is clearly not the case as slope failures are widespread. Slide surfaces

3-37 Clarke (2014) Chapter 3 have been identified by dating and it has been shown that these features are associated with relatively shallow slides and not the large slide features that are evident in the bathymetry. Triaxial tests have indicated a significant increase in the sensitivity of the shear response with increasing vertical stress level (i.e. burial depth), and that below a depth of ~20 m the soil response will be compressive, leading to the build up of pore pressure when subjected to cyclic (earthquake) loading. This is thought to be significant in explaining why the slides have large thicknessess of 50 to 200m.

The slope stability modeling results imply that the sediment forming the EA margin is reasonably strong and inherently stable. Classical limit-equilibrium modeling indicates that submarine landslides should not be a common occurrence on the margin. However, the wide occurrence of upper slope slides across the EA margin indicates that submarine sliding should be considered to be a common characteristic of this passive continental margin. This indicates that one or more of the potential triggering mechanisms can operate in passive margin settings to destabilise the slope. The processes suspected to be most likely include: 1) dramatic reduction of the shear strength of the upper-slope sediments to extremely low values, possibly induced by creep or a build-up of pore-pressure; 2) long- term modification of the slope-geometry i.e., sedimentation on the head of the slope and/or erosion of the toe of the slope; and/or 3) seismic events large enough to trigger sediment liquefaction or a sudden increase of pore-fluid pressure.

Acknowledgments

We would like to acknowledge the P&O crew and scientific crews of the RV Southern Surveyor voyage (12/2008). Funding for this voyage was provided by ARC Australia and ConocoPhillips Pty. Ltd.

References

3-38 Clarke (2014) Chapter 3

Netherlands, pp. 491-502. Clarke, S., Hubble, T., Airey, D., Yu, P., Boyd, R., Keene, J., Exon, N., Gardner, J., 2012. Submarine Landslides on the Upper Southeast Australian Passive Continental Margin – Preliminary Findings, in: Submarine Mass Movements and Their Consequences, in: Yamada, Y., Kawamura, K., Ikehara, K., Ogawa, Y., Urgeles, R., Mosher, D., Chaytor, J., Strasser, M. (Eds.). Springer Netherlands, pp. 55-66. Coulter, S., Phillips, R., 2003. Simulating submarine slope instability initiation using centrifuge model testing, Submarine Mass Movements and Their Consequences. Springer, pp. 29-36. Craig, R.F., 2004. Soil Mechanics, 6th ed. E & FN Spon, London. Gaina, C., Muller, R. D., Brown, B., Ishihara, T., 2003. Microcontinent formation around Australia in The Evolution and Dynamics of the Australian Plate, in: Hillis, R.R., Muller, R. D. (Ed.), Spec. Pap. Geol. Soc. Am., pp. 405-416. Gaina, C.M., Muller, R.D., Royer, J.-Y., Stock, J., Hardebeck, J., Symonds, P., 1998. The tectonic history of the Tasman Sea: a puzzle with 13 pieces. Journal of Geophysical Research 103, 12413–12433. Glenn, K., Post, A., Keene, J., Boyd, R., Fountain, L., Potter, A., Osuchowski, M., Dando, N., Party, S., 2008. NSW Continental Slope Survey – Post Cruise Report in: Australia, G. (Ed.). Haflidason, H., Sejrup, H.P., Nygard, A., Richter, Mienert, J., Bryn, P., Lien, R., Fosberg, C.F., Berg, K. & Masson, D.G., 2004. The Storegga Slide: Architecture, Geometry and Slide Development. Marine Geology 213, 201-234. Head, K.H., 1982. Manual of Soil Laboratory Testing. Pentech Press Limited, Devon. Heiri, O., Lotter, A.F., Lemcke, G., 2001. Loss on ignition as a method for estimating organic and carbonate content in sediments: reproducibility and comparability of results. Journal of Paleolimnology 25, 101-110. Hight, D., Leroueil, S., 2003. Characterisation of soils for engineering purposes. Characterisation and engineering properties of natural soils 1, 255-362. Hubble, T., 2013. Voyage Summary SS2013-V01: Marine Geology and Geohazard Survey of the SE Australian Margin off Northern NSW and Southern Queensland, CSIRO, Hobart. Hubble, T., Rutherfurd, I., 2010. Evaluating the relative contributions of vegetation and flooding in controlling channel widening: the case of the Nepean River, southeastern Australia. Australian Journal of Earth Sciences 57, 525-541. Hubble, T., Yu, P., Airey, D., Clarke, S., Boyd, R., Keene, J., Exon, N., Gardner, J., 2012. Physical Properties and Age of Continental Slope Sediments Dredged from the Eastern Australian Continental Margin – Implications for Timing of Slope Failure Submarine Mass Movements and Their Consequences, in: Yamada, Y., Kawamura, K., Ikehara, K., Ogawa, Y., Urgeles, R., Mosher, D., Chaytor, J., Strasser, M. (Eds.). Springer Netherlands, pp. 43-54. Islam, M., Carter, J., Airey, D., 2004. Comparison of the yield locus and stress-dilatancy function of some critical state constitutive models with experimental data for carbonate sand. Journal of the Institution of Engineers(India), Part CV, Civil Engineering Division 84, 267-274. Jenkins, C.J., Keene, J.B., 1992. Submarine slope failures on the southeast Australian continental slope. Deep Sea Research 39, 121-136. Kramer, S., 1996. Geotechnical earthquake engineering. Practice Hall, New Jersey. Kuo, M., Bolton, M., 2013. The nature and origin of deep ocean clay crust from the Gulf of Guinea. Geotechnique 63, 500-509. Lade, P.V., 1993. Initiation of static instability in the submarine Nerlerk berm. Canadian Geotechnical Journal 30, 895- 904. Lade, P.V., Yamamuro, J.A., 1999. Physics and mechanics of soil liquefaction. AA Balkema. Leroueil, S., Hight, D., 2003. Behaviour and properties of natural soils and soft rocks. Characterisation and engineering properties of natural soils 1, 29-254. Leroueil, S., Vaughan, P., 1990. The general and congruent effects of structure in natural soils and weak rocks. Geotech- nique 40, 467-488. Li, A.J., Lyamin, A.V., Merifield, R.S., 2009. Seismic rock slope stability charts based on limit analysis methods. Comput- ers and Geotechnics 36, 135-148. Liu, M., Carter, J., 2002. A structured Cam Clay model. Canadian Geotechnical Journal 39, 1313-1332. Liu, M.D., Carter, J.P., Airey, D.W., 2010. Sydney soil model. I: Theoretical formulation. International Journal of Geome-

3-39 Clarke (2014) Chapter 3

chanics 11, 211-224. Locat, J., Lee, H.J., 2002. Submarine landslides: advances and challenges. Canadian Geotechnical Journal 39, 193-212. Masson, D.G., Harbitz, C.B., Wynn, R.B., Pedersen, G., Lovholt, F., 2006. Submarine landslides: processes, triggers and hazard prediction. The Philosophical Transactions of the Royal Society A (Phil. Trans. R. Soc. A) 364, 2009-2039. McAdoo, B.G., Pratson, L.F., Orange, D.L., 2000. Submarine landslide geomorphology, US continental slope. Marine Geology 169, 103-136. Minning, M., Hebbeln, D., Hensen, C., & Kopf, A., 2006. Geotechnical and Geochemical Investigations of the Marquês de Pombal Landslide at the Portuguese Continental Margin. Norwegian Journal of Geology 86, 187-498. Mitchell, J.K., Soga, K., 2005. Fundamentals of soil behavior. John Wiley & Sons Ltd, Chichester, UK. Ozkan, M.Y., 1998. A review of considerations on seismic safety of embankments and earth and rock-ﬁll dams. Soil Dy- namics and Earthquake Engineering 17, 439-458. Phillips, R., Byrne, P., 1994. Modelling slope liquefaction due to static loading, 47th Canadian Geotechnical Conference, Halifax, Canada. Puga-Bernabéu, Á., Webster, J., Beaman, R., 2013. Potential collapse of the upper slope and tsunami generation on the Great Barrier Reef margin, north-eastern Australia. Natural Hazards 66, 557-575. Puzrin, A., Germanovich, L., Kim, S., 2004. Catastrophic failure of submerged slopes in normally consolidated sediments. Geotechnique 54, 631-643. Sladen, J., D’hollander, R., Krahn, J., 1985. The liquefaction of sands, a collapse surface approach. Canadian Geotechnical Journal 22, 564-578. Take, W., Bolton, M., Wong, P., Yeung, F., 2004. Evaluation of landslide triggering mechanisms in model fill slopes. Land- slides 1, 173-184. Talling, P.J., Paull, C.K., Piper, D.J., 2013. How are subaqueous sediment density flows triggered, what is their internal structure and how does it evolve? Direct observations from monitoring of active flows. Earth-Science Reviews 125, 244-287. Urgeles, R., Leynaud, D., Lastras, G., Canals, M., Mienert, J., 2006. Back-analysis and failure mechanisms of a large submarine slide on the ebro slope, NW Mediterranean. Marine Geology 226, 185-206. Watts, C.W., Dexter, A.R., 1998. Soil friability: theory, measurement and the effects of management and organic carbon content. European Journal of Soil Science 49, 73-84. White, D., Take, W., Bolton, M., 2003. Soil deformation measurement using particle image velocimetry (PIV) and pho- togrammetry. Geotechnique 53, 619-631. Wright, S., Rathje, E., 2003. Triggering mechanisms of slope instability and their relationship to earthquakes and tsunamis. Pure and Applied Geophysics 160, 1865-1877. Yu, P., 2010. Mass Failure on Passive Margins: Causes and Triggers for Submarine Landslides on the East Australian Con- tinental Margin, School of Geosciences. The University of Sydney, Sydney, p. 186.

3-40 Clarke (2014) Chapter 4 Chapter 4

Eastern Australia’s submarine landslides: implications for tsunami hazard between Jervis Bay and Fraser Island

Samantha Clarke1, Thomas Hubble1, David Airey2, and Steven Ward3

1. Geocoastal Research Group, University of Sydney, Sydney, NSW, Australia

2. School of Civil Engineering, University of Sydney, Sydney, NSW, Australia.

3. University of California, Santa Cruz, CA, United States [email protected]

Abstract

Recently collected high-resolution bathymetric data from the eastern Australian continental slope between Jervis Bay (New South Wales) and Fraser Island (Queensland) has been examined to assess the hazard presented to coastal communities by submarine landslide generated tsunami greater than 5 m. Submarine landslides are ubiquitously present in water depths of ~400 to 3500 m along the entire length of continental margin, but are increasingly prevalent northward of Coffs Harbour without clustering at any particular water depth. Steeper slopes (>4°) in the canyon sections of the margin tend to shed thicker blocks (>100 m) and have the potential to generate larger tsunami waves.

Two-hundred and fifty individual submarine landslide scars greater than one kilometre in width are identified, thirty-six were able to produce a tsunami flow depth >5 m at the coastline. On-shore surge (coastline flow depth), run-up height, and inundation distance has been calculated for each of these submarine landslides using Ward and co-workers method (Chesley and Ward 2006; Ward 2001; Ward 2011). Submarine landslides along the eastern Australian continental margin have the potential to generate tsunami with flow depths at the coast ranging from 3 to 38 m for a landslide velocity of 20 ms-1. Flow depth at the coastline directly relates to landslide thickness. Thin (<100 m) and narrow (<5

4-1 Clarke (2014) Chapter 4 km) landslides produce smaller tsunami, with coastal flow depths of <5 m, and thick (>100 m) and/ or wide (>5 km) landslides generate coastal flow depths of 5-10+ m. The combination of both thick and wide landslides had the greatest potential to generate the largest coastal flow depths >10 m. Water depth strongly influences the onshore tsunami size with larger events generated from shallower water depths between ~500 -1500 m. Maximum flow depth at the coastline is larger for the thicker canyon landslides (50-250+ m) which occur on steeper slopes in comparison to shallow plateau landslides (<50 m) which generally produce waves less than 1 m in height, except where landslide surface area was particularly large (>50 km2). Maximum inundation distances and run-up heights of 1.6 km and 22 m respectively are calculated for landslide velocities of 20 ms-1, with these values can varying significantly depending on local coastal topography.

Thirteen potential submarine landslides able to produce a tsunami flow depth >5 m at the coastline are also identified. Conservative estimates indicate that future submarine landslides with similar characteristics to these could generate tsunami with maximum flow depths at the coastline ranging between five and twenty metres, run-up heights of up to 20 m and inundation distances of up to 1.6 km. These results are similar to the onshore surge produced from the Papua New Guinea submarine landslide tsunami which generated 5 to 10 metre high onshore surges that penetrated up to 1 km inland along a 30 km stretch of coast. That tsunami killed approximately 2000 people.

The largest slide blocks identified have the potential to generate substantial tsunamis that could cause widespread damage on the eastern Australian coast and threaten coastal communities. If a submarine landslide mass similar to that the enormous mid-Neogene olistostromic block found on Tas- man Sea Basin’s abyssal plain (Hill 1992) was to be released from the eastern Australian continental slope today, a tsunami wave of ~30 m would reach the coastline, with inundation distances and run up height reaching approximately 5.6 km and 2.07 m respectively but there is no evidence for a submarine landslide large enough and young enough to have generated a Holocene megatsunami on the central and southern New South Wales coast as postulated in the body of work by Bryant and Young et al.

Keywords: submarine landslides, wave height, southeastern Australia, upper slope, flow depth, run- up, inundation distance

4-2 Clarke (2014) Chapter 4 4.1 Introduction and Aims

Submarine landslides can damage seabed infrastructure (communications cables and buried pipelines), cause subsidence of coastal land, and generate tsunamis (Masson et al. 2006). Examples of large submarine landslide generated tsunamis include the 1929 Grand Banks’ event (Fine 2005), the 1946 Scotch Cap Alaska event (Fryer et al. 2004), the 1958 Litiuya Alaska event (Miller 1960), the 1964 Seaward & Valdez Alaska slide (Haeussler et al. 2007), the 1979 Nice Airport slide in France (As- sier-Rzadkiewiz et al. 2000), the 1998 Aitape Papua New Guinea event (Tappin et al. 2001) and 2002 Stromboli volcanic island event in Italy (Tinti 2005). Some of these submarine generated tsunami have resulted in significant casualties and property damage (e.g. Scotch Cap, Grand Banks, Aitape) but not at the same scale that large-earthquake generated tsunamis can produce, e.g. the devastating 2004 Sumatra Indian Ocean tsunami (Lay et al. 2005) and the March 2011 Japanese events. Submarine landslide generated tsunami are not as well understood as those associated with large plate-boundary earthquakes and consequently present a significant but poorly-quantified hazard (Bardet et al. 2003; Watts 2004; Maretzki et al. 2007; Sue et al. 2011).

The eastern Australia (EA) coast is potentially vulnerable to tsunamis due to the population concentration (~85%) and critical infrastructure within 50 km of the coast (Short and Woodroffe 2009). However, there has been little reason to suspect a local source for the generation of tsunami on the EA coastline. The identification of relatively recent, abundant submarine landslide scars has changed this perception (Boyd et al. 2010; Clarke et al. 2012; Keene et al. 2008) and established that submarine landsliding should be considered a common and ongoing characteristic of this passive continental margin at geological timescales (Clarke et al. 2012; Hubble et al. 2012).

This study uses data collected during three RV Southern Surveyor surveys (SS2006-V10; SS2008- V12; SS2013-V01) of the EA continental margin (Glenn et al. 2008; Boyd et al. 2010; Hubble et al. 2013) to a) characterise slope morphology of the margin between Jervis Bay and Fraser Island, b) determine the geometry of selected landslide features on the continental slope in water depths 500 – 2500 m, and c) quantify the potential tsunami hazard generated by these landslides. Submarine landslides that can generate tsunami inundation depths equal to or greater than 5 m at the coastline are considered the focus of this work because: a) wave this size is accepted worldwide as big enough to affect to more than just the immediate coastline, and will usually cause fatalities and significant coastal

4-3 Clarke (2014) Chapter 4 damage; and b) the beach-barrier systems of the east Australian coast frequently deal with storm wave greater than 4 m (Short and Woodroffe 2009), equivalent to ~3-4 m tsunami surge. Tsunami flow depths at the coastline of around 5 m represent significant inundation which breaches high-stand dune features and most the beach barriers would be overtopped on a high tide.

4.2 Study Area

The section of continental slope considered by this study (Fig. 4.1a) is located 30 km to 70 km offshore of the New South Wales (NSW) and southern Queensland (QLD) coastline along the EA continental margin and it is approximately 1150 km long between Jervis Bay in the south and Fraser Island in the north and is bounded by the 35°20’S and 24°40’S parallels of latitude and 150°55’E to 154°20’E meridians of longitude. Water depths vary from 150 m at the shelf break to 4500 m at the junction between the continental rise and abyssal plain.

The study area has been divided into 4 distinct morphometric regions (Fig. 4.1a, 4.3-4.6): Re- gion 1 - Jervis Bay to Port Stephens (central NSW); Region 2 - Port Stephens to Coffs Harbour (northern NSW); Region 3 - Coffs Harbour to North Stradbroke Island (northern NSW/southern QLD); and Region 4 - North Stradbroke Island to Fraser Island (southern QLD) (c.f. Table 4.1), on the basis of regional geological transitions identified in Keene et al. (2008). These are the boundaries between two major geological terrains present on the east Australian coast: i.e. the Sydney Basin and the New

Table 4.1: Summary of the four regions investigated. Approx. length of Major Bathymetric Region Offshore from Latitude/Longitude continental slope Dataset (km) Jervis Bay to Port 35°20’S : 30°42’S 1 Stephens (central 340 SS2006-V10 NSW) 150°55’E :153°11’E

Port Stephens to Coffs 32°41’S: 30°14’S SS2008-V12 2 Harbour (northern 280 NSW) 153°11’E:153°50’E SS2013-V01 Coffs Harbour to 30°17’S: 27°30’S SS2008-V12 North Stradbroke 3 310 Island (northern 153°26’E:154°30’E SS2013-V01 NSW/southern QLD) North Stradbroke 27°30’S: 24°39’S 4 Island to Fraser Island 310 SS2013-V01 (southern QLD) 153°51’E:153°45’E

4-4 Clarke (2014) Chapter 4

England fold belt, with some outcropping of the Clarence-Moreton Basin and Nambour Basin (c.f. Keene et al. 2008; Fig. 4.1b). Each region presents differences in margin width, slope, and morphological trend, as well as different general slope morphologies.

a 152°E 154°E b a Fraser Island 154°

26°S

Noosa Heads Fig. 6 Fig.

Moreton Island Brisbane a

Stradbroke b Island 25° 28°S Tweed Heads

Bryon Bay Ballina b

Evans Head 5 Fig.

Yamba

30°S 154° Coffs c Harbour 30° Smokey Cape

25° c Fig. 4 Fig. 32°S Tuncurry N 30° Sugarloaf Point 100 km d Newcastle Port Depth (m) d Stephens 35° Norah 0 35° Head 300 700 1100 Sydney 1500 34°S 1900 Botany 2300 Bay 2700 3100

Fig. 3 Fig. 3500 Jervis Bay 4500

Figure 4.1: Location map of the study area a) bathymetry and b) regional geology, showing the continental margin from Jervis Bay (southern NSW) to Fraser Island (southern QLD). The extent of the high-resolution multibeam bathymetry for the continental slope, collected on the RV Southern Surveyor voyages (see text), is coloured according to the depth scale. Box insets mark the location and breakdown of four regions investigated within the study area (south to north): (a) Region 1 – Fig. 4.2; (b) Region 2 – Fig 4.3; (c) Region 3 – Fig. 4.4; (d) Region 4 – Fig. 4.5.

4-5 Clarke (2014) Chapter 4 4.3 General Method Overview

4.3.1 Identification and selection of potentially tsunamogenic submarine landslides Distinct landslide scars on the continental slope between 500 m and 2500 m (sediment wedge is ~2500 m and 2500 m water depth is the limit of high resolution multibeam data) with slide dimensions 50-250 m thick and 1-10+ km wide. A major requirement of this study is the need to understand slides that have the potential to produce a tsunami height at the coastline greater than 5 m and quantify the minimum size of slides that can produce 5+ m tsunami heights. Previous work on the east Australian margin (Clarke et al. 2014) and elsewhere (Harbitz et al. 2013) suggests that slides similar to the Byron Slide (~3.6 km wide, 220 m thick) or the Cudgen Slide (5.3 km wide, 50 m thick) are required. The size of tsunami that these slides can generate is then dependant on the velocity the moving mass achieves. These speeds are difficult to determine or estimate but previous work (c.f. Masson et al. 2006) has determined a range of appropriate velocities. The slide velocities tested range between 20 ms-1 (conservative), 40 ms-1 (reasonable), 60 ms-1 (high), and 80 ms-1 (maximum recorded; Masson et al. 2006). The dimensions of the landslide mass, initial acceleration and maximum velocity of the sliding mass are important when assessing expected tsunami size (Ward 2001). The range of velocities tested has been constrained by minimum and maximum velocity values reported in the literature (Masson et al. 2006). The velocity at which submarine landslides travel after failure is not well defined due to a lack of direct measurements. The 1929 Grand Banks Slide was measured travelling at 25 ms-1 (Fine 2005), while the 2006 SW Taiwan event measured turbidity current velocities between 17 and 20 ms-1 (Hsu 2008). Velocities are based on cable breakages during failure and are measured from the gentle upper slope, approximately 2° and <0.5° respectively. At the higher end, speeds of up to 80 ms-1 have been inferred for some large landslides based on landslide debris travel distances (Masson et al. 2006). Of the range of possible landslide velocities tested we consider 20 ms-1 (60 kmhr-1) a reasonable and conservative (i.e. minimum) value (c.f. Driscoll et al. 2000).

The following iterative process was used to determine the size of submarine landslide able to generate tsunami greater than 5 m:

1. Firstly, a general bathymetric investigation is conducted for each of the four regions to identify slides of appropriate/sufficient size that may have generated 5+ m tsunamis (see Section 4.3.2; Fig. 4.2)

4-6 Clarke (2014) Chapter 4

2. Once these slides have been identified, a basic characterisation of each slide scar is carried out to determine landslide thickness (t), length (L), width (W), and water depth at

landslide centre of mass (h0), as well as distance from the adjacent coastline to head of the landslide source (r). These characteristics are a volume combination i.e. sufficient thickness for a given width and vice versa (sufficient width for a given thickness) and vary with water depth. Consequently, the slope is divided into water depth bands: i) <500 m; ii) 500-1000 m; iii) 1000-1500 m; iv) 1500-2000 m; v) 2000-2500 m; and vi) >2500 m, which are applied within each of the four regions. (See Section 4.3.2)

3. Slide characterisation parameters are used to calculate flow depth at the coastline using Ward’s equations (see Section 4.3.5).

Figure 4.2: DEM of a section of the continental slope in Region 4 coloured by a) depth and b) slope angle. DEM by slope angle highlights headscarps and landslide scars. See Fig. 4.5 for location relative to the coast. Examples of five investigated submarine landslide scars are shown (R4-2 – R4-6; solid blue lines) able to produce maximum flow depths at the coastline >5 m (see Table 4.3 for landslide names) using the proscribed method (see text). Also shown are examples of over 35 headscarps of rejected submarine landslide scars (black dashed lines). Rejected slides are unable to produce a flow depth at the greater than 5 m due to the slide scars being too narrow, too thin, too deep, too far from the adjacent coastline, or a combination of some or all of these factors. Perspective looks west-north- west, landward onto the continental margin. Vertical exaggeration 3x.

4-7 Clarke (2014) Chapter 4

4. Once flow depth at the coastline is determined for each feature, slides are then rejected or accepted based on their ability to produce a flow depth at the coastline of 5 m (see Fig. 4.2). Slides large enough to generate 5 m surges at the coastline are then given an ID number and added to the catalogue (see Table 4.3).

5. Once a slide is given an ID number, adjacent coastal topography is determined, and run-up height and inundation distance characteristics are calculated using Ward’s equations (see Section 4.3.4)

6. Regions were also examined for future potential slide features. Potential slides were either identified in previous works or from the multibeam bathymetry as areas of concern due to features such as tension cracking, erosion and undercutting at the toe of the slope, and/or extensive mass wasting in the surrounding the area. Once potential slides were identified, the same method as above was followed.

4.3.2 Bathymetry and Landslide Geometry - Calculation of Slide Characteristics Approximately 35,000 km2 of bathymetric data was collected during three RV Southern Sur- veyor surveys of the EA continental margin (SS2006-V10 (Glenn et al. 2008); SS2008-V12 (Boyd et al. 2010); and SS2013-V01 (Hubble et al. 2013)) and was acquired using a 30-kHz Kongsberg EM300 multibeam echosounder. The multibeam data was processed and used to produce 50 m gridded digital elevation models (DEM) covering the four regions investigated. The 50 m RV Southern Surveyor data is augmented by the 50 m gridded DEM for offshore Australia (Wilson et al. 2012). This dataset provides almost complete coverage of the entire Australian continental slope from Jervis Bay to the northern tip of Fraser Island. The aerial extent of the high-resolution multibeam coverage for each region is shown in Fig. 4.1, 4.3-4.6.

The DEMs were used to examine the four regions of the EA continental slope. Over 250 individual submarine landslide scars were identified within the study area (40+ from Region 1, 50+ from Region 2, 70+ from Region 3, and 100+ from Region 4) using Fledermaus V7.3.3b software (http:// www.qps.nl/). Of these, 36 individual submarine landslide scars (4 from Region 1, 3 from Region 2, 13 from Region 3, and 16 from Region 4) and 13 potential landslide sites have the potential to produce a tsunami height at the coastline greater than 5 m (see Fig. 4.3-4.6; Table 4.3). Rejected submarine landslides are unable to produce a flow depth at the >5 m as the slide scars are either too narrow, too thin, too deep, too far from the adjacent coastline, or a combination of some or all of these factors.

4-8 Clarke (2014) Chapter 4

Figure 4.2 shows an example of over 35 headscarps of rejected submarine landslide scars (black dashed lines) from within Region 4, along with five selected submarine landslide scars (R4-2 – R4-6; solid blue lines). Landslide thickness (t), length (L), width (W), and water depth at landslide centre of mass

(h0), as well as distance from the adjacent coastline to head of the landslide source (r) were determined for each of the selected features (Table 4.3). Landslide thickness is the maximum thickness within the landslide scar assuming the surface is continuous without the apparent landslide feature (McAdoo et al. 2000). Landslide length is the distance from landslide head to landslide toe. Landslide width is the average of measurements taken every 500 m down the landslide scar, perpendicular to the landslide axis. Water depth is taken from the landslide centre of mass.

4.3.3 Coastal Topography (adjacent to slide sites) Coastal topography was obtained from 1 second Shuttle Radar Topographic Mission (SRTM) derived Digital Surface Models (DSM) Version 1.0 (Geoscience Australia 2012), accessed through the Geoscience Australia (GA) National Elevation Data Framework Portal (NEDF; nedf.ga.gov.au). The dataset used is a 1 arc second (~30m) gridded Digital Elevation Model - Smoothed (DEMs) that represents ground surface topography with vegetation offsets removed and data smoothing to reduce noise (Slater et al. 2006). The coverage of the DEMs over the four regions investigated is shown in Figs. 4.3-4.6.

The 30 m gridded DEMs was used to examine the EA coastal topography adjacent to each identified slide in the four regions using Fledermaus V7.3.3b software (http://www.qps.nl/). A topographic profile (>5+ km) was generated from the coastline adjacent to each identified slide and used to help determine run-up and inundation depths specific to each site (see Fig. 4.7). Fig. 4.3-4.6 show the locations of each topographic profile.

