Geological Control of Physiography in Southeast Queensland

Geological control of physiography in southeast

Queensland: a multi-scale analysis using GIS

Jane Helen Hodgkinson Bachelor of Science (Hons), Geology (Birkbeck University of London, UK)

School of Natural Resource Sciences

A thesis submitted for the degree of Doctor of Philosophy Queensland University of Technology 2009

STATEMENT OF ORIGINAL AUTHORSHIP

The work contained in this thesis has not been previously submitted for a degree or diploma at any other higher education institution. To the best of my knowledge and belief, the thesis contains no material previously published or written by another person except where due reference is made.

Signed…………………………………

Jane Helen Hodgkinson

Date……………………………

iii

ABSTRACT

The study reported here, constitutes a full review of the major geological events that have influenced the morphological development of the southeast Queensland region. Most importantly, it provides evidence that the region’s physiography continues to be geologically ‘active’ and although earthquakes are presently few and of low magnitude, many past events and tectonic regimes continue to be strongly influential over drainage, morphology and topography. Southeast Queensland is typified by highland terrain of metasedimentary and igneous rocks that are parallel and close to younger, lowland coastal terrain. The region is currently situated in a passive margin tectonic setting that is now under compressive stress, although in the past, the region was subject to alternating extensional and compressive regimes. As part of the investigation, the effects of many past geological events upon landscape morphology have been assessed at multiple scales using features such as the location and orientation of drainage channels, topography, faults, fractures, scarps, cleavage, volcanic centres and deposits, and recent earthquake activity. A number of hypotheses for local geological evolution are proposed and discussed. This study has also utilised a geographic information system (GIS) approach that successfully amalgamates the various types and scales of datasets used. A new method of stream ordination has been developed and is used to compare the orientation of channels of similar orders with rock fabric, in a topologically controlled approach that other ordering systems are unable to achieve. Stream pattern analysis has been performed and the results provide evidence that many drainage systems in southeast Queensland are controlled by known geological structures and by past geological events. The results conclude that drainage at a fine scale is controlled by cleavage, joints and faults, and at a broader scale, large river valleys, such as those of the Brisbane River and North Pine River, closely follow the location of faults. These rivers appear to have become entrenched by differential weathering along these planes of weakness. Significantly, stream pattern analysis has also identified some ‘anomalous’ drainage that suggests the orientations of these watercourses are geologically controlled, but by unknown causes. To the north of Brisbane, a ‘coastal drainage divide’ has been recognized and is described here. The divide crosses several lithological units of different age, continues parallel to the

coast and prevents drainage from the highlands flowing directly to the coast for its entire length. Diversion of low order streams away from the divide may be evidence that a more recent process may be the driving force. Although there is no conclusive evidence for this at present, it is postulated that the divide may have been generated by uplift or doming associated with mid-Cenozoic volcanism or a blind thrust at depth. Also north of Brisbane, on the D’Aguilar Range, an elevated valley (the ‘Kilcoy Gap’) has been identified that may have once drained towards the coast and now displays reversed drainage that may have resulted from uplift along the coastal drainage divide and of the D’Aguilar blocks. An assessment of the distribution and intensity of recent earthquakes in the region indicates that activity may be associated with ancient faults. However, recent movement on these faults during these events would have been unlikely, given that earthquakes in the region are characteristically of low magnitude. There is, however, evidence that compressive stress is building and being released periodically and ancient faults may be a likely place for this stress to be released. The relationship between ancient fault systems and the Tweed Shield Volcano has also been discussed and it is suggested here that the volcanic activity was associated with renewed faulting on the Great Moreton Fault System during the Cenozoic. The geomorphology and drainage patterns of southeast Queensland have been compared with expected morphological characteristics found at passive and other tectonic settings, both in Australia and globally. Of note are the comparisons with the East Brazilian Highlands, the Gulf of Mexico and the Blue Ridge Escarpment, for example. In conclusion, the results of the study clearly show that, although the region is described as a passive margin, its complex, past geological history and present compressive stress regime provide a more intricate and varied landscape than would be expected along typical passive continental margins.

The literature review provides background to the subject and discusses previous work and methods, whilst the findings are presented in three peer-reviewed, published papers. The methods, hypotheses, suggestions and evidence are discussed at length in the final chapter.

Keywords

Geomorphology, GIS, Drainage patterns, Stream-ordering, southeast Queensland, Passive margin, Earthquake distribution

vii

LIST OF PUBLICATIONS

REFEREED INTERNATIONAL JOURNAL PAPERS

PAPER 1 Title: The influence of geological fabric and scale on drainage pattern analysis in a catchment of metamorphic terrain: Laceys Creek, southeast Queensland, Australia Authors: Jane Helen Hodgkinson, Stephen McLoughlin, Malcolm Cox Status: Published November 2006 (available on-line from July 2006) Journal: Geomorphology, 81 394-407

PAPER 3 Title: Drainage patterns in southeast Queensland: the key to concealed geological structures? Authors: Jane Helen Hodgkinson, Stephen McLoughlin, Malcolm Cox Status: Published December 2007 Journal: Australian Journal of Earth Sciences, 54 1137-1150

REFEREED CONFERENCE PAPER

PAPER 2 Title: The correlation between physiography and neotectonism in southeast Queensland Authors: Jane Helen Hodgkinson. Stephen McLoughlin, Malcolm Cox Status: Reviewed for DEST purposes, published in conference proceedings, presented with poster (Appendix 1) Conference: Australian Earthquake Engineering Society Conference, Canberra, ACT November 2006

viii

ACKNOWLEDGEMENTS I wish to acknowledge the support and encouragement provided by my principle supervisor, Dr Stephen McLoughlin who gave invaluable advice and comments, who gave his time generously and whose ability and standards I will always strive to emulate. My associate supervisor, Associate Professor Malcolm Cox also provided constructive support, held many useful discussions, and contributed his valuable time for the review of manuscripts, for which I am very grateful.

I am indebted to Dr Andrew Hammond for his help, support and valuable advice, and for our countless beneficial talks and debates. I am grateful to Dr Micaela Preda for sharing her wealth of GIS knowledge with me and for our constructive discussions with regard to local geology and geomorphological analysis. I also acknowledge the helpful and educational discussions and field trips with Mr Bill Ward whose interest in the subject of geomorphology is valuable and encouraging. Further, I wish to express my gratitude to Mr Nate Peterson who provided valuable assistance with GIS methodology.

This research project would not have been possible without the datasets provided by various sources. Special thanks go to Dr Dion Weatherly and Mr Col Lynam at the Earth Systems Science Computational Centre (ESSCC), University of Queensland, for their enormous encouragement with my project, for their collaborative discussions and for providing their valuable earthquake database. Thanks also goes to Geoscience Australia for further earthquake data, and to Pine Rivers Shire Council and the Geological Survey of Queensland at the Department of Mines and Energy (Queensland Government) for providing geological, topographical and drainage data, which were required for GIS analysis. Datasets were also obtained from USGS/NASA via their on-line service, without which, the first paper could not have been written.

I would like to thank all the staff and students in the School of Natural Resource Sciences at QUT, whose help and encouragement have been a great benefit to me in the course of this study. I also thank the staff at the School of Earth Sciences, Birkbeck University of London, among other things, for inspiring me in the

wonderful subject of geology. I also thank all my friends and family both in Australia and the UK, who believed I could do this. I especially thank Jean and David Hodgkinson, Madonna O’Brien, Ben Henderson, Jennifer Baker, Caroline Cole, Jeanette Fleming, John Hodge and everyone from the ‘Whitton School Class of 1982’, whose humour and friendship has been invaluable. A particularly special thank you goes to Jonathan Hodgkinson, my husband, friend, field assistant, fellow student and room-mate at QUT, whose help and encouragement is beyond measure. I also express my gratitude to my late parents to whom I dedicate this work. One of the most important things they taught me was the value of enquiry, without which, science would go nowhere.

TABLE OF CONTENTS

STATEMENT OF ORIGINAL AUTHORSHIP iii ABSTRACT v Keywords vii LIST OF PUBLICATIONS viii Refereed International Journal Papers viii Refereed Conference Paper viii ACKNOWLEDGEMENTS ix INTRODUCTION 1 Setting 4 Scope of study 5 Methods and summary of results 6 LITERATURE REVIEW 10 INTRODUCTION 12 INTRODUCTION TO GEOMORPHOLOGY 14 Surface processes: Weathering and erosion 14 Climate 20 Palaeoclimate and present-day climate in southeast Queensland 22 Regolith and ground cover 23 Effects of sea-level change 24 Anthropogenic influence 26 The role of lithology and rock fabric in geomorphology 30 The role of tectonics and major geological events in geomorphology 38 Neotectonism in Australia 48 Post-Mesozoic tectonism in southeast Queensland 54 GEOMORPHOLOGICAL ANALYSIS 59 Drainage patterns 59 Palaeosurfaces, palaeodrainage and current drainage patterns in southeast Queensland 62 Stream ordering 63 Data analysis 65 Erosion analysis 66 SOUTHEAST QUEENSLAND 70 Introduction to the study area 70 Reasons for selecting the study region 70 Geological history 74 Faulting 85 Sea level influences 86 Terraces 86 Incision 87 Geology of Pine Rivers and Laceys Creek: a fine-scale case study 93 Previous geomorphological studies of southeast Queensland 95 ANALYSIS METHODS USED IN THIS STUDY 102 Digital elevation models (DEMs) 102 Geological data 103 Earthquake data 103 Geographic Information Systems (GIS) and choice of GIS products 103 Remote sensing 106 Spatial analysis 106 Methods of channel analysis 107 Stream ordering 108 SUMMARY 111 References 114 PAPER 1 135

Introduction to Paper 2 163 PAPER 2 165 PAPER 3 183 CONCLUSIONS 215 Summary of results and major findings 216 Hypothesis 1 216 Hypothesis 2 217 Hypothesis 3 218 Additional findings 219 Implications for future research 222 Other observations and general discussion 223 Implications for evaluating the evolution of the landscape 230 Evidence of past geological events in the present landscape 230 National and international significance 234 Evolutionary model 238 Précis of main findings 238 Future work 239 References 240 APPENDICES 243 Appendix 1 - AEES2006 Poster 245 Appendix 2 - A GIS and map-analysis deficiency 247 Appendix 3 - Other software products used 249 Appendix 4 - Statistical analysis of planar features in Laceys Creek 251

xii

INTRODUCTION

Geological processes are a primary control upon landscape physiography, as internal modifications of the Earth often lead to distortion and contortion of the surface. These internal modifications of Earth’s crust are also responsible for the location of hydrocarbon reservoirs, aquifers, mineral accumulations and geological hazards such as active faults. Geological controls on physiography range from large scale processes such as plate movements, folding, faulting and jointing to the finer scale, where variations in the mineralogy of rocks, micro-fractures and porosity can lead to differing susceptibility to weathering. The derived physiography is, therefore, a result of both endogenic processes and exogenic modification by agents of weathering and erosion. These interactive processes may also generate geological hazards, such as unstable slopes or the superposition of valleys on faults (of importance for dam location). Various methods of physiographic analysis can be used to interpret the evolution of the landscape, from which we may deduce the underpinning lithological and structural influences. This type of analysis can greatly enhance understanding of sub-surficial geology and is an effective tool for geohazard identification and resource exploration.

The main objective of this study is to improve understanding of the controls on landscape evolution in southeast Queensland, as this may be an important consideration for future land use. An improved understanding of the influences on local physiographic evolution can then be used to interpret landscape development in similar tectonic settings globally. Key to this study was identification of the extent to which geological processes influence the existing landscape morphology of southeast Queensland. This was undertaken by analysis and interpretation of the genesis of various physiographic features including fine-scale streams, major drainage systems, scarps, valleys and plains at varying scales. Geomorphological analysis was undertaken at both regional and catchment scales to assess the influences of different-magnitude geological features on southeast Queensland’s physiography.

Although landscapes develop from the interaction of both internal and external processes, one of these processes typically dominates the evolution of the system. In particular, this research seeks to determine whether geological features (structure and

lithology) sufficiently correlate with physiographic features to be deemed the dominant influence on the evolution of the landscape in southeast Queensland. An individual index is not sufficient to determine the main control of the landscape, so this study evaluates multiple criteria at various scales, to identify whether geological processes are (or were) dominant in the development of southeast Queensland’s landscape. To establish whether more than one major influence exists, the research uses a sequential approach and the results are presented in three peer-reviewed, published papers.

Setting Around 370-220 Ma, the easternmost part of Australia was accreted onto the older cratonic part to the west. Subsequently, the easternmost part has been modified by epicratonic basin development, rifting and putative hotspot volcanism, which has contributed to a vast array of natural resource deposits in the region and has presumably imposed strong influences on the land’s topography. The area for this study is situated on the eastern margin of the Australian continent (approximately 151° 53', 26° 11'S to 153° 31', 28° 30'S), and covers approximately 41,000 km2. Although the topography of the region is generally subdued by world standards, some steep and mountainous areas also exist. Outcrop availability and accessibility is generally moderate to poor and commonly limited to road cuttings, due to deep soil and extensive vegetation cover. Furthermore, the region is currently undergoing extensive land-use change particularly due to rapid urbanisation and development; the area represents one of the fastest population growth centres in Australia (Australian Bureau of Statistics, 2006).

The region’s geomorphology has been generally well studied, albeit on piecemeal basis (e.g. Marks, 1933; Watkins, 1967; Arnett, 1969; 1971; Donchak, 1976; Beckmann and Stevens, 1978; Lucas, 1987; Cuthbertson, 1990; Childs, 1991). In summary, the western, southern and northern margins of the study area are fringed by highlands (mainly plateaux over 300 m. a.s.l.), and a dissected highland area (the D’Aguilar Ranges) occurs central to the region. Foot hills and coastal plains occur across the remainder of the area, principally in the east, and escarpments are also fairly common across the region. Eastern Australian rivers may be considered to be

shorter than others by global standards, although runoff is generally higher and more variable. This is due to the climate variations dictated by the region’s mid-latitude position and the competing influences of the western Pacific tropical monsoon and temperate frontal systems that impose a Mediterranean climate on the southern half of the continent (Finlayson and McMahon, 1988). By Australian standards, rivers in southeast Queensland are moderate in size with moderate to high discharge that varies seasonally. Many streams are ephemeral and flow mainly in response to heavy summer rainfall events. The main drainage systems of the region display strong northwest-southeast and northeast-southwest trends similar to those of some large faults in the region. Similarly, a strong northwest-southeast trend is evident in the distribution and foliation of rock units located in southeast Queensland, this being related to late Palaeozoic – early Mesozoic convergent margin deformation. Earthquake data over the past 130 years records only two earthquakes of >5 magnitude (Richter scale) in the region and more than 50 that were >2 magnitude (ESSCC, 2006). Although the database provides discontinuous evidence of earthquake activity spatially and temporally, the foci of many earthquakes are clearly positioned in shallow clusters and even a casual examination reveals alignment of many epicentres within discrete corridors.

Scope of study Southeast Queensland has been selected as the study area primarily due to its complex geological history and varied physiography. Despite many investigations of the region, detailed information on the genesis of landscape features at a regional and local scale is deficient. As the human population in the region is rapidly increasing concomitant with residential, commercial and industrial development, a better understanding of the driving forces behind geomorphological change is critical for future landscape management. Understanding the evolution of other regions with a similar climatic setting and complex mix of both convergent- and passive-margin geological histories may also benefit from the methods and results presented in this study.

Methods and summary of results Remote sensing is a valuable tool for studying landscapes that may be inaccessible and/or very large, such as the southeast Queensland region. Remote sensing has been used here to analyse the alignment, position and form of physiographic features including mountains and highlands, valleys, drainage channels, scarps and lowlands. Other datasets were also integrated into the analysis including rock types and fabric, earthquake locations and geological structures. The method further lent itself to analysis at multiple scales from regional to sub-catchment settings.

In the following section a review of the literature provides an introduction to some of the principles of geomorphology. This is followed by a review of the methodologies employed for geomorphological analysis, their advantages, problems and applicability to the southeast Queensland region. The geology of southeast Queensland is then summarized and previous studies of the region’s geomorphology are examined. The literature survey concludes with a review of the specific analytical methods employed in this study, the limitations of the data and an outline of the need for a new methodology for stream ordering in drainage network analysis.

The first part of the results (presented as paper one) analysed whether drainage orientation is affected by its underlying rock fabric, including cleavage, joints, fractures and faults. The hypothesis tested in paper one was: “Complex geological fabric of metamorphic rocks of southeast Queensland has control over orientation of streams at the sub-catchment scale”

This was tested in a sub-catchment where the meta-sedimentary rock types retain some bedding features but are complicated by multiple orientations of cleavage, joints and faults. The aim was to assess the orientation of streams within a catchment developed on two juxtaposed metamorphic rock types and identify the extent to which streams show alignment with the underlying rock-fabric. Having introduced a new stream-ordering method for this study, as other methods were not suitable, a positive result for some stream-orders suggested that there is some geological control on drainage architecture at this scale. The results showed that higher order streams in the sub-catchment had a similar orientation and close spatial relationship with

fractures and faults. Analysis of other faults and drainage channels outside of the sub-catchment also revealed a similar correlation which required further exploration.

A second investigation of geological control on the landscape incorporated a spatial investigation of recent shallow earthquakes across the whole of southeast Queensland, to assess whether the streams and valleys that showed a correlation with ancient faults might be influenced by continuing earth movements along these structures. The hypothesis tested in paper two was: “The location of recent earthquakes in southeast Queensland aligns with geomorphological features such as scarps, mountain ranges and valleys”

The aim of paper two was to identify, firstly, whether earthquake epicentres show spatial alignment and, secondly, whether any ancient geological and geomorphological features (including faults, scarps, river valleys, highland lineaments) align with trends of earthquake epicentres. From this information the study discusses whether neotectonics plays an ongoing role in the evolution of the landscape. The results showed an alignment of low magnitude, shallow earthquakes with some strong geomorphological trends such as large river systems and areas of highlands. This suggests that tectonics may be playing an active albeit minor role in present-day landscape modification. Although the results were positive, the earthquake database and earthquake monitoring is fairly sparse and may not represent a thorough picture of neotectonism in the region. Therefore, the results are not sufficiently comprehensive to reflect the full scale of neotectonic influence on the landscape. Further data is clearly needed to clarify the role of this process.

Strong relationships have been proven to exist between ‘non-random’ drainage patterns and structures underpinning the landscape, especially where tectonic features such as uplifted or down-thrown blocks, and faults and folds are of strong amplitude. Combining the geology-drainage relationships similar to those found in paper one, with the scale and relationships in paper two, a third measure of geological control on the landscape was introduced into the research (paper three). This involved drainage pattern analysis at the regional scale. Drainage orientation and patterns and physiography were compared to the distribution of geological

structures, lithostratigraphic packages of various ages, igneous intrusions and earthquake corridors, to assess the degree to which geological features of various ages and origins influence the modern physiography. The hypothesis tested in paper three was: “Drainage networks across southeast Queensland show repeating, aligned and anomalous patterns that are controlled by a range of geological structures of varying age”

The history of southeast Queensland is tectonically complex and although the affects of some ancient tectonic events on the landscape might still be evident, it is equally possible that over the intervening time, other processes may have obscured some of the original influences on the physiography. However, the results of paper one show a strong relationship exists between fine-scale streams and underlying rock structure and fabric that date back to at least the Late Carboniferous, even though these rocks have been influenced by a range of subsequent geological processes (nearby convergent margin pluton intrusion, regional uplift, passive margin development, and local emplacement of volcanic plugs). The results of paper two revealed common alignment between recent shallow earthquakes and large ancient valleys that also correspond to a Permian-Triassic trend of strike-slip faulting and the orientation of major tectono-stratigraphic unit boundaries. The results of paper three show a variety of associations between geological features and physiography that include volcanoes, rock types, faults, block emplacements and tectonic tilting that occurred over various stages of southeast Queensland’s geological history. Integration of the datasets within a Geographic Information System (GIS) provided a suitable platform for this analysis and led to new insights into the extent to which the physiography of the region is controlled by geological structures such as faulting and rock fabric. Additionally, the results suggest that previously unknown geological structures may control some important physiographic features and that some parts of the region may have been subject to neotectonic modifications. The results ultimately lead to the conclusion that there is a strong geological control over existing morphology providing sufficient evidence that surface or exogenic controls are not the dominant factor influencing the present landscape.

“The greatest obstacle to discovering the shape of the earth, the continents, and the oceans was not ignorance but the illusion of knowledge.” Daniel J. Boorstin 1914 - 2004

LITERATURE REVIEW

INTRODUCTION Tectonism must be acknowledged as a primary control on the structure and physiography of land masses. It is responsible for the position of landmasses, and within them, the location of geological units, faults, joints and other planes of weakness. It also involves uplift of broad regions and basin subsidence. Weathering and erosion further modify the landscape and, as they are acting upon a terrain that has been ‘processed’ by a primary force, are described as secondary processes. The use of the terms ‘primary’ and ‘secondary’ describes the order in which the processes occur and not the relative importance of the process. Although these processes may, and often do, occur concurrently, weathering and erosion are exogenic, acting upon the endogenically processed landscape, hence the term ‘secondary process’ is applied. Primary processes can directly generate landscape features, such as hills or mountains and scarps, resulting from folds, uplift and faults. However, the additional action of secondary processes upon planes of weakness, such as faults, joints and folds, or the differential weathering of juxtaposed lithologies of different durability, can also lead to further landscape alteration. Geological structure and tectonics are intrinsically linked to a range of geomorphological features so, in reverse, it is possible to analyse the landscape to establish how it has been influenced by geological features and tectonic events (e.g. Ellis et al., 1999; Burrato et al., 2003; Vannoli et al., 2004; Delcaillau et al., 2006). Primary controls can have an influence at multiple scales: on entire landmasses, terrane emplacement, fault systems, folding and even microstructures. Tectonics may down-throw rocks, positioning them for deep burial and metamorphism; uplift them, exposing them to weathering and erosion; fracture and weaken them for further preferential weathering; move and deform them one or many times causing complex, multiple-scale changes to a rock’s strength, form and character. These changes may produce landscape features of varying magnitudes either as a direct result of the primary process, or a result of secondary processes acting upon the primary landforms. The product of primary controls at multiple-scales across a complex landscape is the primary focus of this research. The aim is to determine the extent to which geological control over the landscape is reflected in its morphology, by performing analyses of integrated multiple-scale geomorphological, geophysical and

geological datasets using a Geographic Information System (GIS). In particular the work will establish whether underpinning structure or lithology correlate with the positions and orientations of drainage channels as opposed to stream placement being dominantly influenced by secondary processes. Scales analysed vary according to the types of features being investigated. At the finer scale, analysis has been performed on terrains developed on low-grade metamorphic rock that was originally turbiditic sediments, but which was later buried, deformed, uplifted and deformed further. On a broader scale, analysis was performed on the regional structure of an area that has been affected by both convergent and passive margin tectonism and which has been fractured, uplifted, in places down-thrown, and is composed of sedimentary, metamorphic and igneous rocks. The study also focuses on a variety of spatial scales from sub-catchment to multi-catchment regions. The drainage pattern analysis seeks to reveal drainage patterns that cannot be explained by presently defined geological structures but which, by the nature of their patterns, suggest a geological control. The research, therefore, may reveal the locations of previously unknown, unmapped or deep-seated geological structures that may have implications for planning infrastructural developments in the region. The literature review explains how primary and secondary processes lead to geomorphological change, and considers the relevance of geological control over geomorphology. The regional features of southeast Queensland and the finer-scale case study area of Laceys Creek are presented, together with a review of previous work in this branch of the geosciences both within this study area and globally. The review also discusses the methods involved in the current study in the context of methodologies applied elsewhere and of limitations of the available data. The review then considers the importance of identifying geological control where landscape alteration is being evaluated in relation to anthropogenic activity.

INTRODUCTION TO GEOMORPHOLOGY

Geomorphology, from the Greek ge, meaning earth and morfé meaning form, is the study of landforms including their evolution and origin. Geomorphological analysis provides an understanding of past surface-shaping processes and events and helps predict future landscape changes. Additionally, geomorphological studies are increasingly becoming a valuable supplementary method for identifying recent and current tectonic processes. It has become a useful tool for a wide variety of applications including land management (e.g. Gupta and Ahmad, 1999; Andreas and Allan, 2007), engineering of roads and dams (e.g. Seppala, 1999; Graf, 2005), location and management of water resources (e.g Thoms and Sheldon, 2002; Ghayoumian et al., 2007), geohazard analysis including evaluation of slope stability (e.g. Dominguez-Cuesta et al., 2007; Schulz, 2007; Kirby et al., 2008), and it has even been used as an indicator of global warming (Goudie, 2006). Geomorphological features such as mountains, hills, slopes, valleys, gullies, plateaux and drainage channels of all scales, typically form from the interaction of both endogenic and exogenic processes. The following section discusses these processes and their inter-relationships.