4.3.4 Tsunami Calculations The size of tsunami generated by submarine landslides and the size of failure required to produce a 5 m surge at the coastline has been determined using the empirical equations developed by Ward and co-workers (Chesley and Ward 2006; Ward 2001, 2011). Flow depth at the coastline (Fd(0)) is defined as the tsunami height as it reaches the coastline; inundation distance X( max) as the distance inland from the coastline that the tsunami surge reaches; and run-up (R(Xmax)) as the topographic elevation reached at the inundation distance. Equations 4.1-2 use a landslide’s geometric characteristics

4-9 the8 surface is continuous without the apparent landslide feature (McAdoo et al. 2000). Landslide length is the distancethe8 surface from is continuouslandslide head without to landslide the apparent toe. Landslide landslide widthfeature is (theMcAdoo average et ofal. measurements2000). Landslide taken length every is 500the 203 distance from landslide head to landslide toe. Landslide width is the average of measurements taken every 500 203 m down the landslide scar, perpendicular to the landslide axis. Water depth is taken from the landslide centre of 204 m down the landslide scar, perpendicular to the landslide axis. Water depth is taken from the landslide centre of 204 mass. 205 mass. 206205 206 2.4 Coastal Topography (adjacent to slide sites) 2.4 Coastal Topography (adjacent to slide sites) 207 207 Coastal topography was obtained from 1 second Shuttle Radar Topographic Mission (SRTM) derived Digital SurfaceCoastal topographyModels (DSM) was Versioobtainedn 1.0 from (Geoscience 1 second ShuttleAustralia Radar 2012 Topographic), accessed throughMission the (SRTM) Geoscience derived Australia Digital 208 (GA)Surface National Models Elevation (DSM) Versio Data nFramework 1.0 (Geoscience Portal (NEDF;Australia nedf.ga.gov.au 2012), accessed). The through dataset the used Geoscience is a 1 arc Australia second 209208 (~30m)(GA) National gridded Elevation Digital Elevation Data Framework Model - PortalSmoothed (NEDF; (DEMs) nedf.ga.gov.au that represents). The ground dataset surface used istopography a 1 arc second with 210209 vegetation(~30m) gridded offsets Digital removed Elevation and data Mod smoothingel - Smoothed to reduce (DEMs) noise that(Slater represents et al. 2006 ground). The surface coverage topography of the DEMs with 211210 overvegetation the four offsets regions removed investigated and data is shown smoothing in Fig tos. reduce4.3-4.6 noise. (Slater et al. 2006). The coverage of the DEMs 212211 over the four regions investigated is shown in Figs. 4.3-4.6. 213212 The 30 m gridded DEMs was used to examine the east Australian coastal topography adjacent to each identified 214213 The 30 m gridded DEMs was used to examine the east Australian coastal topography adjacent to each identified+ 214 slide in the four regions using Fledermaus V7.3.3b software (http://www.qps.nl/). A topographic profile (>5 215 slide in the four regions using Fledermaus V7.3.3b software (http://www.qps.nl/). A topographic profile (>5+ 215 km) was generated from the coastline adjacent to each identified slide and used to help determine run-up and in- 216 km) was generated from the coastline adjacent to each identified slide and used to help determine run-up and in- 216 undation depths specific to each site (see Fig. 4.7). Fig. 4.3-4.6 show the locations of each topographic profile. 217 undation depths specific to each site (see Fig. 4.7). Fig. 4.3-4.6 show the locations of each topographic profile. 217 218 219218 219 2.5 Tsunami Calculations 2.5 Tsunami Calculations 220 The size of tsunami generated by submarine landslides and the size of failure required to produce a 5 m surge at 220Clarke (2014) Chapter 4 theThe coastlinesize of tsunami has been gen determinederated by submarine using the landslidesempirical andequations the size developed of failure by required Ward toand produce co-workers a 5 m ( Chesleysurge at 221 the coastline has been determined using the empirical equations developed by Ward and co-workers (Chesley 221and an andestimate Ward of2006 landslide; Ward 2001 mass 2011 velocity). Equations to generate 4.1-2 use maximum a landslide’s flow geometric depth characteristicsat the coastline and an(Fd (0)estimate) 222 and Ward 2006; Ward 2001 2011). Equations 4.1-2 use a landslide’s geometric characteristics and an estimate 222 of landslide mass velocity to generate maximum flow depth at the coastline (Fd(0)) (i.e. tsunami height). Func- 223(i.e. tsunami height). Function 4.5 enables maximum run-up (R(Xmax)) and inundation distance (Xmax) of landslide mass velocity to generate maximum flow depth at the coastline (Fd(0)) (i.e. tsunami height). Func- 223 tion 4.5 enables run-up (R(Xmax)) and inundation distance (Xmax) to be estimated and is specific to particular 224to be estimatedtion 4.5 enables and is runspecific-up (R (Xto particular)) and inundation coastline distance profiles (X directly) to be estimatedinland of and the is specificspecific tolandslide particular 224 coastline profiles directly inlandmax of the specific landslide site (semaxe Fig. 4.7). 225site (see coas Fig.tline 4.7). profiles directly inland of the specific landslide site (see Fig. 4.7). 225 226 226 • FlowFlow depth depth at coastline: at coastline: 227 Flow depth at coastline: 228227 (4.1) 𝟒𝟒/𝟓𝟓 𝟏𝟏/𝟓𝟓 (4.1) 228 𝑭𝑭 𝒅𝒅 𝟎𝟎 = 𝑨𝑨𝟎𝟎𝑷𝑷 𝒓𝒓 𝒉𝒉𝟎𝟎 229 𝟒𝟒/𝟓𝟓 𝟏𝟏/𝟓𝟓 𝒅𝒅 𝟎𝟎 𝟎𝟎 229 Propagation𝑭𝑭 𝟎𝟎 = 𝑨𝑨 and𝑷𝑷 𝒓𝒓 beaching𝒉𝒉 factor: 230 • PropagationPropagation and beaching and beaching factor: factor: 230 231 𝟐𝟐 (4.2) 𝟎𝟎.𝟑𝟑𝟑𝟑 𝒗𝒗𝒔𝒔!𝒗𝒗𝒕𝒕 𝟎𝟎.𝟔𝟔𝟔𝟔 𝑾𝑾 𝑳𝑳𝑳𝑳𝑳𝑳𝑳𝑳 !𝟑𝟑.𝟕𝟕𝟕𝟕 𝒗𝒗𝒕𝒕 𝑳𝑳𝑳𝑳𝑳𝑳𝑳𝑳 231 𝒗𝒗 !𝒗𝒗 𝟐𝟐 (4.2) 𝟎𝟎 𝑳𝑳𝑳𝑳𝑳𝑳𝑳𝑳 𝒉𝒉𝟎𝟎 𝟎𝟎.𝟑𝟑𝟑𝟑 𝒔𝒔 𝒕𝒕 𝒓𝒓 𝟎𝟎.𝟔𝟔𝟔𝟔 𝑨𝑨 𝑷𝑷 𝒓𝒓 = 𝟎𝟎. 𝟕𝟕𝟕𝟕𝟕𝟕𝟕𝟕𝟕𝟕 𝑾𝑾 𝑳𝑳𝑳𝑳𝑳𝑳𝑳𝑳 𝒆𝒆!𝟑𝟑.𝟕𝟕𝟕𝟕 𝒗𝒗𝒕𝒕 𝑳𝑳𝑳𝑳𝑳𝑳𝑳𝑳 232 𝟎𝟎 𝑨𝑨 𝑷𝑷 𝒓𝒓 = 𝟎𝟎. 𝟕𝟕𝟕𝟕𝟕𝟕𝟕𝟕𝟕𝟕 𝑳𝑳𝑳𝑳𝑳𝑳𝑳𝑳 𝒉𝒉𝟎𝟎 𝒆𝒆 𝒓𝒓 9 232 9 WhereWhere A00 isis the the initially initially generated generated surface surface elevation elevation,, P(r) is propagationP(r) is propagation factor P at distancefactor P r fromat distance the source r of 233 Where A0 is the initially generated surface elevation, P(r) is propagation factor P at distance r from the source of

233 the wave, vs is landslide speed, vt is tsunami speed at the landslide = √(gh0), Lref is a reference length, taken here from the source of the wave, vs is landslide speed, vt is tsunami speed at0 the landslide = √(gh0), Lref is a 234 the wave, vs is landslide speed, vt is tsunami speed at the landslide = √(gh0), Lref is a reference length, taken here 234 as 1 m. 235reference as 1length, m. taken here as 1 m. 235 236 • Run-up and Inundation depth: 236 Run-up and Inundation depth: 237 Run-up and Inundation depth: 237 PartPart 1 1 (solving(solving inundation depth depth (X max(X)max over) over flat flatground) ground) 238 Part 1 (solving inundation depth (Xmax) over flat ground) 238 (4.3)3) 239 (4.3) 239 𝟏𝟏.𝟑𝟑𝟑𝟑 𝟐𝟐 239 𝒎𝒎𝒎𝒎𝒎𝒎 𝒅𝒅 𝑿𝑿 𝒎𝒎𝒎𝒎𝒎𝒎 = 𝑭𝑭𝒅𝒅 𝟎𝟎 𝟏𝟏.𝟑𝟑𝟑𝟑 𝟏𝟏𝟏𝟏. 𝟕𝟕𝒏𝒏𝟐𝟐 240 𝑿𝑿 𝒎𝒎𝒎𝒎𝒎𝒎 = 𝑭𝑭𝒅𝒅 𝟎𝟎 𝟏𝟏𝟏𝟏. 𝟕𝟕𝒏𝒏 240 Part 2 (solving inundation depth (Xmax) allowing for arbitrary topography) 241 PartPart 2 2 (solving(solving inundation depth depth (X max(X) allowing allowing for arbitraryfor arbitrary topography) topography) 241 max) 242 (4.4) 242 𝟐𝟐 (4.4) 242 𝟐𝟐 4/5 1/5 4/5 1/5 (4.4) 𝒅𝒅𝑭𝑭𝒅𝒅 𝑿𝑿 𝟏𝟏𝟏𝟏.𝟕𝟕𝒏𝒏𝟐𝟐 𝒅𝒅𝒅𝒅 𝑿𝑿 4/5 1/5 4/5 1/5 𝒅𝒅𝑭𝑭𝒅𝒅 𝑿𝑿 𝟏𝟏𝟏𝟏.𝟕𝟕𝒏𝒏𝟎𝟎.𝟑𝟑𝟑𝟑 𝒅𝒅𝒅𝒅 𝑿𝑿 Fd 0 = A0P r 4/5h01/5Fd 0 = A0P r 4/5h01/5 243 𝒅𝒅 𝟎𝟎.𝟑𝟑𝟑𝟑 Fd 0 = A0P r h0 Fd 0 = A0P r h0 243 𝒅𝒅𝒅𝒅 = − 𝑭𝑭𝒅𝒅 𝟎𝟎 𝟎𝟎.𝟑𝟑𝟑𝟑 + 𝒅𝒅𝒅𝒅 F 0 = A P r h F 0 = A P r h 243 𝒅𝒅𝒅𝒅 = − 𝑭𝑭𝒅𝒅 𝟎𝟎 + 𝒅𝒅𝒅𝒅 Part 3 (run-up and inundation distance) 244 PartPart 3 3 (run-up(run-up and and inundation inundation distance distance)) 244 245 𝟐𝟐 (4.5)5) 245 𝟏𝟏𝟏𝟏.𝟕𝟕𝒏𝒏𝟐𝟐 245 𝟏𝟏𝟏𝟏.𝟕𝟕𝒏𝒏𝟐𝟐 (4.5) 𝒎𝒎𝒎𝒎𝒎𝒎 𝟎𝟎 𝟏𝟏𝟏𝟏.𝟕𝟕𝒏𝒏 𝒎𝒎𝒎𝒎𝒎𝒎 𝟎𝟎 𝒅𝒅 [ 𝑹𝑹(𝑿𝑿𝒎𝒎𝒎𝒎𝒎𝒎) − 𝑻𝑻(𝑿𝑿𝟎𝟎)] + 𝑭𝑭𝒅𝒅 𝟎𝟎 (𝑿𝑿𝒎𝒎𝒎𝒎𝒎𝒎 − 𝑿𝑿𝟎𝟎) = 𝑭𝑭𝒅𝒅 𝟎𝟎 246 𝒎𝒎𝒎𝒎𝒎𝒎 𝟎𝟎 𝑭𝑭𝒅𝒅 𝟎𝟎 𝒎𝒎𝒎𝒎𝒎𝒎 𝟎𝟎 𝒅𝒅 246 [𝑹𝑹(𝑿𝑿 ) − 𝑻𝑻(𝑿𝑿 )] + (𝑿𝑿 − 𝑿𝑿 ) = 𝑭𝑭 𝟎𝟎 Where X is distance inland from the coastline (X0 = 0 at coastline), T(X) is topographic elevation at location X. 247 WhereWhere X isis distance distance inland inland from from the coastline the coastline (X0 = 0 ( atX coastline) = 0 at coastline),, T(X) is topographic T(X) is topographicelevation at loc elevaation- X. 247 Land surface roughness is represented by Manning’s coefficient0 n which is taken as 0.015 for very smooth to- Land surface roughness is represented by Manning’s coefficient n which is taken as 0.015 for very smooth to- 248tion at Llocationand surface X. rLandoughness surface is represented roughness by Manning’sis represented coefficient by Manning’s n which is takencoefficient as 0.015 n forwhich very smoothis taken to- 248 pography, 0.03 urbanized/built land, and 0.07 densely forested landscape (Gerardi et al. 2008). Equation 4.3 249 pography, 0.03 urbanized/built land, and 0.07 densely forested landscape (Gerardi et al. 2008). Equation 4.3 249as 0.015 (d eforveloped very bysmooth Ward 2001topography, from Gerardi 0.03 et urbanized/built al. 2008) is applicable land, for and solving 0.07 inundation densely forested depth (X maxlandscape) over flat 250 (developed by Ward 2001 from Gerardi et al. 2008) is applicable for solving inundation depth (Xmax) over flat 250 ground. Flow depth decays faster with X as the Manning coefficient n (i.e. the friction) gets larger. By integrat- 251(Gerardi ground. et al. Flow2008). depth Equation decays faster 4.3 (developedwith X as the by Manning Ward coefficient2001 from n (Gerardii.e. the friction) et al. 2008) gets la rgeris applicable. By integra t- 251 ing equation 4.3, equation 4.4 is modified to allow for arbitrary topography when determining run-up and inun- 252for solving ing equation inundation 4.3, equation depth ( 4.Xmax4 is) modifiedover flat to ground.allow for Flowarbitrary depth topography decays whenfaster determining withX as the run Manning-up and inu n- 252 dation depth. The conversion function 4.5 is generated from equation 4 (which is an integration of equation 4.3). 253 dation depth. The conversion function 4.5 is generated from equation 4 (which is an integration of equation 4.3). 253coefficient n (i.e. the friction) gets larger. By integrating equation 4.3, equation 4.4 is modified to- al Function 4.5 is applied by defining Xmax as the particular X that makes the equality (i.e. left side = right side). 254 Function 4.5 is applied by defining Xmax as the particular X that makes the equality (i.e. left side = right side). 254low for Thearbitrary particular topography Xmax value iswhen determined determining from the run-up site-specific and coastlineinundation profiles depth. adjacent The to conversioneach slide site func (see- 255 The particular Xmax value is determined from the site-specific coastline profiles adjacent to each slide site (see 255 Fig. 4.7, 4.9). These profiles lie perpendicular to the coastline. Run-up (R(X )) and inundation depth (X ) can tion 4.5Fig. is generated 4.7, 4.9). These from profiles equation lie perpendicular4 (which is an to integrationthe coastline. ofRun equation-up (R(X max4.3).)) and Function inundation 4.5 depth is applied (Xmax) can 256 Fig. 4.7, 4.9). These profiles lie perpendicular to the coastline. Run-up (R(Xmax)) and inundation depth (Xmax) can 256 therefore to be estimated specifically for each slide based on their adjacent coastal topography. Flow depth at 257by defining therefore Xmax to as be the estimated particular specifically X that for makes each slidethe equalitybased on their(i.e. adjacentleft side coastal = right topography. side). The Flow particular depth at 257 the coastline (Fd(0)) is also site specific and determined from equation 1. It is important to note that function 258 the coastline (Fd(0)) is also site specific and determined from equation 1. It is important to note that function 258Xmax value is determined from the site-specific coastline profiles adjacent to each slide site (see Fig. 4.7, 4.5 provides an indicative estimate of run-up (R(Xmax)) and inundation depth (Xmax) that holds when certain val- 259 4.5 provides an indicative estimate of run-up (R(Xmax)) and inundation depth (Xmax) that holds when certain val- 259 ues for run-up (R(X )) and inundation depth (X ) are chosen. These values are accepted to be sensible and ues for run-up (R(Xmax)) and inundation depth (Xmax) are chosen. These values are accepted to be sensible4-10 and 260 ues for run-up (R(Xmax)) and inundation depth (Xmax) are chosen. These values are accepted to be sensible and 260 plausible due to comparisons with the coastal profile (Fig. 4.7). 261 plausible due to comparisons with the coastal profile (Fig. 4.7). 261 262 262 Clarke (2014) Chapter 4

4.9). These profiles lie perpendicular to the coastline. Run-up (R(Xmax)) and inundation depth (Xmax) can therefore be estimated specifically for each slide based on their adjacent coastal topography. Flow depth at the coastline (Fd(0)) is also site specific and determined from equation 1. It is important to note that function 4.5 provides an indicative estimate of run-up (R(Xmax)) and inundation depth (Xmax) that holds when certain values for run-up (R(Xmax)) and inundation depth (Xmax) are chosen. These values are accepted to be sensible and plausible due to comparisons with the coastal profile (Fig. 4.7).

4.4 Results

4.4.1 Calculation of Required Slide Block Sizes Calculations were performed using Equation 4.1 (see Section 4.3.4), which varied water depth (m), slide thickness (m), and slide width (km), to identify necessary block characteristics for submarine landslides. The water depths examined were divided into 500 m “depth bands”, a minimum slide velocity of 20 ms-1 is assumed, and an average distance from the coastline of 50 km is applied. Using these characteristics, a tsunami flow depth of 1 m is produced by blocks with a thickness between 50- 200 m and width between 0.4-2 km.

A 5 m tsunami flow depth at coastline under the same initial conditions requires landslide blocks with dimensions outlined in Table 4.2 for each specific water depth. Characteristic examples of necessary minimum sizes are: a) 50 m thick and 8 km wide (thin block); b) 200 m thick and 2 km wide (thick block); and c) 100 m thick and 4.5 km wide (intermediate block) for 1000 m water depth; or a) 50 m thick and 13 km wide (thin block); b) 200 m thick and 3.2 km wide (thick block); and c) 100 m thick and 6.5 km wide (intermediate block) for 2000 m water depth. A 5 m tsunami surge produced by a relatively thin block (i.e. ~50 m or less) needs to be of significant width (>5 km; especially in water depths >1500 m), while a 5 m tsunami surge produced by a relatively thick block (i.e. >150 m) requires a smaller width (<5 km) to produce the same flow depths at coastline as thinner slide at an equivalent water depth.

Table 4.2 shows that slide width needs to increase with depth (for slides of the same thickness) and slides must have a width to thickness ratio greater than 6 in order to produce a coastline flow depth >5 m. The sensitivity analysis provides a set of approximate minimum threshold values for slide width and thickness to apply when trying to identify tsunamogenic submarine landslides from bathymet-

4-11 Clarke (2014) Chapter 4 ric data. Identification of potential Table 4.2: Sensitivity analysis results of Equation 4.1 summarizing parameter combinations required to produce ~5 m tsunamogenic slides focuses firstly flow depth at coastline in water depths between 500-2500 m on slide width, where submarine for landslide blocks with dimensions similar to those outlined in Table 4.3. Initial condition variables tested include: water landslide scars must meet a mini- depth (h0), slide thickness (t), and slide width (W). Landslide -1 mum width criteria for each water velocity (vs) of 20 ms was used for the sensitivity analysis and distance to the coastline was averaged to ~50 km. depth range in order to be consid- Slide Width to Water Depth Slide width ered, which are minimum widths of thickness t thickness (m) W (km) a) 1 km for an 200 m thick slide in (m) ratio 500 50 5 100 500 m water depth; b) 2 km for an 100 2.5 25 200 m thick slide in 1000 m water 150 1.5 10 depth; c) 2.7 km for an 200 m thick 200 1.2 6 1000 50 8 160 slide in 1500 m water depth; d) 3.2 100 4.5 45 km for an 200 m thick slide in 2000 150 3 20 m water depth; and e) 4 km for an 200 2 10 1500 50 12 240 200 m thick slide in 2500 m water 100 6 60 depth. The approximate thickness 150 3.5 23.3 of the slide is then measured (and 200 2.7 13.5 2000 50 13 260 hence the width to thickness ratio 100 6.5 65 found) which then yields the tsu- 150 3.5 30 namogenic potential of the slide i.e. 200 3.2 16.3 2500 50 15 300 the size of the surge. These measured 100 7 70 minimum criteria for landslide slab 150 5 33.3 size indicates that tsunamogenic 200 4 20 Maximum landslide velocity (v ) = 20 ms -1 submarine landslide scar features s should be obvious in the high-resolution multibeam data (i.e. these features are >1 km in width). This assumes that the submarine landslides fail as translational slab slides. A conservative approach was used to identify the minimum size of slide block required to generate a tsunami surges at the coastline.

4-12 Clarke (2014) Chapter 4 N 30 km VE = 3 R1-8 R1-4 Port Stephens R1-3 R1-7 ii Newcastle R1-6

Norah Head

-1500m

ii -500m -2500m Sydney R1-5 R1-2 Botany Bay i

R1-1 -2500m Bay

Jervis

-2500m

b -1500m

-500m

-1500m -500m i Depth (m) 0 300 700 1100 1500 1900 2300 2700 3100 3500 4500 R1-8 R1-4 R1-3 Tuncurry R1-7 Sugarloaf Point N 50 km p Port Stephens R1-6 p R1-4-8 R1-3 Newcastle p R1-5 R1-2 R1-7 Norah Head p

Zoomed in view of Region 1 (white box). The extent of the high-resolution multibeam bathymetry The extent of the high-resolution 1 (white box). for the continental slope is coloured in view of Region a) Zoomed

Region 2 Region -2500m

Botany Bay R1-6 R1-1 -1500m

p -500m p p R1-5 Bay Sydney Jervis a R1-2

R1-1 Region 1 Region Figure 4.3: Figure according to the depth scale. The investigated submarine landslides scars and potential future tsunamogenic submarine landslides are outlined in blue tsunamogenic submarine landslides are submarine landslides scars and potential future to the depth scale. The investigated according adjacent to each slide the coastal profiles lines mark 4.3. Red Table in to slide names shown (R1-1 to R1-8) correspond Numbers and pink respectively. (blue – landslides scars; pink the location of submarine landslides investigated 8. b) DEM of the slope geometry 1, showing of Region in Fig. shown in details of zoomed show exaggeration 3x. Insets Vertical onto the continental margin. landward looks west, – potential landslides sites). Perspective i) R1-1, R1-2, and R1-5; ii) R1-6, R1-7, R1-3, R1-4, R1-8. submarine landslides investigated:

4-13 Clarke (2014) Chapter 4 - N R2-3 10 km VE = 3

-1500m

Coffs Harbour -500m

-2500m R2-2 Smokey Cape ii N VE = 3 2.5 km R2-1

-2500m

-1500m -500m R2-5 Tuncurry R2-4 Sugarloaf Point Port Stephens i

-500m N -2500m

-1500m VE = 3 2.5 km i b N Depth (m) 25 km 0 300 700 1100 1500 1900 2300 2700 3100 3500 4500 R2-1 R2-3 R2-2 R2-4

R2-5 -2500m

p -1500m Coffs Harbour

R2-3 Smokey Cape p -500m R2-1 p Zoomed in view of Region 2. The extent of the high-resolution multibeam bathymetry accord 2. The extent of the high-resolution in view for the continental slope is coloured of Region a) Zoomed R2-2 Sugarloaf Point

Tuncurry Region 2 Region p p R2-5 a R2-4 Figure 4.4: Figure ing to the depth scale. The investigated submarine landslides scars and potential future tsunamogenic submarine landslides are outlined in blue and tsunamogenic submarine landslides are submarine landslides scars and potential future ing to the depth scale. The investigated adjacent to each slide the coastal profiles lines mark 4.3. Red Table in to slide names shown (R2-1 to R2-5) correspond Numbers pink respectively. (blue – landslides scars; pink the location of submarine landslides investigated showing 2 slope geometry, 8. b) DEM of Region in Fig. shown in details of zoomed show exaggeration 3x. Insets Vertical onto the continental margin. landward looks west, potential landslides sites). Perspective i) R2-4 and R2-5; ii) R2-1, R2-2, R2-3. submarine landslides investigated:

4-14 Clarke (2014) Chapter 4 N 10 km VE = 3 Morten Island R3-15 R3-9 Stradbroke Island

N -2500m Brisbane VE = 3 R3-8 iv R3-7

Tweed Heads Tweed

-500m

-2500m -500m

-1500m -1500m

-1500m -500m ii Byron Bay R3-13 iii N VE = 3 N VE = 3 Balina iv R3-6 Evans Head R3-12 N R3-4 VE = 3 R3-5 ? ii Yamba ? R3-3 R3-10 R3-11 R3-2 R3-1 R3-14 -500m

-1500m

Coffs Harbour

-2500m -500m

i -2500m -1500m i iii b N

25 km Region 3 Region R3-9 R3-7 R3-8 R3-4 R3-13 R3-12 R3-15 R3-2 R3-11 R3-14 R3-1 R3-5 R3-6

R3-10

R3-3

-2500m -500m

Byron Bay p -1500m Evens Head Morten Island p p p R3-14 Balina p R3-13 p p p R3-12 R3-10-11 R3-9 p p p R3-8 p R3-7 Yamba p R3-15 p Tweed Heads Tweed R3-6 R3-5 R3-4 R3-3 Zoomed in view of Region 3 (white box). The extent of the high-resolution multibeam bathymetry The extent of the high-resolution 3 (white box). for the continental slope is coloured in view of Region a) Zoomed R3-2 R3-1 Depth (m) 0 300 700 1100 1500 1900 2300 2700 3100 3500 4500 a Figure 4.5: Figure according to the depth scale. The investigated submarine landslides scars and po-tential future tsunamogenic submarine landslides are outlined in blue tsunamogenic submarine landslides are submarine landslides scars and po-tential future to the depth scale. The investigated according adjacent to each slide the coastal profiles lines mark 3. Red Table in to slide names shown (R3-1 to R3-15) correspond Numbers and pink respectively. (blue – landslides scars; pink poten - the location of submarine landslides investigated showing 3 slope geometry, 8. b) DEM of Region in Fig. shown in details of subma - zoomed show exaggeration 3x. Insets Vertical onto the continental margin. landward looks west, tial landslides sites). Perspective i) R3-1 to R3-6; ii) R3-7 R3-9; iii) R3-10 R3-12, and R3-14; iv) R3-13 R3-15. rine landslides investigated:

4-15 Clarke (2014) Chapter 4 N

R4-16

R4-11 10 km VE = 3

R4-5

R4-20 v & vi & v

R4-18

R4-15

R4-10

R4-6

R4-4

-500m

Fraser Island

-500m

-1500m -2500m

R4-14

R4-13

R4-9

R4-3

-1500m -2500m N VE = 3

R4-12

R4-19 iv Double Island Point

R4-2 N N N VE = 3 VE = 3 VE = 3

R4-21 iii

-1500m

-500m -2500m

N -1500m Noosa Heads

R4-16

R4-20 VE = 3 N VE = 3

R4-14

R4-15

R4-1

-500m -2500m

-1500m

R4-13

R4-19 Morten Island Brisbane

R4-17

-2500m -2500m

R4-12

R4-7

-1500m -500m

i -1500m Stradbroke Island iii v i

b -500m

R4-8

N -2500m

R4-8 R4-1 R4-6 -1500m

R4-5 R4-17 R4-3 -500m R4-12 R4-7 R4-9 R4-14 R4-4 R4-11 R4-21 R4-10 R4-18 R4-16 R4-2 p p Morten Island R4-19 p p R4-13 Stradbroke Island R4-15 p R4-20 R4-11 Double Island Point R4-18 R4-10 R4-7-8 Coffs Harbour R4-9 p p p Fraser Island p p p p adjacent R4-3 p R4-2 p Brisbane R4-5 R4-20 R4-21 R4-4 R4-1-17 R4-7-8 R4-6 p R4-12 Depth (m) 0 p p 300 700 1100 1500 1900 2300 2700 3100 3500 4500 p 50 km

R4-19 Region 4 Region R4-14 p R4-16 p R4-13 R4-15 a Zoomed in view of Region 4. The extent of the high-resolution multibeam bathymetry for the continental slope is coloured according to multibeam bathymetry according 4. The extent of the high-resolution in view for the continental slope is coloured of Region 4.6: a) Zoomed Figure re - outlined in blue and pink tsunamogenic submarine landslides are submarine landslides scars and potential future the depth scale. The investigated Fig. in adjacent to each slide shown coastal profiles lines mark 4.3. Red Table in to slide names shown (R4-1 to R3-21) correspond Numbers spectively. (blue – landslides scars; pink potential the location of submarine landslides investigated showing 4 slope geometry, 8. b) DEM of Region in details of submarine landslides zoomed show exaggeration 3x. Insets Vertical onto the continental margin. landward looks west, sites). Perspective i) R4-1 and R4-17; ii) R4-2 to R4-6; iii) R4-7, R4-8, R4-21; iv) R4-9 R4-11, R4-18; v) R4-12 R4-16, R4-19, R4-20 investigated: west-south-west). vi) R4-12 to R4-16, R4-19, and R4-20 (perspective west-north-west); (perspective

4-16 Clarke (2014) Chapter 4

Table 4.3: Catalogue of the submarine landslides investigated within the study area including morphometric parameters for each submarine landslide listed (see Fig. 3-6 for landslide locations).