Surface processes: Weathering and erosion From an exogenic perspective, weathering and erosion (‘surface processes’) act upon uplifted, extruded, exposed or emplaced rock units. Alabyan and Chalov (1998) stated that channel development strongly relies upon water discharge and river slope and emphasised that the greater the stream power, the stronger the branching tendency of a river system. Nevertheless, there are other controls that may be equally important whether from an exogenic or endogenic perspective. Endogenically, for example, lithology and induration controls a unit’s erodibility. Weaknesses and differences in rock strength within a rock unit or between abutting or juxtaposed units, may lead to differential weathering causing physical surficial features to form at varying scales, such as gullies, channels and basins. A plane or line of weakness in a rock may be exploited by surface processes such as fluid flow. Then, as a preferential course of drainage, it may eventually erode to form

a channel. Such channels may become widened or incised, retaining the original orientation, even though it may no longer rely on the original plane of weakness to be the preferred location of flow. Furthermore, the physical feature that originally caused preferential flow may no longer exist although the orientation and drainage pattern that it initiated may persist. Channels, directly following geological weaknesses or following ‘ghost’ features (structures now completely eroded), both represent manifestations of geological control of the landscape. The form and pattern of river channels is affected by a multitude of processes such as: changes in climate that may, for example, lead to increase or decrease in precipitation and vegetation cover; anthropogenically forced changes to ground cover; sea-level changes; and uplift and down-warp of the landscape. These factors are discussed below and although they are typically described as discrete processes, they frequently occur simultaneously. Early workers such as Gilbert (1917) recognized that a river channel may adjust its character following many types of disturbance to the land such as uplift, down-throw, tilting or warping, hence these factors have been the focus of fluvial geomorphological studies for many years. Alteration of these parameters may lead to adjustment of the stream’s longitudinal profile but may equally induce stream widening and downstream aggradation (Doyle and Harbor, 2003). Further incision of the landscape and backfilling of the upper reaches of channels may also ensue (Woolfe et al., 2000). Although erosion may occur over long periods of time, it may also occur suddenly and the causes of abrupt erosion events have been explored widely. For example, Thornes and Alcantara-Ayala (1998) investigated the main cause of mass hill-slope failure that occurs on metamorphic rocks in the mountains of southeast Spain. Anthropogenic activities had been suggested as the cause, although this was later discounted as the events were ‘unpredictable’. They concluded that mass failures depended on the slope material properties, topography, climate and hydrological interactions: where there appeared to be relatively poor resistance and impermeability in the metamorphic rocks (such as phyllites), mass movement was enhanced, both at shallow depths, within the regolith and deep-seated within the bedrock. However, in northern Spain, Calcaterra et al. (1998) noted that natural slopes failed less frequently than man-made slopes of similar angle in low-grade, weathered, metamorphic rocks, suggesting that the natural slopes were more likely to

be closer to equilibrium and better able to withstand variations in weather patterns. In the French Riviera, where high-relief steep slopes with a large variety of rock types occur, shallow, first time landslips (i.e. those that are not repeated) were linked to long periods of heavy rain, and deeper-seated landslips were the result of longer- period gravitational and tectonic effects coupled with weathering (Julian and Anthony, 1996). Two rain thresholds were identified as being necessary to destabilise slopes in the Himalayas during monsoonal rain – the rain thresholds being values of the seasonal accumulation plus the daily total (Gabet et al., 2004). Additionally, it was found that the slope angle controls the amount of daily total rainfall required to destabilise the slope; water storage determines the amount of seasonal rainfall required to destabilise and trigger slope failure. They concluded that thinner regolith on steeper slopes will fail faster than thicker regolith on gentler slopes. In southeast Queensland, Granger and Hayne (2000) reported that the most common trigger for slope failure in the region is an episode of intense rainfall and that antecedent rainfall may be of critical importance. Hoffmann et al. (1976) stated that this is particularly the cause for landslides on the steepest slopes. Granger and Hayne (2000) further stated that long, antecedent rainfall events are most relevant to deep-seated, slow moving landslides and that short, antecedent rainfall periods are relevant to shallow slips and shallow debris flows. Where natural forest cover has been removed, groundwater levels may rise significantly due to a reduction in transpiration. With the increase in groundwater levels, the pore pressure is also amplified leading to a reduction in shear strength of surface materials. In these circumstances, even minor rainfall events may be sufficient to cause failure of rock and/or the overlying soil horizon by raising pore pressure above critical levels. This has been found to occur on the cleared slopes of the Tertiary basalt plateaux and ranges of southeast Queensland where landslides are particularly common (Willmott, 1987). On the Maleny-Mapleton plateau for example, Willmott identified several types of slides including debris slides or flows on scarps and very steep slopes; small rotational slides or slumps on the moderate slopes; and also complex multiple rotational slides that affect broad areas up to 1 km in width. Although the latter type is slow moving, they are typically reactivated in extreme wet seasons. Willmott (1987) also described the basalt terrain as geologically sensitive, resulting from

accumulations of unconsolidated debris that may be easily mobilised; alternating horizons of porous and massive basalts that direct groundwater flow outwards to the slopes; swelling clays within the soil and colluvium that lose strength on saturation; and the presence of soft sediments underlying the basalt, which themselves may fail. From his work, it is clear that surface or endogenic processes are responsible for shaping the landscape, and although the slopes are less stable due to previous deforestation or major rainfall events, the ultimate control of the landslides is typically geological. Rotational slides, both semi-circular and back-tilted, are evident in the Mount Mee area and sites underlain by the Neranleigh-Fernvale Beds through the Brisbane region. Some rotational slides in the area were found to have occurred on slopes as low as 11° (Granger and Hayne, 2000). Soil slumps were also observed on grassed slopes of 11 to 17° in areas underlain by greenschist. They concluded that there may be an increased susceptibility to landslips in the greenschist-derived soils and also colluvial soils derived from banded chert. Willmott and Surwitadiredja (2003) also confirmed that landslides in the Mount Mee area were primarily caused by groundwater seepage and the removal of forest cover on the deep soils that developed on the basalts and greenschists. A debris flow is evident in the western part of Pine Rivers where loose material has been mobilised by torrential rainfall on a steep mountain side (Granger and Hayne, 2000). Such events are fairly common on slopes of greater than 25° and particularly common on slopes that formerly supported rainforests on the Neranleigh-Fernvale Beds. Small landslips have also occurred on the bank of the South Pine River in Cenozoic sediments of the Petrie Formation. On the Bunya Phyllite, rockslides have been observed on the steep banks of the Brisbane River (Granger and Hayne, 2000). A large debris flow occurred in the Laceys Creek catchment, southeast Queensland, after heavy rains in January, 1974, a rainfall event that caused widespread flooding, ground saturation and many other events of mass- wasting in southeast Queensland. The debris flow in Laceys Creek consisted of completely weathered Bunya Phyllite and highly weathered Neranleigh-Fernvale Beds; although it was triggered by saturation, it was aided by ‘an intersecting system of a vertical faults and joints’ leaving a head scarp of 10 m height corresponding to the orientation of the fault and joints (Hofmann et al., 1976). Other landslides that were observed at this time, for example at Mt Nebo, Mt Mee and near Woodford

were on varying slopes, but all were caused by ground saturation. However, the example at Laceys Creek demonstrates that lithological fabric and structure of the underlying rocks may strongly control mass-wasting episodes. Examples of some erosional features that have been identified in the Pine Rivers area of southeast Queensland are highlighted in Table 1. Angle Description of erosion Example in SEQ of slope >25° Small debris slides, rotational slumps, Neranleigh Fernvale debris flows Beds, Brisbane River region, 17-25° Rotational slumps in soil and colluvium on Neranleigh Fernvale concave slopes around gully heads (slumps Beds; Bunya Phyllite can also occur from 11° on greenschist derived or red colluvial soil) 20-25°g Small debris slides, rotational slumps or Rocksberg Greenstone, debris flows in deep pockets of soil or high country such as colluvium western Pine Rivers 11-25° Small rotational slides in colluvium (sand, Undulating plateaux and soil, clay and rock debris) middle country <11° slumps Plateaux, river banks, river terraces and channels Table 1. Examples of erosional features in southeast Queensland (after Hofmann et al., 1976; Granger and Hayne, 2000)

Although rainfall and slope are commonly identified as the combined causes of landslip, Scheidegger (1998) also identified that local, modern tectonic stress was the cause of landslides in the Chinese Himalayas. He stated that, although it is commonly assumed that progressive erosion over-steepens slopes leading to destabilisation and that landslips are commonly triggered by extreme weather or prolonged rainfall, endogenic processes are also of major importance to mass movements and are commonly overlooked. Scheidegger and Ajakaiye (1994) identified that, in Nigeria, apparent ‘erosional features’ such as gullies, have a definite orientation pattern due to endogenic control; where cracks occurred due to

tectonic stress, unstable features were formed that later led to a landslide. They concluded that tectonics was the primary contributor to the landslide, even though it did not trigger the ultimate event. Although mass-wasting occurs episodically, the principle land-forming surface process in southeast Queensland at present is fluvial erosion. Average runoff and peak discharges have been calculated for the Moreton Region (for example Hofmann et al., 1976): peak discharge occurs January to March during which period, flooding incidents and erosion rates are highest. All main streams in the Moreton Region eventually drain to the coast and have relatively steep upper catchments and meandering middle to lower reaches. This is typical of southeast Queensland. In these catchments, erosive forces are influenced by the gradient and length of slope and finer, less cohesive grains will generally erode more readily than coarse-grained and more cohesive soils and rocks (Hofmann et al., 1976) except in cases where ionic bonds generate special cohesive forces in very fine-grained (clay-rich) sediments. Down-cutting is evident in the headwaters of many southeast Queensland drainage systems, particularly in the D’Aguilar Range where gully transverse profiles are commonly V-shaped and bedrock channels form low-order streams (Hofmann et al., 1976); higher order channels typically deepen and widen into alluvial channels on valley floors. North of Brisbane on the Pine River, the ‘Strathpine Terrace’ (Beckmann, 1959), a stream channel and flood plain deposit of Pleistocene age, (deposited when sea levels were interpreted to have been approximately 4.5 – 6 m higher than today) has been incised despite receiving sediment from local streams. Where headwater swales have not kept pace with erosional lowering of the main valley, small hanging valleys also occur in the Strathpine Terrace, near Harrisons Pocket (Hofmann et al., 1976). Gully erosion is evident in the Samford Valley, which is floored by strongly weathered granitic rock. The main modification that occurs in the middle and lower reaches of streams in the Pine Rivers catchments is meander migration and during major floods (such as January 1974) bank erosion has caused channel straightening. Using the modelling software, SedNet, Olley et al. (2006) identified high and medium gulley erosion rates in the upper Lockyer, middle Laidley and western Bremer catchments. However, Caitcheon et al. (2005) identified that down-cutting and gulley erosion are not the only surface processes presently

contributing to stream sediment load, and they reported that hillslope erosion is significant in the middle Lockyer and upper Brisbane catchment regions.

Climate Variations in stream morphology can reflect changes in climate (Kershaw and Nanson, 1993; Nanson et al., 2003; Thomas et al., 2007). The sedimentary sequences caused by varying conditions become stacked, from which, climate change can be ascertained (Read et al., 1991; Steckler et al., 1993). Variations between greenhouse and icehouse conditions have profound affects on the landscape. Sedimentary evidence suggests that low-amplitude and high-frequency sea-level changes during a greenhouse period will lead to moderate to low variations in the location of the shoreline, modest to low continental erosion, and aggrading carbonate ramps that are infrequently sub-aerially exposed and eroded (Séranne, 1999). Conversely, during icehouse conditions, when cooler and warmer, and drier and wetter periods rapidly oscillate, along with high-amplitude sea-level changes, the shoreline location will fluctuate more broadly over time. Furthermore, continental weathering becomes enhanced, providing more terrigenous sediments for transport by streams to the shoreline, producing a prograding terrigenous wedge (Séranne, 1999). In addition to flow regime, channel form and initiation, climate may also directly impact on bedrock form as flow regime also affects the amount and type of sediment supplied to the system. For example, in The Sprongdøla, southern Norway, climate is responsible for frost-shattered bedrock, which itself directly influences the geomorphology, but in turn also provides the sediment source (McEwen and Matthews, 1998). In wet tropical climates, a greater degree of chemical weathering can impose quite a different signature on the sediments derived from the same parent rocks. As climate affects landforms and weathering profiles, the study of the position of landforms and their potential age, can give an indication of palaeoclimate. The amount of rainfall, angle of slope, the surface texture and resulting runoff will affect the amount of erosion and the position and stability of the watertable will control the weathering profile (Gasparini et al., 2004). How ‘heavy’ a rainfall event is, the length of a rainfall event, and even the size of raindrops will have an effect upon the erosion, ensuing slope and channel shape and overall resulting landforms

(Bill Ward, 2006, Pers Comm). Temperature, rainfall distribution and type of events affect the quantity and variety of overland cover. For example, warm regions that receive frequent rainfall may be rich in vegetation that may control the influence of overland flow, infiltration and erosion. Burch et al. (1987) noted that runoff in a catchment that had been completely cleared of forest behaved as if the ground was constantly in an advanced stage of saturation, whereas a similar-sized neighbouring catchment still vegetated by natural remnant eucalypt forest, remained more permeable allowing runoff to be delayed until the soils became wet. High temperatures frequently ‘bake’ and dry exposed soils and this will lead to greater erosion during rainfall events. Infiltration is decreased when heavy precipitation occurs on a hardened soil and overland flow occurs more readily. Such events are typical in many tropical and subtropical zones including southeast Queensland. The predictable outcomes that might be expected from these surface processes have been used by workers in this field to compare weathering, erosion and climatic factors with patterns in the resulting features (Woolnough, 1927 cited by Watkins 1967; Bryan, 1939; Whitehouse, 1940). Climate change may also lead to changes in vegetation type and cover. Rich vegetation covering the landscape can provide stability for slopes via soil-binding roots and reduce the power of rain upon the soil. Furthermore, evaporative losses occur when precipitation is first intercepted by foliage and the magnitude of run-off is much lower resulting in reduced overland flow (Bonnet et al., 2001). This can lead to more gradual and more thorough infiltration than a similar style of rainfall falling on sparsely vegetated land, although is it also argued that leaf debris can prevent infiltration. Nevertheless, runoff from leaf matter still prevents some erosion that otherwise would have occurred if rainfall was directly onto the soil or rocks. The main contributing factors of channel initiation brought on by overland flow (Bonnet et al., 2001) are climate, vegetation coverage, precipitation patterns and ‘weatherability’ of underlying rocks. Channel initiation requires the upper tips of the first-order channels to correspond with a point where erosion can be generated due to sufficient shear-stress induced by overland flow.

Palaeoclimate and present-day climate in southeast Queensland During the late Cenozoic, climate in southeast Queensland varied between long, dry periods and shorter, humid periods. Sheet-wash and scarp-retreat during dry periods resulted in pediments and pediplain, and upper erosion surfaces such as at Stanthorpe, the crest of the D’Aguilar Range and the Mount Mee Plateau (Watkins, 1967). During humid periods thalwegs were accentuated by more linear erosion, pediplains were rejuvenated and topographic relief became more dissected, such as in the Lockyer Valley and upper Logan and Mary river areas (Watkins, 1967). The upper erosion surface is a well-preserved pediplain that has been lateritised, although the laterite was stripped, leaving a surface of silcrete in places, such as Miocene-aged surfaces west of the Great Divide. Closer to the coast, the upper erosion surface of Eocene-Oligocene age is warped with an approximate north-south axis of uplift with the axis rising towards the south. A lateritised pediplain surface at approximately 15- 60 m above sea level, of Pliocene age, shows pediplain and lateritisation conditions continued until the onset of the Pleistocene. Dissection of this surface took place when conditions became more humid, although the pediplain was not completely obliterated. Uplift and warping of the erosion surfaces was brought about by Cenozoic volcanic activity. During this time, the Middle Erosion Surface developed (Watkins 1967) and this surface (at 120-210 m) is present, as discontinuous pediment-like surfaces flanking the D’Aguilar Range (Hofmann et al., 1976). Cenozoic deep-weathering profiles and duricrusts in southeast Queensland suggest that climates were generally more humid and warmer than at present (Grimes, 1988). The Lower Erosion Surface (at 15-80 m) of predominantly Pliocene age is interpreted to have extended into the Pleistocene, during which time the three erosion surfaces became dissected (Watkins, 1967). This may have been due to lower sea- level (Hofmann et al., 1976) or uplift, greater humidity and increased runoff (Watkins, 1967). The strongly fluctuating sea levels of the Pleistocene suggest that base level falls may have been the key driving force for fluvial incision of the Pliocene surfaces. The Lower Erosion Surface is preserved in the upper reaches of the North Pine River approximately 60-80 m a.m.s.l. (Hofmann et al., 1976). More recently, a coastal depositional plain of approximately 3 m a.m.s.l emerged exposing a narrow lower coastal plain (Watkins, 1967).

The present climate in southeast Queensland is subtropical, typically with wet summers and dry winters. It is generally considered to be a humid climate with humidity increasing in the summer. Although Watkins (1967) stated that the Kingaroy-Darling Downs area ‘represents an island of arid climate’, this description is incorrect as mean annual rainfall in the area is approximately 800 mm and summer months are typically humid (Bureau of Meteorology, 2008). The current climate of southeast Queensland causes aeolian, pediplain and fluvial erosional processes to occur, and also processes that lead to the formation of calcrete, silcrete or laterite (duricrust) (Watkins 1967).

Regolith and ground cover Surface processes include both mechanical and chemical weathering: the former where colluvial and fluvial processes physically break down material at the surface; the latter where grains are dislodged from the parent rock by mobilisation of chemical species principally via the action of groundwater movement in southeast Queensland. The depth and type of weathering profile and regolith that forms on bedrock is strongly controlled by climate. For example, a deep profile and duricrust will form where the climate is warm and humid. However, lithology also plays a role in the formation of the weathering profile as the composition, permeability and structure of the rock (such as jointing) will dictate the rock strength and weatherability. The weathering profile may also be controlled by topography: shallow valleys where water preferentially flows will form deeper weathering profiles than steep slopes. However, the intensity of erosion is equally important. If erosion is too fast, weathered material will be quickly removed and a profile will not form. If deposition is too rapid, fresh sediment will cover the surface (Grimes, 1988) potentially accumulating a sediment pile incorporating multiple palaeosols. Deeply weathered land surfaces are a common feature in low latitude regions that have not been affected by Pleistocene glaciation or aeolian erosion and occur in regions such as Asia, central Africa, northern South America and Australia (Taylor and Howard, 1999). Taylor and Howard showed that alternating cycles of chemical and mechanical weathering are closely related to tectonic processes. Mechanical weathering typically occurs in times of tectonic activity and in particular during

crustal uplift; chemical weathering (or deep weathering) typically occurs in times of tectonic quiescence. Taylor and Howard further noted that where the rate of accumulation of weathered products exceeds the rate of removal (recharge-dominant hydrological regime), these areas experience tectonic quiescence (no or low tectonic activity) and persistent deep-weathering. They argued that conversely, where the rate of accumulation of weathered products is less than the rate of removal (run-off dominant regime) tectonics is typically influencing the landscape. To summarise, where there is bare rock or low surface cover that has not been influenced by aeolian or glacial processes, tectonic uplift is dominant. Similarly, Montgomery (2003) identified three different landscape types: chemical weathering dominant landscapes, such as ancient cratons, where chemical weathering exceeds mechanical denudation; low-relief and post-orogenic landscapes, where hill-slope processes and erosion rates are reflected by the mean slope and local relief; and steep terrain and tectonically active landscapes. Broadly speaking, these landscapes relate to ancient cratons, post- orogenic landscapes and active orogens respectively. The models of both Montgomery and Taylor and Howard identify that surface processes are typically controlled by the underlying geological regime. In summary, their work suggests that landscapes are characterised by deep weathering profiles where there is low tectonic activity and conversely, where there is low regolith development, greater tectonic activity is most likely occurring although these represent end-members of the soil- forming spectrum and most landscapes fall between the two.

Effects of sea-level change Sea-level change may be eustatic or it may be local (with reference to areas undergoing active uplift or subsidence). It can be related to ice-sheets loading the crust, or isostatic rebound (where the crust rebounds after an ice-sheet melts); tectonics, where uplifted or downthrown land masses may experience a relative change in sea-level; or climate change which might physically alter the volume of water in the oceans. Slope and stream power are important factors in channel formation. Increased rainfall will provide a greater water supply and increase stream power, which, depending on slope, may lead to deepening or widening of rivers. Numerical modelling has shown that, during eustatic sea-level fall, a strong drainage connection can exist on passive continental margins that links the drainage basin

with the depocentre on the shelf, but bypasses the exposed shelf. This connection between the terrestrial (on-shore) and marine drainage system environments exists as a single, cross-shelf river, causing there to be just one depocentre and preventing inland, catchment-derived sediment from being deposited on the upper continental slope (Meijer, 2002). Although adjustment to a river’s longitudinal profile has been linked with tectonic influence (for example Demoulin, 1998) it can also be a general characteristic of changes in base-level, including anthropogenic and climatic driven changes. It is widely accepted that, as a consequence of sea-level (or any base-level) fall, incision of streams and rivers will occur, due to an enforced new base-level; conversely, with sea-level rise, it is expected that streams will aggrade, or accumulate sediment particularly in the lower part of their course (Miall, 1991). Many sequence stratigraphic models employ this interpretation (for example Vail, 1987; Posamentier et al., 1988; Posamentier and Vail, 1988). Similarly, it is expected that the longitudinal profile of a river system will adjust accordingly, causing a shortening of the curve with a relative rise in sea-level, and falling sea-level causing incision and lengthening of the curve. However, some workers suggest that sea level change may cause only minor alterations to a drainage system and this is discussed further, later. Additionally, other workers have proposed that incision during a lowstand is not necessarily the ‘norm’ (Wood et al., 1993; Woolfe et al., 1998; Woolfe et al., 2000). Woolfe et al. (2000), for example, suggested that drainage of the Herbert River, Queensland, would become further incised should sea-level rise and would aggrade onto the Great Barrier Reef should sea-level fall. These assertions were made on the basis that bed erodibility, stream gradient and flow velocity, all of which may induce incision should they increase, are unlikely to change should there be sea- level fall by less than 100 m. They ascertained that channel incision would be driven sea-ward with alluvial filling of previously incised channels. They also identified that, should sea-level rise, present day incision would migrate landward and fill downstream with increased sediment supply from the newly incised channels upstream. This alternative model is important as the present-day prediction is for sea- level to rise. Should this occur at a sufficient rate to drown the coastline, channel incision may move landward and channel filling will occur at the seaward end of a drainage system. The geomorphology and evolution of the Herbert River has been

well documented. A similarly good understanding of other drainage systems would be useful for assessing the impacts of sea-level changes upon the system in other locations. The time-scale over which changes occur is also important. For example, Thomas et al. (2007) reported that apparent changes in the climate may be accentuated at the catchment scale, if the catchment is highly sensitive to change. Conversely, they found that in their study area in northeast Queensland, although changes at the decade or century scale were not apparent, there was landscape-based evidence for longer-term changes in climate. Rivers may avulse due to changes in sea level although this may also occur due to variations in climate (for example Allen, 1978; Shanley and McCabe, 1991) although previous down-cutting and incision may prevent a river from avulsing (Meijer, 2002). The processes controlling river patterns can be difficult to identify without the integration of a very broad range of climate, eustatic, tectonic, lithological and structural variables. The presence of one indicator is not sufficient evidence to substantiate a claim that there is geological control over a drainage pattern. Therefore, where geological control is hypothesised, multiple indices should be sought (Demoulin, 1998; Holbrook and Schumm, 1999).

Anthropogenic influence Since the quantity and type of vegetation cover will promote or prevent surface erosion, it is important to recognise that anthropogenic influence upon ground-cover may have equally broad effects on the landscape (Bronstert et al., 2002). Changes in land-use may alter vegetation type and quantity, leading to changes in runoff, weathering, erosion and landscape alteration. Increasingly, studies such as the Biospheric Aspects of the Hydrological Cycle (Hutjes et al., 1998), are focussing on the impact of human-induced changes to vegetation cover and how the changes influence the lateral redistribution of water and its transported constituents. Many of these studies reveal subtle impacts on the surface and subsurface environment that might not be predicted from the nature of the anthropogenic processes. The Copper Basin, Tennessee, USA, was subjected to 100 years of logging, mining, grazing, fire, and water acidification and was reforested approximately 50 years ago. A study in the region to measure the impact of land-use change identified that although soil erosion decreased within 10 years of replanting, runoff rates remained high and organic matter in the soil remained low, suggesting that landscape

recovery even in humid areas of high biotic productivity may be slow and changes may have prolonged effects (Allan and Peterson, 2002). A landslip study in the Ocean View plateau of Mount Mee, southeast Queensland, identified that the plateau consisted of ‘bulging’ slopes which indicate slow downhill movement of soil beneath the grass mat, which may lead to land-sliding (Hofmann et al., 1976). The Ocean View area was cleared of native vegetation in the early 1900s and small slumps and landslips are visible despite revegetation of the area having commenced. In the upper Murrumbidgee catchment, southeastern Australia, Olley and Wasson (2003) identified a major adjustment to catchment dynamics since the settlement of Europeans. Results demonstrated that sediment flux altered and gully erosion increased significantly due to the introduction of grazing, damming of rivers and changes in vegetation. Large tracts of land in southeast Queensland have been cleared of trees for up to 100 years, mainly for cultivation and grazing, but also for urban development. This has generated changes in drainage volume and rates and potentially influenced other surface processes. A study of a substantially cleared catchment, Mount Samson Creek, north of Brisbane in southeast Queensland, identified that the creek carried approximately three times the suspended solids than that of neighbouring creeks (Laceys and Baxters creeks) where the catchments of the latter had only partially been cleared (Hofmann et al., 1976). It is evident from the literature that land use changes affect runoff, and many studies have focused on measuring increase in runoff due to deforestation. However, the impact of re- afforestation is also important for considering the impact of land-use change. For example, a study in the Ebro catchment, Spain, identified that annual stream discharge has decreased considerably over the last few decades during which time there has been a substantial increase in forest cover due to re-afforestation and farm- land abandonment (Gallart and Llorens, 2004). Similarly, modelling has also predicted that a reduction in flow would occur in response to large-scale re- afforestation in the Macquarie River, New South Wales, Australia (Herron et al., 2002). This modelling also accounted for climate change variation. Individually, climate variation and anthropogenic influence can both lead to changes in runoff and stream flow. However, anthropogenic changes to the landscape can be compounded by climate change (and vice versa). Some studies have been undertaken to identify the combined influence of climate change, runoff and

evapotranspiration in agricultural regions. For example, modelling has predicted that an ephemeral catchment of cleared farm land, like many of those in southeast Queensland, would experience four times the percentage change in runoff to the percentage change in rainfall, and in a wet or temperate catchment, the percentage change in runoff may be about twice the percentage change in rainfall (Chiew and McMahon, 2002). A common anthropogenic change on landscapes is the formation of man- made lakes. These create an artificial base level to which the catchment above and below the new dam will adjust. Adjustment is complex and may vary depending on the maturity of the catchment prior to damming. A forced base-level rise can have devastating effects upon the catchment upstream. Where sediment may previously have been transported downstream to lower gradient, alluvial and coastal plains, sediment is instead deposited where the stream gradient may have once been relatively steep. Sedimentation will occur at the mouths of the streams that feed into the lake as energy levels decrease rapidly where previously, energy levels would have been higher. Downstream from the dam, braided channels would be expected to narrow and a reduction in channel migration rate of meandering channels would occur (Friedman et al., 1998). Although this was the case for streams downstream of the Flaming Gorge Dam in Utah and Colorado, USA, the magnitude of narrowing was not found to correlate linearly with the distance downstream from the dam. However, it related to the degree to which peak flows had been reduced by the dam (Grams and Schmidt, 2005). Other changes may also occur when the lake level changes intermittently, due to rainfall and water-use variation. Lake Samsonvale in southeast Queensland for example, built and filled in the mid-1970s has caused deposition at the interface between the lake and the feeder-streams. When lake levels fell in 2007, the deposits were exposed and the ‘deltas’ prograded further into the lake. These surfaces have since been drowned due to recent increased rainfall. Similar continuing adjustment would occur at other lakes in southeast Queensland, including Lake Somerset, Lake Wivenhoe and smaller dams situated on agricultural properties in the region. The net effect will equate to a decrease in sediment delivery to the coastal regions, as sediment that would have been transported to the coast is instead trapped behind dams. A report in 2001 identified that the natural flow rates of the rivers hosting lakes Kurwongbah and Samsonvale have been reduced by the

formation of the dams (Abal et al., 2001). Upstream of the dams, flow rates have altered where longitudinal profiles have adjusted to the new base-level and downstream of the dams, flow rates are controlled by the dams. The report also identified that the proliferation of small farm dams in rural areas upstream of the major dams trap and store minor flows individually but, cumulatively, have a large impact upon base flow conditions. Douglas et al. (2003) reported that sediment from the Neara Volcanics which, in the Brisbane River catchment region, all lie upstream of the Lake Wivenhoe dam, are not recognizable in sediment contributions in Moreton Bay. During the January 1974 cyclone and ensuing flood events of southeast Queensland, bank erosion was observed in the lower reaches of Laceys Creek: the stream removed the entire flood plain in some areas that consisted of several hundred lateral metres of a previously cleared bank (Hofmann et al., 1976). Hofmann et al. (1976) also attributed forest clearing to increased gully erosion in the north of Kobble Creek catchment, and additionally identified that indiscriminate gravel excavation on the South Pine River was linked to changes in river flow and increased major bank erosion during floods. Although it may appear that anthropological changes to the landscape cause far-reaching affects, some parts of southeast Queensland are clearly little-affected by land-use change. For example, a study of Coombabah Lake, on the Gold Coast, identified that although natural processes have fluctuated over its >6000 year history, its current state suggests human activities surrounding the catchment have caused no adverse affects (Frank and Fielding, 2004). A review of major catchments of southeast Queensland, including the Bremer, Lockyer and Wivenhoe subcatchments (Caitcheon et al., 2005), identified the major sources of sediment to Moreton Bay and the lower Brisbane River. Using the SedNet modelling package, coupled with erosion process tracing, the results identified gully and stream bank erosion, and also hillslope erosion from both grazing or cultivated lands. From their previous analysis, they identified that soil from forests was deemed to be similar to grazing soil. The majority of sediment that reached Moreton Bay, originated from the Lockyer catchment and a smaller but substantial amount of sediment came from the Bremer catchment. Little sediment that originated north of Wivenhoe and Somerset dams reached the mid-Brisbane

River, probably due to the sediment being captured by the lakes. In the Wivenhoe catchment, sediment was mostly sourced from the lower to middle reaches of Kilcoy Creek primarily as a result of stream bank and gully erosion, although hillslope erosion was predicted as the dominant sediment source in western tributaries of Kilcoy Creek. In the upper Brisbane River, sediment was mainly sourced from grazed hillslopes, whereas downstream of the Lockyer, in the lower Brisbane River, sediment was primarily derived from channel erosion. The results in the Lockyer catchment were less conclusive although they suggested that upper Lockyer sources were mainly eroded grazing land, and the lower Lockyer was a combination of channel erosion and cultivated land erosion. Results from the lower Bremer suggested the dominant sediment source is channel erosion with smaller amounts of sediment derived from cultivated soils. Gully formation, [defined by Caitcheon et al. (2005) as ‘incision of valley floor alluvium since European settlement’ p8] and gully alteration over the past 150 years, were also analysed. Caitcheon et al. concluded that most gully erosion occurs at a slower rate now than when the gully networks originally developed, having potentially reached an approximate state of equilibrium. Gully erosion, as with riverbank erosion, continue to add to bed-load and suspended sediment supply, whereas hillslope sources only supply sediment to the suspended load budget although some new gullies develop in these areas, particularly during times of flood and increased overland flow rates.

The role of lithology and rock fabric in geomorphology Crustal deformation may be instantaneous or may take place over millions of years. Where the crust is weakest, deformation may occur due to compressional or extensional stress often caused by the movement of tectonic plates (Figs. 1,2 and 3). Nevertheless, whether movement is gradual or sudden, each motion can contribute to the generation of various surface features such as gently undulating hills or sheer scarps, although not all tectonic movements produce surficial expressions. The theory of plate tectonics has been dominant since the 1960’s and its processes can explain diverse styles of deformation and movement of Earth’s crust (e.g. Dietz, 1961; Hess, 1962). Through plate motion and the associated crustal stress, landscape features such as fault systems, orogenic belts, hills and valleys may form, which are

altered further by surficial processes including weathering and erosion, concurrently acting upon them. The mineralogical composition and organization of rocks dictates their susceptibility to weathering and erosion; marl or clay-rich beds may erode faster than indurated quartzose sandstone, for example, and if juxtaposed, this will be reflected in physiographic differences in the landscape (Fig. 4). Most major mountain ranges such as the Pyrenees in Spain (Fig. 4), show examples of slope variations that have been caused by the differential weathering of distinct sediment types. Sedimentary rocks typically contain many beds of different lithology caused by the sorting of sediment type and size during deposition. This can lead to dissimilarity in rock strength and chemical stability between beds. When exposed or close to the surface, the variable competence layers will be subjected to differential weathering. Regions with breached folds in sedimentary rock typically show strongly surficial expression of differential weathering of lithologies from whole landscape to outcrop scale (Figure 5). Igneous rocks may also vary in strength depending on their mineralogy. Often intruded into rocks of a different strength, they may weather more slowly than the surrounding rock. Excellent examples of this are expressed in southeast Queensland where the outer flanks of mid-Cenozoic felsic volcanoes have eroded, and the more resistant volcanic necks now remain as the Glass House Mountains (Figure 6) and plugs of the Mt Alford region. Alternatively, the intruded igneous body may erode more quickly than the surrounding rock, forming a basin in the landscape. An example of this type is the bowl-shaped Samford Valley in southeast Queensland, which is formed on the strongly weathered material of the Samford Granodiorite surrounded by hills developed on a more resistant thermally metamorphosed aureole in the Neranleigh-Fernvale Beds (Figure 7). Minerals within a rock may alter over time if heat, temperature and pressure are changed after its initial crystallisation or deposition. Where temperature and/or pressure are increased, mineral alteration may lead to a preferred orientation of minerals, such as mica, and the development of foliation. Foliation is best developed in regionally metamorphosed rocks; this is a planar fabric caused by the parallel alignment of crystals leading to a slatey, phyllitic, schistose or gneissose texture or cleavage. Stress and associated strain may also lead to physical alteration in the structure of the rock causing folds and discontinuities such as faults, shear zones and joints, which

may occur in association with one another even at the microscale. In folded rocks, for example, weaknesses may occur due to tension within the rock-fabric, allowing weathering to exploit these zones. This will present natural planes of weakness that, if eroded, will expose other rocks that may weather differentially. The fractures, cleavage, joints, preferred mineral alignment or fault gouge displayed by a rock are caused by a specific orientation of stress or folding. This may cause repeated and parallel planes of weakness in rocks. Bedding and other sedimentary features may also result in repeated and aligned fabrics. Such bedding and structural features are commonly exploited by differential erosion to produce step- and ridge-like patterns in the landscape (Figure 8). These rock patterns can occur in areas ranging from outcrop to continental scale. Where streams incise to bedrock, they commonly follow bedding and fracture patterns and become deflected along these discontinuities to generate ‘anomalous’ drainage patterns (for example Holbrook and Schumm, 1999). Such patterns are typically expressed by parallel, conjugate, radial concentric or other geometric arrangements of streams, gullies, hill slopes and scarps. Recognition of anomalous drainage patterns is typically a strong index of geological control rather than exogenic or regolith control of stream flow. Surface processes take advantage of the geology to form the landscape’s morphology, but the geology itself is potentially a product of numerous events and tectonic settings. The morphology of some regions may display patterns that were controlled by a previous terrain or rock unit that has since been eroded or altered. In such a case, the remnant morphology still shows there was an original geological control, but to ascertain whether control is recent or ancient, further examination of the geomorphology is required. This involves study of the morphology at varying scales, such as that being undertaken in this research, in order to identify the long- term geological history of a region and the degree and extent to which drainage is being geologically controlled. Stream orientation and control can be used to identify the timing of events. For example, large, aligned streams that both meander and cut across lineaments may provide evidence that the stream orientation was caused by an ancient control no longer present, and that surface processes have since taken over its morphological development. Alternatively, where many low-order streams are found to change direction, this might suggest fairly recent uplift. Drainage patterns will respond to

changes both in topography and base level imposed by uplift and subsidence (Burbank et al., 1996). Changes may include adjustments to channel gradient and width, sinuosity, bed load grainsize and extent of alluvial cover, and bed morphology and roughness (Whipple, 2004). A stream’s orientation may be an indicator that changes have been constrained by an endogenic factor as outlined above. Where 1st order or other low order streams are aligned with others of similar and higher orders, or aligned with visible and underlying rock fabric, it generally indicates an existing or ongoing geological control. Therefore, study of these features at multiple scales is important to understand whether the control is recent or ancient. Depending on the exposed surface and angle of dip of the planes of weakness, the resulting eroded surface may have broad scale implications: for example, the River Torrens in the Mt Lofty Ranges, east of Adelaide is strongly controlled by phyllitic and gneissic cleavage (Twidale, 2004). Planes of weakness resulting from faulting and jointing can be exploited by precipitation seepage. Fluid flow within and over this type of weakened rock may lead to erosion and widening of such planes, providing preferential conduits for further fluid flow. The courses of the Rhône and Rhine rivers in the Alps for example, are trapped by faults, the planes of which display lower erosional resistance than adjacent lithologies. This, combined with enhanced discharge of the rivers and low erosional resistance of their bedrocks probably increased surface erosion relative to neighbouring areas (Schlunegger and Hinderer, 2001). Geological features such as faults, fractures and lithological differences are rarely visible in a continuous manner across large regions due to regolith cover and in cases where groundcover is dense, such as in southeast Queensland, inference of underlying control must be made based on what can be seen and measured. If geological fabric is known in only part of a region, and alignment of that fabric with morphological features is evident, the implication may be that the same geological control exists across the broader region containing that pattern of morphological alignment or repetition.

Figure 1 Simple schematic cartoon to illustrate some structural geology features that may control landforms in a compressional regime

Figure 2 Faulting that may occur in an extensional regime

Figure 3 Folded sandstones in a sheer cliff face, Sandgate, Queensland

Figure 4 Differential weathering causing slope variations, Pyrenees, Spain

Figure 5 Faulted and folded sedimentary rocks displaying differential weathering, Shorncliffe, southeast Queensland

Figure 6 The Glasshouse Mountains, southeast Queensland. The outer part of the volcanoes and the surrounding rocks have been eroded leaving the volcanic plugs protruding from the landscape. (Photograph courtesy of David Hodgkinson)

Figure 7 Aerial view of Samford Valley, southeast Queensland. Samford is situated on an eroded granitic batholith that intruded the regionally metamorphosed Neranleigh-Fernvale Beds. The Samford Granodiorite has eroded preferentially, and has formed a basin surrounded by a thermally metamorphosed aureole developed within the Neranleigh-Fernvale Beds. (Photograph – Google Earth)

Figure 8 Folded strata under compressive stress may form microstructures within the fabric. Extension within the crests of antiformal folds may cause joints and fractures to form which could increase weathering in these zones

Bedrock lithology may influence stream behaviour such as the orientation of flow, degree of meandering and anastamosing, and the magnitude of down cutting. Fine-scale structures such as phyllitic cleavage have been identified as controlling large streams (for example Holbrook and Schumm, 1999; Twidale, 2004) where the rock type is ‘soft’ or weakened and more susceptible to erosion allowing incision, or

where the rock types are more resistant and force the direction of flow. One aim of this research is to identify whether such a control is exerted over streams at multiple- scales, from first order through to major channels.

The role of tectonics and major geological events in geomorphology The influence of structural control over surface features including stream patterns has been recognised as an important asset for better understanding of the geology and structure of large areas (Hills, 1960). Tectonics is a widely accepted cause of geomorphological change, and acts as a primary control upon the shape of the landscape (e.g. Ollier, 1995; Burrato et al., 2003; Vannoli et al., 2004; Delcaillau et al., 2006). The geomorphological approach has been used to analyse and better understand the effects of recent tectonism in several areas (for example Oguchi et al., 2003; Palyvos et al., 2006). In Greece, for example, geomorphology and drainage patterns have been used to identify active normal fault evolution (Goldsworthy and Jackson, 2000). Schlunegger and Hinderer (2001) examined the potential geological controls upon the anomalous drainage patterns in the Alps and concluded that enhanced rates of crustal uplift, associated with frequent small earthquake events were responsible. Drainage pattern anomalies have been used in the Turkana Rift, north Kenya, as ‘key-markers’ to establish the location of large-scale transverse fault zones (Vétel et al., 2004). Barbed drainage such as that seen in the Cowan River, Western Australia (Clarke, 1994) and the Clarence River, New South Wales (Haworth and Ollier, 1992) are distinctive evidence of warping or uplift of the landscape. Drainage is known to commonly follow the orientation and plane of faults that provide a natural channel as outlined above. Faults caused by high magnitude earthquake events may cause large-scale surface ruptures providing a natural location for preferred fluid flow. However, small faults and fractures can also be the site of preferred overland flow. Joints and faults resulting from even minor (low magnitude) earthquake events, may occur in repeating alignment, as stress orientation is typically relatively stable for long periods of time. Small, aligned faults and fractures may eventually merge where fault tips slowly migrate, lengthening the faults and changing the stress dynamics (as discussed in more detail below). Coalescence of faults in this way may lead to the appearance of large faults which may simply

represent a composite system of smaller faults. Therefore, even very small fractures caused by low magnitude earthquake events can, over time, provide a preferred position for overland flow. Attraction of further stream flow along the plane will eventually lead to deepening and widening of a channel. Cowie and Roberts (2001) presented a conceptual model showing multiple stages in fault-growth. Initially, an array of small faults forms across a region and slowly extends in length and throw. As this continues, the faults interact and the overall fault-array profile changes. The displacement-to-length ratios increase over time and when multiple faults have joined with others, central portions of the fault will eventually have greater throw than the distal portions. The temporal and spatial variations of movement along faults and fault arrays may be explained using the noncharacteristic earthquake model, first proposed by Roberts (1996b), which provides a model that not only accounts for spatial variations in cumulative throw but also for ruptures that are shorter than the host fault segment. The model also implies that recurrence intervals vary temporally for an individual locality. Roberts concluded that palaeoseismological evidence from one site along a fault segment should not be used to imply earthquake recurrence at another on the same fault. Roberts and Minchetti (2004) and Roberts et al. (2004), working in the Lazio-Abruzzo Appennines, central Italy, further showed that interactions of multiple fractures along a fault segment are complex, but knowledge of scaling relationships between the fault throws and lengths may assist prediction of throw-rates and with it seismic hazard. The ratio of maximum displacement to length on a fault is an important characteristic for assessing slip rates and for future earthquake prediction (Cowie and Roberts, 2001). The fact that many faults grow by the connection of smaller fault segments (Peacock and Sanderson, 1991; Roberts, 1996a) is an important consideration in a region where low magnitude earthquakes are most typical; numerous small scale earthquakes over time may not directly be a geohazard, but could play a large part in the evolution of landscape morphology. Some earth movements may be slow and steady, leading to potentially large displacements over time. Slow earthquakes have occurred in many regions. They are defined in the literature as discontinuous events that release energy over long periods of up to several months, unlike typical earthquakes that may release energy in just seconds or minutes (e.g. Kanamori and Stewart, 1979; Linde et al., 1996). These events may be

accompanied by earth tremor and may be detected at some very low frequencies. They may provide a link between shallow and deep crustal events. Aseismic creep and slow earthquakes may lead to earth movements equivalent to moderate or large earthquake magnitudes and on such faults, seismicity may account for as little as 2% of total moment release (Amelung and King, 1997; Scholz, 2002). However, in order to resolve aseismic creep, an accurate background deformation rate must be measured to identify regions of anomalously high deformation rates (e.g. Linde et al., 1996; Kitagawa et al., 2006) and background deformation rates have not yet been calculated for southeast Queensland. Earthquakes of > M 5 (M 5.6 for example) are known to cause ground surface displacement by nearly 10 m (for example, Fort Sage Mountains, CA., USA, 1950 cited in Wells and Coppersmith 1994, p.976) and earthquakes of M 7 or greater have caused surface displacements of 10’s or hundreds of kilometres in length (for example Luzon, Phillipines, M 7.8, 1990, surface rupture length 120 km, Wells and Coppersmith, 1994, p.981). It is generally considered that low magnitude events (M<4) do not cause surface displacement. A widely cited reference in support of this hypothesis is the work carried out by Wells and Coppersmith (1994). Wells and Coppersmith used a ‘global’ earthquake database and from the results they inferred that surface displacement is unlikely below M 6. However, they also stated that this inference is based on regression analysis for which standard deviations were large. This qualification is commonly overlooked and, therefore, the likelihood of no surface expression resulting from earthquakes M<4 is typically assumed to be ‘fact’. However, other workers have subsequently provided evidence that significant surface displacement can be produced by earthquakes of much smaller magnitudes. For example, in the Appenines and in Sicily, earthquakes of M 2.7 to M 4 have commonly caused surface ruptures from 100 m to more than 2 km long (Mohammadioun and Serva, 2001; Serva et al., 2002; Michetti et al., 2005). These earthquakes were of shallow depth as are modern earthquakes occurring in southeast Queensland. Michetti et al. (2005) stated that different scaling laws exist between earthquake magnitude and surface faulting parameters as a function of, for example, style of faulting, focal depths, heat flow and other geodynamic settings. It is evident that the dataset used in the study by Wells and Coppersmith may have been lacking

data from a full suite of potential seismic landscapes and, therefore, their results should not be treated as being definitive. Although rivers that incise to bedrock may follow fracture patterns at multiple scales (Howard, 1967; Droste and Keller, 1989; Holbrook and Schumm, 1999), rivers floored by unconsolidated alluvial material may also be affected by geological controls. Such rivers are sensitive to even subtle changes in the gradient of the topographic surface and may respond to tectonic tilting in several ways. In particular, they are deflected by surficial warping (for example Goodrich, 1898; Zernitz, 1932; Howard, 1967; Holbrook and Schumm, 1999). The affects of surficial tilting or warping on drainage include: degradation in foretilted and aggradation in back tilted reaches; channel deflection around zones of uplift; shifts in channel pattern to compensate for tilting changes in bed-load grainsize; changes in frequency of overbank flooding (Holbrook and Schumm, 1999) and anomalous steepening of channel gradients (Kirby et al., 2008). The affects of uplift are most prominent for low-gradient streams. Longitudinal profiles of streams that encounter zones of active subsidence or uplift will traverse or be deflected by the deformed zone (Holbrook and Schumm, 1999). A stream can cross a zone of uplift if the rate of incision outpaces the rate of uplift, and if the river course is already well established at the site of deformation prior to uplift (Holbrook and Schumm, 1999). Warping of the surface may cause terraces and valley floors or longitudinal profiles to become convex causing degradation, or they may become concave, where aggradation will occur (Burnett, 1982; Burnett and Schumm, 1983; Ouchi, 1985; Schumm et al., 1994). The Rio Grande River in Mexico, for example, traverses the dome of the Socorro magma body in New Mexico; the river is aggrading in the down-warped reaches downstream and upstream of the domal axis, and incising where steepened gradients are occurring across the dome (Ouchi, 1985). In the eastern Himalayas, the longitudinal profiles of many rivers vary, although the variability of structure and profiles of the rivers could not be attributed to climatic variation (Baillie and Norbu, 2004). Tectonic factors were concluded to be the main controlling process in their development. Terraces of the River Ganga, India, which is in an active foreland basin, were analysed by Srivastava et al. (2003). Although climate-related sea-level change had previously been suggested as the cause of incision and terrace formation, their results suggested that regional up-warping caused by a significant tectonic event

approximately 6 ka led to the incision of the river by several metres, causing the formation of terraces in many parts of the system. Westaway (2002b; 2002a) concluded that river terrace sequences produced globally during the late Pliocene and early middle Pleistocene were caused by uplift of a thickening continental crust. He suggested that the thickening was caused by flow in the lower crust that was induced by cyclic surface loading caused by both ice-sheet formation and retreat and by sea- level fluctuations. Uplift by up to hundreds of metres has led to the formation of multiple river terraces along some rivers, such as at the Vlatva near Prague and the Thames in southern England. Stepped river terraces in Canterbury, New Zealand, were caused by uplift of approximately 30 m, that preceded a faulting event and led to accelerated river incision and terrace formation (Campbell et al., 2003). Koss et al. (1994) showed from experimental studies that simple rise and fall of base-level can have little effect upon a drainage basin. Most importantly they identified that a change in base-level had a more significant effect on the shelf area, and a considerable lag time occurred before any secondary effects were seen in the drainage basin. Tebbens et al. (2000) studied the River Meuse in northwest Europe, in relation to sea-level rise and concluded that long-term fluvial dynamics were not effected by sea-level rise, although downstream, high-stands were marked by river terraces. Having assessed the formation of river terraces in northwest Europe, Bridgland (2000) identified that, although climatic fluctuations were the driving force behind terrace formation, the link was indirect, as the terraces were only found to have formed where uplift was also experienced. Bridgland concluded that uplift may have been due to isostatic adjustment following unloading and redistribution of sediments off-shore. Berryman et al. (2000) recognised that terraces in the lower Waipaoa River, North Island, New Zealand, were strongly controlled by uplift, but as the down-cutting rate exceeded uplift, climate fluctuations were likely to be the primary control on the terrace formation in the region. Depending on the amount of deformation and direction of tilting of the landscape, channel patterns may be altered either directly or indirectly. Sufficient decreases in slope may cause a channel to transform from braided or meandering to straight or anastamosing, or vice versa. Meandering streams may be altered to braided (Twidale, 1966; Burnett, 1982) or may become anastamosing (Burnett, 1982; Ouchi, 1985). However, it is less likely that complete shifts in pattern will occur and

more likely that subtle changes will develop (Holbrook and Schumm, 1999). For example, an increase in slope might cause a meandering channel to increase sinuosity, or conversely reduce sinuosity in response to a decrease in slope (for example Welch, 1973; Burnett, 1982; Ouchi, 1985; Jorgensen, 1990; Boyd and Schumm, 1995; Holbrook and Schumm, 1999). The low-gradient Mississippi River reduces its sinuosity where it crosses the Lake County uplift. A 10 m high topographic bulge is caused by active deformation and where gradients increase on the down-dip flank, sinuosity increases (Russ, 1982; Schumm et al., 1994; Holbrook and Schumm, 1999). Less intense deformation can also cause a multitude of changes to drainage. A good example of this is in the Sorbas Basin of southeast Spain, where, during the later stages of basin inversion, sheetflood conditions evolved to braided drainage that later changed further to small meandering channels (Mather, 1993). Later still, river capture occurred, cutting off the original sediment source from the rest of the system, and although this caused a decrease in water supply, it also led to an increase in the water-to-sediment ratio. Although sedimentation was reduced initially, the increase in water-to-sediment ratio initiated minor incision that provided a new sediment supply downstream where channels were easily choked and the system was eventually abandoned. The system has since become rejuvenated due to modern conditions (Mather, 1993). Although there are multiple responses to surface tilting and deformation, other controlling factors may have similar results. For example, although sharp deflections in a river pattern may reflect tectonic deformation, they may also be caused by streams coming into contact with particularly resistant material (Holbrook and Schumm, 1999). Nevertheless, in the absence of tectonic forcing, changes in river pattern, stream avulsion and alterations to channel transverse and longitudinal profiles are all potential signs of other endogenic controls and reflect an alternative geological control of the landscape. As discussed above, changes in base-level (or relative sea-level changes) may cause modifications to a drainage system although there may be a considerable lag time between any affects observed at the coast line and influences on the catchment inland. Uplift of bedrock streams that are exposed to base-level fluctuations can be affected in several ways depending on, for example, near-shore bathymetry and the relationship between the rate of base-level change and wave-base erosion (Snyder et al., 2002). It was also reported that channels will