Measured values include: slide thickness (t), slide length (L), slide width (W), water depth (h0), and distance to the coastline (r). Water depth Distance from Slide Slope width W at slide h (m) slide to ID Number Region* Slide Name Adjacent Coastal Location Latitude Longitude Thickness t 0 (m) (centre of adjacent (m) mass) coastline r (m)

R1-1 Region 1 Shovel Slide Sydney/Illawarra 34°26.5364'S 151°24.0720'E 75 5000 1000 40000

R1-2 Region 1 Bulli Slide Sydney/Illawarra 34°24.4354'S 151°32.9894'E 200-500 10000 1700 47000

R1-3 Region 1 Birubi Slide Myall Region 33°7.3396'S 152°35.8212'E 70 3820 1300 54000

R1-4 Region 1 Yacaaba Slide Myall Region 32°52.5398'S 152°44.7720'E 70 3200 925 53000

R1-5 Region 1 Illawarra Potential Slide** Northern Illawarra Region 34°11.9473'S 151°41.0684'E 150 6500 1200 44000

R1-6 Region 1 Newcastle Potential Slide 1** Newcastle Region 33°27.2343'S 152°23.9056'E 100 10000 2200 75000

R1-7 Region 1 Newcastle Potential Slide 2** Newcastle Region 33°13.7436'S 152°22.0871'E 50 11000 800 60000

R1-8 Region 1 Port Stephens Potential Slide** Port Stephens Region 152°48.9835'E 32°56.5689'S 250 7000 2200 62500

R2-1 Region 2 Kempsey Slide Kempsey 31°7.7927'S 153°27.5410'E 250 2100 1800 42000

R2-2 Region 2 Smokey Cape Slide Smokey Cape 31°0.7781'S 153°19.5249'E 100 2400 650 30000

R2-3 Region 2 Scotts Head Slide Scotts Head 30°45.5688'S 153°24.3297'E 150 2300 1800 38000

R2-4 Region 2 Sugarloaf Point Potential Slide 1** Sugarloaf Point 32°32.4231'S 152°58.7542'E 50-100 4800 1100 41000

R2-5 Region 2 Sugarloaf Point Potential Slide 2** Sugarloaf Point 32°27.1865'S 152°58.6168'E 100 5000 800 40000

R3-1 Region 3 North Solitary Island-1 Slide Grafton 29°59.9536'S 153°56.5273'E 200 3200 3400 68000

R3-2 Region 3 North Solitary Island-2 Slide Grafton 29°58.4208'S 153°58.4387'E 200 3400 3350 69000

R3-3 Region 3 North Solitary Island-3 Slide Grafton 29°56.1291'S 153°58.7672'E 300 4460 3300 68000

R3-4 Region 3 Grafton Lower Slope Slide Grafton 29°50.2879'S 154°3.1481'E 300 4300 3400 71000

R3-51 Region 3 Grafton Mid Slope Slide^ Grafton 29°46.1991'S 153°54.4593'E 200 8000 2000 57000

R3-6 Region 3 Ulmarra Slide Ulmarra 29°29.0548'S 153°49.3431'E 80 4000 700 42000

R3-7.12 Region 3 Clarenace Canyon Slide^^ Clarenace 29°10.5134'S 153°57.7814'E 150 7000 2000 44000

R3-7.23 Region 3 Clarenace Canyon Slide^^^ Clarenace - - 300 11500 2000 44000

R3-8 Region 3 Knickpoint Slope-1 Slide Ballina 29°4.2173'S 153°56.7084'E 75 8400 1400 42000

R3-9 Region 3 Knickpoint Slope-2 Slide Ballina 28°57.6697'S 153°56.1474'E 100 13000 1500 38000

R3-10 Region 3 Pre-Byron Slide Byron 28°37.2058'S 153°57.2120'E 150 2255 1500 34000

R3-11 Region 3 Byron Slide Byron 28°37.2058'S 153°57.2120'E 220 3558 1000 34000

R3-12 Region 3 Tweed Canyon Slide Tweed 28°27.4006'S 153°59.3802'E 250 1600 1900 42000

R3-13 Region 3 Cudgen Slide Cudgen 28°14.3814'S 153°58.1528'E 50 5338 600 38150

R3-14 Region 3 Byron Potential Slide** Byron 28°39.4477'S 153°56.8219'E 200 3200 800 30000

R3-15 Region 3 Nerang Plateau Potential Slide** Stradbroke Island (south) 27°44.6877'S 154°4.1602'E 50 8000 800 60000

R4-1 Region 4 Morten Island Slide (Block Base) Morten Island 27°10.0743'S 154°2.8982'E 250 4300 2400 60000

R4-2 Region 4 Bribie Bowl Slide Bribie 26°45.2528'S 153°43.5712'E 125 2465 600 57860

R4-3 Region 4 Sunshine Coast-1 Slide Sunshine Coast 26°37.2387'S 153°54.9225'E 200 4900 1900 80000

R4-4 Region 4 Sunshine Coast-2 Slide Sunshine Coast 26°33.1619'S 153°55.2457'E 250 6500 1700 80000

R4-5 Region 4 Sunshine Coast-3 Slide Sunshine Coast 26°28.6873'S 153°54.8729'E 200 7000 1600 78000

R4-6 Region 4 Sunshine Coast-4 Slide Sunshine Coast 26°35.1506'S 153°57.1847'E 250 6000 2600 85000

R4-7 Region 4 Great Sandy National Park Slide 1 Gympie 26°8.4351'S 153°58.7321'E 150 10000 900 90000

R4-8 Region 4 Great Sandy National Park Slide 2 Gympie 26°9.8681'S 154°1.6510'E 100 13000 1550 90000

R4-9 Region 4 Southern Fraser Island Slide Fraser Island 25°38.6144'S 153°57.8482'E 250 17000 1500 85000

R4-10 Region 4 Wide Bay Canyon Slide Fraser Island 25°27.1517'S 153°56.5601'E 150 8400 1400 77000

R4-11 Region 4 Mid Fraser Island Mid-Slope Slide Fraser Island 25°21.2596'S 153°57.7265'E 160 4884 1400 75000

R4-12 Region 4 North Fraser Island Slide 1 Fraser Island 24°58.9941'S 153°43.6838'E 150 3400 1200 36000

R4-13 Region 4 North Fraser Island Slide 2 Fraser Island 24°52.6960'S 153°38.5905'E 200 6600 560 36000

R4-14 Region 4 North Fraser Island Slide 3 Fraser Island 24°48.2603'S 153°38.6433'E 100 5780 1100 37500

R4-15 Region 4 North Fraser Island Upper-Slope Slide Fraser Island 24°45.5392'S 153°35.4744'E 150-200 2900 750 32000

R4-16 Region 4 North Fraser Island Slide 4 Fraser Island 24°40.7090'S 153°33.7323'E 50 5580 950 28000

R4-17 Region 4 The Block Potenital Slide** Morten Island 27°10.9071'S 154°2.8982'E 500 14000 1500 52000

R4-18 Region 4 Wide Bay Canyon Potential Slide** Fraser Island 25°24.1243'S 153°56.2698'E 150 8000 1100 75000

R4-19 Region 4 North Fraser Island Poential Slide 1** Fraser Island 24°55.7588'S 153°42.2340'E 300-350 5600 800 37500

R4-20 Region 4 North Fraser Island Poential Slide 2** Fraser Island 24°43.0853'S 153°35.1053'E 80-100 4600 890 29500

R4-21 Region 4 Double Island Point Potential Slide** Double Island 25°50.7738'S 153°59.3083'E 150 6000 1300 89000 * see Fig. 1 for Region location **Potential slide 1 Minimum values, full slide scar not imaged (see Fig. 5) 2 Minimum values 4-17 3 Maximum values Clarke (2014) Chapter 4

Figure 4.7: Example of how to determine run-up (R(Xmax)) and inundation distance (Xmax) from coastal topography profiles using coastal elevation and distance inland from the coastline specific for each slide site (see text for details). In this example the Byron Slide R3-11 is shown, with R(Xmax) and -1 -1 Xmax determined for slide velocities 20 ms and 40 ms . Firstly, the slide adjacent coastal topography is found (R3-11p; see a). For the Byron Slide, flow depth at the coastline Fd(0) is approximately 10.3 m and 24.0 m for slide velocities 20 ms-1 and 40 ms-1 respectively (see Table 4 for all values). Fd(0) becomes the right side of Function 5. For the left side of the function to equal the right side, the lowest appropriate combination of Xmax and R(Xmax) values that satisfy the function are chosen using the shoreline profile (see b). In this case, when Xmax = 118 m and R(Xmax) = 9.5 m (red dashed lines), then left hand side equals 10.3, which satisfies the right hand side of Function 5 for slide velocity 20 -1 ms . Similarly, when Xmax = 1898 m and R(Xmax) = 14.02 m (blue dashed lines), which satisfies Func- tion 5 for slide velocity 40 ms-1.Once acceptable minimum values have been deter-mine from the topography that hold in Function 5, then Xmax and R(Xmax) are found.

4-18 Clarke (2014) Chapter 4

4.4.2 Morphometric Characteristics of Slide Regions

The 36 submarine landslide scars (Fig. 4.3-4.6; Table 4.2) are in general U-shaped in cross- section (1-13+ km wide and 50-500 m thick) backed by an amphitheatre shaped crestal zone. In each case, landslide morphology is similar to the classical circular failure profile described by Varnes (1978), but elongated in the downslope, longitudinal profile. Slide thickness is conservative, as calculations do not account for any recent sediment accumulation, therefore underestimating the depth to failure surface. However, the low sedimentation rates (between 0.3 and 1.2 m/10 ka) over the margin (see Chapter 2), the apparent young age (<25 ka) of for some of the slides (Clarke et al. 2012, 2014), and the lack of sediment drape observed in subbottom profiles over slide scars (see Glenn et al. 2008; Chapter 2) suggests that only a small amount of recent sediment has accumulated since the slides took place. Failures are assumed to occur as one complete landslide block, rather than as multiple landslides from the same site. This may generate an overestimated volume of the landslide block and ability of the slide to produce tsunami waves ≥5 m at the coastline. Two to six potential landslide masses have also been identified within each region.

4.4.2.1 Region 1 Identification of submarine landslides from Region 1 was first undertaken by Glenn et al. (2008) investigation of the NSW continental margin. This survey described a number of large submarine failures between Jervis Bay and Coffs Harbour, and identified a number of potential failure sites. Post slide sediment drapes are thicker (>6 m) than the length of the corer deployed. The actual age of the identified failures could not be determined from their data. Glenn et al. (2008) suggested substantial ages of several million years (or even tens of millions). Minimum ages of 40 ka were reported (i.e. pre-Holocene). Radiocarbon ages for these sediments exceed 8850 ka within the top 60 cm of all six gravity cores dated, as well as exceeding 23 ka BP at the base of all cores (depths 2.8 – 4.65 m within core), with four ages >40 ka. Significant areas of the continental slope in this region have been prone to sediment mass wasting over time (Glenn et al. 2008). Glenn et al. (2008) also suggest that downslope sediment movement was not coherent and that the slide materials probably disintegrated.

The high-resolution multibeam bathymetry shows submarine landslide scars in Region 1 that appear subdued, indicating older slide scars. No evidence of submarine landslides is evident above ~800 m water depth and the tsunamogenic slides identified in Region 1 occur in deeper water (>1700

4-19 Clarke (2014) Chapter 4 m), which requires them to be relatively thick (>70 m) in order to generate a 5 m plus surge at the coastline. The slope gently inclined (1-4°), narrow (~20-35 km), and incised by canyons. A change of margin orientation occurs at the boundary between Region 1 to Region 2 (see Fig. 4.1a, 4.3), coincident with the transition between Sydney Basin and New England Fold Belt lithologies. Basement lithologies are Permo-Triassic sedimentary rocks of the Sydney Basin (Fig. 4.1b).

Two slides meet the minimum criteria to produce a 5 m surge at the coastline ≥5 m at slide velocity 20ms-1 and are either moderately thick (>75 m) to thick (>100 m) and wide to very wide (5-10 km). All identified slides are located in water depths >1700 m. For example, R1-1 (Shovel Slide, 7.97 km3) and R1-2 (Bulli Slide, 20 km3) are the two largest slides for central NSW. Both the slides are large to very large: R1-1 is an example of a moderately thick (75 m), moderately wide to wide (5 km) slide, located in 1000 m water depth; R1-2 is an example of an extremely thick (200-500 m), very wide (10 km) slide, located in 1700 m water depth. Full details on the slides from Region 1 are given in Figure 4.3 and Table 4.2.

Four potential slides (R1-5 to R1-8) have been identified that meet the minimum criteria to produce a wave at the coastline ≥5 m for Region 1. These have similar general characteristics to slides that have already failed: thick to very thick (50- 250 m) and wide to very wide (7-11 km), and are located in water depths 800-2200 m.

4.4.2.2 Region 2 Identification of submarine landslides from Region 2 was also undertaken by Glenn et al. (2008) investigation of the NSW continental margin (see Region 1 above). Submarine landslide scars in Re- gion 2 present subdued, rounded morphology similar to the slide scars in Region 1 and are suspected to be of a similar age (Glenn et al. 2008). No evidence of submarine landslide activity is apparent above ~800 m water depth and relatively thick (>100 m) slides are required to generate an onshore surge greater than 5 m. The slope is relatively gentle (3-7°), very narrow (~15-35 km), and canyon incised (slope in canyon incised regions increases to >12° - e.g. offshore Coffs Harbour). A change of margin orientation occurs at the boundary between Region 1 to Region 2 (see Fig. 4.1a, 4.4), coincident with a narrowing of the continental margin. Basement lithologies are Upper Palaeozoic sedimentary, igneous, and metamorphic rocks of the New England Fold Belt (Fig. 4.1b).

Two tsunamogenic slides meet the minimum criteria to produce a wave at the coastline ≥5 m

4-20 Clarke (2014) Chapter 4 for Region 2 at slide velocity 20ms-1, they are thick (100-250 m), narrow (2.1-2.4 km), and located in either shallow (650 m) or deep (1800 m) water depths. Slide R2-1 (Kempsey Slide) is a thick (250 m), narrow (2.1 km) slide thick, located in deep water (1800 m). R2-1 can produce flow depths at the coastline of 5 m, run-up of 2.6 m, and inundation distance of 263 m. Similarly, R2-2 (Smokey Cape Slide) is a thick (100 m), narrow (2.1 km) slide, however it is located in shallow water (650 m). R2-2 can produce flow depths at the coastline of 5.6 m, run-up of <1 m, and inundation distance of 605 m. Full details on the slides from Region 1 are given in Figure 4.4 and Table 4.3.

Two potential slides (R2-4 to R2-5) have been identified that meet the minimum criteria to produce a wave at the coastline ≥5 m for Region 2. These have similar general characteristics to slides that have already failed: thick (50- 100 m) and narrow (4.8-5 km), and are located in 800-1100 m water depth.

4.4.2.3 Region 3 The larger submarine landslides present in Region 3 were identified by Boyd et al.’s (2010) investigation of the northern NSW/southern QLD continental margin, expanded during the SS2013-V01 voyage (Hubble et al. 2013). These surveys identified a number of large submarine failures between Coffs Harbour and North Stradbroke Island ranging from more commonly occurring small slides with volumes of less than 0.5 km3 to large slides that displace more than 20 km3 of sediment . They also identified several potential failure sites. Radiocarbon ages from the sediment indicate that three of submarine landslides in Region 3 are geologically young (< 25 ka) (Boyd et al. 2010; Clarke et al. 2012; Chapter 2).

Slope failures occur in the two dominant slope morphologies present in the region identified by Boyd et al. (2010) and Clarke et al. (2012, 2014) which are the: 1) relatively steep (3-7°) and canyon incised slope; and 2) relatively gentle slope (1-3°) of the Yamba and Nerang Plateau. Significant areas of the continental slope in this region have been prone to sediment mass wasting over time (Boyd et al. 2012; Clarke et al. 2014). Slide scars in Region 3 are more sharply defined than slide scars from Regions 1 and 2, suggesting younger slides and/or a thinner, post-slide sediment drape. Recent submarine landslide activity is obvious on the continental slope between depths of 400 m and 3000 m, with tsunamogenic slides identified in Region 3 also present in 400 m to 3000+ m water depth. Region 3 presents as a younger and more active section of the EA margin, with the number and size of observable past and potential submarine landslides increasing in this region when compare to Region’s 1 and

4-21 Clarke (2014) Chapter 4

2. Tsunamiogenic slides range from thick (50+ m) with varying width (<5 km to 10+ km), to thin (50 m) and wide (>5 km) with large surface area (>50 km2) to generate a 5 m plus surge at the coastline. The slope is narrow (~30-50 km), and alternates between relatively gentle (1-3°) and relatively steep (3-7°) slopes. The gentle plateau sections of the margin (Yamba and Nerang plateau) present abundant, overlapping and intercalated thin slides (<50 m) suggesting repetitive events. The steep slopes are heavily canyon incised, reach local gradients in excess of 30-40°, and have thicker submarine landslide scars (75-250+ m). Basement lithologies are Upper Palaeozoic sedimentary, igneous, and metamorphic rocks of the New England Fold Belt and some outcrops of Clarence-Morten Basin (Fig. 4.1b). A change of margin orientation occurs at the boundary between Region 2 to Region 3 (see Fig. 4.1a, 4.5), coincident with a widening of the continental margin.

Ten tsunamogenic slides meet the minimum criteria to produce a wave at the coastline ≥5 m for Region 3 at slide velocity 20ms-1, that are either thin (<50 m) to very thick (>200 m), and narrow (<5 km) to very wide (>10 km), and located in shallow (600 m) or very deep (3400 m) water depths. Some specific examples are: R3-6, a moderate thickness (80 m), narrow (4.0 km) canyon slide, located in shallow water (700 m); R3-7, a very thick (150-300 m), wide (7-11.5 km) canyon slide, located in deep water (2000 m); R3-11, a very thick (220 m) but narrow (3.6 km) canyon slide, located in intermediate water depth (1500 m); and R3-13 is an example of a thin (50 m), but wide (5.4 km) plateau slide, located in a shallow water (600 m). The complete set of slides and their characteristics are given in Figure 4.5 and Tables 4.2,4.4-4.5.

Two potential slides (R3-14 to R2-15) have also been identified as examples that meet the minimum criteria to produce a wave at the coastline ≥5 m for Region 3. These are located in 800 m water depth and have similar general characteristics to tsunamogenic slides that have already failed from either the plateau or canyon sections of the margin: canyon slides are thick (100-200 m) and narrow (3.2 km) to wide (>5 km); plateau slides are thin (50 m) and wide (8 km).

4.4.2.4 Region 4 Identification of submarine landslides from Region 4 was undertaken by Hubble et al.’s (2013) investigation of the northern NSW and southern QLD continental margin. This survey described a number of large submarine failures between North Stradbroke Island and North Fraser Island, and identified a number of potential failure sites. The continental slope of Region 4 is dissected by three large submarine canyons offshore northern Fraser Island, Wide Bay and Caloundra (i.e. the Fraser

4-22 Clarke (2014) Chapter 4

Canyons, the Wide Bay Canyon, and the Noosa Canyons – see Fig. 4.10). The slope is very narrow (~15-25 km), steep (7-10°), heavily canyon incised, and can reach local gradients in excess of 30-40°. A change of margin orientation occurs at the southern boundary of Region 4 (see Fig. 4.1a, 4.6), coincident with a steeping and widening of the continental margin. The northern boundary of Region 4 is defined by the tip of Fraser Island and the change in margin orientation from NNE to NNW. Basement lithologies are Upper Palaeozoic sedimentary, igneous, and metamorphic rocks of the New England Fold Belt and some outcrops of the Nambour Basin (Fig. 4.1b).

The continental slope in Region 4 has shed several large submarine slides and is currently subject to active canyon incision, as well as generating mass gravity flows and turbidites of suspected Holocene age (Fletcher et al. 2013; Hubble 2013). The morphology of this segment of the continental slope suggests that the large, landslide olistostrome thought to have been generated from a site located near the south of Fraser Island (see Fig. 4.10) was more likely to have been shed from this slope in the Miocene as suggested by Hill et al. (1992) rather than in relatively recent geologic time as suspected by Hubble (2013).

Slide scars in Region 4 are sharply defined consistent with them being younger slides or less recent sediment drape. Recent submarine landslide activity is obvious on the continental slope in water depths from 400 m to 3000 m, and tsunamogenic slides identified in Region 4 occur in 400 m to 3000+ m water depths, however the majority of slides occur between 500-2500 m water depth. Region 4 is also a more active section of the EA margin, again with the number and size of observable past and potential submarine landslides increasing in this region when compare to Region’s 1 and 2.

Sixteen tsunamogenic slides meet the minimum criteria to produce a wave at the coastline ≥5 m for Region 4 at slide velocity 20ms-1, that are either thick (>100 m) to very thick (>200 m), and narrow (<5 km) to very wide (>10 km), and are located in shallow (600 m) or very deep (2600 m) water depths. Some specific examples are: R4-8, a thick (100 m), wide (13 km) slide, located in intermediate water depth (1550 m); R4-2, a thick (125 m), narrow (2.5 km) slide, located in shallow water depth (600 m); R4-9, a very thick (250 m), very wide (17 km) slide, located in intermediate water depth (1500 m); and R4-1, a very thick (250 m), narrow (4.3 km) slide, located in intermediate water depth (1500 m). The complete set of slides and their characteristics are given in Figure 4.6 and Tables 4.2,4.4-4.5.

4-23 Clarke (2014) Chapter 4

Five potential slides (R4-17 to R4-21) have been identified that meet the minimum criteria to produce a wave at the coastline ≥5 m for Region 4. These have similar general characteristics to slides that have already failed: very thick (150- 500 m) and wide to very wide (5-14 km), and are located in water depths 800-1500 m.

4.4.3 Characteristics of Onshore Tsunami Surge Generated by Identified Slides

The characteristics of the onshore surge generated by the submarine landslide masses identified in the four geomorphic regions described in the previous section are presented below. As indicated in the methods section above, the characteristics have been determined using Ward et al.’s equations for a range of possible slide velocities (vs): 20 ms-1 (conservative), 40 ms-1 (reasonable), 60 ms-1 (high), and 80 ms-1 (maximum recorded). Table 4.4/Fig. 4.8a summarises maximum flow depth at the coastline (Fd(0)) for the 36 submarine landslides and 13 potential submarine landslides identified to be large enough to have generated surges >5 m. These are described in order from Region 1 to Region 4.

Table 4.5/Fig. 4.8b summarizes maximum expected inundation distance and run-up for the identified landslides over a range of landslide velocities v( s). While these parameters are dependent on the flow depth at the coastal, inundation distance and run up values are also largely a function of coastal topography (Mori et a. 2011) and the behavior of tsunami on land in eastern Australia also shows a clear regional dependence. Local topography greatly affects the ability of a wave to inundate beyond a zone immediately adjacent to the coastline (Gerardi et al. 2008), and inundation distance and run-up values are estimates which are less reliable than coastal inundation depth values due to the variation of coastal topography which can change markedly over small distances (1 to 5 km) along the east Australia coast. Tsunami energy from surges adjacent to steep coastal topography presents as larger run-up heights rather than longer inundation distance. Geomorphic features such as rivers and inlet bays also affect the behavior of tsunami waves and the amount the surge is able to travel inland (Mori et a. 2011).

Fig. 4.9a-f shows representative examples of coastal topographic profiles for each region, along with the maximum inundation distance (Xmax) / run-up (R(Xmax)) point for the submarine landslide associated with that profile, for slide velocities 20 ms-1 (red circles) and 40 ms-1 (blue triangles). To demonstrate the degree to which local topography can affect maximum inundation distance (Xmax) /

4-24 Clarke (2014) Chapter 4

run-up (R(Xmax)) values, an instructive example of pronounced local variation of coastal topography within a small area can be seen in Fig. 9h. Profile 1 (solid black) shows the coastal topography (perpendicular to the coastline) adjacent to the Great Sandy National Park Slide 1 and 2 (R4-7 and R4-8 respectively). Profile 2 (dashed black) shows the coastal topography 19 km south of these slides. If slide R4-7 occurred 19 km to the south at coastal Profile 2 (rather than at Profile 1), inundation distance increases by 300 m (landslide velocity 20ms-1) and 2390 m (landslide velocity 40ms-1). If slide R4-8 occurred 19 km to the south at coastal Profile 2 (rather than at Profile 1), inundation distance would decrease 220 m (landslide velocity 20ms-1) due to higher elevations present at the coastline (see Fig. 9h) and increase over 1000 m (landslide velocity 40ms-1) due to the surge and inundation characteristics.

4.4.3.1 Region 1 Slabs shed from the four submarine landslide scars identified in Region 1 (R1-1 to R1-4) are calculated to generate maximum flow depths at the coast ranging from 2.9 to 13.9 m for landslide velocities of 20 ms-1 (see Figure 4.8a; Table 4.4). The four slides correspond to the four largest erosional surface slides scars identified by Glenn et al. (2008): the R1-1 Bulli Slide (~20 km3), R1-2 Shovel Slide (7.97 km3), R1-3 Birubi Slide (2.31 km3), and R1-4 Yacaaba Slide (0.24 km3). R1-3 and R1-4 require velocities of >35 ms-1 to generate flow depths >5 m at the coastline, however they were included as they were identified by Glenn et al. (2008) as two of the four largest slides for this section of the margin. The four identified slides for this region range from narrow (<3 km) to wide (10 km) and all relatively thick (70-500 m). The largest volume slide located in this region is the Bulli Slide, with a potential to induce a wave ~14 m at the coastline for a vs equal to 20ms-1. Four potential slide sites are also investigated, three that were identified by Glenn et al. (2008) as sites of potential future failure (R1-5, R1-6, and R1-8). All four potential slides have the ability to induce sizable flow depths

-1 at the coastline (5.5 to 9.7 m) for a vs equal to 20ms .

R1-1 (Shovel Slide) and R1-2 (Bulli Slide) produce surges at the coastline ≥5 m. Both the slides are large to very large: R1-1 is a moderately thick (75 m), moderately wide to wide (5 km) slide located at water depth (1000 m); R1-2 is a very thick (200-500 m), very wide (10 km) slide located in 1700 m water depth. These slides are calculated to produce flow depths at the coastline of 5.2 m and 13.9 m, run-up of 5.25 m and 13.49 m, respectively for R1-1 and R1-2, and inundation distances <100 m due to steep coastal topography (see Tables 4.4 and 4.5). While these two large slides from Region 1 can produce sizable coastal flow depths, local coastal topography prevents these waves surging inland any great distance (see Fig. 9a), but elsewhere in Region 1 where coastal topography is more subdued

4-25 Clarke (2014) Chapter 4 the inundation distances are up to 1 to 2 km for coastal flats and lagoons.

Four potential slides are investigated in Region 1 (R1-5 - R1-8), identified by Glenn et al. (2008) as areas of concern due to features such as tension cracking, erosion and undercutting at the toe of the slope, and/or extensive mass wasting in the surrounding the area. These potential slides have the ability to produce coastal flow depths between 5.5 to 9.7 m at slide velocity 20 ms-1, run-up between 0.99 m and 8.94 m, and inundation distances between 82 m and 1183 m (see Tables 4.4 and 4.5).

4.4.3.2 Region 2 Slabs shed from the three submarine landslide scars identified in Region 2 (R2-1 to R2-3) are calculated to generate maximum flow depths at the coast ranging from 3.7 to 5.6 m for landslide velocities of 20 ms-1 (see Figure 4.8; Table 4.4). These features were originally identified by Glenn et al. (2008). R2-1 and R2-3 require velocities of >40 ms-1 to generate flow depths >5 m at the coastline, however they were included for analysis as they were identified by Glenn et al. (2008) as two of the three largest slides for this section of the margin. The three identified slides for this region are all narrow (<2.4 km) but relatively thick (100-250 m), with the Smokey Cape Slide having the potential to induce the greatest wave at the coastline of 5.6 m due to its location in relatively shallow water 650 m, which compared to 1800 m for the other two slides. Region 2 slides can produce run-up 0.45-3.72 m, and inundation distances of 0-605 m for landslide velocity 20ms-1 (see Table 4.5). Topography adjacent to Region 2 is dominated in general by high (>20 m) coastal sea cliffs, preventing surges moving inland any great distance (see Fig. 9b). However, relatively flat local coastal topography adjacent to R2-2 enables the generated surge to move inland ~600 m.

Two potential slide sites are also investigated, both identified by Glenn et al. (2008) as sites of potential future failure (R2-4 and R2-5), with the ability to induce sizable coastal flow depths of 6.0 to 7.5 m at slide velocity 20 ms-1, run-up of 5.27 m and 7.15 m, and inundation distances of 85 m and 50 m (see Tables 4.4 and 4.5). These potential slides are investigated as areas of concern due to features such as tension cracking, erosion and undercutting at the toe of the slope, and/or extensive mass wasting in the surrounding the area (Glenn et al. 2008).

4.4.3.3 Region 3 Slabs shed from the thirteen submarine landslide scars identified in Region 3 (R3-1 to R3-13) are calculated to generate maximum flow depths at the coast ranging from 3.4 to 20.6 m for landslide

4-26 Clarke (2014) Chapter 4 velocities of 20 ms-1 (see Figure 4.8; Table 4.4). These slides were identified from multibeam bathymetry data (see section 4.3.1; Fig. 4.5) expect for R3-11 (Bryon Slide) and R3-13 (Cudgen Slide), which were originally identified by Boyd et al. (2010). Half of these 13 slides generate an onshore surge ≥5 m

-1 at the coastline for a vs 20 ms . The exceptions are R3-1, R3-2, R3-10, and R3-12. R3-1, R3-2, R3-10, and R3-12, which were investigated due to their large thickness values (>150 m); however they are too narrow (<3.4 km) and located too deep (>1600 m) for the large thickness to produce coastal surges >5 m at slide velocity of 20ms-1. Region 3 slides have the potential to generate maximum inundation distances around 1-1.5 km and run-up >11.5 m for landslide velocity 20ms-1. Topography adjacent to Region 3 is characterized in general by subdued and flat coastal topography, backed by coastal lagoon systems (see Fig. 4.9c), which enables surges to move further inland than Region 1 and 2 surges. There are a number of large slides in Region 3 (see Fig. 4.8) and tsunamogenic slides able to produce a coastline flow depth ≥5 m at slide velocity 20ms-1 have a range of slide characteristics: thin (<50 m) to very thick (>200 m), and narrow (<5 km) to very wide (>10 km), and are located in shallow (600 m) or very deep (3400 m) water depths. The largest slide in the region is the Clarence Canyon Slide (R3-7), able to produce a coastline flow depth ~20 m at slide velocity 20ms-1.

Two potential slides are investigated in Region 3: R3-14 from canyon-incised morphology and R3-15 from plateau morphology. Potential slides were identified from the multibeam bathymetry as areas of concern due to features such as tension cracking, erosion and undercutting at the toe of the slope, and/or extensive mass wasting in the surrounding the area. The Byron Potential Slide (R3-14; see Fig. 4.5) is located immediately adjacent to the Byron Slide and protrudes anomalously out from the shelf in the heavily incised southern canyon section. The extensive mass wasting surrounding this block and apparent tension crack present at the head of the identified mass (Fig. 4.5biii) suggests failure in the future. A break in slope which is the crest of a small mid-slope circular failure located at the 1500 m contour defines the expected toe. R3-14 would generate a coastal surge 10.8 m high, run-up of 9.1 m, and inundation distances of 250 m at a slide velocity 20 ms-1 (see Tables 4.4 and 4.5).

The Nerang Plateau Potential Slide (R3-15; see Fig. 4.5) is located within a relatively smooth section of the plateau (see Fig. 4.1a, 4.5a) and is surrounded by numerous upper slope failures that present a distinctive “hummocky” texture within the failure region and a gently concave landslide shape. This feature presents apparent tension crack features at the head of the identified mass and is roughly equal in size to the already failed Cudgen Slide (R3-13). R3-15 would generate a coastal surge 5.0 m high, run-up of <0.5 m, and inundation distances of 572 m for a slide velocity 20 ms-1 (see Tables 4.4

4-27 Clarke (2014) Chapter 4 and 4.5).

4.4.3.4 Region 4 Slabs shed from the sixteen submarine landslide scars identified in Region 4 (R4-1 to R4-16) are calculated to generate maximum flow depths at the coast ranging from 5.2 to 21.9 m for landslide velocities of 20 ms-1 (see Figure 4.8; Table 4.4) making this area the most hazardous. One slide was first identified by Boyd et al. (2010): the Bribie Bowl Slide, and three slides were identified by Hubble et al. (2013). They are the Mid Fraser Island Mid-Slope Slide, the Wide Bay Canyon Slide, and the North Fraser Island Upper-Slope Slide. There are a number of large slides in this region (see Table 4.3, 4.4), the North Fraser Island Slide 2 generates a coastal surge ~22 m at the coastline. Five potential slide sites are also investigated (R4-17 to R4-21), and they are able to induce sizable flow depths at the coastline (5.9 - 38.0 m). Boyd et al. (2010) originally identified the Block Potential Slide (R4-17) and Hubble et al. (2013) originally identified the Wide Bay Canyon Potential Slide (R4-18).

Region 4 slides generate inundation distances of ~0.75 km, and run-up up to ~12 m for a landslide velocity 20ms-1. The onshore areas of land in Region 4 are mostly protected by the steeply rising topography of Morten and Fraser Island (see Fig. 4.9e).

Five potential slides are investigated in Region 4 (R4-17 and R4-21), with volumes between 3.9 km3 (R4-20) and 105 km3 (R4-17). These can generate a surge depth between 5.9 to 38.0 m at a slide velocity 20 ms-1, run-up of 5.4 m to 34.5 m, and inundation distances of <100 m to 715 m (see Tables 4.4 and 4.5). The Block Slide (R4-17) was identified by Boyd et al. (2010), the Wide Bay Canyon Potential Slide (R4-18) was identified by Hubble et al. (2013), and R4-19, R4-20, and R4-21 were identified from the multibeam bathymetry. All present as areas of concern due to features such as tension cracking, erosion and undercutting at the toe of the slope, and/or extensive mass wasting in the surrounding the area.

4-28 Clarke (2014) Chapter 4

Figure 4.8: a) Map showing the approximate location of the seismic line E-F and olistostromic block/slump mass reported by Hill (1992), offshore Double Island Point (just south of Fraser Island). The location of investigated submarine landslides scars (blue) and potential future tsunamogenic submarine landslides (pink) are outlined on the high-resolution multibeam bathymetry. Numbers (R4- 7-R4-11, R4-18, and R4-21) correspond to slide names shown in Table 3. Also shown is the location of the coastal topography profile R4-21. b) Modified seismic line E-F reported by Hill (1992). The outline of the identified olistostromic block/slump mass is shown in red. c) Representative coastal topography profile R4-21 adjacent to the olisto-strome block. Given a 500 m thick submarine landslide, the Xmax / R(Xmax) points for a slide width of 5 km (solid markers) and 10 km (hollow markers) are shown for slide velocities 20 ms-1 (red circles) and 40 ms-1 (blue triangles).

4-29 Clarke (2014) Chapter 4

Figure 4.9: a) The maximum expected flow depth at the coastline (Fd(0)) for all submarine landslides investigated within the study area (see Table 3 for full submarine landslide list). Fd(0) is calculated -1 -1 over a range of maximum slide velocities (vs): i) 10 ms (purple diamond); ii) 20 ms (red circle); iii) 30 ms-1 (green square); iv) 40 ms-1 (blue triangle). Solid markers represent submarine landslide scars and hollow markers represent potential submarine landslide sites. A full list of the calculated values is shown in Table 4.4. b) The maximum inundation distance (Xmax) / run-up (R(Xmax)) points for all slides for slide velocities 20 ms-1 (red) and 40 ms-1 (blue). Points are broken down by region (see legend) and a full list of the calculated values for all slides is shown in Table 4.5. Solid markers represent submarine landslide scars and hollow markers represent potential submarine landslide sites.

4-30 Clarke (2014) Chapter 4

4-31 Clarke (2014) Chapter 4

Figure 4.10: a-f) Representative examples of coastal topography profiles for each region. The maximum inundation distance (Xmax) / run-up (R(Xmax)) point for each slide is shown on the coastal -1 -1 topographic profile for slide velocities 20 ms (red circles) and 40 ms (blue triangles). Xmax values do not exceed 1.5 km for profiles shown in b-f. Xmax values exceed 1.5 km for profiles shown in g. For profile locations see Fig. 2-5. a) Region 1 – profiles R1-2, R1-3, R1-4, and R1-7; b) Region 2 – profiles R2-1, R2-3, R2-4, and R2-5; c) Region 3 – profiles R3-2, R3-3, R3-4, and R3-6; d) Region 4 – profiles R4-4, R4-6, R4-7, and R4-18; e) Region 4 – profiles R4-1, R4-11, R4-16, and R4-19. The profiles are representative of topography that include Morten Island or Fraser Island (see Fig. 5); f) All regions - representative examples of coastal topography profiles where maximum Xmax values exceed 1.5 km (R1-8, R2-2, R3-13, R3-15, and R4-13). Only 10 slides produce Xmax values > 1.5 km and only when the slide velocity is ≥40 ms-1. g) Example of the (sometimes) high local variation of coastal topography within a small area experienced in the east coast of Australia and how this can change Xmax and R(Xmax) values. Profile R4-7-8 (black solid) is the coastal topography perpendicular to the coastline and adjacent to the Great Sandy National Park Slides 1 and 2 (slide ID R4-7 and R4-8 respectively). Profile R4-7-8-adjacent (black dashed) is the coastal topography perpendicular to the coastline 19 km south of the Great Sandy National Park Slide 1 and 2. Xmax / R(Xmax) points for R4-7 and R4-8 are shown using both Profile R4-7-8 (solid markers) and Profile R4-7-8-adjacent (hollow markers) for slide velocities 20 ms-1 (red circles) and 40 ms-1 (blue triangles).

4-32 Clarke (2014) Chapter 4

Table 4.4: Maximum expected flow depth at the coastline (Fd(0)) for submarine landslides investigat- -1 ed within the study area. Fd(0) is calculated over a range of maximum slide velocities (vs): a) 10 ms ; b) 20 ms-1; c) 30 ms-1; d) 40 ms-1

Maximum expected flow depth at coastline Fd(0) (m) Slide ID 10* ms-1 20* ms-1 30* ms-1 40* ms-1 R1-1 3.1 5.2 8.2 12.2 R1-2 9.2 13.9 20.2 28.3 R1-3 1.9 2.9 4.4 6.4 R1-4 1.8 3.1 4.9 7.4 R1-5 6.0 9.7 14.8 21.5 R1-6 3.8 5.5 7.7 10.5 R1-7 3.7 6.5 10.6 15.9 R1-8 6.6 9.5 13.3 18.1 R2-1 3.3 4.9 7.1 9.9 R2-2 3.0 5.6 9.4 14.3 R2-3 2.5 3.7 5.4 7.5 R2-4 3.7 6.0 9.3 13.6 R2-5 4.3 7.5 12.2 18.4 R3-1 2.5 3.4 4.5 5.8 R3-2 2.6 3.5 4.7 6.1 R3-3 4.5 6.2 8.2 10.7 R3-4 4.3 5.8 7.7 10.0 R3-51 6.6 9.7 13.7 18.9 R3-6 3.1 5.6 9.2 14.0 R3-7.12 5.4 8.0 11.3 15.6 R3-7.23 14.1 20.6 29.3 40.4 R3-8 4.1 6.5 9.6 13.8 R3-9 7.6 11.8 17.4 24.7 R3-10 2.8 4.3 6.3 8.9 R3-11 6.2 10.3 16.3 24.0 R3-12 2.6 3.9 5.5 7.7 R3-13 3.0 5.6 9.5 14.5 R3-14 6.1 10.8 17.5 26.3 R3-15 2.9 5.0 8.2 12.3 R4-1 4.4 6.3 8.7 11.8 R4-2 2.7 5.0 8.5 13.0 R4-3 3.8 5.5 7.9 11.0 R4-4 5.8 8.8 12.7 17.8 R4-5 5.3 8.1 11.9 16.8 R4-6 4.7 6.6 9.0 12.1 R4-7 6.3 10.8 17.3 25.7 R4-8 4.7 7.2 10.6 15.0 R4-9 12.6 19.4 28.8 40.9 R4-10 5.1 8.0 12.0 17.2 R4-11 3.6 5.6 8.3 11.9 R4-12 4.0 6.4 9.8 14.3 R4-13 11.5 21.9 37.3 57.1 R4-14 4.5 7.3 11.3 16.5 R4-15 4.5 7.9 13.0 19.7 R4-16 3.1 5.2 8.2 12.2 R4-17 24.7 38.0 56.2 79.9 R4-18 5.4 8.9 13.8 20.2 R4-19 11.8 20.6 33.4 50.3 R4-20 3.8 6.5 10.5 15.6 R4-21 3.7 5.9 8.9 12.8

*Maximum landslide velocity (vs) 1Minimum values, full slide scar not imaged (see Fig. 5) 2Minimum values 3Maximum values

4-33 Clarke (2014) Chapter 4

Table 4.5: Full list of the inundation distance (Xmax) and run-up (R(Xmax)) values determined for each identified submarine landslide within the study area over a range of maximum slide velocities (vs): a) 10 ms-1; b) 20 ms-1; c) 30 ms-1; d) 40 ms-1. Manning’s coefficient is 0.03.

Inundation distance (Xmax) (m) and Run-up (R(Xmax)) (m)

Slide ID Xmax R(Xmax) Xmax R(Xmax) Xmax R(Xmax) Xmax R(Xmax)

10* ms-1 20* ms-1 30* ms-1 40* ms-1 R1-1 0 3.14 0 5.25 13.9 8.14 40.5 11.92 R1-2 40 8.92 61 13.49 104 19.58 153 27.5 R1-3 0 1.86 0 2.94 12 4.33 58 5.93 R1-4 91 0.71 162 1.44 496 0.55 819 0.99 R1-5 51 5.59 103 8.94 147 13.87 218 20.28 R1-6 38 3.43 82 4.77 418 4.46 506 6.95 R1-7 61 3.12 234 4.61 513 7.02 565 12.48 R1-8 690 0.99 1183 0.99 1918 0.99 2960 0.99 R2-1 9 3.21 263 2.59 637 2.09 919 3.42 R2-2 79 2.22 605 0.45 1031 1.96 1561 4.51 R2-3 0 2.5 0 3.72 0 5.37 0 7.48 R2-4 0 3.65 85 5.27 225 7.64 621 9.62 R2-5 18 4.13 50 7.15 329 10.08 995 12.7 R3-1 25.0 2.2 80.0 2.6 151 3.1 223 3.97 R3-2 0.0 2.6 7.0 3.5 53 4.23 121 5.14 R3-3 49.0 4.1 87.0 5.5 117 7.34 142 9.76 R3-4 71.0 3.6 92.0 5.0 126 6.71 166 8.83 R3-51 160.0 5.3 247.0 7.9 386 11.3 1186 12.17 R3-6 0.0 3.1 0.0 5.6 37 8.97 67 13.6 R3-7.12 92.0 4.7 930.0 0.9 1424 1.7 1535 6.27 R3-7.23 1497.0 4.7 1589.0 11.8 1683 21.01 4270 21.44 R3-8 0.0 4.1 106.0 5.6 1196 1.13 1642 3.41 R3-9 243.0 5.8 419.0 9.0 2658 1.82 4020 3.73 R3-10 50.0 2.2 63.0 3.7 83 5.62 105 8.18 R3-11 80.0 5.5 118.0 9.5 1447 7.58 1898 14.02 R3-12 23.0 2.4 34.0 3.5 49 5.11 68 7.14 R3-13 53.0 2.5 81.0 4.9 120 8.66 1606 4.57 R3-14 86.0 5.4 250.0 9.1 381 15.25 652 22.97 R3-15 271.0 0.0 572.0 0.0 1090 0 1877 0 R4-1 48.0 4.0 81.0 5.7 97 8.04 123 11 R4-2 46.0 2.2 69.0 4.4 103 7.72 1219 5.1 R4-3 49.0 3.3 71.0 4.9 100 7.18 412 8.19 R4-4 48.0 5.4 68.0 8.3 94 12.12 155 16.95 R4-5 50.0 4.9 77.0 7.5 169 10.78 264 15.23 R4-6 60.0 4.1 85.0 5.9 872 2.67 1331 3.29 R4-7 293.0 3.9 326.0 8.6 370 15.08 389 23.72 R4-8 159.0 3.3 300.0 4.8 325 8.34 356 12.78 R4-9 646.0 8.4 793.0 15.0 1086 23.37 1238 35.38 R4-10 47.0 4.7 74.0 7.5 115 11.25 170 16.18 R4-11 44.0 3.1 68.0 5.0 93 7.62 127 11.05 R4-12 26.0 3.8 42.0 6.1 65 9.37 94 13.7 R4-13 142.0 10.5 347.0 20.0 847 33.46 5172 36.69 R4-14 58.0 3.9 101.0 6.5 168 10.15 369 14.34 R4-15 38.0 4.1 69.0 7.4 121 12.23 972 14.2 R4-16 57.0 2.5 113.0 4.2 209 6.66 1149 4.65 R4-17 447.0 22.4 715.0 34.8 1111 51.78 1388 74.99 R4-18 33.0 5.2 55.0 8.5 87 13.26 236 18.92 R4-19 113.0 11.0 164.0 19.7 262 32.21 1491 44.18 R4-20 37.0 3.5 65.0 6.0 547 6.67 661 11.6 R4-21 34.0 3.4 55.0 5.4 86 8.27 217 11.42

*Maximum landslide velocity (vs) 1Minimum values, full slide scar not imaged (see Fig. 5) 2Minimum values 3Maximum values

4-34 Clarke (2014) Chapter 4 4.5 Discussion The material present above demonstrates that blocks shed from submarine scars have the potential to generate significant tsunami on the EA coast with flow depths of at least 5 m in height, maximum run-up heights of ~20 m, and maximum inundation distances of ~1.6 km. These estimates are based on relatively conservative estimates of landslide velocity (20 ms-1) given a) the relatively steep slopes down which sliding material has moved and b) the lack of evidence for a depositional site where the landslide mass ceased moving in mapped areas downslope of the landslide, indicating that the landslide material has travelled at least 15-20 km from its site of origin.

Flow depth at the coastline directly relates to landslide thickness at a particular location. The thin (<100 m) and narrow (<5 km) landslides tend to produce smaller tsunami, with coastal flow depths of <5 m. The thick (>100 m) or wide (>5 km) landslides tended to generate coastal flow depths of 5-10 m. The combination of both thick and wide landslides had the greatest potential to generate the largest coastal flow depths >10 m. Water depth at the slide site determines the flow depths and the highest surges are generated by slides located in water depths between ~500 -1500 m. Maximum flow depth at the coastline is larger for the thicker canyon landslides (e.g. Byron Slide R3-11, flow depth ~10.3 m; see Table 4.4/Fig. 4.8) which occur on steeper slopes in comparison to shallow plateau landslides which generally produce waves less than 1 m in height, except where landslide surface area was particularly large (e.g. Cudgen Slide: surface area ~50 km2, flow depth ~5 m).

These findings enable examination, and in some cases resolution (or partial resolution), of several issues and questions of interest related to historical tsunami occurrence on the eastern Australian margin and the debate about the Bryant et al. megatsunami hypothesis.

4.5.1 Comparison to related prior work Very little tsunami modelling for the eastern Australian coastline has been published, particularly work about submarine landslide induced tsunami. Geoscience Australia has produced an ‘offshore tsunami hazard map’ for Australia (http://www.ga.gov.au/hazards/tsunami/offshore-tsunami-hazard- for-australia.html) that estimates the maximum tsunami amplitude at the 100 m bathymetric contour, as simulated tsunami heights at this depth or greater are considerably simpler to determine and less sensitive to shallow bathymetry (Burbridge et al. 2008). Estimates are based primarily on earthquake sources (Burbidge et al. 2008). Talukder and Volker (2014) calculated initial wave heights above the

4-35 Clarke (2014) Chapter 4 mass centroid for submarine landslide induced tsunamis for the Shovel (R1-1) and Bulli Slides (R1-2) offshore the central NSW coastline. They report similar heights (10–25 m) to this study for the initial wave height and conclude that submarine landslides may well provide sources for local tsunamis (Ta- lukder and Volker 2014). This work did not attempt to calculate coastal inundation characteristics.

Several tsunami hazard assessments have been conducted for Coastal Councils Sydney in order to assess the hazard and risk of tsunami induced coastal inundation (see Dall’Osso and Dominy- Howes 2010 and references therein). While useful, this work focuses the consequences of a tsunami event, rather than the source.

4.5.2 Tsunamogenic submarine landslide frequency estimates TThis study has identified 36 distinct submarine landslide scars capable of generating a tsunami flow depth >5 m at the coastline. A crude maximum recurrence interval for these events can be estimated based on the approximate age of the submarine landsides. Sediment no older than the Oligocene, and mostly mid-Miocene to early Pliocene and some much younger than early Pliocene, has been sampled from headscarps of the lower slope (>2000 m) along the EA margin (see Hubble et al. 2012, 2013). All but eight of the tsunamogenic slides discussed in this work occur in sediment deposited above the lower slope material. Therefore the 36 tsunamogenic slides have been constrained as occurring over the last ~5 Ma. The recurrence interval of east Australia submarine landslides able to produce a tsunami flow depth >5 m at the coastline is therefore calculated to be one event every 150 ka. Given the prevalence of slides on the EA margin, if we were to consider smaller submarine landslides producing smaller tsunamis with flow depths ~2 m, it is suspected that the prevalence of submarine landside induced tsunami would be 10 to 100 times more common. If this is the case, then a recurrence interval for the whole margin this would be about 15 ka to 1.5 ka, given that Clarke et al. (2012) has identified 3 slides younger than 25 ka and there are a number of other slides of similar morphometric appearance (i.e. of probable similar age). While an estimate of recurrence interval between 15 ka and 1.5 ka seems reasonable, it is obvious that further work to constrain this very crude estimate is required.

4.5.3 Tsunami record of eastern Australia and implications for the megatsunami hypothesis Evidence for historical tsunamis that have inundated the EA coastline is limited. A recently published tsunami catalogue (Dominey-Howes 2007) identifies historical events on the EA coast, of which

4-36 Clarke (2014) Chapter 4

16 (~30%) have no identified cause (run-ups between <0.1 m to 6 m asl). Nevertheless, our results show that upper slope landslides similar to those investigated in this study are a plausible source for the tsunamis documented in Dominey-Howes (2007) catalogue. Further investigations of the margins submarine landslides is required to identify a) the depositional site for the slide material to constrain maximum speed calculations and b) determine their age and frequency of occurrence. A growing body of evidence (Hubble 2013) is strengthening the contention that shedding submarine landslides of the size investigated in this study should be considered to be a common, ongoing characteristic of the EA continental margin such that future failures are very likely to occur. Constraining estimates of landslide velocity is critical as coastal flow depth, inundation distance and run-up all increase with landslide velocity. We have used conservative estimates of landslide velocity and it is quite possible that these landslides can generate significantly more destructive events than we have suggested.

In addition to these historical events a number of paleotsunami events have been reported for the EA coast and paleotsunami catalogues are available for both Australia and New Zealand (see Dominey-Howes 2007; Goff and Dominey-Howes 2009; Switzer et al. 2011). Within the Austral- ian catalogue, mega-paleotsunami events have also been proposed for the Holocene (e.g. Bryant et al. 1992; Bryant et al. 1996; Young et al. 1996; Bryant et al. 1997: Nott 1997; Bryant 2001; Bryant and Nott 2001) for this part of the Australian coast on the basis of large stranded boulders and apparently Holocene marine deposits anomalously located well above the apparent limit of modern-day storm surge events. According to Bryant et al. and colleagues (1996, 1997, 2001a,b) palaeo-megatsunami events in Australia have been “truly catastrophic”. This body of work propose that megatsunami have caused a number of extreme events along the NSW coast (study sites located primarily along the NSW coastline, see Courtney et al. 2012), producing tsunami waves with the ability to, amongst other things, 1) overtop coastal cliffs (some in excess of 80 m above sea-level) in areas such as Jervis Bay, Wollongong, and Sydney (Bryant et al. 2001 and references therein); 2) generate flow depths at the coastline greater than 10-20 m (Bryant et al. 1997; Bryant et al. 2001); 3) transport massive boulders (over 6 m in diameter and weighing >280 tonnes) and place them on top of coastal cliffs ranging from 40 m to 130+ m elevation (Bryant et al. 2001); and 4) travel around 10 km inland in some cases (Bryant et al. 2001). In addition to reported mega-paleotsunami deposits, a number of other paleot- sunmai deposits have been identified on the NSW coast. Goff and Dominey-Howes (2009) identified 5 paleotsunami deposits that could correlate in both Australia and New Zealand, and Switzer et al. (2011) report a possible paleotsunami deposit (2900 cal bp) at Bateman’s Bay on the NSW coast that is possibly coeval with a tsunami event which is also recorded in other paleotsunami deposits at several

4-37 Clarke (2014) Chapter 4 other sites along the NSW coast.

The palaeo-megatsunami hypothesis has progressively been placed under increasing scrutiny (see Dominey-Howes 2007; Courtney et al. 2012). Bryant et al. identified over 60 paleo-tsunami sites along the NSW southern and central coast and 60 of these sites have been reinspected and re-examined by Courtney et al. (2012). In the context of this work, 44 of the Courtney et al.’s sites all occur within Region 1, with the other 16 sites being in other areas. In order for the central and southern NSW coast to generate a megatsunami similar to those hypothesised (i.e. 50-100 m), they must generate submarine landslide masses that move as a single intact block and at speeds above 60ms-1; these velocities are equivalent to the highest velocities proposed for the extremely large landslide masses such as the Storrega Slide event, which is a proposition that is difficult to accept (e.g. >60 -1ms for the Bulli Slide, is the largest submarine landslide identified in Regions 1 and 2, which at 20 km3 is 100 orders of magnitude smaller than Storrega). It is far more probable that the southern and central NSW submarine landslides identified in this work are too small to be able to produce megatsunami waves. When these modelling finding are considered in the context of the pre-Holocene ages established for the slides by 14C dating (see Chapter 2; Glenn et al. 2008) the southern and central NSW slides could not have generated a megatsunami during the Holocene.

The Bulli Slide (R1-2) is the only slide from Region 1 capable of producing a maximum probable flow depth >20 m at the coastline but it can only do this if the slide velocity was ≥30 ms-1. Ad- ditionally, Glenn et al. (2008) suggest that the Bulli Slide transformed into a turbidity flow soon after failure, reducing the likelihood that it resulted in a megatsunami wave. The Illawarra Potential Slide (R1-5) could also produce a maximum flow depth >20 m at the coastline but again only if the slide block achieved a velocity ≥40 ms-1. The only submarine landslides capable of potentially generating a megatsunami are located in Regions 3 and 4, further to the north. Of these only 3 actual slides (R3- 7.2, R4-9, R4-13) and 2 potential slides (R4-17, R4-19; see Table 4.4) were capable of generating a maximum probable flow depth >20 m at the coastline with a slide velocity 20 ms-1.

Bryant et al. have quoted submarine landslides as a possible cause of the megatsuami events on more than one occasion (see Bryant et al. 1996, 2001). Assuming the validity of the megatsunami hypothesis, we ask: 1) What are the minimum submarine landslide dimensions required to produce a maximum expected flow depth at the coastline of 20 m, 50 m, and 100 m, and how do simulated minimum initial slide values required to generate mega-tsunami compare to dimensions of existing

4-38 Clarke (2014) Chapter 4 slide scars from the region; 2) What velocity must slide blocks coming from already identified submarine landslide scars need to travel at in order to generate a mega-tsunami wave surge (aka flow depths at the coastline) with the ability to overtop coastal cliffs 20 m, 50 m, and 100 m high, and potentially move massive boulders to top of these cliffs; and 3) Assuming a conservative slide velocity of 20 ms-1, how would much must the slide dimensions of the Bulli Slide need to increase in order to generate a mega-tsunami wave surge with the ability to overtop coastal cliffs 20 m, 50 m, and 100 m high?

The minimum block dimensions required for a submarine landslide to generate a mega-tsunami along the NSW coast with the ability to overtop coastal cliffs greater than 20 m, 50 m, and 100 m are given in Table 4.6.