lengthen over time if rock-uplift rates exceed wave-base erosion; stream-profile concavity and steepness may also be affected. However, as previously discussed, onshore influences of base-level change may be minor and are unlikely to affect the whole of a catchment. Some drainage and morphological features are common to the various tectonic settings. The Corinth Graben, Greece, is an area of rapid subsidence and the region is bounded by asymmetric listric faults and characterised by antecedent drainage, reverse drainage and drainage flowing parallel to faults with strong control by transfer faults (Zelilidis, 2000). In the northwestern Himalayas, India, a collisional tectonic zone, a drainage system typical of such settings has developed many drainage channels parallel to the range confined by antiforms. Antecendent drainage is also typical of settings where ancient rivers have maintained their original course throughout folding and uplift events and the Sutlej River is such an example, although it has been partially diverted and is now trapped within and parallel to the folded ranges (Malik and Mohanty, 2006). A rifted continental margin or passive margin, typically displays a low-relief, highly weathered upland area or range usually with simple landward drainage, and a deeply incised, high relief coastal area demarcated by a seaward-facing escarpment where simple drainage typically flows directly to the coast. Between the range and the escarpment, drainage is typically complex and may be influenced by the location of normally faulted blocks parallel to the escarpment and coast (for example Ollier, 1985; Seidl et al., 1996). High sediment loadings along passive margins may also generate extensive listric growth faulting that may further influence stream position, stream character and sediment accumulation. Passive margins with these characteristic morphological features include the coastal zone flanking the East Brazilian Highlands, the area to the east of the Drakensburg Mountains of South Africa (Ollier, 1985) and the Atlantic Coastal Plain flanking the Blue Ridge escarpment of the Appalachian Mountains, eastern North America (Spotila et al., 2004). In northeast Brazil, recent work (Bezerra et al., 2008) has revealed that geomorphological evolution of the passive margin does not comply with the pediplanation model (for example King, 1956) that assumes uniform regional uplift and concomitant development of erosion surfaces, whilst the uplifted margins remained uplifted following rifting. After rifting, some subsidence commonly occurs driven by conductive heat loss and thermal contraction of the

lithosphere (McKenzie, 1978). This is also generally assumed to be uniform across the rifted margin. Bezzera et al. (2008) recognised that in contrast to the model, the passive margin of northeast Brazil has subsided locally rather than uniformly, and that the sedimentary basins along the passive margins are no longer tectonically inactive. They concluded that the present day compressive regime in northeast Brazil has reactivated faults that have induced the subsidence. As would be expected, other passive margins also show extensive evidence of subsidence following rifting such as in West Greenland, the Antarctica Margin, north of Victoria Land (Bezerra et al., 2008) and the Gulf of Mexico that has continued to subside since rifting occurred in the Triassic and where basin fill has provided one of the most valuable sources of oil in the world (e.g. Sharp and Hill, 1995; Galloway et al., 2000). The Queensland and Marion plateaux off-shore at the northeast Australian passive margin display signs of subsidence far in excess of the expected thermal subsidence (Müller et al., 2000). This has been attributed to accelerated tectonic activity for at least the last 9 Ma and continues at present due to the margin putatively overriding a slab burial ground. The passive margin of the northern South China Sea has a complex geological history similar in some ways to that of Queensland having experienced both accretion and rifting processes. Thermal subsidence followed rifting and as a consequence, large carbonate platforms and reefs that had formed on the new continental edge were drowned (Lüdmann and Wong, 1999) although active tectonics have not been identified in this region as a cause for increasing subsidence. Holdgate et al. (2008) analysed the east Victoria Highlands, southeast Australia, and identified that uplift and divide migration has occurred through the Cenozoic (following the Tasman rifting event in the Late Cretaceous) and continues to the present day. However, this area has a complex history having experienced interactions with several tectonic plates and uplift has been attributed to block faulting and reactivation of basement faults due to the present compressive stress regime. In contrast to subsidence, plate boundaries may be ‘underplated’ through intrusion of large, sill-like magma bodies, such as those in the Karoo province of southern Africa (Cox, 1980). In the Dead Sea transform and rift zone, sediment loading has lead to arching of the Galilee region, northern Israel, which has created a north-south water divide. The widespread and complex uplift and tilting of the landscape has resulted in abandoned and incised channels, reversed drainage and incised meanders (Campbell

et al., 2003). The Maraetotara Plateau, New Zealand, is an extensional setting, where drainage is disrupted by normal and listric faulting, horst ridges, blocks, graben and depressions that cause drainage to be reversed on back-tilted blocks and captured in fault and depression zones, although some antecedent drainage may remain where the orientation has been able to prevail despite tectonic changes in the landscape (Pettinga, 2004). The thrust-faulted terrain of Hawkes Bay, New Zealand, is characterised by surface expressions of reverse and thrust faults, isoclinal and recumbent folds and hills that are complexly deformed. The regional uplift has resulted in drainage patterns aligned parallel to structural grain as well as rejuvenated, incised streams and uplifted and tilted marine terraces. Back tilting near the coast is considerable at up to 30° and gravitational collapse is widespread in the region (Pettinga, 2004). As illustrated, different tectonic settings tend to display a variety of drainage characteristics. However, some characteristics such as terraces and parallel or reversed drainage may occur in more than one setting. Frankel and Pazzaglia (2005) used the model of the passive margin escarpments to assess escarpments developed in active mountain fronts and their work successfully showed that models are not always specific to a given setting. Similarly, some localities clearly display characteristics of more than one setting as a result of complex and varied geological histories involving compression, extension and transcurrent movements. Eastern Australia is a passive margin tectonic setting and the principal geomorphological features include the Great Divide, and approximately parallel to this, the Great Escarpment. In the south of eastern Australia, these two features lie close to one another, but north of Brisbane they run in a north-northwest orientation away from the coast, with the divide moving farther inland and the two become spatially distant from one another. Although the Great Divide is not a significant mountain range by relative height, its position is critical to characterising the coastal drainage of eastern Australia. In southeast Queensland, where the divide is relatively close to the coast line, the position of the Great Divide causes coastal catchments to be small with limited sediment budgets (Jones, 2006a). The Great Divide acts as a drainage divide, separating drainage that flows westward from that which flows to the Pacific (Ollier and Stevens, 1989). Generally, a great escarpment would migrate away from the coast due to erosion. However, escarpment migration may depend on

several factors such as the presence of incising bedrock channels on the escarpment and high elevation relative to the continental hinterland (Spotila et al., 2004). Alternatively, escarpments may degrade under circumstances where streams adjust to base-level changes or where erosion in valleys exceeds that on escarpment-face interfluves (Spotila et al., 2004). The inability of escarpment streams to incise due to small drainage areas would also lead to escarpment degradation rather than migration. The Blue Ridge Escarpment flanking the passive margin of eastern North America is interpreted to have survived for an exceptionally long time, despite climate changes, as it has undergone long-term, slow but steady erosion (Spotila et al., 2004). The morphology of the central Namibian margin is generally characteristic of passive margins, with a well-defined escarpment, inland drainage divide and inland plateau. However, Cockburn et al. (2000) identified that the escarpment has retreated anomalously slowly. They conclude that this is unlikely to be simply due to highly resistant lithologies or climatic changes. They state that there is an important link between the location of the drainage divide and the location of major escarpments and, therefore, proposed that there has been flexural isostatic rebound that has ‘pinned’ the inland drainage divide in place and controlled the location and evolution of the escarpment. Although the great escarpment of southeastern Australia is also well preserved, in the Queensland region it is less well preserved and in some places is cryptic (Ollier, 1982). The rivers of eastern Australia display an unusual pattern, probably due to migration of the Great Divide where westward-flowing headwaters have been captured and reversed (Taylor, 1911). The divide and escarpment are thought to have migrated due to both erosion from the coastal streams and also migration of the axis of uplift away from the coast (Watkins, 1967). The former is probably true: streams may now be incising into the scarp causing westward migration. Migration of the axis of uplift is also possible and this would suggest its position is now to the west of the main scarp, which would cause the streams to its east to more freely flow towards the ocean, perhaps over the edge of the retreating scarp and to the coast. Further examination of this model is undertaken later in this study: topographic and drainage maps show that streams rarely flow directly eastwards to reach the coast in the southeast. If the axis of uplift has now migrated to the west, flow should now be towards the ocean, rather than away from it. Should this be the case, small, low

order, ‘youthful’ streams, would display modern drainage conditions and flow towards the coast. However, as described in paper 3, some streams change direction and flow away from the scarp, which suggests perhaps renewed uplift in this area. Stream networks do not wholly flow to the east in this region unless they are on the low-lying coastal plain. Those on higher ground flow in diverse orientations before finally joining other rivers then flowing to the coast. This suggests that uplift is actively continuing in this region. The fact that small, relatively poorly developed and, therefore, young streams are turning away form the scarp suggests tilting to the east may have occurred in the recent past when the small streams as headwaters started to flow, but have since changed direction due to reverse tilting towards the west. This is further investigated later in this thesis.

Neotectonism in Australia In the past, the general view has been that Australia is relatively ‘tectonically inert’ (Sandiford, 2003). However, work in this area has increased and results show tectonic influence is somewhat greater than previously acknowledged; earthquake hazards in Queensland for example, are much greater than previously thought (Mora, 1996). Twidale and Bourne (2004) more recently confirmed that tectonic forces not only modified the Australian landscape in the past, but continue to do so. The Australian continent is situated within the Indo-Australia Plate, which is presently under compressional stress. Modification of the land will occur in order to accommodate shortening of the continental mass where the crust is weakest. Folds, faults and joints resulting from such tectonic movement, have long been known to produce a variety of distinctive land surface features such as scarps and diverted river channels (e.g. Hobbs, 1904; 1911; Zernitz, 1932; Strahler, 1960; 1966; Twidale, 1980; Scheidegger, 1998; Scheidegger, 2002; Ericson et al., 2005). The orientation of the compressional stress field across the continent varies due to the multiple plate boundaries surrounding Australia (Hillis, 1998; Hillis et al., 1999; Hillis and Reynolds, 2000; Hillis and Reynolds, 2003; Nelson and Hillis, 2005). Intraplate earthquakes typically occur less frequently and at shallower depths than those at plate boundaries, although, relative to other intraplate regions, Australian earthquake activity is moderate to high (Global Seismic Hazard

Assessment Program, 1992-1997: Fig 9; Cuthbertson and Jaumé, 1996; Clark and McCue, 2003).

Figure 9 Stress Map of Australia showing variation of stress orientation across the continent: note stress orientation in the southeast Queensland Region ('Brisbane') is northeast (after Hillis and Reynolds, 2000). Explanation to the key: regime: 'NF' – normal faulting; ‘SS’ – strike-slip faulting; ‘TF’- thrust faulting; ‘U’- unknown faulting regime. Quality ranked as A, highest quality to D, lowest quality (Clark and McCue, 2003)

Geohazard maps may be well-constrained where recent earthquakes can be related to known geological structures and where earthquake monitoring stations are closely spaced. However, evidence is lacking to confidently correlate the geology of Australia with many Australian earthquakes. Consequently, the extent of current tectonic activity along most structures in the region is poorly known (after Zoback and Zoback, 1991; Zoback, 1992). Accurate location of an earthquake requires the event to be recorded by several stations and this is particularly important for small events, such as those typical in Queensland, where accurately locating earthquakes, is a difficult task (e.g. Levshin and Ritzwoller, 2002). An estimated earthquake location may be inaccurate, especially when derived from early records. Therefore, care must be taken when relating them to physiography such as scarps, slopes, hills

or valleys. An earthquake focus (location within the Earth) rarely aligns precisely with surface-features such as mapped faults, scarps and joint systems due to the dip on a fault. Surface expressions of small magnitude earthquakes in southeast Queensland are rare. Nevertheless, as the distance from epicentre to surface-feature will increase as the focal-depth increases, the epicentre and hypocentre of shallow earthquakes are often closely related. The accurate calculation of focal-depth however, is often a challenge (e.g. Kondorskaya et al., 1989; Ogata et al., 1998; Husen et al., 1999; Bondár et al., 2004) and as this parameter may be largely inaccurate, it should be treated with caution. The error margin for epicentre location increases, as the focus depth increases. Therefore, epicentres for the more shallow foci should be more accurately placed. A procedure for more accurately calculating epicentres has been tested and results showed that, for local networks (0° to 2.5°), where at least 10 stations occur within a 250 km range, a minimum of one should be within 30 km in order to provide an epicentre location within 5 km accuracy, with a 95% confidence level (Bondár et al., 2004). Earthquakes in southeast Queensland are typically low magnitude and of shallow origin. The monitoring system is closely spaced to identify the location of very small earthquakes mainly around dams. Therefore, it is reasonable to assume that because the epicentre and focus are spatially relatively close, the error margin for the accurate location of the epicentre is reasonably small. To allow for earthquake location errors, it would be practical to assign earthquakes to a corridor of approximately 30 km width, to describe a zone approximately within which the earthquake epicentres are situated. In southeast Queensland, few well-located events have been identified although it is generally accepted that earthquake activity in the region occurs mainly onshore. Focal mechanism analysis has identified reverse faulting and northeast- southwest compression in the region (Cuthbertson, 1990; Cuthbertson and Murray, 1990; ESSCC, 2006). Large earthquakes in Australia usually occur where little or no recent activity has been recorded previously (Brown and Gibson, 2004), nevertheless, work continues to target regions that experience relatively high recorded seismicity, such as the Flinders Ranges in southern Australia (e.g. Celerier et al., 2005). A multiple tiered approach to modelling Australian-style earthquakes was recommended by Brown and Gibson (2004). They proposed an approach that must include identification and mapping of active faulting on a local scale and

gravity, topography and seismicity surveys on a regional scale. In the USA and in New Zealand, active-fault source-data is included in earthquake hazard assessment and has already proven to be valuable; this may be equally beneficial to Australian hazard assessment (Clark and McCue, 2003). Holdgate et al. (2008) reported that earthquake epicentres in part of the East Victoria Highlands, southeast Australia, show good correspondence to local faults such as the Mt. Beauty and Buffalo faults, which may correspond to the current axis of maximum uplift in the region. Their analysis of the region charted the uplift history of the area and concluded that Cenozoic uplift and tilting by several hundred metres is evident and uplift continues to the present day. Earthquake monitoring in Queensland, by international standards is generally sparse and has only operated over recent decades (Cuthbertson and Jaumé, 1996). However, Queensland earthquakes have been recorded since the late 1800’s (Rynn, 1987). In 1937 the first international monitoring station for Queensland was opened in Brisbane. This was followed by the Charters Towers station in 1957 (Cuthbertson and Jaumé, 1996) and subsequently, a seismic monitoring network developed. After 1977, this expanded considerably, as detailed monitoring was implemented around the large dams, Lake Somerset and Lake Wivenhoe, which were integrated with other seismographs into a state-wide network, monitored by the University of Queensland (UQ) from 1993 (the QUAKES Centre). Funding was withdrawn in 1998 and the operational instrumentation at UQ has been gradually decommissioned. There are 22 Queensland Government seismograph stations that have continued to collect data since 2000 and these are under commercial contract to ES&S (Victoria). Although the monitoring network and historical earthquake database contracts are currently under review by the Queensland State Government, the QUAKES Centre, which evolved into the Earth Systems Science Computational Centre (ESSCC) continues to be involved in the study of earthquake modelling and prediction (Col Lynam at ESSCC, pers. comm. 2006). As discussed, earthquake locations are commonly inaccurate due to the temporal and spatial variations in the earthquake datasets and this is a problem particularly where small magnitude earthquakes are detected by only a few stations. However, the ‘localised’ recording stations proximal to southeast Queensland dam-sites form a concentrated network that allows the

location of epicentres and foci for small magnitude earthquakes to be recorded and reasonably well constrained for the study region. A relatively high incidence of earthquakes in southeast Queensland have been recorded close to the dams. This may be due to the concentration of the monitoring network within the dam vicinity, or they may be reservoir-induced earthquakes that typically can occur due to crustal loading or a change in pore-pressure leading to the premature release of stress (e.g. Gibson, 1997; Scholz, 2002). However, ‘no abnormal activity’ was reported following preliminary analysis and only limited induced seismicity was identified in the vicinity of the monitored reservoirs in southeast Queensland (Cuthbertson, 1995; Cuthbertson et al., 1998). Some slightly higher magnitude earthquakes were located in distal areas (Fig. 10) in alignment with the large fault system that underlies the largest reservoirs in the region. This suggests that the fault system may be releasing stress in locations other than near the reservoirs. The Queensland earthquake database represents only a very short period of time geologically (less than 140 years) and it is possible that the recorded earthquakes may not be a representative sample of all seismic activity in the region. As most of the earthquakes recorded are the result of the more detailed monitoring system now in place, the database may represent a bias towards smaller earthquakes only occurring more recently. Prior to the installation of the new network, fewer small earthquakes may have been recorded as only seismographs situated close to the epicentre would have detected such small magnitude events. Following the installation of the more condensed network, although more very small events have been recorded, the network has not been consistent for very long periods of time. Therefore, it is important to recognise that the database is not a fair representation of earthquakes in Queensland, even over the short term. For evidence of the frequency of all orders of earthquakes and resulting deformation, we must look beyond our monitoring equipment for other evidence of tectonic activity. Considering the size of local fault systems in southeast Queensland, it is possible that the maximum magnitude earthquake that may occur in the region could be 6.5 to 7 magnitude (D. Weatherley, ESSCC, University of Queensland, pers. comm. 2007). However, as previously discussed, it is also possible that a large fault could result from many smaller events where small faults have conjoined over time.

Figure 10 Earthquake epicentres, main drainage, major faults and joint systems in southeast Queensland. Blue circles indicate earthquakes clustered at reservoir locations. Red ovals indicate earthquake clusters on same fault system distally located from reservoirs

Although the southeast Queensland earthquake database contains evidence of earthquake events in the region in the very recent past, physical evidence of these particular events on the landscape surface may be obscure. With respect to aseismic creep, an accurate background deformation rate has not yet been calculated for southeast Queensland and, therefore, it is not possible to identify regions of anomalously high deformation rates (e.g. Linde et al., 1996; Kitagawa et al., 2006). If there has been deformation in the lower crust, it has not been measured.

Post-Mesozoic tectonism in southeast Queensland There is good evidence of surface displacement in southeast Queensland that occurred during the Cenozoic. Substantial subsidence during the Cenozoic led to the formation of some spatially small but significantly deep sedimentary basins flanking Palaeozoic blocks of southeast Queensland. In the Petrie and Oxley basins for example (Fig 11), accumulation of continental sediments has been estimated to exceed 300 m in thickness (Houston, 1967; Cranfield et al., 1976). The subsidence and vertical displacements required to provide accommodation space for these thick sedimentary piles is substantial and faulting may be at least partly responsible for the formation of these basins (Houston, 1967; Cranfield et al., 1976). Mid-Cenozoic volcanic eruptions that constitute the Main Range Volcanics, (e.g. Stevens, 1965; Murphy et al., 1976; Cranfield and Scott, 1993) emplaced extensive basalt sheets that rest upon several raised land surfaces in southeast Queensland (Fig. 12). The occurrence of these highland terrains, for example along the Blackall Range implies peneplanation in the late Mesozoic-Paleogene and later uplift. In the northeast of the region, the emplacement of clusters of Eocene and Oligo-Miocene basaltic and felsic volcanics are most likely to have been associated with tectonism, thermal doming, and crustal loading. Several of the volcanic vents are aligned (see Fig. 12), suggesting structural controls on their emplacement. Sussmilch (1933) identified that the Tertiary age basalts cap both the peneplains at high altitudes and some low lying strata at the level of the coastal plain; from this he concluded that the peneplains have been uplifted relative to what is now the coastal plain and also that the scarp that divides the coastal plain and the highlands is a fault scarp. Further, Sussmilch stated that Dr W. H Bryan pointed out to him that the ‘Neranleigh beds occur both in the Coastal Plain and in the adjoining Mt. Tambourine Horst…’. He concluded that this is evidence that the morphology is not due to differential erosion.

Figure 11 Map of the southeast Queensland region identifying locations of: Cenozoic basins; Wellington Point, North Pine Fault and West Ipswich Fault where Cenozoic faulting was identified (2001); location of South D’Aguilar block in relation to paleocurrent flow direction (red arrows) in the Clarence-Moreton Basin (Jurasssic) sediments (Cranfield et al., 1976); spot heights shown to clarify uplift of South D’Aguilar Block relative to surrounding area

Figure 12 Distribution of the Cenozoic Main Range (and associated) Volcanics (shown in red)

Renewed movement along ancient fault zones has disturbed Cenozoic sediments, such as at Wellington Point, Ipswich (West Ipswich Fault) and Strathpine on the North Pine Fault (Cranfield et al., 1976 p 115);Fig. 11). A small fold and severe disturbance of the Petrie Formation was reported at Strathpine where a series

of north-northwest trending faults were observed with dips up to 60° (Jones, 1927; Houston, 1967). Minor faulting has also been identified in the Oxley Group deposits (Houston, 1967). Palaeodrainage patterns in the Bundamba Group Jurassic sediments of the northeast Clarence-Moreton Basin, flow consistently towards the D'Aguilar Block (O’Brien and Wells 1994; Fig. 11). Parts of the present-day surface of the South D’Aguilar block are approximately 300-600 m above sea level. This implies that the D’Aguilar Block has been uplifted by at least this amount, relative to the surrounding blocks since cessation of Jurassic-Cretaceous sedimentation in the Clarence-Moreton Basin. The geological history of southeast Queensland has been dynamic, substantial changes have occurred since the Jurassic and certainly through Cenozoic time; therefore, the region should not be considered as geologically dormant. Orientation of the Great Moreton Fault system (Fig. 11) through southeast Queensland is aligned with the location of recent, small magnitude earthquakes (see paper 2) and orientation is approximately northwest-southeast. It is obvious both from the length and extent of displacement that the fault is not a result of recent, small magnitude events. As previously discussed, the probability that such events will cause any surface displacement is low. It might be concluded that very large earthquake events in the ancient past caused the large fault system to occur. However, having considered the discussion above, it should not be discounted that many small magnitude and frequently occurring events over a long time frame, could have collectively produced the significant discontinuity and surface expression evident in this region. Although it is possible that a large event did occur in the past, the system may be the result of multiple, small magnitude earthquakes that could individually produce small movements or simply cause rock strength to decrease. Either way, this may encourage preferential weathering along the orientation of the fault, or along associated small joint systems that occur in specific orientations related to the stress in the area. Such planes of weakness would lead to weathering and erosion and encourage aligned drainage along this fault system. Presently active or recent faulting is not necessary for the entrapment of the drainage segments along the faults. The Brisbane River is a large, well developed watercourse and is situated close to, and well aligned with the northern sector of the Great Moreton Fault for most of its upper reaches. This is a typical example of a bedrock-based river

following faults and joints (Howard, 1967; Droste and Keller, 1989; Holbrook and Schumm, 1999), but the location of the Brisbane River along the line of the faults makes no assumption about the size of the earthquake event(s) that caused the main fault or adjoining fault sections. If a large earthquake event was the cause of the Great Moreton Fault, entrapment caused by surface rupture may have been the origin for the river, although today we have no evidence of this. In southeast Queensland, where earthquake activity is low in number and magnitude and where drainage has already developed throughout the area in line with many existing faults, further earthquake activity of low or moderate magnitude is most likely to just enhance existing zones of weakness already exploited by drainage. The location and magnitude of small earthquakes in the region suggests that if similar future events occur, they will not cause surface displacements and, therefore, will not actively entrap new watercourses. The events may provide a basis for the location of the drainage to remain where it already is by increasing the size of joints, faults or rock weakness via which fluid flow can continue to take advantage. Over time, the ongoing occurrence of localised, shallow earthquakes that may align with the presently existing fault and drainage system may result in extending the course of the rivers upstream. Large earthquakes have not been recorded in the recent geological past in this region and are not considered to be ‘characteristic’ at present day, although this should not be taken to mean that large events will not happen in the future.

GEOMORPHOLOGICAL ANALYSIS

Drainage patterns As rock characteristics can effect drainage patterns, it is of no surprise that the interpretation of drainage patterns has proven to be a powerful tool for analysing and better understanding structural geology (for example, Hills, 1960; 1963; Twidale, 2004). Drainage patterns have been studied and used as an instrument for a wide range of theoretical and applied geological investigations (e.g. Strahler, 1966; Schumm and Khan, 1972; Twidale, 2004; Vétel et al., 2004; Delcaillau et al., 2006) and the significance of channel patterns has been reviewed at length (Hobbs, 1911; Zernitz, 1932; Twidale, 2004). Following an extensive review of river patterns, Twidale (2004) concluded that most river patterns are determined by structure and slope, and diversions from and anomalies within these patterns are commonly caused by active faults and folds. This theory has been tested successfully in several projects (for example: Ellis et al., 1999; Schlunegger and Hinderer, 2001; Vétel et al., 2004). Non-random channel orientations may be controlled by endogenic influences such as geological structure or ancient stress regimes (Scheidegger, 1979a; Ellis et al., 1999). Random or dendritic patterns are less likely to be geologically influenced (Zernitz, 1932) and more likely to be influenced by exogenic controls such as climate. Some rivers show correlation and repeating alignment with straight or gently arcuate structural features such as faults, bedding and fold axes to produce drainage systems with a distinctive pattern (Zernitz, 1932; Twidale, 2004). Regular channel patterns (e.g. trellis, rectangular and parallel drainage networks Twidale, 1980) can be a particularly valuable guide to underlying geological structures in areas of poor outcrop. Initially following slope, rivers will later adjust to structure as they incise into bedrock, although, as slope may be controlled by active tectonism, even low order channels, prior to incision, may be geologically controlled. Kirby et al. (2008). have shown that analysis of stream gradients can also reveal underlying structure and seismic hazards. Earlier geomorphological analysis of the Longmen Shan Mountain Range, Sichuan Proince, China, had revealed abrupt

changes in river segment gradients through the region but few other signs of active faulting. The region experienced a 7.9 magnitude earthquake in May 2008. The change in river profile gradients strongly coincided with the fault trace indicating that long-term uplift had preceded the May 2008 event. Kirby et al. (2008) concluded that this is the most compelling evidence to-date that the landscape ‘encodes information about the rates and patterns of tectonic activity’. Such studies may be particularly relevant for detecting hidden or blind faults in areas of extensive soil or vegetation cover. As discussed in a previous section, rock fabric can influence drainage patterns but this is not restricted to regional scales (e.g. Twidale, 1972; Scheidegger, 1979a; Scheidegger, 1979b; Ackermann et al., 1997; Eyles et al., 1997; Eyles and Scheidegger, 1999; Beneduce et al., 2004). Although it has long been assumed that the relationship occurs as a result of zones of weakness in the bedrock becoming enhanced by weathering and erosion processes, most data have been inadequate to confirm this (Ericson et al., 2005). Large numbers of fracture traces may be accurately mapped and measured where exposures are abundant and regolith cover does not obscure structural details (e.g. Scheidegger, 1979b), although this method is less suitable for assessing the structure-drainage relationship in catchments where natural surface features are concealed by regolith, vegetation and human infrastructure. Some studies have analysed the diverse structural influences on drainage patterns. The relationship between joints and channels in parts of a granite- dominated catchment in the Sierra Nevada, USA, was demonstrated using high- quality aerial photographs and correlation was assisted by the high degree of bedrock exposure. Results showed that, in places, this relationship was evident irrespective of regional slope (Ericson et al., 2005). In New South Wales, Australia, a finer scale study has shown a correlation between gully orientation and bedrock joints, and despite the presence of regolith and vegetation cover, results showed the strength of this correlation is strongly dependent on the bedrock lithology (Beavis, 2000). A study in the Macaronesian Islands, revealed a good correlation between stream orientations and ridge trends, which suggests a common tectonic control on the landforms (Scheidegger, 2002). On a coarser scale, in the Chinese Himalayas, tectonic fabric has been shown to exert a first-order control on landscape patterns

(Scheidegger, 1998). In central Italy, slopes and valley trends were found to strongly correspond with Appenine and anti-Appenine lineaments (Alexander and Formichi, 1993). Tectonic controls are also the main influence on valleys, gorges and many other large-scale landforms (Scheidegger, 2001). A relationship between trellis-style channel networks and geological controls was suspected for many decades, and although some workers found this to be weak (1971; Mock, 1976), more recently it has been quantitatively demonstrated (Abrahams and Flint, 1983). It has been suggested that bedrock fractures and metamorphic cleavage are, respectively, the likely controls on the incised meanders of the Shenandoah River, USA, and the channel-and-gorge systems of the meandering River Torrens, Mt Lofty Ranges, Australia (Twidale, 2004). As first described by Powell (1875), the base level is the lowest limit to which rivers can erode – the ultimate base level being sea-level. Channel slope and longitudinal profile, first discussed by Gilbert (1877), might theoretically reach a state of equilibrium where the channel bed was neither aggrading nor degrading. Longitudinal profiles and hill slopes on which overland flow occurs typically erode to form a concave-upward shape (Watkins 1967). However, the idealistic longitudinal profile is rarely reached: hillslopes commonly vary in shape due to ongoing changes in tectonics, fluctuations in sea-level and climate. For example, where sudden, heavy and prolonged rainfall occurs in regions of desiccating heat and heavy rain events, such as in regions of New Zealand and southeast Queensland, hillslopes are more likely to form in a convex upwards shape (B. Ward. 2006, Pers. Comm.). Some geomorphological features such as peneplanation surfaces are typical results of certain palaeoclimatic regimes. Demoulin (1998) concluded that although tectonics can influence a longitudinal profile, it is not possible to assign a specific profile to a specific tectonic regime because so many additional factors also influence the profile, including lithology, climate and history of an area. Meijer (2002) stated that where a longitudinal profile may have adjusted to a base-level previously, if that base-level is reached again, drainage channel alteration will not be equal to the first time that particular base-level was reached, as some of the ‘work’ was already done in the past. This may be similar for any other changes forced on the landscape such as those caused by tectonic processes. Therefore, care should be taken if assigning a longitudinal profile to a particular setting, or if using the affects

of past changes as a forecast for future changes. Fuzzy-logic may be suitable for use in such predictive modelling. Asymmetry of longitudinal profiles and stream cross- sections should also be used with caution as an indicator of setting. Asymmetry of both cross-stream and downstream planes of a river is independent; asymmetry in one plane does not automatically imply asymmetry in the other which indicates that the drivers of each are unique (Rayburg and Neave, 2008). This suggests that, where streams may appear to be symmetrical, some asymmetry may also exist as it can vary greatly with changes in flow stage. An index of cross-sectional asymmetry was formulated by Knighton (1981). However, the index did not prescribe individual cross-sectional types to specific flow and drainage pattern regimes, hence it is difficult to apply. Tectonic adjustment may alter river height; for example if uplift is not met with sufficient down-cutting, a river may broaden and become shallower, or if down-cutting occurs rapidly, incision may narrow and deepen a river. However, a river’s asymmetry may adjust seasonally or with longer term climate change and alone, may not be indicative of tectonic setting. It has been suggested that all of southeast Queensland’s current drainage resulted from headward capture of drainage previously flowing westward and also the migration of the axis of uplift away from the coast (for example, Gasparini et al., 2004). Under modern landscape controls, old streams may continue to take precedence and maintain their original course, and smaller or younger streams may be initially diverted away from an older stream before eventually joining it.

Palaeosurfaces, palaeodrainage and current drainage patterns in southeast Queensland Drainage patterns in southeast Queensland are typically non-dendritic. Many rivers and small streams of all orders display combinations of angular, radial, dendritic, parallel and trellis style drainage patterns and this is explored further in paper 3. Non-dendritic drainage usually indicates that a geological control exists. A study of the Rocksberg catchment, north of the North Pine River, southeast Queensland, revealed that larger channels commonly followed the orientation of both faults and foliar weaknesses in the phyllitic rocks (Arnett, 1971). The study concluded that lithology, vegetation and aspect did not have any significant influence on drainage and slope structure. In the Caboolture region, northeast of Rocksberg (Roy et al.,

1980), strong tributary alignment was identified as being controlled by the location of geological units as well as a zone of relative weakness where greenstones and phyllite are interbedded along the line of a postulated fault zone. Roy et al. (1980) concluded that these geological influences were not apparent at the individual site level of soil-slope associations. It was also evident that soil and slope, and profile aspects, appeared to have no control over drainage orientations (Arnett, 1969; Arnett, 1971; Roy et al., 1980). These examples in the study area indicate that stream orientation and drainage pattern analysis are an important part of assessing geomorphological control.

Stream ordering Before drainage patterns can be validly characterized, stream segments are typically ordered within a network. Each stream ‘order’ is designed to represent a category within a hierarchical relationship of segments constituting a stream network. The ordering system usually designates a number that should be relative to other segments within the hierarchy. The three most commonly used ordering systems were proposed by Horton (1945), Strahler (1957) and Shreve (1967), in which, the uppermost ‘fingertip’ streams are typically designated as first order. Order numbers increase downstream as further tributaries join each segment, although each ordination system ranks downstream segments differently. In some cases the stream order does not reflect the true position of the stream in the network; in other cases, the stream order does not reflect its relationship with other streams of similar position in the network. This can be problematical for some applications but is often overlooked. Ordering systems are discussed in greater detail in a later section. A study undertaken in New Jersey (Ackermann et al., 1997) identified that ‘low Strahler order channels’ followed underlying bedrock structures more frequently than higher order streams. It was concluded that the higher order streams had greater power and, therefore, were able to cut across structures and did not follow the structures as frequently as lower order streams. Although streams of greater power may have the ability to cut across underlying structure, the conclusions made in that study assumed that all higher order streams in the catchment had greater power than all lower order streams. However, the streams designated in that study as ‘higher order’ may not have been topologically equivalent and may not represent

sufficient similarity for such comparison. Furthermore, the Strahler ordering system does not account for the stream power when orders are assigned. The order numbers are relatively meaningless in relation to ‘size’ or ‘power’ of a stream. For example, in a Strahler system, a 4th order stream may have many times greater or lesser power than others of similar rank in the network. Although it would be a reasonable conclusion that a high powered stream may cross-cut a structure whereas a low powered stream cannot, these results should not be discussed in terms of stream orders, simply because no ordering system incorporates stream power within its definitions; at best a stream ordering system can compare relative position of stream segments across a single network. However, even this cannot fully be achieved with the present ordering systems and highlights the need for a new ordering system. Another example of a study in which the misuse of stream orders may have lead to false conclusions is a study by Demoulin (1998) who tested for tectonic control upon longitudinal profiles of rivers in the Ardenne, Belgium. Demoulin measured the profiles of 24 rivers of 3rd to 5th order having used the Strahler ordering system. However, there was no discussion as to why these orders were selected suggesting the only thing these streams had in common was their order number. However, using the Strahler system, 3rd, 4th and 5th order streams can be topologically dissimilar and very different in character. For example, within the 24 streams used by Demoulin, the stream lengths varied from approximately 21 km to 65 km and the stream gradients ranged from 1.61° to 18.44°. As the Strahler stream orders do not identify a particular style, similar type or topological position of each stream in the area, characteristics of the streams such as longitudinal profile may be influenced by different controls to different extents and, therefore, this stream set may not have been a suitable sample for this study. Ideally, streams with similar topological placement in the catchment, would be a better measure if comparing the characteristics of a group of streams to identify a particular control over their longitudinal profiles. A range of characteristics for each stream was measured and used within Demoulin’s (1998) analysis. Although the conclusion correctly identified that it is essential to consider a number of parameters as well as each parameters’ significance, the use of these streams as a relevant set for this analysis is debatable. The value of commonly used ordering systems is discussed further in the ‘Methods’ chapter.

Data analysis Statistical analysis of data requires a suitable methodology. Traditional statistical methods are not applicable to all datasets and other, qualitative, methods are often utilised. Analysis of the landscape by comparing the orientation of drainage with the orientation of geological structure requires comparison of multiple sets of linear features. This may be undertaken on a case-by-case basis by comparing the orientation of streams with the orientation of separate planar rock fabric features such as faults, scarps, hillslopes, cleavage and joints. The orientation of these features may be described as axial data (or double or zero-headed vectors), whereas streams may be described as vectorial data (or uni-directional vectors). Due to the nature of circularly derived measurements, where 0° = 360°, standard statistical procedures cannot confidently be applied to directional data (e.g. Jones, 1968). This causes a problem as, for example, a stream with orientation towards 358° may be close in orientation towards a fault of 2° although statistical analysis would suggest otherwise. The same problem also arises using the orientations as axial data. The problem is then compounded where clusters of faults of multiple orientations are being compared with several hundred channel reaches, also of multiple orientations. Recognition of clear patterns in these datasets has proven unsuccessful in trial analysis for this research. Although statistical analysis of directional data has been explored to some extent in a variety of scientific disciplines where orientation data naturally occur, such as determination of circular averages for palaeocurrent trends derived from measurement of cross-stratification (for example Krumbein, 1939; Krieger Lassen et al., 1994), difficulties remain in identification of multiple trends, groups and clusters that may exist in directional datasets, specifically where multiple clusters are known to exist and where both axial and vectorial data are being used. Although some progress has been made on this subject (e.g. Fisher, 1996; Jones, 2006b), the methods do not lend themselves to the comparison of channel orders across a drainage network where lineations in underlying rock fabric cause multiple clusters controlled by multiple processes. Parametric orientation statistics in relation to earth sciences have been discussed by Kohlbeck and Scheidegger (1985) who noted that statistical methods typically seek to describe a mean value, a deviation from that mean, or a closeness of fit between data sets. For datasets such as those

used in this study, however, it is more important to seek correlations between the datasets, than to find averages and standard deviations within it, in order to analyse the multi-modal (multiple cluster) nature of the data. For the purposes of this study, it would be necessary to compare vectorial stream orientation, which is probably also multi-modal, with axial, planar rock fabric that are known to be multi-modal. A MATLAB® (a registered trademark of the Mathworks, Inc.) script has recently been prepared specifically to deal with the vectorial data, and axial data may be dealt with using the process described by Krumbein (1939). However, comparison of the two types of data is problematical as they both require different treatment. The analysis is further complicated by the known (and unknown) multi-modal aspect of the datasets and ‘general inclusive computer programs’ do not currently exist to analyse such datasets using all potential scripts and processes (Jones, 2006b). For this study, simple graphical analysis, in the form of rose diagrams, is the most utilitarian procedure and provides the most readily understandable comparison of parameters to the reader.

Erosion analysis As previously discussed, surface processes play a large part in the evolution of landscape morphology. Erosion and deposition throughout the landscape leads to commonly slow, but obvious changes. To identify the degree of erosion and erosion susceptibility typical in different parts of catchments or regions, it is common to calculate an index based on several measurable characteristics. Erosion index maps have been developed using various methods worldwide, each with varying degrees of success, although well-tested and reliable input data is the main requirement. One such method of index mapping is the SEIMS network (Soil and Environment Interaction based Mapping System; (Selvaradjou et al., 2007) Using ‘predictors’ including soil organic carbon content, slope, altitude, soil-water storage- capacity, soil erodibility, soil crusting, land cover percentage, rainfall, temperature, and mean annual potential evapotranspiration, a cumulative index or EIS (Equilibrium Index of System) is calculated and used as the Erosion Index. This method was used for Europe as all parameters had previously been calculated and obtained from other sources such as the European Soil Data Centre (ESDAC).

Although the method proved to be suitable for Europe, a full set of similar datasets is not available for southeast Queensland and presently cannot be applied to this region. Soil erosion estimation may be calculated using several erosion models such as the Revised Universal Soil Loss Equation (RUSLE) (Renard et al., 1997), Water Erosion Prediction Project (WEPP; (Flanagan and Nearing, 1995), and the Soil and Water Assessment Tool (SWAT; (Arnold and Allen, 1992; Arnold et al., 1998). The models were designed to quantify the amount of soil erosion from various areas and identify areas that are vulnerable to soil erosion. However, they typically did not address the sediment delivery ratio and, therefore, could not estimate the sediment delivered to a given downstream part of a catchment, hence a further model was developed specifically for this purpose: Sediment Assessment Tool for Effective Erosion Control (SATEEC; (Lim et al., 2005). Additionally, although still a favoured method, RUSLE only calculates sheet and rill erosion, does not predict the effects of concentrated runoff and assumes that rain energy is directly related to erosion yield. The latter may be problematic as soil characteristics, such as texture and permeability that both affect susceptibility to erosion can change temporally, especially where climate and land-use fluctuate. Nevertheless, RUSLE is used widely and calculates:

A=R*K*LS*C*P where: A is the computed spatial average soil loss and temporal average soil loss per unit area (for example tonnes/acre/year); R is the rainfall-runoff erosivity factor; K is the soil erodibility factor (the soil loss rate per erosion index unit is measured on an area defined as 22.1 m long of uniform 9% slope on a continuous clean-tilled fallow); L is the slope length factor (ratio of soil loss from the field slope length to soil loss from a 22.1 m length under identical conditions); S is the slope steepness factor (ratio of soil loss from the field slope gradient to soil loss from a 9% slope under identical conditions); C is the cover management factor (ratio of soil loss from an area with specified cover and management to soil loss from an identical area under identical conditions); and

P is the support practice factor (ratio of soil loss with a support practice such as contouring, strip cropping or terracing to soil loss with straight-row farming up and down slope). Slope and slope-length (S and L) are typically considered together, reflecting the terrain at a given site. Ouyang and Bartholic (2001) developed an on-line GIS based soil erosion prediction method for regions in the USA where digital soil data is available and can be added to the K value of a chosen soil survey. Although some of these values may be calculated for southeast Queensland, C and P values may be somewhat vague as to date, they have not been calculated for the entire region. All erosion indices rely upon accurate and specific data for input and where it is not available, generalised values may not be suitable. Probably the most complete study of the erosion source and processes for southeast Queensland, was a project undertaken by Caitcheon et al. (2005) and presented as a commercial-in-confidence report for Moreton Bay Waterways and Catchments Partnership. The region studied was described as the western catchments and comprised the Bremer, Lockyer and Wivenhoe catchments. The study’s aim was to identify the main causes of erosion and main sediment and nutrient (phosphorous and nitrogen) sources to Moreton Bay, using the SedNet model and further testing using spatial source and erosion process tracing. Hillslope erosion estimation was calculated using RUSLE, although it was stated that most hillslope-eroded soil remains trapped on the hillslope, with little delivered to the stream. Although detailed gully mapping for the region was unavailable, they used calculations for gully density, previously calculated by Prosser et al. (2003) who used a statistical data-mining tool called Cubist. The authors concluded that in areas such as southeast Queensland, where limited data is available, complementary methods should be combined for best results. In addition to the erosion delivery processes being analysed, Caitcheon et al. (2005) analysed nutrient load (phosphorous and nitrogen) sources. Due to the scope of the report, it did not provide an interpretation with respect to the causes of the dominant processes. As discussed earlier, Taylor and Howard (1999) and Montgomery (2003) provided evidence that landscape types may be dominated by specific erosion processes: where chemical weathering dominates, the landscape is most likely to be tectonically quiescent and where mechanical weathering dominates, the landscape may be presently or recently

tectonically active. In their report, Caitcheon et al. (2005) reported that the sediment sources are hillslope, gully and river bank erosion and nutrient sources are dissolved loads in runoff water and point sources such as sewage treatment plants. Where nutrient loads are being mobilised and transported they may be related to, or may include the process of chemical weathering. However, whether the nutrient loads may be caused primarily by chemical weathering is not conclusive. Mechanical weathering is clearly evident in the region and although this may suggest the region is not tectonically quiescent, the results may have been accentuated by anthropogenic influence. Although it is intuitive that chemical weathering would dominate the seasonally moist subtropical landscape of southeast Queensland, to cleary demonstrate whether mechanical or chemical weathering/erosion processes are dominant, the results of Caitcheon et al. (2005) would require further information and analysis to separate the anthropogenic influences on the levels of erosion and the source of nutrient loads. It is necessary, therefore, to use other means to determine whether southeast Queensland has a tectonically active or quiescent landscape, such as the methods used in this thesis. This will provide further information in order to allow future landscape erosion studies to establish the level of anthropogenic influence on the landscape. Soil erosivity indices are valuable for land-use management, planning, conservation and environmental education, and may provide a snap-shot of present susceptibility to erosivity in the region. Where this study ultimately aims to identify the overall geological control of the landscape, an erosivity index is more suitable for measuring the effects of surface processes upon the landscape. An erosion index map, whilst valuable for many purposes, will better describe the effect that surface processes might have on today’s land-surface. Whilst it could be argued that an erosivity index will measure the susceptibility of a geologically controlled landscape to erosion, an index ultimately describes exogenic control and erosion potential at a given point in time. The research completed by Caitcheon et al. (2005), may be a suitable alternative to an erosivity index given the paucity of data in the region. In particular, a full soil mapping database is not currently available for the region and broadly generalised values for some other factors would also be required. Although factors such as slope and slope length can be calculated in GIS, values for C and P in the region are not available for a RUSLE calculation across the entire region; a

calculation using the SEIMS method for example, would be lacking many other factors such as soil organic carbon content, soil water storage capacity and soil crusting. The author is presently engaged in using the Self-Orgainising Map (SOM) method (Kohonen, 2001) within the software SiroSOM (for example Fraser and Hodgkinson, 2008) to assist soil-mapping in forestry areas of southeast Queensland, as detailed soil information is not widely available. This may provide a new method for broad-scale soil suitability mapping across the region and provide useful information for an erosion index map in the future. Presently, however, insufficient data is available to produce a reliable erosion index map for southeast Queensland, based on present methods. Should this data become available, an erosivity index will only describe erosivity potential and is unlikely to describe the degree of geological influence on present landscape morphology: therefore, this method is beyond the scope of this thesis.

SOUTHEAST QUEENSLAND

Introduction to the study area The study area (Fig 13a,b) corresponds broadly to the region covered by the Moreton 1:500 000 Geology Map (Whitaker and Green, 1980). The climate is subtropical, typically with warm, dry winters (March – October) and hot, wet summers (November to February). Rainfall is strongly seasonal and highly variable causing water resources to be limited. Artificial reservoirs are required to provide the majority of the region’s water supply. The largest artificial reservoirs are built in the biggest river valleys in the region and these are typically situated in faulted zones. Seismicity over the past 130 years has been relatively high compared to other intraplate regions although most earthquakes are of low magnitude. Over this period, 56 earthquakes of >2 magnitude were recorded in the region. Seventeen of these were >3 magnitude and two were >5 magnitude (ESSCC, 2006).

Reasons for selecting the study region The primary reason the area was selected for this study is that it has varied morphology, complex geology and a complex geological history lending itself to a

broad scope of geomorphological analyses at varying scales. The region currently constitutes a passive margin tectonic setting although for much of its Palaeozoic and early Mesozoic history it was under compressive stress within a convergent plate margin setting. Although most geological knowledge of the region was gained when it was of greater interest as an ‘unexplored’ area, the region is reasonably well understood geologically and is known to be extremely complex but requires further examination. However, recent and more detailed study has been lacking at a regional and local scale and therefore, is open to further exploration and interpretation. Morphologically, the region consists of plateaux, scarps, rugged and even terrains, lowland, coastal flats and gently sloping hills. This region provides a challenge to the question whether the landscape morphology is strongly or weakly geologically controlled. A less complex region with less morphological variation and more simple geological composition may have more clearly revealed the answer to this question. The current study provides a new and more in-depth assessment of the landscape that developed on rocks that formed in a mixture of accretionary, volcanic and passive margin settings. As this particular study aims to identify whether geological factors control the form of the landscape, the results would be of use to workers studying other areas of comparable geological terrain such as mixed accretionary and passive margins. Brazil and China for example, have complex geological histories and their coastal regions now represent passive margins but they do not conform to the simpler geology of more ‘classical’ passive margins such as the Atlantic Coastal Plain of the United States. This work may provide a basis for future prediction of landscape and coastline alteration that may occur due to both land use and climate change. The southeast Queensland region is generally well understood with respect to its geological history and the region is reasonably well studied with respect to surface processes. However, the interrelationships of these parameters are less well understood and require further analysis. The region provides a basis for multiple scale studies as the variety of rock types, ages, processes and stresses is broad. The study area has also been selected as it is quoted as being one of the fastest growing population centres in Australia (Australian Bureau of Statistics, 2008). Approximately 4.2 million people live in Queensland (estimated 31 Dec 2007), 66% of whom live in the southeast of the state. The region is currently undergoing extensive urbanisation and development. There is

increasing demand on water and other resources, and locating suitable settlement areas for the expanding population is essential. Knowing what drives the shape of this landscape in addition to surface processes, is important for development and future planning. This study will provide a better understanding of how geology controls the landscape of southeast Queensland and in so doing, may provide new insights into the geological processes of the region. Although previous studies show relationships between faults, fracture patterns and stream orientation, this does not imply that all channels or valleys are endogenically controlled. To identify those channels and landscape elements that are geologically controlled, analyses are required both at a local and regional scale. As discussed previously, tectonic processes are responsible for many landscape features, but it would also be valuable to asses the current state of seismicity in the region, as the location of recent earthquakes may identify areas that are perhaps under the influence of recent stress changes. Earthquake risk in southeast Queensland has been assessed based on past events (Granger and Hayne, 2000). For a better understanding of the current tectonic regime, studies to correlate mapped geological structures with the delineation of seismically active zones has been attempted (Cuthbertson, 1990; Cuthbertson and Murray, 1990). As earthquake monitoring is ongoing, assessment of the alignment of earthquake epicentres in relation to known faults will improve, and with continued monitoring and more data, lineaments will be modified and refined (Cuthbertson, 1990); a more up to date review using modern techniques and a larger dataset would be of value.

Figure 13 Location map a) Australia, b) southeast Queensland, c) North Pine River and Laceys Creek catchments and drainage

Geological history The geology of southeast Queensland (Fig. 14) is the result of a complex series of compressional and extensional events from the late Palaeozoic to Mesozoic followed by Cenozoic block-faulting, intrusions and volcanism. During the Late Carboniferous, the Australian continent was part of Gondwana. At this time, in association with a west-dipping subduction zone, the Connors-Auburn Volcanic Arc, an Andean-type volcanic chain, together with a central forearc basin (the Yarrol Basin) and an accretionary prism in the east (Wandilla Slope and Basin) developed (Day et al., 1978; Plumb, 1979; Murray and Whitaker, 1982; Day et al., 1983; Fergusson and Leicht, 1993). The postulated locations of these events in relation to the present shoreline are shown in figure 15. This province formed the northern sector of the New England Fold Belt. Shortening of the accretionary prism and deformation and local obduction of oceanic crust (Day et al., 1978; Plumb, 1979) led to low grade metamorphism and uplift in the southeast Queensland sector of the New England Fold Belt (e.g. Fleming et al., 1974; Cranfield et al., 1976; Holcombe, 1978; Murphy et al., 1979; Murray et al., 1979). The location of rocks that relate to this episode are presently situated as shown in figure 16 During the Early Permian, andesitic volcanism resumed (Day et al., 1978; Day et al., 1983) and the convergent tectonic regime persisted throughout the remainder of the Permian and most of the Triassic, forming the Hunter-Bowen Orogeny. Associated forearc and backarc subsidence allowed the formation of widespread shallow seas and new sediments were deposited on the older, Carboniferous, metamorphosed terranes. Small plutons intruded into the old accretionary wedge and offshore a new subduction zone developed (Willmott, 2004). During the Middle Triassic an extensional event commenced, causing further volcanics and granitic intrusions (e.g. Evernden and Richards, 1962; Webb and McDougall, 1967; Cranfield et al., 1976). Tight folding, metamorphism, uplift and the development of mountainous terrain followed (Cranfield et al., 1976; Plumb, 1979; Willmott, 2004). The present location of rocks emplaced from 286 – 265 Ma is shown in figure 17.

Figure 14a Geological map of Moreton District, southeast Queensland. 1:500 000. Extract of map sheet compiled by W. G. Whitaker and P. M Green, from data available at June 1978, printed 1980. Regional Mapping Section, Geological Survey of Queensland. Key to rock units see next page.