Table 4.6: Summary of the minimum slide width values required for a range of water depths (500 – 2500 m) and slide thicknesses (50 m – 500 m) required for a submarine landslide to generate a maximum expected flow depth at the coastline -1 -1 (Fd(0)) of 20 m, 50 m, and 100 m at slide velocities of 20 ms and 40 ms . Fd(0) values of 20 m, 50 m, and 100 m are approximately equivalent to a megatsunami that could reach the EA coast with the ability to overtop coastal cliffs greater than 20 m, 50 m, and 100 m. These values provide an indication as to the minimum submarine landslide block dimensions required to generate such a wave.

Water depth Maximum expected flow Slide Width to at slide h (m) Slide width depth at coastline F (0) (m) thickness t 0 d thickness (centre of W (km) (m) ratio mass) 20* ms-1 40* ms-1 500 500 1 9.1 24.1 2 5 32.9 87.3 10 10 57.2 151.9 20 20 99.6 264.5 40 1000 1 5.8 13.6 2 5 21.2 49.1 10 10 36.9 85.6 20 20 64.2 149.0 40 1500 1 4.7 9.9 2 5 17.0 35.8 10 10 29.7 62.4 20 20 51.7 108.6 40 2000 1 4.1 8.0 2 5 14.9 29.1 10 10 25.9 50.6 20 20 45.0 88.2 40 2500 1 3.7 6.9 2 5 13.5 25.0 10 10 23.5 43.5 20 20 40.9 75.8 40

*Maximum landslide velocity (v s )

4-39 Clarke (2014) Chapter 4

In summary, slides have to be at least 5 km wide: 1) at least 250 m thick in order to produce 20 m flow depth at the coastline; 2) very thick (>500 m) and occur in water depths between 500 - 1500 m to produce 50 m flow depth at the coastline; or 3) very thick (>500 m), occur in water depths between 500 - 1500 m and move quickly (>40 ms-1) to produce 100 m flow depth at the coastline.

If these required slide geometries to generate a megatsunami in Table 4.6 are compared to the existing submarine landslide scar geometries (see Table 4.3), it is apparent that there are no slides in central or southern NSW, with the required dimensions that have been identified. The majority of the slides actually present cannot generate flow depths at the coastline > 15 m (slide velocity 20 ms-1) and are therefore unlikely candidates to generate megatsuanmis. Three slides (R3-7.2, R4-9, and R4-13) do have the ability to generate flows depths at the coastline of around 20 m (slide velocity 20 ms-1), however all of these slides are in the northern Regions 3 or 4, not in Region 1 where Bryant et al. requires them to be located. No slide scar or potential slide identified in this work is able to generate a flow depth at the coastline of >40 m with a slide velocity 20 ms-1. One slide (R4-17) was able to generate a flow depth at the coastline of >50 m with a slide velocity 30 ms-1, and another slide (R4-13) was able to generate a flow depth at the coastline of >50 m with a slide velocity 40 ms-1. There are no identified slide scars anywhere on the EA margin that are able to generate a flow depth at the coastline of >100 m.

In order for the existing submarine landslides scar geometries to produce a megatsunami wave heights (i.e. 20 m, 50 m, 100 m), the slide velocity must be much greater. For example for the largest of the Region 1 slide scar, the Bulli Slide (R1-2), to result in a megatsunami able to overtop coastal cliffs greater than 20 m, 50 m, and 100 m the slide block must be travelling at a velocity of 30, 60, and 100 ms-1 respectively. A velocity of 30 ms-1 is not unreasonable, but values of 60 ms-1 are getting close to the maximum velocity values in the literature, and 100 ms-1 (300 kmhr-1) is more than the largest proposed velocity value of 80 ms-1 (see Masson et al. 2006). Aside from the Bulli Slide, no other slide in Region 1 (Shovel, Birubi, or Yacaaba Slides) is capable of generating waves equal to or greater than 50 m at the coastline. They can produce a 20 m coastal wave if the slide velocity is increased suf- ficiently: the Shovel Slide (R1-1) block must travel at 56 ms-1, while the Birubi (R1-3) and Yacaaba (R1-4) Slides must travel at 92 ms-1.

Bryant et al. hypothesize mega-paleotsunamis occurring mostly within Region 1 of the southern

4-40 Clarke (2014) Chapter 4 and central NSW coastline, but there is little evidence of tsunamogenic submarine landslides in the offshore, at least with respects to an extant, identifiable submarine landslide scar of an appropriate age to provide a landslide source for Bryant and co-workers hypothesised event. There may be some other source for a megatsunami; the velocity estimates for slide motion may be underestimated or seriously inadequate; or some other process that generates the megatsunami deposits such as the boulder debris etc. found on the cliffs. But if the modelling presented by this study is accepted to be reasonable then submarine landsliding from the EA margin should be rejected as a cause of megatsunami.

4.6 Conclusions

In this work, we have analysed the potential tsunami hazard associated with submarine landsliding events on the east Australian continental margin between Jervis Bay and Fraser Island.

Specific conclusions are as follows:

1. The first inventory and set of maps identifying the submarine landslide locations along ~1500 km the east Australian continental slope is presented. Evidence of submarine landsliding exists along the entire length of continental margin, however the distribution of submarine landslides increases noticeably northward of Coffs Harbour. Subma- rine landslide scars are evident in water depths of ~400 to 3500 m with no apparent clustering at any particular water depth. Thicker submarine landslides (>100 m) generally have the potential to generate significant tsunami waves.

2. Over 260 individual submarine landslide scars are identified, and 36 distinct submarine landslides able to generate a tsunami flow depth >5 m at the coastline

3. Thirteen potential submarine landslides able to generate a tsunami flow depth >5 m at the coastline

4. Flow depth (2.9 – 21.0 m), run-up (0.1 – 20.0 m) and inundation distance (0.1 – 1600 m) were calculated assuming landslide velocities of 20 ms-1 for 49 submarine landslides.

5. The reoccurrence of submarine landslides with similar characteristics to those shed from the margin in the geologically recent past would therefore be expected to generate tsunami with maximum flow depths between five and twenty meters at the coastline, run-up of up to 20 m and inundation distances of up to 1.5 km. 13 potential subma-

4-41 Clarke (2014) Chapter 4

rine landslides have been identified as examples of future failures within the study area. For example, a potential landslide mass adjacent to the Byron Slide has been identified. If it was to fail it could generate maximum flow depths of ca.10 m at the coastline, with inundation distances of ca. 1 km, for a conservative landslide velocity of 20 ms-1. If these assumptions are correct, a tsunami this size would cause significant damage and possibly loss of life.

6. The hazard of tsunamogenic submarine landslides increases northward of Coffs Har- bour, with the number and size of both past tsunamogenic submarine landslides and potential becoming much greater offshore northern NSW and southern QLD.

7. Sensitivity analysis established a set of minimum size criteria to identify of potential tsunamogenic submarine landslides from bathymetric data based on slide width, thickness, and water depth. Potential tsunamogenic submarine landslides should be obvious in high-resolution multibeam data (> 1 km in width) and have a width to thickness ratio greater than 6.

8. The maximum recurrence interval of east Australia submarine landslides able to produce a tsunami flow depth >5 m at the coastline is calculated at ~one event every 150 ka, assuming the onset of current submarine landsliding to be 5 Ma

9. Evidence for submarine landslides able to induce megatsunami waves is limited for the central NSW continental slope. Tsunamogenic submarine landslide scars reported for the central NSW are not big enough to have shed a block able to generate a megatsunami similar to paleo-megatsunamis hypothesised for this section of margin. For the paleo-megatsunami hypothesis to hold, the submarine landslide masses must move as a single intact block and faster than reasonable estimates suggest (>40 ms-1); the candi- date slides are too old to have occurred during the Holocene.

It is difficult to establish the onshore tsunami surge characteristics for the submarine landslides reported here with certainty. Nevertheless this work has established that they are certainly volumetrically large enough and occur at shallow enough water depths (400-2500 m) to generate substantial tsunamis that could cause widespread damage on the east Australian coast and threaten coastal communities. Further investigation and modelling is required.

4-42 Clarke (2014) Chapter 4

References

Assier-Rzadkiewicz, S., Heinrich, P., Sabatier, P. C., Savoye, B. & Bourillet, J. F., 2000. Numerical modelling of a landslide- generated tsunami: the 1979 Nice event. Pure and Applied Geophysics 157, 1717–1727. Bardet, J.P., Synolakis, C.E., Davies, H.L., Imamura, F., Okal, E.A., 2003. Landslide Tsunamis: Recent Findings and Re- search Directions Pure and Applied Geophysics 160, 1793–1809. Boyd, R., Keene, J., Hubble, T., Gardner, J., Glenn, K., Ruming, K., Exon, N., 2010. Southeast Australia: A Ce-nozoic Continental Margin Dominated by Mass Transport, in: Mosher, D.C., Moscardelli, L., Baxter, C.D.P., Urgeles, R., Shipp, R.C., Chaytor, J.D., Lee, H.J. (Eds.), Submarine Mass Movements and Their Consequences. Springer Netherlands, pp. 491-502. Bryant, E., 2001. Tsunami: The Underrated Hazard. Springer London, Limited. Bryant, E.A., Nott, J., 2001. Geological Indicators of Large Tsunami in Australia. Natural Hazards and Earth System Sci- ences 24,, 231-249. Bryant, E.A., Young, R.W., Price, D.M., 1992. Evidence of tsunami sedimentation on the Southeastern coast of Australian. Journal of Geology 100, 753-765. Bryant, E.A., Young, R.W., Price, D.M., 1996. Tsunami as a Major Control on Coastal Evolution, Southeastern Australia. Journal of Coastal Research 12, 831-840. Bryant, E.A., Young, R.W., Price, D.M., Wheeler, D.J., Pease, M.I., 1997. The impact of tsunami on the coast-line of Jervis Bay, southeastern Australia. Physical Geography 18, 440-459. Burbidge, D., Cummins, P., Mleczko, R., Thio, H., 2008. A Probabilistic Tsunami Hazard Assessment for West-ern Aus- tralia. Pure and Applied Geophysics 165, 2059-2088. Chesley, S.R., Ward, S.N., 2006. A quantitative assessment of the human and economic hazard from impact-generated tsunami. Natural Hazards 38, 355-374. Clark, D.J., 2010. Large Earthquake Recurrance in New South Wales: Implications for Earthquake Hazard, Seismic Engi- neering - design for management for geohazards, Australian Geomechanics Society, Sydney. Clarke, S., Hubble, T., Airey, D., Yu, P., Boyd, R., Keene, J., Exon, N., Gardner, J., 2012. Submarine Landslides on the Upper Southeast Australian Passive Continental Margin – Preliminary Findings Submarine Mass Movements and Their Consequences, in: Yamada, Y., Kawamura, K., Ikehara, K., Ogawa, Y., Urgeles, R., Mosher, D., Chaytor, J., Strasser, M. (Eds.). Springer Netherlands, pp. 55-66. Clarke, S., Hubble, T., Airey, D., Yu, P., Boyd, R., Keene, J., Exon, N., Gardner, J., Ward, S., 2014. Morphology of Aus- tralia’s Eastern Continental Slope and Related Tsunami Hazard, Submarine Mass Movements and Their Conse- quences. Springer, pp. 529-538. Courtney, C., Dominey-Howes, D., Goff, J., Chagué-Goff, C., Switzer, A.D., McFadgen, B., 2012. A synthesis and review of the geological evidence for palaeotsunamis along the coast of southeast Australia: The evidence, issues and potential ways forward. Quaternary Science Reviews 54, 99-125. Dominey-Howes, D., 2007. Geological and historical records of tsunami in Australia. Marine Geology 239, 99-123. Dominey-Howes, D., Goff, J., 2009. Hanging on the line – on the need to assess the risk to global submarine tel-ecom- munications infrastructure – an example of the Hawaiian “bottleneck” and Australia. Nat. Hazards Earth Syst. Sci. 9, 605-607. Driscoll, N.W., Weissel, J.K., Goff, J.A., 2000. Potential for large-scale submarine slope failure and tsunami generation along the U.S. mid-Atlantic coast. Geology 28, 407-410. Fine, I.V., Rabinovich, A. B., Bornhold, B. D., Thomson, R. E. & Kulikov, E. A., 2005. The Grand Banks landslide-generated tsunami of November 18, 1929: preliminary analysis and numerical modeling. Marine Geology 215, 45-57. Fletcher M., T. Hubble, S. Clarke, D. Airey, P. Yu, 2014. Submarine Landslide Morphology of Box Slides Present on the Continental Slope Offshore Fraser Island, Queensland, Australia. Proceedings of 2014 Spring Meeting, EGU, Vienna, Austria Fryer, G.J., Watts, P., Pratson, L.F., 2004. Source of the great tsunami of 1 April 1946: a landslide in the upper Aleutian

4-43 Clarke (2014) Chapter 4

forearc. Marine Geology 203, 201-218. Gerardi, F., Barbano, M.S., De Martini, P.M., Pantosti, D., 2008. Discrimination of Tsunami Sources (Earth-quake versus Landslide) on the Basis of Historical Data in Eastern Sicily and Southern Calabria. Bulletin of the Seismological Society of America 98, 2795-2805. Glenn, K., Post, A., Keene, J., Boyd, R., Fountain, L., Potter, A., Osuchowski, M., Dando, N., Party, S., 2008. NSW Continental Slope Survey – Post Cruise Report in: Australia, G. (Ed.). Goff, J., Dominey-Howes, D., 2009. Australasian palaeotsunamis — Do Australia and New Zealand have a shared trans- Tasman prehistory? Earth-Science Reviews 97, 147-154. Haeussler, P.J., Lee, H.J., Ryan, H.F., Labay, K., Kayen, R.E., Hampton, M.A., Suleimanip, E., 2007. Submarine Slope Failures Nears Seward, Alaska, During the M9.2 1964 Earthquake, in: Lykousis, V., Sakellariou, D., Locat, J. (Ed.), Submarine Mass Movements and Their Consequences, pp. 269–278. Harbitz, C., Løvholt, F., Bungum, H., 2013. Submarine landslide tsunamis: how extreme and how likely? Natural Hazards, 1-34. Hill, P.J., 1992. Capricorn and northern Tasman Basins: structure and depositional systems. Exploration Geo-physics 23(1/2), 153-162. Hsu, S.-K., 2008. Turbidity Currents, Submarine Landslides and the 2006 Pingtung Earthquake off SW Taiwan. TAO : Terrestrial, atmospheric, and oceanic sciences 19, 767. Hubble, T., 2013. Voyage Summary SS2013-V01: Marine Geology and Geohazard Survey of the SE Australian Margin off Northern NSW and Southern Queensland, CSIRO, Hobart. Hubble, T., Yu, P., Airey, D., Clarke, S., Boyd, R., Keene, J., Exon, N., Gardner, J., 2012. Physical Properties and Age of Continental Slope Sediments Dredged from the Eastern Australian Continental Margin – Implications for Timing of Slope Failure. Submarine Mass Movements and Their Consequences, in: Yamada, Y., Kawamura, K., Ikehara, K., Ogawa, Y., Urgeles, R., Mosher, D., Chaytor, J., Strasser, M. (Eds.). Springer Netherlands, pp. 43-54. Jenkins, C.J., Keene, J.B., 1992. Submarine slope failures on the southeast Australian continental slope. Deep Sea Research 39, 121-136. Keene, J., Baker, C., Tran, M., Potter, A., 2008. Geomorphology and Sedimentology of the East Marine Region of Aus- tralia, in: Australia, G. (Ed.), Canberra. Lay, T., and others, 2005. The Great Sumatra-Andaman Earthquake of 26 December 2004. Science 308, 1127-1133. Maretzki, S., Grilli, S., Baxter, C.D.P., 2007. Probabilistic SMF tsunami hazard assessment for the upper east coast of the United States. Masson, D.G., Harbitz, C.B., Wynn, R.B., Pedersen, G., Lovholt, F., 2006. Submarine landslides: processes, triggers and hazard prediction. The Philosophical Transactions of the Royal Society A (Phil. Trans. R. Soc. A) 364, 2009-2039. McAdoo, B.G., Pratson, L.F., Orange, D.L., 2000. Submarine landslide geomorphology, US continental slope. Marine Geology 169, 103-136. Miller, D.J., 1960. The Alaska earthquake of July 10, 1958: Giant wave in Lituya Bay. Bulletin of the Seismo-logical So- ciety of America 50, 253-266. Mori, N., Takahashi, T., Yasuda, T., Yanagisawa, H., 2011. Survey of 2011 Tohoku earthquake tsunami inunda-tion and run-up. Geophysical Research Letters 38, L00G14. Nott, J., 1997. Extremely high-energy wave deposits inside the Great Barrier Reef, Australia: determining the cause - tsunami or tropical cyclone. Marine Geology 141, 193-207. Short, A.D., Woodroffe, C.D., 2009. The Coast of Australia. Cambridge University Press, New York. Slater, J.A., Garvey, G., Johnston, C., Haase, J., Heady, B., Kroenung, G., Little, J., 2006. The SRTM data “fin-ishing” process and products. Photogrammetric Engineering and Remote Sensing 72, 237-247. Sue, L.P., Nokes, R.I., Davidson, M.J., 2011. Tsunami generation by submarine landslides: comparison of physi-cal and numerical models. Environmental Fluid Mechanics 11, 133-165. Switzer, A.D., Mamo, B.L., Dominey-Howes, D., Strotz, L.C., Courtney, C., Jones, B.G., Haslett, S.K., Everett, D.M., 2011. On the Possible Origins of an Unusual (Mid to Late Holocene) Coastal Deposit, Old Punt Bay, South-East Australia. Geographical Research 49, 408-430. Talukder, A., Völker, D., 2014. The Tsunami Generation Potential of Shovel and Bulli Slides in the Continental Margin SE

4-44 Clarke (2014) Chapter 4

Australia, in: Krastel, S., Behrmann, J.-H., Völker, D., Stipp, M., Berndt, C., Urgeles, R., Chaytor, J., Huhn, K., Strasser, M., Harbitz, C.B. (Eds.), Submarine Mass Movements and Their Consequences. Springer In-ternational Publishing, pp. 539-548. Tappin, D.R., Watts, P., McMurtry, G.M., Lafoy, Y., Matsumoto, T., 2001. The Sissano, Papua New Guinea tsunami of July 1998 - offshore evidence on the source mechanism. Marine Geology 175, 1-23. Tinti, S., 2005. The 30 December 2002 landslide-induced tsunamis in Stromboli: sequence of the events recon-structed from the eyewitness accounts. Natural Hazards and Earth System Sciences 5, 763-775. Ward, S.N., 2001 Landslide tsunami. Journal of Geophysical Research 106, 11201-11215. Ward, S.N., 2011. Tsunami, in: Gupta, H.K. (Ed.), Encyclopedia of earth sciences series. Springer, Netherlands, pp. 1473- 1493. Watts, P., 2004. Probabilistic predictions of landslide tsunamis of Southern California. Marine Geology 203, 281-301. Wilson, O., Buchanan, C., Spinoccia, M., 2012. 50m Multibeam Dataset of Australia 2012, in: NASA (Ed.). Ge-oscience Australia, Canberra Young, R.W., Bryant, E.A., Price, D.M., 1996. Catastrophic wave (tsunami?) transport of boulders in southern New South Wales, Australia. Zeitschrift Fur Geomorphologie 40, 191-207.

4-45 Clarke (2014) Chapter 5 Chapter 5

Conclusions

5.1 Outline

This thesis has presented an investigation into the submarine landslides occurring along eastern Australia’s upper continental margin, with a focus on investigating the causes, timing, and mechanisms responsible for these features. It has focused on analysing gravity core samples and interpreting of high-resolution multibeam and subbottom profiles collected by Boyd et al. (2010) and Hubble et al. (2013). The age, morphology, composition, and origin of particular submarine landslides on the eastern Australian upper continental margin offshore New South Wales/Queensland has been described and, for the first time, the mechanical characteristics of sediments from the eastern Australian upper continental slope has been presented. Results have been used to investigate two important questions: when and why have slope failures occurred. The hazard posed by these submarine landslides has also been examined by investigating their potential to generate tsunamis along this margin has also been evaluated. This chapter summarizes the results from this thesis as a whole and highlights important outcomes of this research. Suggested topics for future research are presented at the end of this chapter in Section 5.5.

5.2 Specific aims and objectives

This thesis has four primary aims and objectives: (1) define the age, morphology, composition and origin of particular submarine landslides on the eastern Australian upper continental margin recently discovered offshore New South Wales/Queensland; (2) determine the geomechanical characteristics of sediments from the upper eastern Australian continental slope; (3) investigate when and why have slope failures occurred; and (4) improve the understanding of the hazard posed by these submarine landslides by investigating their potential to generate tsunamis along this margin. This approach attempts to provide a sequenced, integrated explanation for the slope failures of the EA continental margin from a number of distinct and different, but related methods. Due to the large number of failures present in the area studied, five representative failure scars were chosen to be the focus of this

5-1 Clarke (2014) Chapter 5 study. Gravity core samples were available for all five sites.

Background information presented in the first part of the study establishes context and summarised the settings and causes of submarine failure as well as previously undertaken studies on the EA continental margin. The second part of this study involved investigating the sediment characteristics and age of the submarine landslides, examining the physical properties of sediment taken in or adjacent to five distinct submarine landslides, obtaining a detailed 14C age record of the sediments in order to investigate the age of the submarine landslides in the region, and establishing a sedimentation rate for this section of the margin. The third part of the study involved investigating geotechnical characteristics of the submarine landslide sediments on the eastern Australian margin in order to establish and quantify the geomechanical properties of the sediment from the margin and to model the currently stability of various sections of the continental slope based on slope morphology and geomechanical properties. The geomechanical modelling methods used in this study employed the same general approach employed for geomechanical evaluations of soil slope stability in terrestrial settings. It has been slightly modified to accommodate the scale of submarine landslides and the effect of constant saturation of the slope sediments. The final part of the study examines the tsunami hazard related to submarine landslides for Australia’s eastern continental margin. It explores and estimates the coastal flow depth, run-up height, and inundation distance of tsunami waves generated by the submarine landslides that might occur along the margin. The morphology of the upper continental slope along the eastern Australian margin is presented throughout the study in order to establish the geomorphology of the identified submarine landslides from the upper continental slope and present a quantitative characterization of these slides.

5.3 Summary of main conclusions

Stable continental margins can be very prone to submarine landslides and some of the largest slides on record have been shed from these relatively passive terrains. Despite this fact and the obvious accompanying tsunami hazard, slides from passive margins are relatively poorly studied and documented when compared to other settings, such as the flanks of volcanic islands, active subduction- zone margins and submarine fans associated with large river deltas. Regional mapping located on the upper slope of the eastern Australian continental margin reveals several large-volume submarine slides in water depths of 400 to 1200 m. Multibeam bathymetry data and the subbottom profiles have been

5-2 Clarke (2014) Chapter 5 used to identify and describe various morpho-structural features and interbed geometry of the upper slope sediments. The unstable nature of the margin is obvious, with the entire continental slope being characterized by numerous mass wasting processes. Additionally, despite limited seismic penetration, there is generally little evidence of burial of the features by post-slide sediment, which is confirmed by small amounts of material identified in the cores (see Chapter 2; Clarke et al., 2012). Visually identified transition surfaces (boundary surfaces) are identified in five cores at depths of 0.8 to 2.2 meters below the present-day seafloor, which separate looser material from stiffer more compacted material. It is considered that the distinct density contrasts across the boundary surfaces are highly significant. The inference that these boundary surfaces represent failure surfaces of submarine landslides is reasonable and consistent with all the available data.

Overall, the widespread occurrence of upper slope slides across the eastern Australian margin indicates that submarine sliding should be considered to be a common characteristic of this passive continental margin. The experimental results imply that the sediment forming the margin is reasonably strong and inherently stable. Classical limit-equilibrium modeling indicates that submarine landslides should not be a common occurrence on the margin. This indicates that pre-conditioning or some triggering mechanisms must be involved on this passive margin setting which destabilises the slope and enables failure. The processes suspected to be most likely include: 1) dramatic reduction of the shear strength of the upper-slope sediments to extremely low values, possibly induced by creep or a build-up of pore-pressure; 2) long-term modification of the slope-geometry i.e., sedimentation on the head of the slope and/or erosion of the toe of the slope; and/or 3) seismic events large enough to trigger sediment liquefaction or a sudden increase of pore-fluid pressure.

The following statements provide a synthesis of this study’s findings from each chapter:

5.3.1 Important results – Chapter 2

Visually identified transition surfaces (boundary surfaces) are identified in five cores taken from within, or adjacent to, five submarine landslides at depths of 0.8 to 2.2 meters below the present-day seafloor, which separate looser material from stiffer more compacted material. These boundary surfaces show distinct gaps in AMS 14C ages of at least 25 ka and are suspected to represent the removal of material, possibly a slide plane detachment surfaces. Subbottom profiles across three submarine landslides

5-3 Clarke (2014) Chapter 5 indicate that the youngest identifiable seismic reflectors located upslope of three slides terminate on and are truncated by slide rupture surfaces and that the studied slides are geologically recent. There is no obvious evidence in the subbottom profiles for a post-slide sediment layer draped over or otherwise burying slide ruptures or exposed slide detachment surfaces. This suggests that these submarine landslides are geologically recent, and that the boundary surfaces are either: a) erosional features that developed after the occurrence of the landslide in which case the boundary surface age provides a minimum age for landslide occurrence or b) detachment surfaces from which slabs of near-surface sediment were removed during landsliding in which case the age of the sediment above the boundary surface indicates approximately when landsliding occurred. While an earthquake triggering mechanism is favoured for the initiation of submarine landslides on the eastern Australian margin, this causal mechanism cannot be conclusively demonstrated.

5.3.2 Important results – Chapter 3

Geomechanical test data from the upper slope (<1200 m) cores show the sediment from the study area to be as remarkably similar to sediment collected from other sections of the margin e.g. offshore Sydney, and the older compacted mud layers from lower in the stratigraphic section. The sediment is characterised by high shear strengths, low clay content and high void ratios, and brittle behaviour. It seems a major earthquake, or toe erosion could have contributed to initiate slides in this region. The identified boundary surfaces are interpreted to represent detachment surfaces or slide plane surfaces. The soil properties of the sediments show no evidence of weak clay layers, although they do contain significant but relatively small amounts of clay (<25%). Compression testing indicates that the sediment above and below the boundary surface is apparently slightly overconsolidated. The friction angles of the sediments are in the range of 30o - 40o, so that conventional soil mechanics would suggest the slopes have high factors of safety. However, this is clearly not the case as slope failures are widespread. Slide surfaces have been identified by dating and it has been shown that these features are associated with relatively shallow slides (~5-10 m) and not the large slide features that are evident in the bathymetry. Triaxial tests have indicated a significant increase in the brittleness of the shear response with increasing vertical stress level (i.e. burial depth), and that below a depth of ~20 m the soil response will be compressive, leading to the build up of pore pressure when subjected to cyclic (earthquake) loading. This is thought to be significant in explaining why the slides have large thicknessess of 50 to 200m. To date, no conclusive triggering mechanism has been identified for initiating submarine

5-4 Clarke (2014) Chapter 5 landslides in the region although the brittle nature of the sediments and grainsize distribution make the slopes susceptible to liquefaction during oscillatory shaking, favouring an initation mechanism related to seismic shaking.

5.3.3 Important results – Chapter 4

Evidence of submarine landsliding is shown along the entire length of the east Australian continental margin. Thirty-six submarine landslide along the eastern Australian continental margin are identified that are able to produce a tsunami flow depth >5 m at the coastline. Flow depths at the coast range from 3 to 38 m for a landslide velocity of 20 ms-1. Thin (<100 m) and narrow (<5 km) landslides produce smaller tsunami, with coastal flow depths of <5 m, and thick (>100 m) and/or wide (>5 km) landslides generate coastal flow depths of 5-10+ m. The combination of both thick and wide landslides had the greatest potential to generate the largest coastal flow depths >10 m. Maximum inundation distances and run-up heights of 1.6 km and 22 m respectively are calculated for landslide velocities of 20 ms-1, with these values can varying significantly depending on local coastal topography. The reoccurrence of submarine landslides with similar characteristics to those shed from the margin in the geologically recent past would therefore be expected to generate tsunami with maximum flow depths between five and twenty meters at the coastline, run-up of up to 20 m and inundation distances of up to 1.5 km. The hazard of tsunamogenic submarine landslides increases northward of Coffs Harbour, with the number and size of both past tsunamogenic submarine landslides and potential becoming much greater offshore northern NSW and southern QLD. Tsunamogenic submarine landslide scars reported for the central NSW are not big enough to have shed a block able to generate a megatsunami similar to paleo-megatsunamis hypothesised for this section of margin.

5.4 Significance

This project has resulted in one of the most comprehensive morphological and sedimentological data sets ever obtained for the eastern Australian continental slope. Using recently acquired high- resolution multibeam and sub-bottom profile (Topas) data, along with twelve precisely located gravity cores, collected during the RV Southern Surveyor (SS2008-V12) examination of the eastern Australian continental margin (Boyd et al. 2010) the age, physical and mechanical properties of sediments collected from within, and adjacent to, submarine landslides on the upper continental margin, the mor-

5-5 Clarke (2014) Chapter 5 phology and geometry of the submarine landslides, and the potential tsunami hazard they represent has been investigated.

Studies of submarine landslides are particularly important for marine research as they provide information on a range of issues such as margin development, sediment transport processes, and potential future coastal hazards. By examining the occurrence of submarine landslides on the east Austral- ian continental slope, this study provides a basis for other studies of submarine landslides on passive continental margins, on the composition, sedimentology, morphology, evolution, and structure of the east Australian margin, and for the numerical modelling of submarine landslide processes and related tsunami generation. The information gained in the is study will be invaluable to governing bodies wishing to better understand the implications of submarine landsliding on the east Australian margin and devise appropriate management strategies for hazard mitigation.

5.5 Future research

This study has conducted a substantial investigation into the submarine mass failures observed on the eastern Australian continental margin. The results presented here highlight the potential for submarine landslides to be considered an important and ongoing process taking place along the eastern Australian continental margin, influencing its development and present morphology. Submarine landslides should also be considered a source of potential future hazard for the eastern Australian seaboard, given their ability to induce significant tsunami surges at the coastline.

Despite this, the expanse of the study area, as well as the complexities of the processes involved in producing such failures means that there is a need for more thorough examination of the margin and the processes to which it is subjected. While there has been an increasing amount of literature dedicated to assessing submarine mass failures, there are a very limited number of studies focused on the Australian context. A number of the triggers are not well understood; furthermore, there has been no study examining the role of toe erosion on the stability of continental margins.

A number of issues that require further investigation. The results offered by this particular study suggest the following as areas in need of close attention. The investigation of the tsunamogenic potential of submarine landslides in Chapter 4 highlighted the importance of accurately estimating param-

5-6 Clarke (2014) Chapter 5 eters such as slide velocity to the assessment of potential tsunami hazard. Further detailed exploration of the slides, the continental margin, and slides masses that should be present in the abyssal plain sediment pile offshore from the eastern Australian continental margin is required. By cataloguing slide geometry and locating the slump masses associated with submarine landsliding, better estimates of slide velocity can be made, as well as the acceleration involved to displace these masses. Investigations into the location the slump masses would also help determine whether these submarine landslides happen as a single event or multiple events. These investigations would greatly increase the certainty of tsunami hazard assessments.

The investigation of submarine landslide age in Chapter 2 and reoccurrence intervals in Chapter 4 clearly highlights the need for further work investigating the timing and frequency of submarine landslides on the eastern Australian continental margin. This would require specifically collected, and longer, sediment cores taken within submarine landslide scars and debris for additional dating of the slide events, and also deep penetration seismic data in order to hopefully identify multiple, older, buried slides within the sedimentary record to determine slide event frequency.

Chapter 4 highlights the need for further work investigating the potential existence of onshore tsunami deposits. Onshore tsunami deposits have the potential to be linked to offshore processes such as submarine landslides, helping to constrain issues such as the timing, frequency, and size of submarine landslide events, as well as the potential hazard submarine landslides represent for the eastern Australian coastline.

Another factor of submarine landsliding of the eastern Australian continental slope that needs to be improved is understanding the behaviour of the slopes due to both their inherent geological characteristics and their response to external triggers. Chapters 2 and 3 showed that margin structure and sediment characterisation are important when trying to determine possible slope behaviour and what potential triggers might induce sliding on the eastern Australian margin. Additional work needs to be done to accurately characterise the underlying geology of the margin at depth, identifying any possible planes of weakness, different sediment types, and the three-dimensional distribution and bedding structures of the sediment. This would require specifically collected seismic data of the study area, and would enable more complex and sophisticated models to be set up. The result would be a clearer understanding of the structure of the slopes, and more accurate predictions and interpretations of their behaviours. It is clear that if the processes affecting submarine landsliding are to be fully understood,

5-7 Clarke (2014) Chapter 5 and subsequently modelled, the affect of external triggering mechanisms (especially that of earthquake loading) and environmental controls also requires further investigation. In particular the complexities in how various triggers can interact and act on the slopes cumulatively.

References

Boyd, R., Keene, J., Hubble, T., Gardner, J., Glenn, K., Ruming, K., Exon, N., 2010. Southeast Australia: A Cenozoic Continental Margin Dominated by Mass Transport, in: Mosher, D.C., Moscardelli, L., Baxter, C.D.P., Urgeles, R., Shipp, R.C., Chaytor, J.D., Lee, H.J. (Eds.), Submarine Mass Movements and Their Consequences. Springer Netherlands, pp. 491-502. Clarke, S., Hubble, T., Airey, D., Yu, P., Boyd, R., Keene, J., Exon, N., Gardner, J., 2012. Submarine Landslides on the Upper Southeast Australian Passive Continental Margin – Preliminary Findings. Submarine Mass Movements and Their Consequences, in: Yamada, Y., Kawamura, K., Ikehara, K., Ogawa, Y., Urgeles, R., Mosher, D., Chaytor, J., Strasser, M. (Eds.). Springer Netherlands, pp. 55-66. Hubble, T., 2013. Voyage Summary SS2013-V01: Marine Geology and Geohazard Survey of the SE Australian Margin off Northern NSW and Southern Queensland, CSIRO, Hobart.

5-8 Clarke (2014) Appendix 1 Appendix 1

Published Works by the Author Incorporated into the Thesis

1. Clarke, S., D. Airey, P. Yu, and T. Hubble (2011), Submarine landslides on the southeastern Australian margin, in AGS Coastal and Marine Geotechnics - Foundations for Trade, edited by Australian Geomechanics Society.

• Incorporated in Chapter 3

2. Clarke, S., T. Hubble, D. Airey, P. Yu, R. Boyd, J. Keene, N. Exon, and J. Gardner (2012), Submarine Landslides on the Upper Southeast Australian Passive Continental Margin – Preliminary Findings, in: Submarine Mass Movements and Their Conse- quences (5th), edited by Y. Yamada, K. Kawamura, K. Ikehara, Y. Ogawa, R. Urgeles, D. Mosher, J. Chaytor and M. Strasser, pp. 55-66, Springer Netherlands.

• Incorporated in Chapter 1, 2, 3

3. Clarke, S., T. Hubble, D. Airey, P. Yu, R. Boyd, J. Keene, N. Exon, J. Gardner, and S. Ward (2014), Morphology of Australia’s eastern continental slope and related tsunami hazard, in: Submarine Mass Movements and Their Consequences (6th), edited by S. Krastel, J.H Behrmann, D. Volker, M. Stripp, C. Berndt, R. Urgeles, J. Chaytor, K. Huhn, M. Strasser, and C.B. Harbitz, pp. 529-538, Springer Netherlands.

• Incorporated in Chapter 4

A1-1 Clarke (2014) Appendix 1

1 SUBMARINE LANDSLIDES ON THE SOUTH-EASTERN AUSTRALIAN MARGIN

S. Clarke1, D.W. Airey2, P. Yu1 and T. Hubble1 1 School of Geosciences, University of Sydney 2 School of Civil Engineering, University of Sydney

ABSTRACT High-spatial resolution bathymetric data acquired during two recent RV Southern Surveyor voyages have identified several distinct large sediment slides varying in volume from <0.5 km3 to 20 km3 on the upper continental slope of the southeastern Australian margin. Gravity cores, up to 5 m long, have been obtained from areas within and outside the identified slide features, and are believed to intersect the slide planes in some cases at depths of between 85 cm and 220 cm below the present-day seabed. The paper will provide a brief review of the factors believed responsible for submarine landslides, and discuss the relevance of these factors to the southeastern Australian margin. The paper will use the data from the recent ship surveys (SS2008/12 off the southern Queensland / northern New South Wales coastline and SS2006/10 off the mid New South Wales coastline) to show the size and magnitude of submarine landslides that have been identified, present results from soil characterisation studies of the recovered cores, and present the results of mechanical tests. These data will be used to consider the stability of the continental shelf sediments and to consider the tsunami risk that they pose.

1 INTRODUCTION The consequences of submarine landslides include damage to seabed infrastructure (communications cables and buried pipelines), subsidence of coastal areas and the generation of tsunamis (Masson et al., 2009). Our failure to understand the causes of these phenomena means that submarine landslides present a significant risk to coastal and offshore development, and have on occasion resulted in the halting of offshore developments. It has been established that large submarine landslides can produce tsunamis, such as the earthquake triggered submarine slides in 1929 (Grand Banks, USA; Fine et al., 2005) and 1988 (Aitape, Papua New Guinea; Tappin et al., 2001) which both resulted in significant casualties. The large loss of life and damage to infrastructure from the Indian Ocean tsunami of 2004 (Lay et al., 2004) and the recent Japanese event have increased interest in the tsunamigenic potential of large submarine slides. Australia is vulnerable to tsunamis with 85% of the population living within 50 km of the coast and much of the critical infrastructure located close to the coast. It has been suggested (Dominy-Howes, 2008) that the maximum credible tsunami could cover Manly in 6 m of water, and while the possibility of such an event has major implications for risk assessment and siting of critical infrastructure, the likelihood cannot be sensibly determined. The geological record contains many examples of submarine landslides, which can vary in scale from minor shallow slides to very large slides, such as the Storegga slides off the Norwegian coast which have a total volume of over 3000 km3 (Haflidason et al., 2005). Statistics on known landslides on the eastern continental slope of North America, which has geological similarities to Australia’s eastern margin, have been published by Masson et al. (2006). These show that between 30oN and 45oN there are 152 large landslides affecting an area of nearly 40000 km2. Most failures occur on slopes of between 1o and 7o, with the greatest number of failures occurring on slopes of 3.5o. The area affected by failures decreases as the slope increases, and the depth of water at the slide head ranges from 250 m to 2500 m, with the greatest number of failures occurring at water depths of around 1000 m. Despite extensive literature on the nature and causes of submarine landslides, their dynamics and triggering processes are not well understood (Locat and Lee, 2002; Bardet et al., 2003). One of the principal reasons for this is the limited data on the physical and mechanical properties of the sediments, particularly from the slide plane, as these materials have not traditionally been collected. In Australia, studies of the southeastern (SE) Australian continental slope (Jenkins and Keene, 1992; Glenn et al., 2008; Boyd et al., 2010) have been very limited until recently. Evidence of submarine landslides on the SE Australian margin was first reported by Jenkins and Keene (1992), but it was not until high resolution, multibeam bathymetric data became available (Glenn et al., 2008; Boyd et al., 2010) that the true distribution of these slides could be established. The recent collection of high-resolution multibeam echo-sounding and sub-bottom profiling data has provided a detailed view of mass-transport features over a 900 km length of the margin. A wide range of slide features has been detected as well as a series of canyons which cut through the slope sediments. The submarine slides range in scale from hundreds of small slides with volumes of <0.5 km3 up to the largest documented slide which has a volume of 20 km3 (Boyd et al., 2010).

Australian Geomechanics Society Sydney Chapter Symposium October 2011 27

A1-2 Clarke (2014) Appendix 1

SUBMARINE LANDSLIDES ON THE SOUTH-EASTERN AUSTRALIAN MARGIN CLARKE ET AL.

2 SUBMARINE LANDSLIDES

2.1 SOUTHEASTERN AUSTRALIAN MARGIN The SE Australian continental margin stretches 1500 km north from Bass Strait to the Great Barrier Reef (Boyd et al., 2004). The margin which is by world standards narrow, deep and sediment deficient, was formed by rifting in the Cretaceous period between 90M and 65M years ago (Gaina et al., 1998). Since then, margin subsidence has been relatively minor. The continental shelf ranges between 14 and 78 km wide and is relatively flat with a thin sediment cover. The sediment reaches a peak thickness of about 500 m at the edge of the shelf, which occurs at depths ranging between 55 m and 180 m. The continental slope is the region from the shelf edge to the Tasman Abyssal Plain where the water depth is around 4500 m. The continental slope ranges from 28-90 km wide and has average slopes in the range from 2.8° to 8.5°. The sediment cover generally reduces from the shelf edge to the Abyssal plain, and is absent from the lower slope off southern NSW (Boyd et al., 2004). Figure 1 shows the regions where detailed bathymetric studies have been conducted in the last 5 years and from which the sediments have been recovered that are discussed later. Two ship surveys have been conducted, one of the continental slope off Brisbane (SS2008-12), and the other off Sydney (SS2006/10). The surveys consisted of both sub-bottom profiles and echo-sounder records to provide a detailed picture of the seafloor and reveal the underlying geology. In addition 26 gravity cores were obtained from these regions from areas within and adjacent to several slide features, and further sediment was dredged from deeper water. An overview showing the bathymetry for both of the studied areas is given in Figure 2. At this scale it is possible to see that a number of canyons cut into the slope sediments and most of these are off the major rivers. Further details of some of the slides are shown in Figures 3 and 4.

Figure 1: Location map of the two study areas along the southeastern Australian coastline. Blue insets a) and b) mark the location of the bathymetric maps presented in Figure 2. Close inspection of the bathymetric data reveals several distinct large sediment slides varying in volume from <0.5 km3 to 20 km3 on the upper slope (water depths < 1200 m) of the SE Australian margin (Boyd et al., 2010). The large slides typically comprise a distinct U-shaped trough in cross-section (3-6 km wide and 20-250 m deep) backed by an amphitheatre shaped crestal zone. This slide morphology is similar to the classical circular failure profile described by Varnes (1978), but they are elongated in longitudinal profile. The sides and head walls of the scarps are relatively steep with slopes of up to 17o. The largest slides are the Bulli (Figure 3c) and Shovel Slides, near Sydney, on slopes of around 4.5° that are up to 13 km long and 5 km wide with volumes of 20 km3 (Glenn et al., 2008) and the Byron slide (Figure 3b), off Byron Bay, with a volume of 3 km3 and slope of 6.5o

28 Australian Geomechanics Society Sydney Chapter Symposium October 2011

A1-3 Clarke (2014) Appendix 1

SUBMARINE LANDSLIDES ON THE SOUTH-EASTERN AUSTRALIAN MARGIN CLARKE ET AL.

(Clarke et al., 2011). Sub-bottom profiling data from multiple sections across the continental slope and in particular across the slides show the sediment is built up of well stratified beds (Glenn et al., 2008), which have been suggested to be evidence of past slide events. In most locations, sediments derived from the slides cannot be detected on the slope and it appears that the slide material has been transported to the abyssal plain. However, in a small number of locations (Figure 3a) where the slopes are less steep (< 2o), the slide debris flow deposits have remained on the slope and contain blocks up to 350 m wide and 50 m high. Figure 2 shows a large number of canyons that cut into the continental slope sediments. These have been categorised into large box canyons, and smaller narrow linear canyons. The 46 large box canyons are on average 14 km wide, 20 km long and over 600 m deep, they stretch from the middle slope to the abyssal plain, and have slopes up to 17° on the walls. Narrow linear canyons occur in the upper slope sediments, most located in central NSW off major river systems such as the Shoalhaven, Hunter and Tweed or off Fraser Island. Well developed examples are 800-1900 m wide, 120-320 m deep and extend downslope for 14-22 km. Canyon wall slopes are up to 34°, the steepest slopes found on this margin (Boyd et al., 2010).

Figure 2: Bathymetric maps of the a) the southern Queensland / northern New South Wales continental slope and b) the mid New South Wales continental slope. Data for these maps were collected on two RV Southern Surveyor voyages: SS2008/12 off the southern Queensland / northern New South Wales coastline (Boyd et al., 2010) and SS2006/10 off the mid New South Wales coastline (Glenn et al., 2008). Insets mark the location of the close-up slopes images presented in Figure 3 and Figure 4.

Australian Geomechanics Society Sydney Chapter Symposium October 2011 29

A1-4 Clarke (2014) Appendix 1

SUBMARINE LANDSLIDES ON THE SOUTH-EASTERN AUSTRALIAN MARGIN CLARKE ET AL.

Figure 3: Digital elevation model (DEM) of the slope geometry for four slide sites (outlines denoted by black line): (a) Coolangatta 1 and Coolangatta 2 Slides, (b) Byron Slide, (c) Bulli Slide. Also shown are locations of the three gravity cores (GC8, GC9, GC12) referenced in this study, collected on the RV Southern Surveyor SS2008/12 voyage.

30 Australian Geomechanics Society Sydney Chapter Symposium October 2011

A1-5 Clarke (2014) Appendix 1

SUBMARINE LANDSLIDES ON THE SOUTH-EASTERN AUSTRALIAN MARGIN CLARKE ET AL.

Figure 4: Digital elevation model (DEM) of a section of the mid-slope within the study area. Note the abundance of slump/slide scars presenting arcuate crests (crest outlines denoted by black dashed lines). Modified from Hubble, 2011. Further observations from the bathymetry include the widespread slope failures on the mid-slope, shown in Figure 4, which demonstrates an average slope of around 8o and the widespread relatively shallow failures observed on the Nerang plateau shown in Figure 3a, where the average slope is < 2o. There are also circular depressions, referred to as pock marks, off Newcastle which are believed to be associated with gas leakage from the underlying Permian coal measures (Glenn et al., 2008). The figures reveal evidence of widespread erosional features on the SE Australian continental slope. This is different from other margins of similar age, for example the US Atlantic and Gulf Coasts, where sediment deposition is the more dominant process. However, both margins exhibit extensive slides and other erosional features. The 500 m thickness of the sediments on the SE Australian margin has been taken as evidence of a previous period of deposition (Davies 1979, Boyd et al., 2004), but the sediment thickness is substantially less than other margins. This can in part be explained by the dryness of the Australian continent, the relatively subdued highlands and its small rivers. When the resulting low sedimentation is combined with ongoing subsidence of the abyssal plain caused by initial crustal thinning and later thermal cooling, which has caused the gradients on the margin to slowly increase, the result has been a retrogressive gravity failure over all of the lower slope and much of the upper slope. Thus it is considered that the present state of sediment instability, where the overlying sediment wedge is continually undercut by slope failure over geological time, is the cause for modern episodes of failure (Glenn et al., 2008).

2.2 TRIGGERS The literature on submarine landslides summarised by Masson et al. (2009) and Locat and Lee (2002) lists a variety of causes for their initiation. These include: earthquakes, storm wave loading, erosion and in particular slope over-steepening, rapid sedimentation leading to under-consolidation, the presence of weak layers, gas hydrate dissociation, sea-level changes, glaciations/isostatic uplift, volcanic activity and diapirs. It is also widely accepted that a combination of these factors is required to initiate a landslide, especially where these occur on very shallow slopes. There is data indicating that several large landslides have coincided with earthquakes, (e.g. Tappin et al., 2001; Bardet et al., 2003; Masson et al., 2006; Synolakis et al., 2002). The role of weak layers, oriented parallel to the sedimentary bedding, has long been used to explain the scale of some large slides, but more recently the importance of weak layers in controlling sliding at all scales has been noted (Masson et al., 2009). Despite this Masson et al. (2009) also commented that “we know very little about the nature and characteristics of these weak layers, since they have rarely been sampled and very little geotechnical work has been done on sediments recovered from them”. An important consideration is the brittleness of the sediments. Weak layers need to lose strength rapidly and pore pressure needs to rise for effective stresses to reduce. Masson et al. (2009) suggest that clay rich sediments with high water content and high plasticity are required for this to occur.

Australian Geomechanics Society Sydney Chapter Symposium October 2011 31

A1-6 Clarke (2014) Appendix 1

SUBMARINE LANDSLIDES ON THE SOUTH-EASTERN AUSTRALIAN MARGIN CLARKE ET AL.

‐20

‐40

‐60 (m)

‐80 Level

Sea ‐100

‐120 Relative

‐140 ‐450 ‐400 ‐350 ‐300 ‐250 ‐200 ‐150 ‐100 ‐50 0 Years Before Present (Thousands)

Figure 5: Relative sea level history (after Waelbroeck et al., 2002) Lee (2009) has shown that landslides were more frequent during and just after the last period of glaciation than they are today. One of the suggested reasons for this is that glacial periods coincide with periods of low relative sea level. Figure 5 shows the relative sea level curve for Australia for the last 0.5 million years, which indicates that the sea level has been over 100 m below its present level on several occasions, and the times of these minima are associated with glacial periods. The lowered sea level can increase the likelihood of sliding because it results in the shoreline migrating closer to the shelf edge, leading to increased erosion and higher rates of sedimentation offshore, which now occurs directly on the slope. The lower water pressures (and possibly changed temperatures) can lead to release of gas from gas hydrates increasing pore pressures and reducing strength, and related changes to stress levels in the crust can increase seismic activity. Increased groundwater flows from underlying rocks can occur also contributing to reduced strength. It has also been suggested that changes to deep ocean currents are associated with glaciations and erosion from these currents can contribute to slope steepening (Hubble et al., 2011). Another mechanism that has been postulated to explain submarine slides is that of creep, slow down slope movements due to gravity stresses that may lead to failure of the sediment mass or to failure on a weak layer at depth. It has been demonstrated that thick deposits on steep slopes can fail by this process (Silva and Booth, 1984). However, as noted by Hampton et al. (1996) proof that creep is significant on continental slopes is elusive. Observations of the widespread occurrence of submarine slides suggests that weak clay layers cannot be a major cause, and tend to favour earthquakes as the triggering mechanism. Nevertheless, it is widely accepted that neither the submarine sliding process nor the slide triggering mechanisms are very well understood (e.g. Locat and Lee, 2002; Mosher et al., 2009), and this is particularly so in the cases of submarine mass failures that do not appear to be linked to seismic activity.

2.3 SEDIMENT PROPERTIES From the recent ship cruises 26 gravity core samples have been obtained, 14 from the region off Sydney and 12 from the region off Brisbane. For most of the gravity cores the soil has been logged and basic properties, particle size distribution, mineralogy and densities have been obtained. The results from the Sydney region have been reported in detail by Glenn et al. (2008) and only typical results are reported here. The basic classification tests have been supplemented by a limited number of triaxial, oedometer, shear box and vane shear tests to investigate the mechanical behaviour of the sediments. The mechanical tests have been performed to investigate the landslide triggering processes and in particular to determine the collapse potential of the sediment and the influence of composition and stress level on this behaviour. A summary of the classification data, which is limited to material from the upper 5.3 m of the sediments as this was the maximum depth of penetration of the gravity corer, is included in Table 1. This shows that the continental slope sediments are predominantly comprised of silt sized material, with about 15% clay, variable amounts of sand sized particles and significant organic content. The sediments contain a significant amount of

32 Australian Geomechanics Society Sydney Chapter Symposium October 2011

A1-7 Clarke (2014) Appendix 1

SUBMARINE LANDSLIDES ON THE SOUTH-EASTERN AUSTRALIAN MARGIN CLARKE ET AL.

carbonate grains derived from the remains of living organisms and also significant amounts of terrigeneous material, believed to be transported by the wind from the interior of the continent. Although there is some variability in the composition from core to core there is a broad similarity in the particle distributions all along the margin. Table 1: Summary of available classification data Clay Silt Sand Carbonate Organic LL Ip Moisture (%) (%) (%) (%) (%) content (%) Brisbane 10-20 50-65 15-40 20 8 46.5 9 50-100 Sydney 8-25 30-80 10-60 35 ? 44 18 55-85 Several of the gravity cores were obtained from within detected landslide features in an attempt to penetrate through the base of the slides to assist in constraining the slide depths and dates. In most locations this was unsuccessful as the recent sediment drape overlying the slide surface was thicker than the gravity corer could penetrate. However, in three locations off SE Queensland a distinct boundary in the sediments could be detected at depths between 87 cm and 220 cm. The sedimentalogical and geotechnical properties of the sediments and their variation with depth from one of these cores (GC12; see Figure 3b for core location) are shown in Figure 6. It can be seen from the figure that there is a distinct change in density and moisture content, as well as appearance of the material at a depth of 87 cm. It is also noticeable that this change in density is not associated with any significant change in the grading, carbonate content or organic content of the material.

Figure 6: Characterisation of sediment from core GC12, showing physical properties with depth below seabed. Bulk radiocarbon dates are also shown. The presumed slide plane is indicated with a dashed black line at 87 cm depth below seabed (Modified from Clarke et al., 2011) Additionally, a bulk radiocarbon (14C) age was determined for sediment sampled directly above the slide plane for this core, returning a date of 15.8 ka for the recent material just above the inferred slide surface (Clarke et al., 2011). This date is consistent with sliding occurring around the time of the most recent sea level low shown in Figure 5. Dating for the deeper sediment could not be determined because its age exceeded that for which 14C dating is reliable. The dating has also enabled the rate of sedimentation to be determined, providing values between 0.3 and 1.2 m/10,000 years. As the rate of deposition is expected to have been higher in the past, these rates of sedimentation suggest that sliding must have been a geologically common event since the formation of the margin 60 million years ago as the current sediment deposit is less than 500 m thick. Using the values of Cc given below, the change in moisture content at the inferred slide plane can be interpreted as representing a slide depth of anywhere between 10 m and 200 m. The depth reconstructed at the GC12 site by replacing the material apparently missing from the U-shaped trough, i.e. by maintaining the continuity and shape of the adjacent slope and projecting it above the GC12 site, is approximately 250 m. Thus while it is possible to date a possible slide at approximately 16,000 years there is insufficient information to determine whether this is

Australian Geomechanics Society Sydney Chapter Symposium October 2011 33

A1-8 Clarke (2014) Appendix 1

SUBMARINE LANDSLIDES ON THE SOUTH-EASTERN AUSTRALIAN MARGIN CLARKE ET AL.

the date of the main slide at this location and further mechanical and dating studies are in progress to further constrain the result.

2.4 GEOMECHANICAL BEHAVIOUR Limited geomechanical data are available from the vicinity of the slides close to Sydney and from the region off Byron Bay. This has consisted primarily of one-dimensionally consolidated undrained triaxial tests, with some additional shear box tests to evaluate residual strength properties for the Sydney sediments. Figure 7 shows typical compression plots from 1-D compression tests on undisturbed specimens. Results of three specimens from one of the cores (GC9; see Fig 3a for location), from SE Queensland are shown together with a typical specimen from a core off Sydney. Although there is some variability in the responses the similarity of the response of the specimen from Sydney and SE Queensland is remarkable. Based on these very limited data it appears that the grading, mineralogy and compressibility of specimens on the continental slope are similar along most of the 1500 km of the SE Australian margin. The specimens show high compressibility with Cc values ranging from 0.3 to 0.65. This is considerably higher than would be expected from their remoulded index test results, as the correlation Cc = IpGs/2 would suggest Cc values of 0.13 to 0.26 for plasticity indices of 10% to 20%. It can also be noted that the moisture contents in the upper 5 m are significantly higher than the liquid limit and vane shear tests have indicated that the specimens have significant sensitivities (>2). These data indicate that the slope sediments are structured and, while the cause of the sensitivity has not been established, it could be related to the relatively high organic content of up to 8%, which is known to be a factor in sensitivity in other soils.

1.5

1.4

1.3 Void ratio, e ratio, Void

SEQ - GC9 - T1

1.2 SEQ - GC9 - T3

SEQ - GC9 - T6

Sydney - GC03

1.1 10 100 1000

Mean effective stress, p' (kPa)

Figure 7: Response of core specimens to 1-D compression The responses of 3 specimens to undrained shearing in triaxial compression are shown in Figure 8, and the associated effective stress paths are shown in Figure 9. To enable comparison of the tests the deviator stress and excess pore pressures have been normalised by the vertical effective stress at the start of shearing. Specimens GC9-T1 and GC9-T3 were adjacent specimens and have similar compressibilities, as seen from Figure 7, but they responded differently to shearing. Specimen GC9-T1 at the higher stress level (σ′v = 620 kPa) shows a more brittle response with the peak deviator stress attained at a very small strain, after which the resistance rapidly decreases to its ultimate value. In contrast, specimen GC9-T3 at the lower stress (σ′v = 167kPa) does not reach a maximum until relatively large strain. From the pore pressure responses and the effective stress paths it can be seen that this difference is a consequence of a transition from dilative to compressive behaviour as the stress level increases. The more compressible (Sydney) specimen shows a response similar to the higher stress GC9-T1 even though the stress level (σ′v = 220kPa) is similar to GC9-T3. The shear response of GC9-T6, which has similar compressibility to the Sydney specimen also shows the tendency for increasing brittleness with increasing stress level, however, this result is not considered reliable due to non-uniform deformation during shearing. This pattern of reducing dilation and increasing brittleness with stress level can explain why deeper failure surfaces develop.