Figure 14b Key to rock units - geological map of Moreton Scanned map also available online at Geoscience Australia : http://www.geoscience.gov.au/geol250k/250dpi/moreton.jpg

During Middle to Late Triassic dextral movement on the Demon Fault, a large meridional transcurrent fault, thought to extend over 550 km, displaced basement rocks by up to 23 km (Wellman et al., 1994). The fault extends from northern New South Wales under the Clarence Moreton Basin and extends to north of Brisbane. From the Late Triassic to Early Cretaceous, extensional epicratonic basins, such as the Nambour, Clarence-Moreton, and Maryborough basins formed and accumulated paludal, braided river and deltaic sediments (Day et al., 1983). During the Early Cretaceous, subduction lead to the emplacement of calc-alkaline volcanics in the Maryborough Basin (Stevens, 1969; Plumb, 1979). This was followed by extension, possibly coupled with a eustatic rise in sea level (Vail et al., 1977; Day et al., 1983) causing large areas, such as the Maryborough Basin, to be filled with thick paludal and deltaic sediments. The locations of rocks that relate to the interval from 265 – to 140 Ma are shown in figure 18. Although it is generally assumed that Gondwana moved northwards during Permian, Triassic and Jurassic times, it has alternatively been suggested that the continent moved southwards during this period, which would have had major climatic and tectonic effects upon the landscape (McKellar, In press). The eastern fringe of Gondwana began to break up approximately 120 million years ago (Veevers and Evans, 1975; Powell et al., 1976; Branson, 1978; Veevers, 2001; Willmott, 2004). Of particular significance to the region, between 70 and 45 million years ago, crustal doming initiated fracturing of the crust along the eastern margin of Australia that lead to the opening of the Tasman and Coral seas. Passive continental margins such as the east coast of Australia are typically characterised by broad, low-relief, high elevation plateaux coupled with a dissected coastal zone. The process of rifting causes uplift, which in turn initiates coastal erosion and this typically creates seaward facing escarpments and coastal plains (e.g. Ollier, 1982; Seidl et al., 1996; Ollier and Pain, 1997). The Great Dividing Range, now referred to as the Great Divide, is considered to be the product of crustal up-warping during rifting. The older term ‘Great Dividing Range’ is rarely used because in many places the drainage divide between easterly and westerly flowing streams it is situated in relatively flat terrain (Ollier and Stevens, 1989). To the east of the Great Divide, or in some places, superimposed on it, lies the Great Escarpment of eastern Australia (Ollier, 1982). Ollier argued that the Great Escarpment can be traced almost continuously along the east coast of Australia. However, in some places the

escarpment is absent or indistinct. Significant uplift probably occurred along the flanks of the Tasman rift system resulting in inversion of the local Mesozoic basins. The Clarence-Moreton Basin consists of infill that was emplaced during the Mesozoic, a time when the present eastern coast of Australia did not yet exist and the continental rocks extended well to the east. At this time drainage in southeast Queensland was mainly towards the northwest. When the basin was inverted after the mid-Cretaceous, the Great Divide was formed across the basin and drainage modifications occurred, such as the capture and reversal of part of the Condamine River by Clarence River (Haworth and Ollier, 1992). Steep escarpments that were established during this uplift event retreated and were later modified by erosion and tectonic activity. In eastern Australia as a whole, it is generally accepted that the Great Divide separates the western area characterised by simple dendritic drainage, from the east where a substantial number of drainage patterns reflect endogenic controls (Ollier and Haworth, 1994; Ollier and Pain, 1997).

Figure 15 Postulated position of subduction zone features in relation to the present southeast Queensland coastline

Figure 16 Present position of rocks 370 - 300 Ma

Figure 17 Present position of rocks from 286 - 265 Ma

Figure 18 Present position of rocks from 265 - 140 Ma

Figure 19 Present position of rocks and sediments from 70 - 22 Ma

Figure 20 Present position of rocks and sediments from 6 - 0 Ma

Rocks emplaced 70–22 Ma are distributed widely but discontinuously across the studied region (Figure 19). Localised subsidence during the Paleogene was responsible for the formation of the small Oxley, Petrie and Booval basins in which clay, limestone, silt, oil shale and basalt accumulated in lacustrine and paludal environments. Erosion during the Neogene shaped the modern, subdued topography of the region (Willmott, 2004). From 30 to 2 million years ago, Australia migrated northwards, and is thought to have moved over one or more hotspots; this caused eruption of localised volcanoes such as the Glasshouse Mountains (Jensen, 1903; Jensen, 1906; Stevens, 1971; Willmott, 2004). Other mid-Cenozoic volcanics were emplaced across the region in places such as the Main Range and the Lamington- Moogerah areas(Stevens, 1965; Stevens, 1966). Mount Warning (also commonly called the Tweed Shield Volcano), which is situated on the Queensland, New South Wales boarder, was originally up to 100 km in diameter. K/Ar dates indicate the eruption occurred between approximately 20.5 and 22.3 (Ewart et al., 1980). The eruptive centres of the Main Range Volcanics are not defined and it has been suggested that lava erupted from multiple vents and fissures (Ewart et al., 1980). K/Ar ages indicate eruption dates of approximately 22.5-24.5 M.y. (Webb et al., 1967). The Maleny Basalts at the Blackall Range have a broader eruptive date range of 21-34 M.y. although an approximate age of 25.2 M.y. has been obtained from a megacrystal andesine and anorthoclase from one of the youngest flows (Ewart et al., 1980). The Blackall Range is isolated from the Main Range volcanics and Mount Warning but is in close proximity to the Glass House Mountains for which, a date of 25.4 M.y has been obtained from K-Ar dating (Webb et al., 1967). Ewart et al. (1980) proposed a genetic link between the Glass House Mountains and the Maleny Basalts. Neogene deposits are distributed throughout the coastal area (Figure 20). Between approximately 6 million and 400,000 years ago, small basaltic volcanoes erupted in the Bundaberg and Gayndah areas, although their origins are not well understood (Willmott, 2004): their age-trend does not conform to the southward younging of other hotspot-related volcanism of eastern Australia (Robertson, 1985; Sutherland, 1985; 2003). Since that time, both erosion and deposition continued and alluvial deposits and shallow marine have accumulated on flood plains, deltas, estuaries, spits, sandbars, coastal dune systems and back-barrier lagoons. The coastal plain is largely underlain by Devonian-Carboniferous Neranleigh-Fernvale Beds to

the south of Brisbane, and Triassic to Jurassic mudstones and sandstones of, for example, the Kin-Kin Beds in the north and Landsborough Sandstone in the central coastal region.

Faulting Little has been published regarding age constraints on fault activity in southeast Queensland (Humphries, 2003). Recent geotechnical core logging for civil engineering works in southeast Queensland has identified some discrepancies in the published geological maps (Brisbane City Council, City Design, Ground Engineering, pers. comm. 2006). Of particular importance is the Buranda Fault (Bryan and Jones, 1954) that may have caused up to 11 km of sinistral terrain displacement through the Brisbane Gap following a line close to the present course of the lower Brisbane River. The position and even the very existence of this fault has long been a source of contention. Nevertheless, the proposed location of the Buranda Fault marks an important discontinuity between metamorphic rocks in the southeast and northwest of the Brisbane region. Evidence of slickensides, fault breccias, and ‘abnormal strikes’ were given for the location of the fault (Bryan and Jones, 1954; Hill and Denmead, 1960 p.134) although its existence appears to have since been questioned as the fault is no longer included on modern geology maps, such as the 1:500,000 Geology of Queensland Map (Geological Survey of Queensland (2003). The ages of most major faults in the region are only loosely constrained by the minimum ages of their bounding strata. Murray (1988) suggested that the Gympie Province (Fig. 13), together with New Caledonia and part of New Zealand once belonged to a single Permian volcanic arc complex, which was later broken apart by large-scale strike-slip faulting. This process emplaced the Gympie Province tectonically to its present position in about the Middle Triassic. Murray et al. (1987) reviewed the Geology of the Gympie district and proposed that the North Pine Fault is potentially a terrane boundary. They further suggest that the northern section of the fault is veiled by the Esk Trough (Fig. 13), although there is little evidence for this and they also recommended more work on this subject may better constrain the geological history of this area.

Sea level influences Approximately 2 million years ago, when sea level was up to 120 m below the present level during ice-house conditions, the eastern Australian shelf was exposed, weathered and incised and as ice melted, sea levels rose and drowned the lowlands. Williams et al. (1998) suggested that 6500 years ago sea-level was 1 m higher than at present and Flood (1981) suggested that present sea-level was reached about 3000 years ago. However, Ward and Hacker (2006), working on the Brisbane coastal region, proposed that the sea reached its present level between 6500 and 6000 years ago. They described the sediments of the area as ‘an alluvial landscape veiled by marine sediments’ and discussed the evolution of area in great detail. Their work also identified oscillating levels of the shore-line in the past 6000 years around the Brisbane Airport region and they assign these adjustments to relative changes in sea- level due to settlement and drainage of soft ground.

Terraces Some rivers in southeast Queensland display terraces and incision that represent responses to changes in relative base-level, which may have been caused by eustatic sea-level change or local crustal uplift. In the Gold Coast region, for example, an extensive study identified over 15,000 ha of lower and upper river terraces, the latter of which included 627 ha of older age terraces greater than 10,000 years old (Whitlow, 2000). The older terraces are found in the lower Logan and Albert rivers and parts of the Nerang valley, for example. The younger Pleistocene upper terraces, consisting of gravel, sand, silt and clay, account for 36% of alluvial deposits in the Gold Coast region and are found in places such as the Coomera and Nerang valleys and the Pimpama cane-lands. Some 7,700 ha of Pleistocene, low level alluvial terraces, consisting of gravel, sand, silt and clay grading into floodplain alluvium were reported as accounting for 40% of Gold Coast alluvial deposits and are found for example in the lower Logan and Albert rivers, Hotham Creek and the Currumbin Valley. The terraces provide economic opportunities for gravel extraction and commercial extraction presently occurs in the region, for example in the Upper Coomera and Pimpama terraces. A similarly broad study for the rest of southeast Queensland is not yet available although other gravel terraces are evident in similar settings to those described in the Gold Coast region. Terraces north of Brisbane such

as on the Pine River at Dohles Rocks, North Pine River at Four Mile Creek and Lawnton and also west of Petrie, and the South Pine River at the Strathpine Flats and southwest of Bald Hills, have all been described in some detail by Hoffmann (1980). Ages of the terraces in the region are vague but Hoffmann proposed an age of approximately 120,000 to 130,000 years for the Strathpine Terrace when sea-level was approximately 5 m higher than at present. The 25 m DEM used in the present study is too coarse to derive cross-sections of small streams resolving the details of the terraces identified by Hoffman. However, Hoffman presented detailed cross- sections of several reaches of the North and South Pine rivers showing the extent and variation of deposits in the lower Pine River valleys and terraces identified in the area. The Strathpine Terrace is at an elevation of 5 m near the North Pine River mouth and the terrace extends along most of the North Pine River. The terrace has been incised by up to 18 m and Hoffman suggested a general northward migration of the ancient Pine River during deposition of the Strathpine deposits. The Lawnton Terrace is approximately 6 km from the present shoreline and lies between Four Mile Creek and the North Pine River. The terrace is approximately 2 to 6 m in elevation. Some terrace remnants occur along the South Pine River. Hoffman’s review concluded that the Strathpine deposits are the result of larger streams during the late Pleistocene that were later benched and incised by subsequent fluvial erosion. Terraces across southeast Queensland represent the locations of previous flood plains and their abandonment and incision indicates a change in fluvial regime and relative sea-level fall although as discussed previously, uplift and eustatic sea-level change may both have been involved.

Incision Another indication of stream habit and morphology changing as a response to relative sea-level fall is the down-cutting and incision of meandering streams, whilst the meanders are retained. Incised meanders are present in many catchments and typical examples include low order streams such as Terrors Creek at Ocean View on Mount Mee, and high order rivers such as portions of the lower Brisbane River. Figure 21a,b shows a cross-sectional view across a tight meander in the lower Brisbane River and reveals the incised nature of this part of the river; the typical level of the coastal plain being 8-10 m above normal river level in this area. The banks are relatively steep and

only slightly asymmetrical – features typical of relatively rapid incision. Hoffman (1980) made a similar observation of incised meanders into bedrock in the Pine Rivers area and concluded it was superimposed from a Cenozoic or Mesozoic erosion surface. Incised meanders indicate rapid relative fall in base-level (B. Ward pers. comm.) and although this may have been eustatic sea-level fall, it may also have been due to tectonic uplift (for example Gardner, 1975; Campbell et al., 2003). Although Hoffmann suggested an age of 120,000 to 130,000 years for the Strathpine Terrace, he argued this on the basis of the known age of a sea-level highstand at this elevation. His argument was counter to Beckmann’s (1959) provisional suggestion that the terrace was only 50,000 to 30,000 years BP at which time sea-levels are now known to have not been high enough. However, a more definitive age of sediments would be required to fully rule out the possibility that the terrace formed as a result of uplift and confirm that it was a result of eustatic sea-level change. Although the 25 m DEM used in this study was too coarse to derive fine-scale detail of, for example, minor terraces, the DEM is able to display coarser detail and has been used here to generate the longitudinal profiles of the Brisbane and North Pine rivers as examples of both long and short catchments in the region (Figs 22 a-c). Hoffman (1980) suggested that the hypsometric curve of the Pine River drainage basin indicates that the basin has almost reached the final stage of landscape evolution. However, this hypothesis is somewhat over-simplified, as the hypsometric curve (Fig. 22c) does not show a smooth sigmoidal curve but in contrast, shows a series of short drops and shelves. These are likely to have been caused by the variation in bedrock types and local faulting, and in the lower reaches may also be eustatically influenced. Similar analyses of the stages of landscape evolution for all other river systems in southeast Queensland have not been undertaken. As the conclusion regarding the Pine River system, reached by Hoffman (1980), was oversimplified, further study on other rivers in the region is required to identify local, regional and global influences.

Figure 21a Location of A-A' meander cross-section superimposed on 25 m DEM of region

Cross section across A-A' meander - Lower Brisbane River

10 m. a.s. l.

0 0 500 1000 1500 2000 2500 3000 3500 4000

Figure 21b Cross section A-A' across meander showing incision status of meander in the Brisbane River

Longitudinal profile of Upper Brisbane River to lake

700

600

500

400

300 m. a. s. l. 200

100

0 0 5450 10900 16350 21800 27250 32700 38150 43600 49050 54500 59950 65400 70850 76300 81750 87200 92650 98100 103550 distance from top of river to lake

Figure 22a Schematic longitudinal profile of the upper Brisbane River from headwaters to mouth of Lake Somerset

longitudinal profile of Brisbane River from lake to shore

90 80 70 60 50 40

m. a. s. l. s. m. a. 30 20 10 0 0 20000 40000 60000 80000 100000 120000 140000 distance from lake

Figure 22b Schematic longitudinal profile of the lower Brisbane River from Lake Wivenhoe to the shoreline

longitudinal profile of North Pine River source to sea

140

120

100

0 0 5000 10000 15000 20000 25000 30000 35000 40000

distance along river

Figure 22c Schematic longitudinal profile of North Pine River, Southeast Queensland

Catchments in southeast Queensland are typically 30-60 km in length, although the Logan, Bremer and Brisbane River catchments are more than twice this length. The average catchment length in southeast Queensland is approximately 68 km and the shortest are the North Pine, South Pine and Caboolture rivers each of which, is approximately 30 km in length. The Brisbane River catchment is over 160 km long: above Lake Wivenhoe the river has an average grade of 0.57116% but below the lake its gradient drops to an average of 0.059771% grade. The average

gradient of North Pine River is 0.336838%. The upper 7000 m average gradient is 0.76861%; the middle 18000 m average gradient is 0.20174% and the average grade of the lower 11000 m of the river is 0.15899%. The steep upper catchment of the Brisbane River in comparison with the less-steep lower portion reflects the high topography of the D’Aguilar range that the upper catchment drains. The steep but very short North Pine River is also a product of the uplifted South D’Aguilar block, over which the upper and middle river flows. The middle section has a less-steep gradient reflecting the less resistant and more eroded rock types such as the Bunya Phyllite, juxtaposed to the more resistant rocks of Mount Mee (e.g., the Rocksberg Greenstone). Typically, the lowest order streams in southeast Queensland flow over regolith but middle order streams in the steeper parts of the catchments have incised to bedrock. Some less-steep middle order streams such as the Laceys Creek main trunk, are ‘armoured’ with very coarse gravel to boulder sized clasts and higher order streams have beds of finer sediment. Streams that have incised into bedrock contain relatively young, mobile sediments (Hofmann, 1980). Controls on the orientation of streams in the Laceys Creek catchment is explored in detail in paper 1. Where southeast Queensland streams are not confined by incision and are able to meander, they are typically asymmetrical due to the nature of meandering streams, although even along some straight sections such as on the North Pine River near Dayboro, the river valley is also asymmetrical (Fig 23). In this case, the asymmetry is most likely caused by the juxtaposition of rocks of differing erosional resistance and in particular, the river valley has preferentially eroded away from the North Pine Fault at this location into softer rock.

Section NE to SW across Upper North Pine River near Dayboro

120 North Pine River 100

m. a. s. l. 40

0 0 500 1000 1500 2000

Figure 23 Section across North Pine River near Dayboro showing asymmetry – the North Pine Fault is situated approximately on the northeast bank of this section although surface expression of the fault has not been seen at this location. Asymmetry is most likely caused by the location of the fault and differential weathering of juxtaposed units controlled by the fault.

Geology of Pine Rivers and Laceys Creek: a fine-scale case study The North Pine River (Fig 13c) drains the northern two-thirds of the Pine River Drainage Basin: the southern one-third is drained by the South Pine River. As stated above, Hofmann (1980) described the hypsometric curve of the Pine River Drainage Basin and defined it as having almost reached the final stage of landscape evolution (the monadnock stage, after Strahler 1957). Major streams in the Pine Rivers catchments commonly meander, although the majority of the meanders are incised into bedrock, suggesting that the stream pattern is superimposed from an older erosion surface that has since been uplifted causing down-cutting faster than the stream can adjust its drainage pattern. Hofmann (1980) suggested this may have continued from a Paleogene or Mesozoic erosion surface. Nevertheless, he also stated that the drainage pattern is fault-controlled and identified that the North Pine River strongly follows the course of the North Pine Fault. Part of the North Pine River was dammed in 1976, forming a reservoir, Lake Samsonvale. The Pine Rivers catchment drains an area exposing diverse rock units that include the: Rocksberg Greenstone (11% of area); Kurwongbah Beds (5% of area); Bunya Phyllite (34% of area); Neranleigh-Fernvale Beds (21% of area); hornfels (1% of area); Permo-Triassic volcanics and Triassic granitoids (13%); Landsborough Sandstone (5%); Petrie Formation (10%) and Cenozoic basalt (<1%) (Hofmann, 1980).

Laceys Creek is the largest subcatchment of the North Pine River system. It is situated approximately 50 km north of Brisbane, southeast Queensland (Figs 13 b,c) and is located upstream of the artificial Lake Samsonvale. Lake Samsonvale is an important part of southeast Queensland’s reticulated water supply. Laceys Creek catchment is located within the South D’Aguilar Block of the New England Fold Belt. The block is characterised by north-northwesterly trending geological assemblages representing a volcanic arc, backarc/forearc basins and subduction complexes (Murray et al., 1987; Coney, 1992; Little et al., 1992; Holcombe et al., 1997b,b; Betts et al., 2002) and consists of steeply dipping, north-northwesterly to south- southeasterly striking, pre-Permian meta-sediments and meta-volcanics (Denmead, 1928; Belford, 1950; Bryan and Jones, 1962; Tucker, 1967; Wilson, 1973; Cranfield et al., 1976). The two dominant geological units in the Laceys Creek catchment are the Bunya Phyllite and the Neranleigh-Fernvale Beds, which are of similar sedimentary origin, but have different fabrics and metamorphic grades. The

Rocksberg Greenstone is overlain by the Bunya Phyllite, which crops out in a north- northwest-trending belt, situated on the southwest flank of the Rocksberg Greenstone in the Pine Rivers catchment: the transition between the Rocksberg Greenstone and Bunya Phyllite is represented by the intercalation of meta-volcanics and meta- sediments. The Bunya Phyllite is strongly foliated, consists of arenites and lutites metamorphosed to greenschist facies, and is a dominantly pelitic suite of metasediments with some intercalated meta-basic volcanic rocks. Bands of alternating quartzose and micaceous rock are crossed by veins of quartz with accessory graphite and calcite. Grain size is typically less than 0.01 mm. The original sediments of the Bunya Phyllite have been interpreted as marine deposits of a relatively deep-water environment. They may represent the oceanward facies equivalent of the Neranleigh- Fernvale Beds and may be older than Carboniferous in age (Cranfield et al., 1976). The Bunya Phyllite forms rugged slopes within the Laceys Creek catchment, with elevations of 80 to 420 m a.s.l. and covers an area of 12.9 km2.

The Neranleigh-Fernvale Beds cover an area of 68.9 km2 in the Laceys Creek catchment (Fig. 13c). They are exposed against the southwestern margin of the Bunya Phyllite and they form rugged hills up to 100 to 660 m a.s.l. Although an accurate age for the Neranleigh-Fernvale Beds has not been determined, it is placed between the mid-Devonian (Fleming et al., 1974) and mid-Carboniferous (Green, 1973). The unit was subjected to low-grade metamorphism towards the end of the Carboniferous (Cranfield et al., 1976). The Neranleigh-Fernvale Beds consist of conglomerates, radiolarian cherts, argillaceous rocks, arenites, basic volcanics and minor limestone (Cranfield et al., 1976). Within the Pine Rivers area, thinly bedded siltstone and shale are widespread, In some places, these are sheared and phyllitic. Arenites are also well developed in the area. Some original bedding is well preserved and cross-bedding is also common. The lithologies and bedforms suggest that deposition occurred in a relatively deep-water, marine environment that experienced periodic turbidity flows. The Neranleigh-Fernvale Beds have been regionally metamorphosed to greenschist subfacies, but to a lower grade than that of the Bunya Phyllite (Winkler, 1967). Triassic intrusions such as the Mount Samson and Samford granodiorites thermally metamorphosed the surrounding Neranleigh-Fernvale Beds. The exact stratigraphic relationship between the Neranleigh-Fernvale beds and the Bunya Phyllite has not been clearly established. The Neranleigh-Fernvale Beds are

considered to have been thrust over the Bunya Phyllite and a major shear zone has been located along their contact (Cranfield et al., 1976). Although several north- northwest trending fold axes are mapped (Denmead, 1928; Belford, 1950; Cranfield et al., 1976), precise structural relationships are not clear (Cranfield et al., 1976). The pre-Permian rocks were folded along north-northwesterly trending axes during the New England orogenic event in the Late Carboniferous and further deformation occurred during the Late Permian and Early Triassic. This was followed by block faulting and the emplacement of granitic intrusions, several of which are located in the North Pine River catchment. Further folding took place along north and northwest trending axes during the late Middle Triassic (Cranfield et al., 1976; Holcombe et al., 1993, 1997a). Due to the multiple episodes of deformation that affected the area, diverse bedding, cleavage, joint and fault orientations have developed within the units. During the Paleogene, small intermontane sedimentary basins, formed through eastern Australia, such as the Petrie Basin in the North Pine catchment, within which, approximately 300 m of sediment accumulated; the Petrie Basin may be partly erosional and partly tectonically controlled (Cranfield et al., 1976). Sediments of Paleogene age are largely undisturbed although rejuvenated faults such as the North Pine Fault (Cranfield et al., 1976) have locally tilted bedding. The close proximity of the highlands to the coast provides short and typically steep catchments. On steeper parts of the catchment, a thinner weathering profile has formed. Erosion may occur on steeper or more exposed areas, causing a thinner or complete lack of a weathering profile, regolith or soil. Streams developing in these areas will form on or in close proximity to bedrock and orientation will be more likely influenced by the lithology. Behind the relatively steep coastal rise, the hinterland is moderately dissected by faults and block boundaries, lending themselves to locations of preferential drainage. Streams in that area are, however, generally less steep than the coastal drainage catchments and may accumulate more sediment and be prone to more ground cover. However, the rock fabric may still bear an influence over the drainage pattern due to the variation in ground cover and age of drainage channels.

Previous geomorphological studies of southeast Queensland A range of geomorphological studies of southeast Queensland have been undertaken although these have been conducted mainly on a piecemeal basis (e.g. Taylor, 1911;

Marks, 1933; Watkins, 1967; Arnett, 1969; 1971; Donchak, 1976; Beckmann and Stevens, 1978; Lucas, 1987; Cuthbertson, 1990; Childs, 1991; Ollier and Haworth, 1994). In summary, previous work has described southeast Queensland as consisting of foot hills and coastal plains to the east and plateaux and highlands (over 300 m a.s.l.) in the west, south and north and escarpments common across the whole region; the DEM shows the distribution of the high and lowlands quite clearly (Fig. 24). The main drainage systems have been described as displaying strong trends of northwest- southeasterly and northeast-southwesterly that are similar to the orientation of many faults (Fig. 25).

Figure 24. 25 m DEM of the study region showing the main drainage and key locations

Figure 25. Faults (brown lines) and main drainage (blue lines) in southeast Queensland

Using Landsat images of the northern part of the region, Childs (1991) showed that the faults correspond with channel orientation and the main ranges and drainage systems are strongly concordant with the bedrock geology. However, some major faults indicated on geological maps could not be identified on the Landsat images and, therefore, may lack surface expression (Humphries, 2003): they may have unfavourable illumination for Landsat (Childs, 1991) or it may be due to the scale at which the images were processed. Taylor (1911) observed a clear association between regional river patterns and structural geology. He identified that headwaters of westward-flowing streams were captured and reversed, due to westward migration of the Great Divide in eastern Australia. To the east of the Great Divide in Queensland, rivers typically flow parallel to the coast along major structural lines (Beckmann and Stevens, 1978). Endogenic control is further observed where Cenozoic volcanism has disturbed part of the drainage pattern. For example, in the Clarence River system drainage pattern, eastern Australia, stream reversal is postulated, on the basis of the orientation of barbed tributaries suggesting that the Clarence River possibly once flowed north joining what is now the Condamine River (Haworth and Ollier, 1992). South of Brisbane, in New South Wales, the Coastal Range acts as a barrier and prevents drainage from crossing the range from west to east (Ollier and Haworth, 1994). South of Brisbane, the Great Divide, the Great Escarpment, the coastline and the continental shelf trend south-southwest, whereas these trends change in orientation at the Clarence-Moreton Basin and north of Brisbane are aligned north- northwest (Ollier, 1985; Ollier and Haworth, 1994). Ollier (1982; 1985) and Ollier and Stevens (1989), described the Great Escarpment in detail and they identified that the escarpment controls drainage in eastern Australia. Ollier (1982) also identified that despite the Great Escarpment being mainly continuous throughout the length of the east coast of Australia, it appears to be ‘absent or obscured’ in some parts, particularly in southeast Queensland. Smaller shore-parallel escarpments have also been identified in the Gympie region. Late Miocene to early Pliocene uplift in the region initiated the development of an erosion surface that was tilted in the late Pliocene, forming northwest dipping cuestas (Murphy et al., 1976). On the eastern edge of the cuestas and continuing close to either end of the Maleny – Mapleton reach of the Great Escarpment, two low, coastal scarps formally named the Glass House Scarp and the Como Scarp (Coaldrake, 1960; 1961) have been described as old

coastline features; each are described as forming local drainage divides (Murphy et al., 1976; Cranfield, 1994). Coaldrake (1960) described the scarps as being a clear break in the pattern of soil and drainage and stated that although both incise into Mesozoic sandstones, the two scarps are of different ages: he suggested that the Como Scarp is a stranded Pleistocene shore line, and that the Glasshouse Scarp is older, suggesting that neither are part of the Great Escarpment. Tilting of the erosion surface may have been responsible for the reversal of some small creeks in the Mary River catchment (Murphy et al., 1976). However, no mechanisms for the late Miocene to early Pliocene uplift and late Pliocene tilting were proposed by Murphy et al. (1976). Further geological control on drainage is evident in other parts of eastern Australia, such as the Victorian Central Highlands where Cenozoic basaltic eruptions covered and preserved a paleodrainage network where streams clearly altered direction where they encountered a system of parallel and intersecting normal faults (Holdgate et al., 2006). The interpretation of remnant landforms has led to a better understanding of landscape evolution through deep time in some regions. For example, the catchments of the Fitzroy and Burdekin Rivers are now known to have enlarged towards the west during the Cenozoic, capturing other streams, which caused an increase in sediment transport to the coast (Jones, 2006a). The movement of catchment boundaries over time may have enormous implications for hydrogeology, groundwater chemistry, and soil and regolith properties (Ollier, 2001). An interpretation of Brisbane River’s physiography and its surrounding catchments suggested that the escarpments were shaped by erosion and that structural lines of faulting, jointing and zones of preferential weathering in the area, coincide with drainage patterns to some degree (Marks, 1933; Beckmann and Stevens, 1978). In an extensive geomorphological review of the Moreton District, southeast Queensland, Sussmilch (1933) discussed river channel positions, although not drainage patterns per se. Sussmilch described the general geomorphology and also discussed the relationship between the complex series of horsts that separate the eastern coastal plain from the continuous high western plateau. Sussmilch also first described the Brisbane Gap, which is a low-lying division between the Beenleigh (then ‘Tambourine’) and D’Aguilar Blocks. Sussmilch observed that the main drainage systems that flow through the Brisbane Gap, specifically the Brisbane and Logan rivers, do not flow along the lowest part as might occur if the division between

100

the two blocks was simply the result of erosion; the two rivers respectively flow along the northern and southern edges of the gap and both are incised into the bedrock. Between these is Tingalpa Creek, which is a relatively short coastal stream The southern margin of the Brisbane Gap may be an east-west fault, previously suggested by Denmead (1928) and its northern boundary is a fault-scarp along the southern margin of the D’Aguilar Block. Sussmilch (1933) further stated that although the Stanley River starts close to the coast, it flows in a general southwest direction to join the Upper Brisbane River near Esk. However, he did not suggest any explanation for this anomalous drainage pattern. For the upper Brisbane River catchment, it has been proposed that, due to back-cutting of the Stanley River during Late Miocene-Early Pliocene times, drainage was reversed, potentially having been assisted by minor tilting to the west, although no evidence to support this hypothesis was given (Beckmann and Stevens, 1978). Beckmann and Stevens also suggested that several coastal rivers that now discharge directly into Moreton Bay, including the Caboolture and Pine Rivers, may have previously flowed into the Brisbane River when sea-level was much lower than at present. Landsat images of the Sunshine Coast area reveal many straight, arcuate and sub-arcuate patterns in landform lineaments (such as scarps, highlands and valleys), that do not coincide directly with the location of lithological units when compared with geological maps (Lucas, 1987). Although differential erosion is an important factor of landform development in the Brisbane region, Marks (1933) undertook a review to ascertain whether other explanations for the location of hills and valleys, and river patterns in particular were viable. The results showed that some drainage divides exhibited ‘a complete disregard’ for the underlying geology and that some rivers clearly follow the site of fault zones. From the above précis of some of the work already conducted on geomorphology in southeast Queensland, it is clear that although there is patchwork evidence of strong geological controls on drainage and topography in the region, a thorough and integrated understanding of the extent to which the landscape is endogenically controlled does not yet exist. This thesis aims to enhance understanding of the physical controls on the evolution of the landscape in this region.

101

ANALYSIS METHODS USED IN THIS STUDY

The main methods used in this study include spatial analysis of datasets on a geographic information system (GIS), stream pattern analysis and stream ordering; basic field work was also undertaken to map geological structures and to ground-truth previously mapped geological and topographic features.

Digital elevation models (DEMs) Datasets were obtained from public sources including Queensland Government (2003; 2005), Pine Rivers Shire Council (2004), NASA (2004), Geoscience Australia (2006), and ESSCC at the University of Queensland (2006). The topographic datasets available early in the project from local sources (DME, Geoscience Australia, Pine Rivers Shire Council) proved to be unsuitable as they were mainly of coarse resolution. Furthermore, the resolution varied spatially (from 10 to 50 m contours) across the southeast Queensland region. The DEMs analysed early in the project were initially derived from traditional topographic maps and errors were found such as one contour that was marked as 50 m and 100 m in separate locations. Therefore, these maps were deemed unreliable and not fully digitised. At best they provided only a broad digital model of the terrain. In order to integrate elevation data with other datasets in a GIS, a digital elevation model (DEM) is required. To perform multi-scale spatial analysis such as the measurement of orientation of channels or to extract detailed longitudinal profiles or cross-sections of rivers, a reliable and high-resolution DEM is required. Satellite Radar Tomography Mission data (SRTM: NASA, 2004) became available from NASA during the course of the project providing a consistent and, therefore, more suitable digital elevation model for the region. Nevertheless, the resolution was still only 25 m and could not be relied upon for computer-generation of a stream-network at the sub-catchment scale required. Therefore, a full stream network was manually digitised and later ground-truthed and adjusted as necessary, in order to ensure the network was of sufficient quality prior to performing detailed analysis upon it. Later in the study, a more reliable digital dataset for the topography of the region was obtained (Queensland Government, 2005) and replaced the use of the SRTM data for the latter part of the research as it covered a broader area, was

102

more flexible within the GIS programs and less memory intensive. The two elevation models were comparable in quality.

Geological data Additional digital maps of southeast Queensland were obtained from the Department of Natural Resources, Mines and Water (Queensland Government, 2003) that provided coarse-resolution geological and drainage features. Detail and accuracy of these maps limited their use depending on the scale required. Although geological data is well represented, recent core logging suggests that there are significant inaccuracies in the current geology maps (pers. comm. Brisbane City Council, 2006). Nevertheless, it was the most complete geological representation available for the area of study at a scale that was suitable for this work. The geological dataset included fault, fracture and cleavage measurements across the region. However, the data was focussed in disparate areas and did not include sufficient measurements in the Laceys Creek catchment in particular. Therefore, it was necessary to expand the dataset and field work was undertaken in the Laceys Creek catchment to provide additional rock fabric orientation measurements for cleavage, joints and fractures.

Earthquake data Earthquake epicentre data available for southeast Queensland from Geoscience Australia (2006) is relatively limited, and for the purposes of this study was augmented by a more comprehensive dataset available from Earth Systems Science Computational Centre at The University of Queensland (ESSCC 2006). The majority of earthquakes in southeast Queensland have low magnitudes (< 2 M). The dataset included all reported seismic ‘activity’ although for some of the reports, the epicentre was uncertain and, therefore, not included in the analysis for this study. For completeness, the remaining data were used in their entirety irrespective of magnitude.

Geographic Information Systems (GIS) and choice of GIS products A geographic information system (GIS) is a computer-based tool that is designed to allow geographically referenced data to be captured, stored, displayed and edited.

103

Spatial information can then be integrated and analysed within the system. MapInfo and ESRI® ArcMapTM are commonly used GIS packages each with their own advantages (e.g. Santos et al., 2000; Cartaya et al., 2006; Sarup et al., 2006; Singh and Phadke, 2006). For example, direct digitising is more straightforward in MapInfo than in ArcMap. Although ArcMap has more functions for multi-dimensional analysis than MapInfo, the addition of the Vertical Mapper facility to MapInfo enhances its capabilities. However, specific properties, such as DEMs, slope maps and shaded elevation model maps, may be created in both products. GIS provides a basis for collating, viewing and then analysing, statistically or visually, georeferenced and spatial data that might have been collected at varying scales, densities and the types of information may equally be disparate (e.g. Childs, 1991; Oguchi et al., 2003; Delcaillau et al., 2006; Palyvos et al., 2006; Tejero et al., 2006). For example, topographic, geochemical and geophysical data can be imported into GIS to view spatial extent and variation individually or to apply multiple-criteria analysis upon them (Marinoni, 2005; e.g. He et al., 2007). GIS tools also have the capability to perform detailed analysis of topographic data such as slope or orientation of valleys and lineaments providing new maps for further interpretative analysis. GIS can perform multiple scale analyses and provide a viewing platform for detailed analysis of large, remote and inaccessible places. It has been employed for surface analysis of acutely inaccessible places such as Mars (Baker, 2004) and Jupiter’s satellite, Europa (Riley et al., 2006). GIS also has the capability of performing a wide range of spatial analysis techniques, although it does not automatically take into consideration surface area for its calculations. This may cause some computational problems where surface area rather than plan area may affect the results (Appendix 2). Despite the ability of GIS to perform multi-scale analyses, the system is, to a great extent, limited by the input data: for example, generalisations of the land surface are built into digital elevation models (DEM) and if they are greater than the resolution of the landscape processes that are being studied, results may not be fully representative and should be treated with caution (Pain, no date). Slope angles, for example, may be inaccurate if the scale of DEM is not sufficient to describe the landscape of the catchment under scrutiny. An appropriate scale for the landscape and processes under scrutiny must be determined to ensure sufficient detail lies within the digital data prior to analysis.

104

Using variously obtained datasets within GIS it is possible to identify geological linearity and associations with marked changes in terrain elevation, drainage anomalies and alignment with known neotectonic structures. If these features are taken separately, they cannot provide conclusive evidence of geological control over landscape. However, where there is synchroneity and/or superposition of several indices, and if such combinations do not imply other interpretations, then it may signify geological control over the feature, and possibly a neotectonic fault zone (Goldsworthy and Jackson, 2000; Ganas et al., 2005; Palyvos et al., 2006). GIS is flexible with regard to the type of spatial data used, ability to combine and integrate data of varying types and scales, extent of view and ability to vary the scale of analysis. Therefore, GIS is a suitable tool for this project, as spatial analysis at varying scales is required, combining elevation, drainage and geological datasets together with the locations and magnitudes of earthquakes. Due to various conflicting licensing issues throughout the course of this study, it was necessary to employ the use of two GIS analysis packages: MapInfo and ArcGIS. Although ArcGIS was initially equipped with more advanced features than MapInfo, both packages have similar capabilities that were required for the analysis. However, MapInfo required coupling with Vertical Mapper, a MapInfo ‘add-on’ that provides a vertical analysis tool, to bring its capability in line with ArcGIS. A further ‘add-on’ was also employed in MapInfo that enabled the program to measure and export the orientation of polylines, necessary for the analysis of faults and stream orientations used in paper 1. Neither ArcGIS nor MapInfo had an automated ordering system that would order the network using the preferred ordering system devised specifically for this work. Therefore, to ensure a suitable comparison was made when analysing the orientation features across and throughout the network, streams were ordered manually and then the orientations of each stream order were measured and exported separately. As with broad scale studies (e.g. Scheidegger, 1979b; Beavis, 2000), analysis at a fine-scale requires the orientation of stream channels to be compared with the orientation of structures such as faults, joints and rock fabric. To increase resolution of fine-scale channels, further drainage may be manually digitised where contour deflections indicate concave-downhill patterns and ground-truthing may then be undertaken by field work, aerial photography and comparison with existing maps. Catchment and subcatchment boundaries may be digitised following topographic

105

highs around each drainage basin. To quantitatively measure the channels and their orientations, channels may be ‘straightened’ following the method of Scheidegger (1979b) by digitising axes from node to node throughout the network. The orientation of each channel must be measured and although neither MapInfo nor ArcGIS has a sufficient facility for this, ‘qik-orientate-345’ (Lawley, 1997), was available as a ‘freeware’ addition to MapInfo (Appendix 3) and this was used for this work.

Remote sensing Remote sensing is a method for analysing spatial information from a distance. Data may be collected aerially and consist, for example, of images of spectral or geophysical data. Many examples of remote sensing exist and have been utilised to analyse a range of parameters over broad areas. An example in the southeast Queensland region is the land-cover and land-use change monitoring of a rapidly urbanising coastal environment – the Maroochy and Mooloolah River catchments (Phinn and Stanford, 2001). Their study was designed to incorporate the input of multiple resource monitoring groups at project planning, implementation and completion phases. Various data types were integrated to provide a tool for a wide group of resource managers and allow them to make better application of information within the land development process. Although remote sensing is not a new technology, it is becoming more popular since the development of Geographical Information Systems (GIS). Using GIS, Allan and Peterson (2002) were able to model the implications of land-use planning in Victoria, which resulted in a new decision support tool.

Spatial analysis Spatial analysis is the study of data and information that relates to an area of interest. In order to analyse particularly large areas, it is now common to use remotely obtained datasets and this is referred to as remote sensing, and the most recent method for such collection is by aeroplane or by satellite. Remote sensing in its various forms has been used for over 100 years to identify geomorphological features and is becoming an increasingly important tool. Early works include those of Hobbs (1901; 1904; 1911), who utilised broad-scale maps (‘111 miles to one inch’) to identify

106

physiographic lineaments and their relationship with regional faulting. A 3- dimensional relief model of Australia was built (University of Melbourne) during the 1940’s that provided early information on the relationship between topography and tectonics. This model revealed strong lineaments that, at the time, could not have been discovered in any other way across an area as large as a continent (Hills, 1956). Since that time, the use of traditional methods such as stereo-pairs and photo-mosaics have been replaced by digital photographs, satellite images, more accurate topographic maps and digital elevation and terrain models, that can now be amalgamated within products such as GIS to further analyse the landscape. Other analysis tools such as GOCAD®(Paradigm) and SiroSOM (Fraser and Dickson, 2007) may also be used for spatial analysis depending on the user’s specific requirements. Structurally controlled surface features such as scarps, aligned hills and drainage patterns have been identified and analysed using remote sensing (Berger, 1985; Harding and Berghoff, 2000; e.g. Philip, 2007). Similarly, satellite imagery was successfully used to asses recent tectonics in the Turkana Rift, North Kenya, when it was combined with both drainage patterns and seismic reflection analyses (Vétel et al., 2004).

Methods of channel analysis Using the definition of Twidale (1980), the term ‘structure’ has been used in relation to geomorphological control, to include rock-grain and texture, in addition to joints, faults, bedding, cleavage and folds. Zernitz (1932) and Twidale (2004) presented extensive reviews of drainage patterns and their meanings, and conclude that geological control is the major influence over many patterns. Geologically controlled drainage patterns commonly include trellis, rectilinear, dendritic, radial, centripetal, parallel and sub-parallel, for example (Twidale, 2004), but dendritic drainage indicates an overall lack of structural control (Zernitz, 1932). As discussed previously, due to the location of the Great Divide and the Great Escarpment to the west and the ocean to the east, it may be assumed that drainage would normally occur from the highlands towards the coast, generally from west to east. Comparison of the location and orientation of features such as faults, fractures or rock fabrics, with the location and orientation of streams, valleys and slopes, may identify relationships between those geological features and the morphology of the landscape. This has been undertaken in several broad scale studies for example by Beavis (2000) and

107

Scheidegger (1979b). In order to facilitate such as study at a finer scale, the resolution of mapped drainage must be increased to ensure the number of channels is of suitable magnitude to be compared with the geological features such as cleavage and joint systems. Although some programs can automatically digitise a drainage network, such as the Soil and Water Assessment Tool – “SWAT” (USDA, 2008), the result will rely on the level of detail in the underlying topographic model. To increase drainage resolution where an automated system may not select sufficient streams, further drainage may be manually digitised where contour deflections indicate concave- downhill patterns and ground-truthing may then be undertaken by field work, aerial photography and comparison with existing maps.

Stream ordering To determine whether different structural features typically control stream segments of different magnitude, it is necessary to partition the channels into similar channel- order ‘sets’ and group channel segments on the basis of similarity of scale and position within the drainage network. Therefore, a hierarchical ordering system is required. It is generally accepted that all finger-tip tributaries or channels in a network, are designated as 1st order. Identification of first-order channel head locations can be problematic (Heine et al., 2004) and lead to spurious orientation measurements for these reaches, so the orientation data of first order channels should be treated with caution. Some channel-ordering methods, such as the Horton (1945) and Strahler (1954) methods, may be manipulated in various ways. However, similarly derived dimensionless numbers might also be treated in a similar way. Shreve (1966; 1967) noted that Strahler’s and Horton’s Laws would be expected from any topologically random distribution. This argument was later confirmed and it was established that, from the properties the laws describe, no conclusion can be drawn to explain the structure or origin of the stream network (Kirchner, 1993). However, ordering systems continue to be used as ranking systems for practical purposes Scheidegger (1965), Woldenberg (1967), Walsh (1972), and Orme (2002) proposed mathematically derived ordering systems although the complex results lead to operational difficulties (Gardiner, 1975). The most commonly used ordering systems

108

are those of Horton (1945), Strahler (1957) (now commonly referred to as the Horton- Strahler method) and Shreve (1967) and each is discussed further, herein (Fig. 26). In summary, the Horton (1945) method may designate a high order to any scale of channel, from fingertip tributaries through to major rivers, which prevents successful comparison between orders grouped in this way (Fig. 26a). The Strahler ordering system (1957), derived from the Horton method, (now referred to as the Horton-Strahler method) fails to allocate a new order at every node. By disregarding tributaries in this way, the method effectively ignores the presence and influence (discharge and capacity) of some channels in the system (Fig. 26b). The Shreve (1967); numbering system, to some extent provides order numbers that relate to channel magnitude or position, although it also leads to situations where similarly positioned channels are assigned to substantially different orders and the relationship between channels in each order may be vague in some parts of the network (Fig. 26c). For the purposes of this study, a stream ordering system was required to ensure that stream segments of equivalent orders are compared with structural and rock fabric characteristics to identify the level of congruence and, thus, geological control on that network. The controls on entrapment of a watercourse may, over time, become concealed. However, the orientation of the channel may remain aligned with the original plane or line of weakness. To ascertain recent or antecedent geological control, it is important to identify the orders of channels that correspond to measured and known physiography and structural features. It is, therefore, important to ensure that the ordering system identifies similarly placed streams throughout the network. For the reasons discussed above, previous ordering systems are unsuitable for this work and a new ordering system has been developed and is presented in Paper 1.

109

Figure 26 Differences in resulting orders among commonly used stream ordering systems as discussed in the text: a) Horton (1945); b) Strahler (1957); c) Shreve (1967)

110

SUMMARY

The author wholly acknowledges that climate and surface processes influence the shape of the landscape. However, for the purposes of this study, it is assumed that those processes may enhance underlying structural and fabric anomalies where mechanical and chemical weathering exploit sites of weaker and more susceptible rock types. Various workers have identified that morphological features such as gullies, valleys, hillslopes and ridges, crests and scarps, may follow repeating alignment where geological or endogenic control exists. Therefore, one of the aims of this research is to identify tectonic and endogenic influences on the landscape. It may not be possible to fully discriminate between these influences, and the influences that are, for example, climate driven because the processes act simultaneously: the latter acting upon features produced by the former. To separate the effects of the processes that occur in tandem would suggest they act alone, but as already acknowledged, these processes act together. Without weakened, uplifted, folded, emplaced, variably altered or different strength rocks, weathering effects would be minor. The primary focus of this study is to discover how endogenic forces control the existing landscape of southeast Queensland. Southeast Queensland is geologically complex and is the result of many cycles of geological processes. Landmasses are dynamic: over geological time they are continually moving, uplifting or being downthrown; becoming folded, extended and compressed; becoming heated or buried and metamorphosed; becoming accreted to other landmasses, eroded and redeposited. This equates to the ‘rock-cycle’. However, the importance of these processes is often taken for granted. If the driving force of plate tectonics were to cease, and weathering processes were to continue, the landmasses would probably, almost entirely erode into the ocean basins and surficial processes would rely on dune formation or impactors to rejuvenate the landscape morphology. Therefore, the overarching driving force behind the shape of our landscape is geological processes. They may be described as the endogenic force that positions rocks upon which, exogenic forces may act. This work integrates pre-existing datasets and uses a method that can be re- applied to other areas, most importantly inaccessible places. In many regions, large, pre-existing bodies of knowledge exist in the form of maps and datasets, which can be digitised and used in a GIS allowing integration of each data-type for simultaneous

111

analyses. Newly acquired data can also be included in such analyses. Datasets may include for example, maps, digital elevation, geology, and seismic data that although in places may not be accurate or extensive, should at least be analysed alongside existing maps and datasets. This would provide multiple-layer visualisation and analysis that may be easily manipulated to change scales including more, or less data, as required. As statistical analysis of circular data for clustered and multiple-feature orientations is currently not reliable, graphical analyses of the data can be used successfully for comparative studies. Numbering systems of streams do not lend themselves easily to comparison of similar streams across a network and therefore an improved method is required. A modification to current analytical methods would integrate multiple geological features, to compare their position and possible association with physiography. This review has described some of the processes involved in geomorphological change and has highlighted the significance of primary, or endogenic controls, that establish the framework that is modified by secondary, or exogenic processes. It is evident from previous work that there is a general acceptance of an intrinsic relationship between geology and the landscape. Increasingly, more work is being conducted to either quantitatively or qualitatively prove this relationship and the extent to which landscapes are controlled by geology. Variations in spatial and temporal scales should be considered, and analysis such as hindcasting may be used as a support to predictive forecasting. Further work on this subject may provide valuable insights, especially if a new approach is taken by integrating multiple facets of the geological control, such as rock fabric and lithological variation, joints, faults, folds and tectonics, at multiple scales. Where geology has a significant influence over a landscape, a better understanding of both the geomorphological and geological systems provides a suitable framework upon which studies of anthropogenic influence can be based (Preda and Cox, 2002). A measure of anthropogenic or climate-change impacting on a landscape can be more easily discerned if the ‘natural’ level to which the landscape is being controlled has first been identified. It may then be possible to more accurately estimate how much erosion is anthropogenically or climatically driven if it is first recognised that, for example, erosion due to uplift or faulting is naturally very high. Similarly, if we identify that the landscape is tectonically and geologically inert,

112

producing little to no change in output over time, then major or sudden landscape changes are more likely be either anthropogenically driven or caused by climate change. Although the influence that tectonics and geology in general has over the landscape and in particular drainage channels is sometimes obvious, this is seldom considered when the causes and impacts of surface processes are being studied. Long term and often the more subtle effects of geology, may be taken for granted or not considered as necessary parts of a landscape-change study. This may be because surface processes such as erosion and weathering can be more easily observed ‘in action’ and more easily measured on a short term basis, whereas geological processes become part of the ‘background’ and are usually very slow. Surface process studies would be better equipped to make a judgement on anthropogenic and climatic influences if the internal driver of the landscape has also been considered. Both anthropogenic and climatic changes will continue and will increase, and measurement of their influences will require knowledge of the underlying system: therefore, this study addresses the importance of geological influence over the landscape of southeast Queensland as an example of a geologically complex, yet tectonically, a relatively quiescent region.

113

References Abal, E., Moore, K., Gibbes, B. and Dennison, B., 2001. State of South-east Queensland Waterways Report. Moreton Bay Waterways and Catchments Partnership. Abrahams, A.D. and Flint, J.-J., 1983. Geological controls on the topological properties of some trellis channel networks. Geological Society of America Bulletin, 94(1): 80-91. Ackermann, R.V., Seidl, M.A. and Schlische, R.W., 1997. First-order regional structural controls on modern drainage networks within the Newark Rift Basin, New Jersey. Geological Society of America Abstracts with Programs, 29(1): 25. Alabyan, A.M. and Chalov, R.S., 1998. Types of river channel patterns and their natural controls. Earth Surface Processes and Landforms, 23(5): 467-474. Alexander, D. and Formichi, R., 1993. Tectonic causes of landslides. Earth Surface Processes and Landforms, 18(4): 311-338. Allan, I. and Peterson, J., 2002. Spatial modelling in decision support for land-use planning: a demonstration from the Lal Lal Catchment, Victoria, Australia. Australian Geographical Studies, 40(1): 84-92. Allen, J.R.L., 1978. Studies in fluvial sedimentation: an exploratory quantitative model for the architecture of avulsion-controlled alluvial suites. Sedimentary Geology, 21: 129-147. Amelung, F. and King, G.C.P., 1997. Earthquake scaling laws for creeping and non- creeping faults. Geophysical Research Letters, 24: 507-510. Andreas, M. and Allan, A., 2007. Incorporating geology and geomorphology in land management decisions in developing countries: A case study in Southern Costa Rica. Geomorphology, 87(1-2): 68-89. Arnett, R.R., 1969. Soil-slope relationships in the Rocksberg Basin, south-east Queensland. M.A. Thesis, University of Queensland, Brisbane, 210 pp. Arnett, R.R., 1971. Slope form and geomorphological process: an Australian example. Slopes, Form and Process. Special Publication, Institute of British Geographers, 3: 81-92. Arnold, J.G. and Allen, P.M., 1992. A comprehansive syrface-groundwater flow model. Journal of Hydrology, 142: 47-69. Arnold, J.G., Srinivasin, R., Muttiah, R.S. and WIlliams, J.R., 1998. Large area hydrologic modelling and assessment: Part I. Model development. JAWRA, 34(1): 73-89. Australian Bureau of Statistics, 2006. Regional Population Growth, Australia, 2004- 05, Cat. no. 3218.0, ABS, Canberra. Australian Bureau of Statistics, 2008. Australian Demographic Statistics, Dec 2007: 3101.0, ABS, Canberra. Baillie, I.C. and Norbu, C., 2004. Climate and other factors in the development of river and interfluve profiles in Bhutan, Eastern Himalayas. Journal of Asian Earth Sciences, 22(5): 539-553. Baker, L., 2004. GIS analysis of Martian valley networks, Abstracts with Programs - Geological Society of America, Rocky Mountain Section, 56th annual meeting; Cordilleran Section, 100th annual meeting. Geological Society of America, Boise, United States, pp. 79.