34 Australian Geomechanics Society Sydney Chapter Symposium October 2011

A1-9 Clarke (2014) Appendix 1

SUBMARINE LANDSLIDES ON THE SOUTH-EASTERN AUSTRALIAN MARGIN CLARKE ET AL.

1.2 SEQ -GC9 -T1 SEQ -GC9 -T1 SEQ -GC9 -T3 1 SEQ -GC9 -T3 SEQ -GC9 -T6 SEQ -GC9 -T6 Sydney - GC02 0.8 Sydney - GC02

0.6

Normalised stress Normalised 0.4

0.2

0 0 0.05 0.1 0.15 0.2 Axial strain Figure 8: Normalised deviator stress and pore pressure responses from 1-D compressed specimens.

600

500

400

300

200 Deviator stress (kPa) stress Deviator

SEQ - GC9 - T1 100 SEQ -GC9 - T3 SEQ - GC9 - T6 Sydney - GC02

0 0 100 200 300 400 500 Mean effective stress, p' (kPa) Figure 9: Effective stress paths from triaxial tests From Figure 9 it can be seen that all specimens approach a similar ultimate frictional resistance, which for the specimens shown ranged from 37o to 40o. Cyclic shear box tests showed no evidence of any lower residual frictional strength (Glenn et al., 2008).

3 ANALYSIS

3.1 LANDSLIDE INITIATION Geomechanical modeling of three of the submarine landslides has been undertaken using the slope stability program GEO-SLOPE/W (2007) to examine the influence of cohesion, friction angle and slope geometry on the stability (Clarke et al, 2011). As the friction angle is around 40o and the slopes are from 3o to 6o static analyses predict very high factors of safety. Analyses have also been conducted to investigate the effects of earthquake loading by including a factor for seismic accelerations in the standard pseudo-static limit-equilibrium calculations. Selecting an appropriate value for the seismic coefficient acting on the failure mass can be especially difficult (Seed and Martin 1966). A very crude investigation of seismic loading on the slopes indicates that lateral and vertical accelerations of 0.3 g (ah = 0.3 g, av = -0.3 g), the upper limit of those used to investigate the stability of earth dams during earthquakes (Seed and Martin, 1966; Ozkan, 1998), would be sufficient to destabilise the slopes of the seafloor in the present study. While this approach has been widely used to assess the

Australian Geomechanics Society Sydney Chapter Symposium October 2011 35

A1-10 Clarke (2014) Appendix 1

SUBMARINE LANDSLIDES ON THE SOUTH-EASTERN AUSTRALIAN MARGIN CLARKE ET AL.

stability of submarine slides to earthquake events, its applicability to such large volumes of soil is questionable and the approach is of limited value in understanding the mechanisms leading to failure. Puzrin et al. (2004) have argued that it is unlikely that a failure can develop over distances of several kilometres instantaneously, and that a progressive failure mechanism must be considered. On land progressive failures are often observed to result from oversteepening of the toe of a slope, where failure at the toe leads to a retrogressive failure that migrates upslope. This mechanism is also considered to be responsible for the large Storrega submarine slide. Puzrin et al. (2004) have suggested an alternative progressive failure mechanism that involves a weakened zone propagating down slope. The basis of the analysis can be explained by considering Figure 10. The starting point is that a zone of elevated pore pressures develop (Figure 10a), possibly owing to an earthquake, where the soil reaches a state of failure for which the mobilisable soil resistance is lower than the stresses at equilibrium from the weight of the overlying soil (τr < τg Figure 10b). If the length of this failed zone is sufficient a global and catastrophic failure will occur. However, if the zone of failure is more limited the soil will tend to move downslope into the currently unfailed region. If the failure plane can propagate because the energy released is greater than that needed to progress the failure then the shear plane can grow, and if conditions are unfavourable, it may continue to advance until it reaches a length where global failure results. For significant energy release to occur the ultimate resistance of the soil needs to be lower than that required to resist the gravitational stresses, and the soil needs to respond in a brittle manner. The triaxial test data shown above display the type of brittle behaviour that can potentially lead to this type of mechanism.

Figure 10: (a) Zone of elevated pore pressures, (b) Slope failure mechanism (after Puzrin et al., 2004) The analysis of Puzrin et al. (2004) suggests that failure begins upslope, but depending on the soil type and behaviour progressive failure may be limited or not occur and it is possible that the resulting length of the failure surface may be less than the critical value required for a catastrophic failure. There is some evidence for this from a number of head scarps present on the SE Australian margin where the soil below has not moved significantly. Glenn et al. (2008) suggest that these features represent the sites most likely to fail in the future. However, if the analysis of Puzrin et al. (2004) is correct, the soil movements make these sites less likely rather than more likely to lead to failure.

3.2 TSUNAMI GENERATION The viscous drag on the overlying ocean due to the movement of the slide block is responsible for the tsunami generation. Many studies have been conducted into tsunami generation and propagation, but only the work of Ward (2001) will be mentioned here. Ward (2001) presents results of tsunami generation and propagation based on classical tsunami theory and assuming linear wave theory. This theory uses a rigid seafloor overlain by an incompressible, homogeneous and non-viscous ocean subjected to a constant gravitational field. Figure 11 provides an indication of the size of the tsunami from analysis of a rectangular block of length L (km) and width W (km) sliding down an inclined plane for a water depth of 1000 m. The tsunami velocity is given by √(gh) where h is the water depth, and in 1000 m of water this is 99 m/s. There are no reliable data on the speed of submarine slides, although turbidity currents from the 1929 Grand Banks slide travelled at 25 m/s, and based on travel distances of slide debris speeds of up to 80 m/s have been inferred for some large slides (Masson et al., 2006). The fracture mechanics approach proposed by Puzrin et al. (2010) enables an estimate of the initial velocity to be determined and values of around 10 m/s were estimated for some reported slide geometries. For the largest slides on the SE Australian margin the water depth is around 1000 m, the maximum slide thickness is 200 m, and assuming a maximum velocity of 20 m/s a peak tsunami wave height of around 20 m can be estimated from Figure 11. As the maximum dimensions of the sliding blocks on the SE Australian margin are 20 km x 5 km some reduction to this height may be appropriate. Further reductions in wave height will occur as any waves approach the coast.

36 Australian Geomechanics Society Sydney Chapter Symposium October 2011

A1-11 Clarke (2014) Appendix 1

SUBMARINE LANDSLIDES ON THE SOUTH-EASTERN AUSTRALIAN MARGIN CLARKE ET AL.

Peak Tsunami Height/ Slide Height

0 1.0 2.0 0 1.0 2.0 Slide velocity/Tsunami velocity Figure 11: Effects of slide velocity and slide dimensions on peak tsunami height (after Ward, 2001).

4 SUMMARY The paper has shown evidence of large scale mass wasting phenomena on the SE Australian continental slope, including the existence of many landslide features. Slides are evident all along the margin from Shoalhaven to the Sunshine Coast, and on slopes ranging from 1o to 9o. The soil properties of the upper sediments are similar along the margin and show no evidence of weak clay layers, although they do contain significant amounts of clay. The friction angles of the sediments are in the range of 37o - 40o, so that conventional soil mechanics would suggest the slopes have high factors of safety. However this is clearly not the case, as slope failures are widespread. Triaxial tests have indicated a significant increase in the brittleness of the shear response with stress level, and this is thought to be significant in explaining why the slides have thicknessess of 50 m to 200 m. The largest slope failures have a volume of 20 km3 and have the potential to generate significant tsunami waves. The dating of the slides suggests that the most recent failures occurred at the time when the sea level was at its minimum during the last glaciation. There are several reasons why the likelihood of slides should increase at these times, however the cause of the slides on the SE Australian margin is not well understood. While the likelihood of slides appears to be lower in inter-glacial periods there are examples of earthquake caused submarine slides that have occurred recently and the possibility that a large submarine slide could occur any day cannot be discounted.

5 ACKNOWLEDGEMENTS We would like to acknowledge the P&O crew and scientific crews of the RV Southern Surveyor voyage (12/2008). Funding for this voyage was provided by ARC Australia and ConocoPhillips Pty Ltd. Funding for the radiocarbon age determinations was provided by Professor Ron Boyd and Newcastle University, Australia, and Ron Boyd and Jock Keene led the ship cruises.

6 REFERENCES Bardet J.P, Synolakis, C.E, Davies, H.L, Imamura, F, Okal, E.A (2003) Landslide Tsunamis: Recent Findings and Research Directions. Pure App Geophys 160:1793–1809. Boyd, R., Ruming, K. and Roberts, J.J. (2004), Geomorphology and surficial sediments on the southeast Australian continental margin, Australian Journal of Earth Sciences, 51, 743-764. Boyd R, Keene J, Hubble T, Gardner J, Glenn K, Ruming K, Exon N (2010) Southeast Australia: A Cenozoic Continental Margin Dominated by Mass Transport. In: Mosher DC, Moscardelli L, et al. (eds) Submarine Mass Movements and Their Consequences. vol 28. Advances in Natural and Technological Hazards Research. Springer Netherlands, pp 1-8. Clarke S, Hubble T, Airey D.W, Yu P, Boyd R, Keene J, Exon N, Gardner J (2011). Submarine landslides on the upper East Australian continental margin - preliminary findings. In: Submarine Mass Movements and Their Consequences (this issue). Advances in Natural and Technological Hazards Research. Springer Netherlands. Davies P.J (1979) Marine geology of the continental shelf off Southeastern Australia. BMR Bulletin 195:51 Dominey-Howes, D. (2007), Geological and historical records of tsunami in Australia, Marine Geology, 239, 99- 123.

Australian Geomechanics Society Sydney Chapter Symposium October 2011 37

A1-12 Clarke (2014) Appendix 1

SUBMARINE LANDSLIDES ON THE SOUTH-EASTERN AUSTRALIAN MARGIN CLARKE ET AL.

Fine, I.V., Rabinovich, A.B., Bornhold, B.D., Thomson, R.E. & Kulikov, E.A. (2005), The Grand Banks landslide-generated tsunami of November 18, 1929: preliminary analysis and numerical modeling, Marine Geology, 215, 45-57. Gaina C.M, Muller R.D, Royer J-Y, Stock J, Hardebeck J, Symonds P (1998) The tectonic history of the Tasman Sea: a puzzle with 13 pieces. J Geophys Res103:12413–12433 GEO-SLOPE (2007), Stability Modeling with SLOPE/W 2007: An Engineering Methodology, GEO-SLOPE International Ltd., Canada. Glenn K, Post A, Keene J, Boyd R, Fountain L, Potter A, Osuchowski M, Dando N, Party S (2008) NSW Continental Slope Survey – Post Cruise Report Geoscience Australia, Record 2008/14, 160pp. Haflidason, H., Sejrup, H.P., Nygard, A., Richter, Mienert, J., Bryn, P., Lien, R., Fosberg, C.F., Berg, K. & Masson, D.G. (2004), The Storegga Slide: Architecture, Geometry and Slide Development, Marine Geology, 213, 201-234. Hampton, M. A., J. Locat, and H. J. Lee (1996), Submarine landslides, Reviews of Geophysics, 34, 33–59. Hubble T, Yu P, Airey D, Clarke S, et al. (2011). Physical properties and age of continental slope sediments dredged from the Eastern Australian Continental Margin implications for timing of slope failure. In: Submarine Mass Movements and Their Consequences (this issue). Advances in Natural and Technological Hazards Research. Springer Netherlands. ‐ Jenkins C.J, Keene J.B (1992) Submarine slope failures on the southeast Australian continental slope. Deep Sea Research 39:121-136. Lay T., Kanamori H., Ammon C.J., Nettles M., Ward S.N., Aster R.C., Beck SL. Bilek S.L. Brudzinski M.R., Butler R. DeShon H.R., Ekström G., Satake K. and Sipkin S. (2005), The Great Sumatra-Andaman Earthquake of 26 December 2004, Science, 308, 1127-1133. Lee, H. J. (2009), Timing of occurrence of large submarine landslides on the Atlantic Ocean margin, Marine Geology, 264(1-2), 53-64. Locat J, Lee, H.J. (2002) Submarine landslides: advances and challenges. Canadian Geotech J 39:193-212 Masson, D. G., Harbitz, C. B., Wynn, R. B., Pedersen, G., Lovholt, F. (2006), Submarine landslides: processes, triggers and hazard prediction, The Philosophical Transactions of the Royal Society A (Phil. Trans. R. Soc. A), 364, 2009-2039. Ozkan M.Y (1998) A review of considerations on seismic safety of embankments and earth and rock-ll dams. Soil Dynamics and Earthquake Engineering 17:439-458 Puzrin AM, Germanovich LN, Kim S (2004) Catastrophic failure of submerged slopes in normally consolidated sediments. Géotechnique, 54 (10):631-643 Puzrin, A.M., Saurer, E. & Germanovich, LN. (2010), A dynamic solution of the shear band propagation in submerged landslides, Granular Matter, 12 (3), 253-265. Seed H.B, Martin G.R (1966) The seismic coefficient in earth dam design. Journal of Geotechnical Engineering Silva A.J, Booth J.S (1984), Creep behavior of submarine sediments. Geo-Marine Letters 4 (3):215-219. doi:10.1007/BF02281709 Synolakis, C.E., Bardet, J-P., Borrero, J.C., Davies, H.L., Okal, EA., Silver, E.A., Sweet, S., and Tappin, D.R. (2002): “The slump origin of the 1998 Papua New Guinea Tsunami,” Proc. R. Soc. London, A 458, 763-789. Tappin D.R, Watts, P., McMurtry, G. M., Lafoy, Y., and Matsumoto, T. (2001) The Sissano, Papua New Guinea tsunami of July 1998 – Offshore evidence on the source mechanism. Marine Geology 175:1-23 Varnes D.J (1978) Slope Movements and Types and Processes. Landslides: Analysis and Control, Special Report, Vol 176. Transportation Research Board, National Academy of Sciences, Washington. Waelbroeck, C., L. Labeyrie, E. Michel, J. C. Duplessy, J. F. McManus, K. Lambeck, E. Balbon, and M. Labracherie (2002), Sea-level and deep water temperature changes derived from benthic foraminifera isotopic records, Quaternary Science Reviews, 21(1-3), 295-305. Ward, S. N. (2001 ), Landslide tsunami, Journal of Geophysical Research, 106, 11201-11215.

38 Australian Geomechanics Society Sydney Chapter Symposium October 2011

A1-13 Clarke (2014) Appendix 1 2 Submarine landslides on the upper southeast Australian passive continental margin - preliminary findings

Samantha Clarke1, Thomas Hubble1, David Airey1, Phyllis Yu1, Ron Boyd2, 5, John Keene1, Neville Exon4, James Gardner3 and Shipboard Party SS12/2008 1. University of Sydney, Sydney, NSW, Australia. [email protected] 2. University of Newcastle, Newcastle, NSW, Australia

3. CCOM, University of New Hampshire, Durham, NH, United States

4. Earth and Marine Sciences, Australian National University, Canberra, ACT, Australia 5. ConocoPhillips, Houston, TX, United States

Abstract

The southeast Australian passive continental margin is narrow, steep and sediment- deficient, and characterized by relatively low rates of modern sedimentation. Upper slope (<1200m) sediments comprise mixtures of calcareous and terrigenous sand and mud. Three of twelve sediment cores recovered from geologically-recent, submarine landslides located offshore New South Wales/Queensland (NSW/QLD) are interpreted to have sampled failure surfaces at depths of between 85 cm and 220 cm below the present-day seabed. Differences in sediment physical properties are recorded above and below the three slide-plane boundaries. Sediment taken directly above the inferred submarine landslide failure surfaces and presumed to be post-landslide, returned radiocarbon ages of 15.8 ka, 20.7 ka and 20.1 ka. The last two ages correspond to adjacent slide features, which are inferred to be consistent with their being triggered by a single event such as an earthquake. Slope stability models based on classical soil mechanics and measured sediment shear-strengths indicate that the upper slope sediments should be stable. However, multibeam sonar data reveal that many upper slope landslides occur across the margin and that submarine landsliding is a common process. We infer from these results that: a) an unidentified mechanism regularly acts to reduce the shear resistance of these sediments to the very low values required to enable slope failure, and/or b) the margin experiences seismic events that act to destabilise the slope sediments.

Keywords: Mass-failure, multibeam, seafloor geomorphology, continental slope.

A1-14 Clarke (2014) Appendix 1

2 1. Introduction

The geological record contains many examples of submarine landslides. They have been often associated with locations presenting inclined seafloor slopes, areas where sedimentation rates are high, sediments are fine-grained, or seafloor rocks are weakened by fractures. Despite extensive literature on the nature and causes of submarine landslides, their dynamics and triggering processes are not well understood (Locat and Lee 2002; Bardet et al. 2003). One of the principal reasons for this is the limited data on the physical and mechanical properties of sediments from the slide plane, as these materials have not traditionally been collected in historical studies. In Australia, there are only a few studies that have investigated submarine landslides on the Australian continental slopes (e.g. Jenkins and Keene 1992; Glenn et al. 2008; Boyd et al. 2009). While evidence for submarine landslides on the SE Australian margin was first reported by Jenkins and Keene (1992), it was not until high resolution, multibeam bathymetric data became available (Glenn et al. 2008; Boyd et al. 2009) that the true distribution of these slides could be established. Preliminary investigations from our survey on the RV Southern Surveyor (SS12/2008) indicate that many large submarine landslides occur on the SE Australian coastal margin between Byron Bay and Tweed Heads (Fig. 1). Radiocarbon ages from the sediment indicate that they are also geologically young (< 20 ka) (Boyd et al. 2009, 2010).

1.1 Study Area

The study area is located along the SE Australian continental margin off northern New South Wales and southern Queensland (Fig. 1). Mapping of the seabed and sampling occurred approximately 30 km to 70 km seaward of the present coastline between Point Cartwright in the north to Clarence Head in the south, in water depths of 150- 4500 m. Three submarine landslides identified by Boyd et al. (2010) on the upper continental slope are described in this work. They are known as the Coolangatta 1 slide (headwall depth 500 m; centre of mass water depth 600 m), the Cudgen slide (headwall depth 550 m; centre of mass water depth 600 m), and the Byron slide (headwall depth 500 m; centre of mass water depth 800 m) (Figs. 1, 2). A total of eight gravity cores were recovered from these landslides.

2. Data and Methods

2.1 Bathymetry and Slide Geometry

High-spatial resolution bathymetric data were acquired during the RV Southern Sur- veyor voyage (SS12/2008) using a Simrad EM300 multibeam echosounder. Sub-bottom profile data (Topaz) were also acquired in water depths less than 1000 m (Boyd et al.

A1-15 Clarke (2014) Appendix 1

3 2010). The raw multibeam data was processed and merged to produce a single complete dataset for the entire study area (for details see Boyd et al. 2010; Fig 1). This dataset was then displayed using Fledermaus v7.1.2 software. Several distinct large sediment slides varying in volume from <0.5 km3 to 20 km3 are identified on the upper slope (< 1200 m) of the SE Australian margin (Boyd et al. 2010). The three slides (Fig. 2II A-C), which are the focus of the present study, typically comprise a distinct U-shaped trough in cross-section (3-6 km wide and 20-250 m deep) backed by an amphitheatre shaped crestal zone. In each case, this slide morphology is similar to the classical circular failure profile described by Varnes (1978), but they are elongated in longitudinal profile.

Figure 1 I) Location map of showing the location and bathymetry of the study area (dashed line). Insets mark the three slides (north to south): (A) Coolangatta 1 Slide, (B) Cudgen Slide, (C) By- ron Slide. II) The three slide sites showing the locations of 8 gravity cores (GC) collected on the RV Southern Surveyor (SS2008/12) voyage.

2.2 Sediment properties

The three gravity cores used in this study (GC8, GC11 and GC12) were visually logged and sub-sampled for: grainsize and mean grainsize distribution (analysis using GRADISTAT 4.0, Blott and Pye 2001), carbonate and organic carbon content (loss on ignition method, following Heiri et al. 2001), dry bulk unit density, unit weight, void ratio, specific gravity, and water content. The mechanical behaviour of the sediment was also investigated using geomechanical tests to determine the collapse potential of the sediment, and the influence of composition and stress level on this behaviour. Monotonic triaxial tests were performed on ~126 cm-long core sections taken from adjacent gravity cores from the same feature in order to generate input parameters (i.e., cohesion and friction angle) required for geomechanical modeling. These tests included one-dimensional consolidation in triaxial apparatus at a range of stress paths, followed by undrained shearing to failure.

A1-16 Clarke (2014) Appendix 1

Figure 2 Digital elevation model (DEM) of the slope geometry for the three slide sites (outlines denoted by black line): (A) Coolangatta 1 Slide, (B) Cudgen Slide, (C) Byron Slide. Also shown are locations of the three gravity cores (GC8, GC11, GC12) in this study.

3. Results and Interpretation

3.1 Sediment Properties

Two distinct sediment units occur in cores GC8, GC11 and GC12 (Fig. 3): a younger, sediment drape overlies an older, sediment unit. These two units are separated by a distinct physical boundary located below the present-day seabed at 87 cm in GC12, 90 cm in GC8 and 220 cm in GC11. The boundary features are all clearly distinguishable and while they all easily identified by a colour change it was the distinct, down-core increase in stiffness identified during the initial logging of the core that indicated their potential genetic significance (Fig. 3II). In all three cases these boundaries are interpreted to represent the basal surfaces of submarine landslides. This interpretation is based on distinct down-core changes in bulk density (increase from 851 kg m-3 to 1062 kg m-3 in GC12), unit weight (increase from 15.0 kN m-3 to 16.4 kN m-3 in GC12) and moisture content (decrease from 80% to 57% in GC12) across the inferred slide-plane boundary (Fig 4). The bulk density determined for the sediment sampled below the identified boundary in GC12 is consistent with a burial and compaction by column of material at least 200 m thick, if comprised of similar sediments that are present above the boundary. In contrast, the sediment sampled just above the boundary feature presents a density consistent with its present-day depth (i.e. approximately 1 m) below the sea floor.

A1-17 Clarke (2014) Appendix 1

5 The down-core changes in unit weight and water content are also consistent with 200 m of burial. The depth reconstructed at the GC12 site by replacing the material apparently missing from the U-shaped trough, i.e. by maintaining the continuity and shape of the adjacent slope and projecting it above the GC12 site, is approximately 250 m.

Figure 3 I) Images of the three gravity cores (GC) analysed in this study that present boundary features: GC8, GC11 and GC12. See Figures 1 and 2 for the locations of the gravity cores. II) Close up of the boundary feature in each core. The inferred slide plane is indicated with a dashed black line. Bulk radiocarbon ages for each core are also shown in yellow (ky = thousand years before present, RCD = radiocarbon dead).

A1-18 Clarke (2014) Appendix 1

6 Figure 4 Core log for GC12 showing physical properties with depth below seabed. Bulk radiocarbon dates are also shown. The presumed slide plane is indicated with a dashed black line at 87 cm depth below seabed. GC8 and GC11 show similar trends.

While there is an obvious change in colour, stiffness, density and water content across the boundary identified in GC8, GC11, and GC12, there is no significant difference in the texture and composition of the sediment between the two units, with grainsize, carbonate and organic carbon content remaining relatively invariant (Fig. 4). Se- diment in both units comprises poorly sorted, medium to coarse sandy silts, with low clay and high silt and sand contents. Carbonate and organic carbon content also remain uniform across the boundary.

3.2 14C Radiocarbon Ages

Bulk radiocarbon ages were used to establish minimum age constraints on the timing of the submarine landslide slide plane material recovered within three gravity cores (GC8, GC11 and GC12). Radiocarbon ages were determined at the Radiocarbon Labor- atory, University of Waikato, New Zealand. Conventional 14C yrs BP were converted into calibrated calendar ages using OxCal V3.10 using marine calibration curve Ma- rine09 (Bronk Ramsey 2005; Marine data from Reimer et al. 2009) following Stuiver and Polach (1977) and are reported here with 2σ errors. Seven radiocarbon ages were determined from bulk sediments collected from the three gravity cores (Table 1). The ages indicate that a significant time gap exists across the boundary within the cores, which confirms our interpretations of two distinct sediment units. Sediment sampled directly above one slide plane (Byron Slide, GC12) dates at 15.8 ka, while sediment directly below the slide plane dates at 47.4 ka or radiocarbon dead (see Fig. 3II). Importantly, radiocarbon ages from sediment just above the inferred slide surfaces in cores GC8 and GC11 returned pene-contemporaneous dates of 20.7 ka and 20.1 ka, respectively (Cudgen and Coolangatta I slides respectively). This material is inferred to be the basal layer of post-landslide sediment. These two slides are adjacent features separated by 13 km; their similar ages are consistent with their being simultaneous events. Although bulk radiocarbon ages measured from marine carbonates can provide basic constraints on a core’s chronology, this sampling method is not ideal for obtaining pre- cise age information. Bulk samples can introduce uncertainties in interpretation because they can be derived from different sources of different age (Harris et al. 1996). Specifi- cally, bulk sediment samples generally contain older re-worked material which results in the calculated age of sedimentation being older than the true age. In addition, because bulk carbonate contains both planktonic and benthic foraminifera, the 400 year reservoir correction cannot be applied to these samples.

A1-19 Clarke (2014) Appendix 1

Table 1 14C dating results of bulk sedimentary organic carbon samples. All samples were taken from above the inferred slide plane boundary except for sample SS2008-12/GC12/1B-88. Conven- Median cali- 2σ calibrated age 14C Depth tional 14C brated range Lab Code Core error* (cm) age age (2σ) (BP) (95.4% probability) (BP) (BP) SS2008- GC8 6 4,157 ±45 4,229.5 4,079.5 – 4,379.5 12/GC8/1C-6/D

SS2008- GC8 85 17,732 ±95 20,699.5 20,249.5 – 21,149.5 12/GC8/2B-85/D

SS2008- GC11 3 3,763 ±44 3,699.5 3,569.5 – 3,829.5 12/GC11/1C-3

SS2008- GC11 206 17,417 ±91 20,149.5 19,849.5 – 20,449.5 12/GC11/3A-206 SS2008- GC12 5 3,207 ±51 3,024.5 2,859.5 – 3,189.5 12/GC12/1B-5

SS2008- GC12 81 13,463 ±77 15,799.5 15,149.5 – 16,449.5 12/GC12/1B-81 SS2008- GC12 88 44,288 ±1205 47,399.5 45,149.5 – 49,649.5 12/GC12/1B-88 *Quoted errors are 1 standard deviation

4. Modeling

Geomechanical modeling of the landslides was undertaken using the slope stability program GEO-SLOPE/W (2007) to examine the influence of several parameters on the stability of the slope profiles (cohesion, friction angle, and slope geometry). The profiles of adjacent unfailed sections of the landslide region were used to indicate the pre- failure geometry of the upper slope (see McAdoo et al., 2000). For each slope stability calculation, a dimensionless parameter, the Factor of Safety (FoS), is determined. It is defined as the ratio of the restoring forces to the disturbing forces (stable slope, FoS>1; unstable slope, FoS<1; critically stable slope, FoS=1). Static modeling of the slopes associated with each of the three landslides indicate that they are all inherently very stable (Fig. 5). This apparently inherent stability is inconsistent with the wide presence of slope failures in the study area. We infer from these data that that the landslides are the result of external triggers, either short term or long term, which reduce the shear strength of the sediments, and/or change the geometry of the slope such that the margin is frequently destabilised. Back analysis modeling further indicates a dramatic reduction of sediment shear strength is required for slope failure to occur. Values well below the measured strength for the materials (c = 0kPa, Ф = 40°) are required. The average range of values used for modeling are summarised in Table 2. Table 3 summaries the resultant FoS values for each scenario. For failure to occur, the friction angle value must drop to values less than 8°, with cohesions of zero.

A1-20 Clarke (2014) Appendix 1

Figure 5 Affect to the Factory of Safety (FoS) values for the slides under static conditions using measured strength values for the materials (c = 0 kPa, Ф = 40°) as input parameters: A) Coolan- gatta1 Slide B) Cudgen Slide C) Byron Slide.

Table 2 Numerical input parameters used for modeling the slides with GEO-SLOPE/W. The friction angle (Ф) represents the friction component of the soil strength and the apparent cohesion (c’) represents the cohesive component of the soil strength. Parameter Unit Input value range

Unit weight (γ) kN/m³ 15 - 17

Apparent cohesion (c’) kPa 0 - 22

Friction angle (Ф) ° 0 - 40

Table 3 Back analysis GeoSlope outputs: a summary of the factors of safety (FoS) for the Byron slide arising from reducing c and Ф is shown in Table 2. Critical FoS are underlined. The Coo- langatta1 and Cudgen slides follow the same trends. Cohesion Friction an- Site Scenario description FoS (lowest) (kPa) gle (°) 40 6.19 30 4.26 Residual cohesion, decreasing 0 15 1.98 friction angle 7.5 0.97 Byron Slide 3.75 0.48 11 8.8 Peak friction angle, decreasing 5.5 7.8 40 cohesion 2.75 7.28 1.375 6.98

A1-21 Clarke (2014) Appendix 1

9 Modeling the response of submarine slopes to seismic shaking using pseudo-static methods, which include a factor for seismic accelerations in classical limit-equilibrium calculations, is useful but of limited value here (Seed and Martin 1966). Selecting an appropriate value for the seismic coefficient acting on the failure mass can be especially difficult (Seed and Martin 1966). A very crude investigation of seismic loading on the slopes indicate that lateral and vertical accelerations of 0.3 g (ah = 0.3 g, av = -0.3 g), the upper limit of those used to investigate the stability of earth dams during earthquakes (Seed and Martin 1966; Ozkan 1998), would be sufficient to destabilise the slopes of the seafloor in the present study. Focusing only on the magnitude of the critical acceleration required to destabilise the slopes does not consider a number of other potential mechanisms that could contribute to the wide occurrence of slides on the SE Australian margin. For example, our preliminary analysis treats the landslide mass as a rigid block on a plane and does not consider the substantial effects of soil liquefaction, which is likely given the saturated loose to medium cohesion of the sediments comprising the landslides (cf., Ozkan 1998).

5. Conclusions

Regional mapping of the SE Australian continental margin reveals several large- volume submarine slides in water depths of 400 to 1200 m. Gravity cores recovered from three of these landslides reveal that each slide contains a distinct, bulk-density contrast defined boundary between 85 and 220 cm below the seabed. We consider that the distinct density contrasts across the boundaries between the upper and lower units of GC8, GC11 and GC12 are highly significant. The inference that these boundaries represent failure surfaces of submarine landslides is reasonable and consistent with all the available data. Also important, two adjacent but separated landslides have penecontemporaneous ages of 20.7 ka and 20.1 ka directly above the slide plane. These ages are consistent with a common external trigger. The experimental results imply that the sediment forming the margin is reasonably strong and inherently stable. Classical limit- equilibrium modeling indicates that submarine landslides should not be a common occurrence on the margin. The wide occurrence of upper slope slides across the SE Australian margin indicates that submarine sliding should be considered to be a common characteristic of this passive continental margin. This indicates that one or more of the potential triggering mechanisms can operate in passive margin settings to destabilise the slope. The processes suspected to be most likely include: 1) dramatic reduction of the shear strength of the upper-slope sediments to extremely low values, possibly induced by creep or a build-up of pore-pressure; 2) long-term modification of the slope-geometry i.e., sedimentation on the head of the slope and/or erosion of the toe of the slope; and/or 3) seismic events large enough to trigger sediment liquefaction or a sudden increase of pore-fluid pressure.

6. References

Bardet JP, Synolakis, CE, Davies, HL, Imamura, F, Okal, EA (2003) Landslide Tsuna- mis: Recent Findings and Research Directions. Pure App Geophys 160:1793–1809.

A1-22 Clarke (2014) Appendix 1

10 Blott SJ, Pye K (2001) Gradistat: a grain size distribution and statistics package for the analysis of unconsolidated sediments. Earth Surf Proc Landforms 26:1237-1248. doi:10.1002/esp.261 Boyd R (2009) SS12/2008 Voyage Summary: Marine Geology and Geohazard Survey of the SE Australian Margin off Northern NSW and Southern Queensland. CSIRO, Boyd R, Keene J, Hubble T, Gardner J, Glenn K, Ruming K, Exon N (2010) Southeast Australia: A Cenozoic Continental Margin Dominated by Mass Transport. In: Mosher DC, Moscardelli L, et al. (eds) Submarine Mass Movements and Their Con- sequences. vol 28. Advances in Natural and Technological Hazards Research. Springer Netherlands, pp 1-8. Bronk Ramsey C (2005) Improving the resolution of radiocarbon dating by statistical analysis. In: Levy TE, Higham, T. (ed) The Bible and Radiocarbon Dating: Archaeo- logy, Text and Science. Equinox, London, pp 57-64 Glenn K, Post A, Keene J, Boyd R, Fountain L, Potter A, Osuchowski M, Dando N, Party S (2008) NSW Continental Slope Survey – Post Cruise Report vol Geoscience Australia, Record 2008/14, 160pp. Harris PT, O'Brien PE, Sedwick P, Truswell EM (1996) Late Quaternary history of sedimentation on the Mac. Robertson Shelf, East Antarctica: problems with 14C-dating of marine sediment cores. Papers and Proceedings, Royal Society of Tasmania 130:47-53 Heiri O, Lotter AF, Lemcke G (2001) Loss on ignition as a method for estimating organic and carbonate content in sediments: reproducibility and comparability of results. J Paleolimnology 25 (1):101-110 Hubble T, Yu P, Airey D, Clarke S, et al. (2011). Physical properties and age of continental slope sediments dredged from the Eastern Australian Continental Margin - implications for timing of slope failure. In: Submarine Mass Movements and Their Consequences (this issue). Advances in Natural and Technological Hazards Re- search. Springer Netherlands. Jenkins CJ, Keene JB (1992) Submarine slope failures on the southeast Australian continental slope. Deep Sea Research 39:121-136 Locat J, Lee, H.J. (2002) Submarine landslides: advances and challenges. Canadian Geotech J 39:193-212 McAdoo BG, Pratson LF, Orange DL (2000) Submarine landslide geomorphology, US continental slope. Mari Geol 169:103-136. doi:Doi: 10.1016/s0025-3227(00)00050-5 Ozkan MY (1998) A review of considerations on seismic safety of embankments and earth and rock-ﬁll dams. Soil Dynamics and Earthquake Engineering 17:439-458 Reimer PJ, Baillie, M. G. L., Bard, E., Bayliss, A., Beck, J. W., ..., Weyhenmeyer, C. E. (2009) Intcal09 and Marine09 radiocarbon age calibration curves, 0-50,000years cal BP. Radiocarbon 51 (4):1111-1150 Seed HB, Martin GR (1966) The seismic coefficient in earth dam design. Journal of Geotechnical and Engineering ASCE 92:25-58 Stuiver MP, Henry A (1977) Discussion: Reporting of 14C Data. Radiocarbon 19:355- 363 Varnes DJ (1978) Slope Movements and Types and Processes. Landslides: Analysis and Control, Special Report, vol 176. Transportation Research Board, National Academy of Sciences, Washington.

Acknowledgments

We would like to acknowledge the P&O crew and scientific crews of the RV Southern Surveyor voyage (12/2008). Funding for this voyage was provided by ARC Australia and ConocoPhillips Pty Ltd. Funding for the radiocarbon age determinations was provided by Professor Ron Boyd and Newcastle University, Austral- ia. This paper benefitted from reviews by Dr Andrew D. Heap and Dr Julie Dickinson.

A1-23 Clarke (2014) Appendix 1 3 Chapter 47 Morphology of Australia’s Eastern Continental Slope and Related Tsunami Hazard

Samantha Clarke, Thomas Hubble, David Airey, Phyllis Yu, Ron Boyd, John Keene, Neville Exon, James Gardner, Steven Ward, and Shipboard Party SS12/2008

Abstract Morphologic characterisation of five distinct, eastern Australian upper continental slope submarine landslides enabled modelling of their tsunami hazard. Flow depth, run-up and inundation distance has been calculated for each of the five landslides. Future submarine landslides with similar characteristics to these could generate tsunami with maximum flow depths ranging 5–10 m at the coastline, maximum run-up of 5 m and maximum inundation distances of 1 km.

Keywords Submarine landslides • Wave height • Southeastern Australia • Upper slope • Flow depth • Run-up • Inundation distance

S. Clarke ( ) • T. Hubble • D. Airey • P. Yu • J. Keene Geocoastal Research Group, University of Sydney, Sydney, NSW, Australia e-mail: [email protected] R. Boyd School of Environmental and Life Sciences, University of Newcastle, Newcastle, NSW, Australia ConocoPhillips, Houston, TX, USA N. Exon Earth and Marine Sciences, Australian National University, Canberra, ACT, Australia J. Gardner CCOM, University of New Hampshire, Durham, NH, USA S. Ward Earth and Marine Science, University of California at Santa Crux, Cal. United States Title of Team: Shipboard Party SS12/2008

S. Krastel et al. (eds.), Submarine Mass Movements and Their Consequences, Advances 529 in Natural and Technological Hazards Research 37, DOI 10.1007/978-3-319-00972-8 47, © Springer International Publishing Switzerland 2014

A1-24 Clarke (2014) Appendix 1

530 S. Clarke et al.

47.1 Introduction and Aims

Submarine landslides can damage seabed infrastructure, cause subsidence of coastal land, and generate tsunamis (Masson et al. 2006). Examples of large submarine landslide generated tsunamis include the 1929 Grand Banks’ event (Fine et al. 2005), the 1946 Scotch Cap Alaska event (Fryer et al. 2004), and the 1998 Aitape Papua New Guinea event (Tappin et al. 2001). Submarine landslide generated tsunami are not as well understood as those associated with large earthquakes and consequently present a significant but poorly-quantified hazard (Sue et al. 2011). The eastern Australia (EA) coast is potentially vulnerable to tsunamis due to the population concentration ( 85 %) and critical infrastructure within 50 km of the coast (Short and Woodroffe 2009). However, there has been little reason to suspect a local source for the generation of tsunami on the EA coastline. The identification of relatively recent, abundant submarine landslide scars has changed this perception (Boyd et al. 2010; Clarke et al. 2012; Keene et al. 2008) and established that submarine landsliding should be considered a common and ongoing characteristic of this passive continental margin (Clarke et al. 2012; Hubble et al. 2012). This study uses data collected during the RV Southern Surveyor (SS12/2008) survey of the EA continental margin (Boyd et al. 2010) to (a) characterise slope morphology of the margin between Noosa Heads and Yamba, (b) determine the geometry of selected landslide features, and (c) quantify the size of tsunami potentially generated by these lands. The study area (Fig. 47.1) is located along the EA continental slope between Noosa Heads to Yamba. We focus on five distinct landslide scars from the upper continental slope identified by Boyd et al. (2010) and one adjacent potential submarine landslide site. They are the: (1) Bribie Bowl Slide; (2) Coolangatta2 Slide; (3) Coolangatta1 Slide; (4) Cudgen Slide; (5) Byron Slide; and (6) Potential

Fig. 47.1 Digital elevation model (DEM) of the slope geometry of the EA continental margin. Insets show details of actual (black lines) and potential landslides (red outline)

A1-25 Clarke (2014) Appendix 1

47 Morphology of Australia’s Eastern Continental Slope. . . 531

Table 47.1 Summary of the submarine landslide parameters

Canyon region Plateau region Landslide Average slope angle >3° Average slope angle <3° name Bribie Bowl Bryon Potential Slide (adjacent to Cudgen Coolangatta1 Coolangatta2 Slide Slide Bryon Slide) Slide Slide Slide Parameter (m) Thickness (t) 125 220 200 50 20 45 Length (L) 3,000 3,700 3,800 7,500 8,300 1,400

Width (W) 2,465 3,558 3,268 5,338 2,286 1,558 Water depth 600 1,000 800 600 600 900 (h0) See Fig. 47.1 for landslide locations

Slide mass (c.f. Table 47.1). The ﬁve scars are representative of the failures that occur in the two contrasting slope morphologies in the study area: (1) the relatively steep (3–7ı) and canyon incised slope (Bribie Bowl Slide, Bryon Slide and Potential Slide) and (2) the relatively gentle sloping (1–3ı) Nerang plateau (Coolangatta1&2 Slides and Cudgen Slides). At least one gravity core was recovered from each of these landslides.

47.2 Data and Methods

47.2.1 Bathymetry and Landslide Geometry

Approximately 13,000 km2 of bathymetric data was acquired using a 30-kHz Kongsberg EM300 multibeam echosounder. The multibeam data was processed to produce a 50 m gridded digital elevation model (DEM) (Boyd et al. 2010). The DEM was used to examine the six individual sites using Fledermaus V7.3.3b software (http://www.qps.nl/). Landslide thickness (t), length (L), width (W), and water depth at landslide centre of mass (h0), as well as distance from the adjacent coastline to head of the landslide source (r) were determined for each feature (Table 47.1). Landslide thickness is the maximum thickness within the landslide scar assuming the surface is continuous without the apparent landslide feature (McAdoo et al. 2000). Landslide length is the distance from landslide head to landslide toe. Landslide width is the average of measurements taken every 500 m down the landslide scar, perpendicular to the landslide axis. Water depth is taken from the landslide centre of mass.

47.2.2 Tsunami Calculations

The size of tsunami generated by potential of slope failures identiﬁed here has been assessed using empirical equations developed by Ward et al. (Chesley and

A1-26 Clarke (2014) Appendix 1

532 S. Clarke et al.

Ward 2006; Ward 2001; Ward 2011). Equations 47.1 and 47.2 use a landslide’s geometric characteristics and an estimate of landslide mass velocity to generate maximum flow depth at the coastline (Fd(0)) (i.e. tsunami height). Function 3 allows run-up (R(Xmax)) and inundation distance (Xmax) to be estimated where left and right sides are equal only at a particular Xmax values, determined from site-specific coastline profiles perpendicular to the coastline and adjacent to each landslide site. Flow depth at coastline:

4=5 1=5 F d .0/ ŒA0P .r/� h0 (47.1) D Propagation and beaching factor:

0:36 2 0:69 W Lref vs vt Lref 3:74 v A0P .r/ 0:7345t e� t (47.2) D Lref h h i r � �� 0 � � � Run-up and Inundation depth:

16:7n2 ŒR.X max/ T .X 0/� .X max X 0/ F d .0/ (47.3) � C F d .0/ � D

Where A0 is the initially generated surface elevation, P(r) is propagation factor P at distance r from the source of the wave vs is landslide speed, vt is tsunami speed at the landslide p(gh0), Lref is a reference length, taken here as 1 m, X is distance D inland from the coastline (X0 0 at coastline), T(X) is topographic elevation at location X. Land surface roughnessD is represented by Manning’s coefﬁcient n which is taken as 0.015 for very smooth topography, 0.03 urbanized/built land, and 0.07 densely forested landscape (Gerardi et al. 2008).

47.3 Results

47.3.1 Morphometric Characteristics of Individual Landslides

The five landslides (Fig. 47.2) are U-shaped in cross-section (3–6 km wide and 20– 220 m deep) backed by an amphitheatre shaped crestal zone. In each case, landslide morphology is similar to the classical circular failure profile described by Varnes (1978), but elongated in the downslope, longitudinal profile. The Bribie Bowl, Byron and Potential Slides are located within the steeper canyon regions (average slope 3–7ı) and are thicker (>100 m) compared with landslides developed on the adjacent Nerang plateau. Both canyon landslides present an average slope of approximately 12ı along the majority of the failure plane, increasing to 33ı at the head scarp. In plan view the crown scarps present

A1-27 Clarke (2014) Appendix 1

47 Morphology of Australia’s Eastern Continental Slope. . . 533

Fig. 47.2 Detailed DEM for each landslide showing slope geometry and cross-section proﬁles across slide (black line S-N) and down slide (black line W-E). Landslide outlines (black a-e and red f) and cross-sections for the Potential Slide (dark blue) are shown, along with tension cracks at the landslide head and slope break at the toe (dashed red). Reference proﬁles marked in light blue

distinctive semicircular shape, but the detached landslide slabs are essentially planar blocks, as the exposed failure surface is planar with a declination roughly parallel to the adjacent unfailed slope. The Coolangatta1&2 and Cudgen Slides are located on the shallower Nerang plateau region (average slope 2–3ı). These landslides are thinner (<50 m) and representative of the numerous upper slope failures that occur on this very gently dipping plateau (Boyd et al. 2010). All three landslides present a “hummocky” texture within the failure region, a gently concave landslide shape, and average slopes of approximately 3.5ı within the failure plane, up to 7.5ı at the head scarp. The potential landslide mass identified adjacent to the Byron Slide protrudes anomalously out from the shelf in the heavily incised southern canyon section (Fig. 47.1). The extensive mass wasting surrounding this block and apparent tension crack features at the head of the identified mass (Fig. 47.2f) suggests it’s possible future failure. The tension crack represents the landslide head, while a break in slope that presents as extant circular failures cresting at the 1,500 m contour defines the expected toe.

A1-28 Clarke (2014) Appendix 1

534 S. Clarke et al.

Table 47.2 Summary of the maximum ﬂow depths

Maximum ﬂow depth at coastline Fd(0) (m)

Average slope angle >3ı Average slope angle <3ı Maximum landslide Bribie Byron Potential Cudgen Coolangatta1 Coolangatta2 1 velocity (vs) (ms� ) Bowl Slide Slide Slide Slide Slide Slide 10 2:7 6:2 6:2 3:0 0.6 0.7 20 5:0 10:3 10:9 5:6 1.2 1.2 30 8:5 16:3 17:8 9:5 2.0 1.8 40 13:0 24:0 26:7 14:5 3.0 2.8

Failures are assumed to occur as one complete landslide block, rather than as a number of multiple landslides from the same site. This may generate an overestimated volume of the landslide block.

47.3.2 Calculated Characteristics of Landslide-Generated Tsunami

Table 47.2 summarises maximum flow depth at the coastline for the five landslides and the potential landslide using a range of maximum landslide velocities (vs). The dimensions of the landslide mass, initial acceleration and maximum velocity of the sliding mass are important when assessing expected tsunami size (Ward 2001). The range of velocities tested has been constrained by minimum and maximum velocity values reported in the literature (Masson et al. 2006). The velocity at which submarine landslides travel after failure is not well defined due to a lack of direct measurements. The 1929 Grand Banks Slide was measured 1 travelling at 25 ms� (Fine et al. 2005), while the 2006 SW Taiwan event measured 1 turbidity current velocities between 17 and 20 ms� (Hsu 2008). Velocities are based on cable breakages during failure and are measured from the gentle upper slope, approximately 2ı and <0.5ı respectively. At the higher end, speeds of 1 up to 80 ms� have been inferred for some large landslides based on landslide debris travel distances (Masson et al. 2006). We have tested a range of possible 1 landslide velocities; however we consider 20 ms� a reasonable and conservative (i.e. minimum) value (Driscoll et al. 2000). The calculations demonstrate that submarine failures along the EA continental margin have the potential to generate tsunami with flow depths at the coast ranging 1 from 1.2 to 10.9 m for landslide velocities of 20 ms� . In particular, the Bribie Bowl, Byron, Cudgen and the Potential Slides all generate flow depths greater than 5 m at 1 the coastline for landslide velocities of 20 ms� (Table 47.2). Flow depth at the coastline directly relates to landslide thickness at a particular location. The thinner landslides considered, Coolangatta1&2 Slides (thickness 20 m and 45 m respectfully), produce smaller tsunami with inundation depths of 1 m. The

A1-29 Clarke (2014) Appendix 1

47 Morphology of Australia’s Eastern Continental Slope. . . 535

Table 47.3 Summary of the maximum inundation distance (Xmax) and run-up (R(Xmax)) Maximum Inundation landslide distance (Xmax) Average slope angle >3ı Average slope angle <3ı velocity (vs) (m) & Run-up Bribie Byron Potential Cudgen Coolangatta1 Coolangatta2 1 (ms� ) (R(Xmax)) (m) Bowl Slide Slide Slide Slide Slide Slide

10 Xmax 26 500 500 24 27 31 R(Xmax) 2.4 2.08 2.08 2:75 0:15 0:15 20 Xmax 398 898 901 50 54 53 R(Xmax) 1.5 4.1 4.8 5:2 0:4 0:4 30 Xmax 880 1,720 1,978 100 130 120 R(Xmax) 1.95 5.95 6.3 8:8 0:42 0:37 40 Xmax 1,715 3,252 3,805 732 220 220 R(Xmax) 1.9 6.9 7.36 10 0:73 0:39

Cudgen Slide (50 m thick) and Bribie Bowl Slide (125 m thick) generated 5–6 m inundations, while the Byron Slide (220 m thick) and Potential Slide (estimated 200 m thick) generated inundation depths of 10 m. Maximum flow depth at the coastline is larger for the thicker canyon landslides (e.g. Byron Slide and Potential Slide 10 m; Bribie Bowl Slide 5 m) which occur on steeper slopes in comparison to shallow plateau landslides which generally produce waves less than 1 m in height, except where landslide surface area was particularly large (e.g. Cudgen Slide: surface area 50 km2, flow depth 5 m). Table 47.3 summarizes maximum expected inundation distance and run-up for the identified landslides over a range of landslide velocities (vs). Local topography greatly affect the ability of a wave to inundate past the immediate coastline (Gerardi et al. 2008) and these values are estimates which are less reliable than coastal inundation depth. The results show that submarine landslides along the EA continental margin have the potential to generate inundation and run-up distances up to 1 km and 5 m 1 1 respectively for landslide velocities of 20 ms . Using 20 ms landslide velocity as a benchmark, the two shallow Coolangatta1&2 Slides generate maximum inundation distances around 50 m and run-up about <0.5 m. The Cudgen Slide generates inundation distances around 50 m, however with much greater run-up at around 5 m. The Bribie Bowl Slide produced inundation distances around 400 m and run-up about 1.5 m, while the Byron and Potential Slides from the canyon regions generates maximum inundation distances around 1 km and run-up about 4 m.

47.4 Discussion

This study demonstrates that blocks shed from submarine scars have the potential to generate signiﬁcant tsunami on the EA coast with ﬂow depths of at least 10 m in height, run-up of 5 m and maximum inundation distances of 1 km. These estimates

A1-30 Clarke (2014) Appendix 1

536 S. Clarke et al.

are based on relatively conservative estimates of landslide velocity given (a) the relatively steeply inclined slopes down which sliding material has moved and (b) the lack of evidence for a depositional site where the landslide mass ceased moving in mapped areas downslope of the landslide, indicating that the landslide material has travelled at least 15–20 km from its site of origin. Evidence for historical tsunamis that have inundated the EA coastline is limited. A recently published tsunami catalogue (Dominey-Howes 2007) identifies historical events on the EA coast, of which 16 ( 30 %) have no identified cause (run- ups between <0.1 to 6 m asl). Nevertheless, our results show that upper slope landslides similar to those investigated in this study are a plausible source for the tsunami documented in catalogue (Dominey-Howes 2007). A growing body of evidence (Hubble 2013) is strengthening the contention that shedding landslides of the size investigated in this study should be considered to be a common, ongoing characteristic such that future failures are very likely to occur. Constraining estimates of landslide velocity is critical as coastal flow depth, inundation distance and run-up all increase with landslide velocity. We have used conservative estimates of landslide velocity and it is quite possible that these landslides can generate significantly more destructive events than we have suggested. Evaluating possible triggering conditions for submarine sliding is also critical, as these conditions are not well understood (Masson et al. 2006). However, the majority of documented twentieth century submarine landslide tsunami events are related to moderate earthquakes (generally >M7). Such events are rare on the stable Australian continent and suggested recurrence intervals for such events are at multi- millennial timescales (Clark 2010).

47.5 Conclusions

The morphometric characterization of five distinct, geologically young, submarine landslides on the eastern Australian upper continental slope has enabled an estimation of tsunami size their related landslide masses probably generated and an insight into tsunami hazard that might be expected on the eastern Australian coast. Flow depth (1.2–10.9 m), run-up (0.4–4.8 m) and inundation distance (50–901 m) 1 were calculated for five landslide sites assuming landslide velocities of 20 ms . The reoccurrence of submarine landslides with similar characteristics to those shed from the margin in the geologically recent past would therefore be expected to generate tsunami with maximum flow depths between 5 and 10 m at the coastline, run-up of up to 5 m and inundation distances of up to 1 km. In particular, a potential landslide mass adjacent to the Byron Slide has been identified. If it was to fail it could generate maximum flow depths of ca. 10 m at the coastline, with inundation distances of ca. 1 1 km, for a conservative landslide velocity of 20 ms . If these assumptions are correct, a tsunami this size would cause significant damage and possibly loss of life.

A1-31 Clarke (2014) Appendix 1

47 Morphology of Australia’s Eastern Continental Slope. . . 537

Acknowledgments We would like to acknowledge the P&O crew and scientiﬁc crews of the RV Southern Surveyor voyage (12/2008). Funding for this voyage was provided by ARC Australia and ConocoPhillips Pty Ltd. This paper beneﬁtted from reviews by Dr Geoffroy Lamarche and Dr Julie Dickinson.

References

Boyd R, Keene J, Hubble T, Gardner J, Glenn K, Ruming K, Exon N (2010) Southeast Australia: a Cenozoic continental margin dominated by mass transport. In: Mosher DC et al (eds) Submarine mass movements and their consequences, vol 28, Advances in natural and technological hazards research. Springer, Dordrecht, pp 491–502 Chesley SR, Ward SN (2006) A quantitative assessment of the human and economic hazard from impact-generated tsunami. Nat Hazards 38(3):355–374 Clark DJ (2010) Large earthquake recurrence in New South Wales: implications for earthquake hazard. Earthquake hazard project. Geoscience Australia, Canberra Clarke S, Hubble T, Airey D, Yu P, Boyd R, Keene J, Exon N, Gardner J (2012) Submarine landslides on the upper southeast Australian passive continental margin – preliminary ﬁndings. In: Yamada Y, Kawamura K, Ikehara K et al (eds) Submarine mass movements and their consequences, vol 31, Advances in natural and technological hazards research. Springer, Dordrecht, pp 55–66 Dominey-Howes D (2007) Geological and historical records of tsunami in Australia. Mar Geol 239:99–123 Driscoll NW, Weissel JK, Goff JA (2000) Potential for large-scale submarine slope failure and tsunami generation along the U.S. mid-Atlantic coast. Geology 28(5):407–410 Fine IV, Rabinovich AB, Bornhold BD, Thomson RE, Kulikov EA (2005) The Grand Banks landslide-generated tsunami of November 18, 1929: preliminary analysis and numerical modeling. Mar Geol 215:45–57 Fryer GJ, Watts P, Pratson LF (2004) Source of the great tsunami of 1 April 1946: a landslide in the upper Aleutian forearc. Mar Geol 203(3–4):201–218 Gerardi F, Barbano MS, De Martini PM, Pantosti D (2008) Discrimination of tsunami sources (earthquake versus landslide) on the basis of historical data in eastern Sicily and southern Calabria. Bull Seismol Soc Am 98(6):2795–2805 Hsu S-K (2008) Turbidity currents, submarine landslides and the 2006 Pingtung earthquake off SW Taiwan. TAO: Terr Atmos Ocean Sci 19(6):767 Hubble T (2013) Voyage summary SS2013-V01: marine geology and geohazard survey of the SE Australian Margin off Northern NSW and Southern Queensland. CSIRO, Hobart Hubble T, Yu P, Airey D, Clarke S, Boyd R, Keene J, Exon N, Gardner J (2012) Physical properties and age of continental slope sediments dredged from the eastern Australian continental margin – implications for timing of slope failure. In: Yamada Y et al (eds) Submarine mass movements and their consequences, vol 31, Advances in natural and technological hazards research. Springer, Dordrecht, pp 43–54 Keene J, Baker C, Tran M, Potter A (2008) Geomorphology and sedimentology of the East Marine Region of Australia. Record 2008/10. Geoscience Australia, Canberra, 262 pp Masson DG, Harbitz CB, Wynn RB, Pedersen G, Lovholt F (2006) Submarine landslides: processes, triggers and hazard prediction. Philos Trans R Soc A 364:2009–2039 McAdoo BG, Pratson LF, Orange DL (2000) Submarine landslide geomorphology, US continental slope. Mar Geol 169(1–2):103–136 Short AD, Woodroffe CD (2009) The coast of Australia. Cambridge University Press, New York Sue LP, Nokes RI, Davidson MJ (2011) Tsunami generation by submarine landslides: comparison of physical and numerical models. Environ Fluid Mech 11(2):133–165

A1-32 Clarke (2014) Appendix 1

538 S. Clarke et al.

Tappin DR, Watts P, McMurtry GM, Lafoy Y, Matsumoto T (2001) The Sissano, Papua New Guinea tsunami of July 1998 – offshore evidence on the source mechanism. Mar Geol 175(1–4):1–23 Varnes DJ (1978) Slope movements and types and processes. Landslides: analysis and control, special report, vol 176. Transportation Research Board, National Academy of Sciences, Washington Ward SN (2001) Landslide tsunami. J Geophys Res 106:11201–11215 Ward SN (2011) Tsunami. In: Gupta HK (ed) Encyclopedia of earth sciences series. Springer, Dordrecht, pp 1473–1493

A1-33 END OF

DOCUMENT