114

Beavis, S.G., 2000. Structural controls on the orientation of erosion gullies in mid- western New South Wales, Australia. Geomorphology, 33(1-2): 59-72. Beckmann, G.G., 1959. The post-Tertiary history of Brisbane and surroundings, University of Queensland Brisbane. Beckmann, G.G. and Stevens, N.C., 1978. Geological history of the Brisbane River system. Proceedings of the Royal Society of Queensland, 89: 77-85. Belford, D.J., 1950. The Neranleigh Series. Unpublished Honours Thesis, University of Queensland, Brisbane. Beneduce, P., Festa, V., Francioso, R., Schiattarella, M. and Tropeano, M., 2004. Conflicting drainage patterns in the Matera Horst Area, southern Italy. Physics and Chemistry of the Earth, Parts A/B/C, 29(10): 717-724. Berger, Z., 1985. Detection and analysis of structurally controlled surface features using remotely sensed data. Environ. Res. Inst. Mich. : Ann Arbor, MI, United States, United States, pp. 205. Berryman, K., Marden, M., Eden, D., Mazengarb, C., Ota, Y. and Moriya, I., 2000. Tectonic and paleoclimatic significance of Quaternary river terraces of the Waipaoa River, east coast, North Island, New Zealand. New Zealand Journal of Geology and Geophysics, 43: 229-245. Betts, P.G., Giles, D., Lister, G.S. and Frick, L.R., 2002. Evolution of the Australian lithosphere. Australian Journal of Earth Sciences, 49(4): 661-695. Bezerra, F.H.R., Brito Neaves, B.B., Corrêa, A.C.B., Barreto, A.M.F. and Suguio, K., 2008. Late Pleistocene tectonic-geomorphological development within a passive margin - The Cariatá trough, northeast Brazil. Geomorphology, 97(3- 4): 555-582. Bondár, I., Myers, S.C., Engdahl, R. and Bergman, E.A., 2004. Epicentre accuracy based on seismic network criteria. Geophysical Journal International, 156: 483-496. Bonnet, S., Besnard, M. and Van den Driessche, J., 2001. Drainage network expansion of the Salagou drainage basin (S. France): and example of relief response to recent climate change? Terra Nova, 13: 214-219. Boyd, K.F. and Schumm, S.A., 1995. Geomorphic evidence of deformation in the northern part of the New Madrid seismic zone. U.S. Geological Survey Professional Papers, 1538-R: 35. Branson, J.C., 1978. Evolution of sedimentary basins from Mesozoic times in Australia's continental slope and shelf. In: E. Scheibner (Editor), The Phanerozoic structure of Australia and variations in tectonic style. Tectonophysics,, pp. 389-412. Bridgland, D.R., 2000. River terrace systems in north-west Europe: an archive of environmental change, uplift and early human occupation. Quaternary Science Reviews, 19(13): 1293-1303. Brisbane City Council, 2006. City Design, Ground Engineering, Brisbane, pp. Core logging showing that geology maps incorrect in many parts of Brisbane. Bronstert, A., Niehoff, D. and Bürger, G., 2002. Effects of climate and land-use change on storm runoff generation: present knowledge and modelling capabilities. Hydrological Processes, 16(2): 509-529. Brown, A. and Gibson, G., 2004. A multi-tiered earthquake hazard model for Australia. Tectonophysics, 390(1-4): 25-43. Bryan, W.H., 1939. The red earth residuals and their significance in south-eastern Queensland. Proceedings of the Royal Society of Queensland, 50: 21-32.

115

Bryan, W.H. and Jones, O.A., 1954. Contributions to the geology of Brisbane No. 2. The structural history of the Brisbane Metamorphics. Proceedings of the Royal Society of Queensland, 65: 25-50. Bryan, W.H. and Jones, O.A., 1962. Contributions to the geology of Brisbane No. 3. The bedded cherts of the Neranleigh-Fernvale Group of south-east Queensland. Proceedings of the Royal Society of Queensland, 73: 17-36. Burbank, D., Meigs, A. and Brozovic, N., 1996. Interactions of growing folds and coeval depositional systems. Basin Research, 8: 199-223. Burch, G.J., Bath, R.K., Moore, I.D. and O'Loughlin, E.M., 1987. Comparative hydrological behaviour of forested and cleared catchments in southeastern Australia. Journal of Hydrology, 90(1-2): 19-42. Bureau of Meteorology, 2008. Climate statistics for Australian locations. Commonwealth of Australia. Burnett, A.W., 1982. Alluvial stream response to neotectonics in the Lower Mississippi Valley, Unpublished M.S. Thesis. Colorado State University Fort Collins, CO, pp. 160 (Cited by Holbrook and Schumm 1999). Burnett, A.W. and Schumm, S.A., 1983. Alluvial-river response to neotectonic deformation in Louisiana and Mississippi. Science, 222(4619): 49-50. Burrato, P., Ciucci, F. and Valensise, G., 2003. An inventory of river anomalies in the Po Plain, Northern Italy: evidence for active blind thrust faulting. Annals of Geophysics, 46(5): 865-882. Caitcheon, G., Rustomji, P., Read, A., Douglas, G. and Palmer, M., 2005. Targeting sediment and nutrient source areas in the Bremer, Lockyer and Wivenhoe catchments, CSIRO Land and Water, (Commercial-in-confidence). Calcaterra, D., Gili, J.A. and Iovinelli, R., 1998. Shallow landslides in deeply weathered slates of the Sierra de Collcerola (Catalonian Coastal Range, Spain). Engineering Geology, 50(3-4): 283-298. Campbell, J.K., Nicol, A. and Howard, M.E., 2003. Long-term changes to river regimes prior to late Holocene coseismic faulting, Canterbury, New Zealand. Journal of Geodynamics, 36(1-2): 147-168. Cartaya, S., Mendez, W. and Pacheco, H., 2006. Adaptation and implementation of a zoning model for mass movement susceptibility through of a geographical information system. Interciencia, 31(9): 638-646. Celerier, J., Sandiford, M., Hansen, D.L. and Quigley, M., 2005. Modes of active intraplate deformation, Flinders Ranges, Australia. Tectonics, 24(6): Art. No. TC6006. Chiew, F.H.S. and McMahon, T.A., 2002. Modelling the impacts of climate change on Australian streamflow. Hydrological Processes, 16(6): 1235-1245. Childs, I., 1991. Earthquake risk, Landsat imagery and fault zones in the Bundaberg area of central eastern Queensland. Queensland Geographical Journal, Series 4, 6: 59-70. Clark, D. and McCue, K., 2003. Australian paleoseismology: towards a better basis for seismic hazard estimation. Annals of Geophysics, 46(5): 1087-1105. Clarke, J.D.A., 1994. Evolution of the Lefroy and Cowan palaeodrainages, Western Australia. Australian Journal of Earth Sciences, 41: 55-68. Coaldrake, J.E., 1960. Quaternary history of the coastal lowlands of southern Queensland. Journal Geological Society of Australia, 7: 403-408.

116

Coaldrake, J.E., 1961. The ecosystem of the coastal lowlands ("Wallum") of southern Queensland. Bulletin of the Commonwealth Scientific and Research Organisation, 283. Cockburn, H.A.P., Brown, R.W., Summerfield, M.A. and Seidl, M.A., 2000. Quantifying passive margin denudation and landscape development using a combined fission-track thermochronology and cosmogenic isotope analysis approach. Earth and Planetary Science Letters, 179(3-4): 429-435. Coney, P.J., 1992. The Lachlan Belt of eastern Australia and circum-Pacific tectonic evolution. Tectonophysics, 214(1-4): 1-25. Cowie, P.A. and Roberts, G.P., 2001. Constraining slip rates and spacings for active normal faults. Journal of Structural Geology, 23(12): 1901-1915. Cox, K.G., 1980. A model for flood basalt vulcanism. Journal of Petrology, 21(4): 629-650. Cranfield, L.C., Schwarzbock, H. and Day, R.W., 1976. Geology of the Ipswich and Brisbane 1:250 000 Sheet Areas. Geological Survey of Queensland Report 95, Queensland Department of Mines. Cranfield, L.C. and Scott, M., 1993. Geology of the Gympie 1:100 000 Special Area (9445, 9545), Southeast Queensland, 20, Department of Minerals and Energy, Queensland. Cranfield, L.C., 1994. Maryborough, Queensland, Sheet SG56-6, Department of Minerals and Energy, Geological Survey or Queensland, Brisbane, Queensland. Cuthbertson, R.J., 1990. The seismo-tectonics of southeast Queensland. Bureau of Mineral Resources, Geology and Geophysics, Australia, Bulletin, 232: 67-81. Cuthbertson, R.J. and Murray, C.G., 1990. Seismo-tectonics of eastern Queensland, Pacific Rim Congress 90. An International Congress on the Geology, Structure, Mineralization, Economics and Feasibility of Mining Development in the Pacific Rim, 6-12 May, 1990. Australasian Institute of Mining and Metallurgy, Parkville, Victoria, Gold Coast, Queensland, Australia, pp. 109- 16. Cuthbertson, R.J., 1995. Earthquake activity to 1994, The University of Queensland, Brisbane. Cuthbertson, R.J. and Jaumé, S.C., 1996. Earthquake hazard in Queensland, QUAKES Report #1. University of Queensland, Brisbane, pp. 13-39. Cuthbertson, R.J., Lynam, C. and Jaumé, S.C., 1998. Seismicity of Queensland - 1995 to 1997, QUAKES Report #3. The University of Queensland, Brisbane, pp. 13 - 22. Day, R.W., Murray, C.G. and Whitaker, W.G., 1978. The eastern part of the Tasman Orogenic Zone. Tectonophysics, 48: 327-364. Day, R.W., Whitaker, W.G., Murray, C.G., Wilson, I.H. and Grimes, K.G. (Editors), 1983. Queensland Geology, A companion volume to the 1:2 500 000 scale geological map (1975). Geological Survey of Queensland, Publication 383. Delcaillau, B., Carozza, J.-M. and Laville, E., 2006. Recent fold growth and drainage development: The Janauri and Chandigarh anticlines in the Siwalik foothills, northwest India. Geomorphology, 76(3-4): 241-256. Demoulin, A., 1998. Testing the tectonic significance of some parameters of longitudinal river profiles: the case of Ardenne (Belgium, NW Europe). Geomorphology, 24(2-3): 189-208.

117

Denmead, A.K., 1928. A survey of the Brisbane schists. Proceedings of the Royal Society of Queensland, 39: 71-106. Dietz, R.S., 1961. Continent and ocean basin evolution by spreading of the sea floor. Nature, 190(4779): 854-857. Dominguez-Cuesta, M.J., Jimenez-Sanchez, M. and Berrezueta, E., 2007. Landslides in the Central Coalfield (Cantabrian Mountains, NW Spain): Geomorphological features, conditioning factors and methodological implications in susceptibility assessment. Geomorphology, 89(3-4): 358-369. Donchak, P.J.T., 1976. A structural study of the greenstone-phyllite sequence of the Petrie-Dayboro area. B.Sc. Honours Thesis, University of Queensland, Brisbane, (Unpublished). Douglas, G., Palmer, M. and Caitcheon, G., 2003. The provenance of sediments in Moreton Bay, Australia; A synthesis of major trace element and Sr-Nd-Pb isotopic chemistry, modelling and landscape analysis. Hydrobiologia, 494(1- 3): 145-152. Doyle, M.W. and Harbor, J.M., 2003. Modelling the effect of form and profile adjustments on channel equilibrium timescales. Earth Surface Processes and Landforms, 28(12): 1271-1287. Droste, J.B. and Keller, S.J., 1989. Development of the Mississippian-Pennsylvanian unconformity in Indiana. Department of Natural Resources, Geological Survey, Occasional Papers(Paper 55): 11 pp. Ellis, M.A., Densmore, A.L. and Anderson, R.S., 1999. Development of mountainous topography in the Basin Ranges, USA. Basin Research, 11: 21-41. Ericson, K., Migon, P. and Olvmo, M., 2005. Fractures and drainage in the granite mountainous area: A study from Sierra Nevada, USA. Geomorphology, 64(1- 2): 97-116. ESSCC, 2006. Earth Systems Science Computational Centre, Earthquake Database, University of Queensland, Brisbane. Evernden, J.F. and Richards, J.R., 1962. Potassium-argon ages in eastern Australia. Journal of the Geological Society of Australia, 9: 1-50. Ewart, A., Baxter, K. and Ross, J.A., 1980. The petrology and petrogenesis of the Tertiary anorogenic mafic lavas of southern and central Queensland, Australia - possible implications for crustal thickening. Contributions to Mineralogy and Petrology, 75: 129-152. Eyles, N., Arnaud, E., Scheidegger, A.E. and Eyles, C.H., 1997. Bedrock jointing and geomorphology in southwestern Ontario, Canada: an example of tectonic predesign. Geomorphology, 19(1-2): 17-34. Eyles, N. and Scheidegger, A.E., 1999. Neotectonic jointing control on Lake Ontario shoreline orientation at Scarborough, Ontario. Geoscience Canada, 26(1): 27- 30. Fergusson, C.L. and Leicht, E.C., 1993. Late Carboniferous to Early Triassic tectonics of the New England Fold Belt, eastern Australia. In: P.G. Flood and J.C. Aitchison (Editors), New England Orogen, Eastern Australia. Department of Geology and Geophysics, University of New England, Armidale, pp. 53-59. Finlayson, B.L. and McMahon, T.A., 1988. Australia v the World: A comparative analysis of streamflow characteristics. In: R.F. Warner (Editor), Fluvial Geomorphology of Australia. University of Sydney, Australia, pp. 1-16. Fisher, N.I., 1996. Statistical Analysis of Circular Data. Cambridge University Press, Cambridge, UK, 295 pp.

118

Flanagan, D.C. and Nearing, M.A. (Editors), 1995. USDA Water erosion prediction project - hillslope profile and watershed model documentation. NSERL Report No. 10. United States Deparment of Agriculture, Agricultural Research Service, National Soil Research Laboratory, West Lafayette, Indiana, USA, 298 pp. Fleming, P.J.G., Day, R.W., Murray, C.G. and Whitaker, W.G., 1974. Late Palaeozoic invertebrate macrofossils in the Neranleigh-Fernvale Beds. Queensland Government Mining Journal, 75: 104-107. Flood, P.G., 1981. Carbon-14 dates deom the coastal plains of Deception Bay, southeastern Queensland. Queensland Government Mining Journal, 82(19-23). Frank, T.D. and Fielding, C.R., 2004. Sedimentology and geochemistry of an urban coastal lake system: Coombabah Lake Nature Reserve, Gold Coast, Queensland. Australian Journal of Earth Sciences, 51(2): 261-271. Frankel, K.L. and Pazzaglia, F.J., 2005. Tectonic geomorphology, drainage basin metrics and active mountain fronts. Geografia Fisica e Dinamica Quaternaria, 28: 7-21. Fraser, S.J. and Dickson, B.L., 2007. A new method for data integration and integrated data interpretation: Self-Organising Maps. In: B. Milkereit (Editor), Proceedings of Exploration 07: Fifth Decennial International Conference on Mineral Exploration, Toronto, Canada, pp. 907-910. Fraser, S.J. and Hodgkinson, J.H., 2008. Data mining geoscientific data using self organising maps: a new approach to geological data analysis, 44th Congresso Brasileiro de Geologica, Brazil. Friedman, J.M., Osterkamp, W., R., Scott, M.L. and Auble, G.T., 1998. Downstream effects of dams on channel geometry and bottomland vegetation: Regional patterns in the Great Plains. Wetlands, 18(4): 619-633. Gabet, E.J., Burbank, D.W., Putkonen, J.K., Pratt-Sitaula, B.A. and Ojha, T., 2004. Rainfall thresholds for landsliding in the Himalayas of Nepal. Geomorphology, 63(3-4): 131-143. Gallart, F. and Llorens, P., 2004. Observations on land cover changes and water resources in the headwaters of the Ebro catchment, Iberian Peninsula. Physics and Chemistry of the Earth, Parts A/B/C, 29(11-12): 769-773. Galloway, W.E., Ganey-Curry, P.E., Li, X. and Buffler, R.T., 2000. Cenozoic depositional history of the Gulf of Mexico basin. AAPG Bulletin, 84(11): 1743-1774. Ganas, A., Pavlides, S. and Karastathis, V., 2005. DEM-based morphometry of range- front escarpments in Attica, central Greece, and its relation to fault slip rates. Geomorphology, 65: 301-319. Gardiner, V., 1975. Drainage Basin Morphometry. Department of Geography, Wolverhampton Polytechnic, 48 pp. Gardner, T.W., 1975. The history of part of the Colorado River and its tributaries, an experimental study, Four Corners Geological Society guidebook. 8th Field Conference (Canyonlands). pp. 87-95. Gasparini, N.M., Tucker, G.E. and Bras, R.L., 2004. Network-scale dynamics of grain-size sorting: Implications for downstream fining, stream-profile concavity, and drainage basin morphology. Earth Surface Processes and Landforms, 29(4): 401-421. Geoscience Australia, 2006. http://www.ga.gov.au/oracle/quake/quake_online.jsp.

119

Ghayoumian, J., Mohseni Saravi, M., Feiznia, S., Nouri, B. and Malekian, A., 2007. Application of GIS techniques to determine areas most suitable for artificial groundwater recharge in a coastal aquifer in southern Iran. Journal of Asian Earth Sciences, 30(2): 364-374. Gibson, G., 1997. Earthquakes and dams in Australia, Earthquakes in Australian cities - can we ignore the risks? Proceedings of a seminar held by the Australian Earthquake Engineering Society, University of Queensland, Brisbane, Australia. Gilbert, G.K., 1877. Report on the Geology of the Henry Mountains. U.S. Geographical and Geological Survey of the Rocky Mountains Region., Government Printing Office, Washington, DC. Gilbert, G.K., 1917. Hydraulic-mining debris in the Sierra Nevada. US Geological Survey Professional Paper, 105: (Cited in Doyle and Harbor 2003). Global Seismic Hazard Assessment Program, 1992-1997. Global Seismic Hazard Map, http://www.seismo.ethz.ch/GSHAP/global/. Goldsworthy, M. and Jackson, J., 2000. Active normal fault evolution in Greece revealed by geomorphology and drainage patterns. Journal of the Geological Society, 157: 967-981. Goodrich, H.B., 1898. Geology of the Yukon Gold District, Alaska. United States Geological Society Annual Report 18, Part III, Washington, DC. Goudie, A.S., 2006. Global warming and fluvial geomorphology. Geomorphology, 37th Binghamton Geomorphology Symposium - The Human Role in Changing Fluvial Systems, 79(3-4): 384-394. Graf, W.L., 2005. Geomorphology and American dams: The scientific, social, and economic context. Geomorphology, 71(1-2): 3-26. Grams, P.E. and Schmidt, J.C., 2005. Equilibrium or indeterminate? Where sediment budgets fail: Sediment mass balance and adjustment of channel form, Green River downstream from Flaming Gorge Dam, Utah and Colorado. Geomorphology, 71: 156-181. Granger, K. and Hayne, M. (Editors), 2000. Natural hazards and the risks they pose to South-East Queensland. Australian Geological Survey Organisation and Bureau of Meteorology, Commonwealth of Australia. Green, D.C., 1973. Radiometric evidence of the Kanimblan orogeny in southeast Queensland and the age of the Neranleigh-Fernvale Group. Journal of the Geological Society of Australia, 20: 153-160. Grimes, K.G., 1988. Cainozoic geology, south-east Queensland. In: L. Hamilton (Editor), Field Excursion Handbook for the Ninth Australian Geological Convention. Geological Society of Australia, Inc. Queensland Division, Brisbane, pp. 53-80. Gupta, A.K. and Ahmad, R., 1999. Geomorphology and the urban tropics: building an interface between research and usage. Geomorphology, 31(1-4): 133-149. Harding, D.J. and Berghoff, G.S., 2000. Fault scarp detection beneath dense vegetation cover: airborne Lidar mapping of the Seattle Fault Zone, Bainbridge Island, Washington State, Proceedings of the American Society of Photogrammetry and Remote Sensing Annual Conference, Washington, D.C. Haworth, R.J. and Ollier, C.D., 1992. Continental Rifting and Drainage Reversal - the Clarence River of Eastern Australia. Earth Surface Processes and Landforms, 17(4): 387-397.

120

He, J., Zhang, Y. and Shi, Y., 2007. A multi-criteria decision support system of water resource allocation scenarios. Lecture Notes in Computer Science, 4798: 593- 598. Heine, R.A., Lant, C.L. and Sengupta, R.R., 2004. Development and comparison of approaches for automated mapping of stream channel networks. Annals of the Association of American Geographers, 94(3): 477-490. Herron, N., Davis, R. and Jones, R., 2002. The effects of large-scale afforestation and climate change on water allocation in the Macquarie River catchment, NSW, Australia. Journal of Environmental Management, 65(4): 369-381. Hess, H. (Editor), 1962. History of ocean basins. Petrologic Studies. Geological Society of America, Boulder, Colorado, 599-620 pp. Hill, D. and Denmead, A.K. (Editors), 1960. The Geology of Queensland. Melbourne University Press, Melbourne, 474 pp. Hillis, R.R., 1998. The Australian stress map. Exploration Geophysics, 29(3-4): 420. Hillis, R.R., Enever, J.R. and Reynolds, S.D., 1999. In situ stress field of eastern Australia. Australian Journal of Earth Sciences, 46(5): 813-825. Hillis, R.R. and Reynolds, S.D., 2000. The Australian Stress Map. Journal of the Geological Society, 157: 915-921. Hillis, R.R. and Reynolds, S.D., 2003. In situ stress field of Australia. Geological Society of Australia Special Publication, 22: 43 - 52. Hills, E.S., 1956. A contribution to the morphotectonics of Australia. Journal of the Geological Society of Australia, 3: 1-15. Hills, E.S., 1960. Morphotectonics and the geomorphological sciences witrh special references to Australia. Quarterly Journal of the Geological Society, 117(1:465): 77-89. Hills, E.S., 1963. Elements of Structural Geology. Methuen, London, 483 pp. Hobbs, W.H., 1901. The river system of Connecticut. Journal of Geology, ix: 469- 484. Hobbs, W.H., 1904. Lineaments of the Atlantic border region. Geological Society of America Bulletin, 15: 483-506. Hobbs, W.H., 1911. Repeating patterns in the relief and structure of the land. Geological Society of America Bulletin, 22: 123-176. Hofmann, G.W., Trezise, D.L.a. and Muller, P.J., 1976. Extractive Resources and Urban Land Suitability in the Pine Rivers Shire and City of Redcliffe, Moreton Region, Queensland. Hofmann, G.W., 1980. Quaternary sediments and geological history of the Pine Rivers area, Southest Queensland. Queensland Government Mining Journal, 81(948): 502-512. Holbrook, J. and Schumm, S.A., 1999. Geomorphic and sedimentary response of rivers to tectonic deformation: a brief review and critique of a tool for recognizing subtle epeirogenic deformation in modern and ancient settings. Tectonophysics, 305. Holcombe, R.J., 1978. Structure and tectonic history of the Brisbane metamorphics in the Brisbane area. Journal of the Geological Society of Australia, 24: 475-490. Holcombe, R.J., Little, T.A., Sliwa, R. and Fielding, C.R., 1993. Tectonic elements and evolution of SE Queensland during the Late Palaeozoic to Early Mesozoic. In: P.G. Flood and J.C. Aitchison (Editors), New England Orogen, Eastern Australia. Department of Geology and Geophysics, University of New England, Armidale, pp. 517-527.

121

Holcombe, R.J., Stephens, C.J., Fielding, C.R., Gust, D., Little, T.A., Sliwa, R., Kassan, J., McPhie, J. and Ewart, A., 1997a. Tectonic evolution of the northern New England Fold Belt: the Permian-Triassic Hunter Bowen event. In: P.M. Ashley and P.G. Flood (Editors), Tectonics and Metallogenesis of the New England Orogen. Geological Society of Australia. Special Publication 19, pp. 52-65. Holcombe, R.J., Stephens, C.J., Fielding, C.R., Gust, D., Little, T.A., Sliwa, R., Kassan, J., McPhie, J. and Ewart, A., 1997b. Tectonic evolution of the northern New England Fold Belt: Carboniferous to Early Permian transition from active accretion to extension. In: P.M. Ashley and P.G. Flood (Editors), Tectonics and Metallogenesis of the New England Orogen. Geological Society of Australia. Special Publication 19, pp. 66-79. Holdgate, G.R., Wallace, M.W., Gallagher, S.J., Witten, R.B., Stats, B. and Wagstaff, R.E., 2006. Cenozoic fault control on 'deep lead' palaeoriver systems, Central Highlands, Victoria. Australian Journal of Earth Sciences, 53(3): 445-468. Holdgate, G.R., Wallace, M.W., Gallagher, S.J., Wagstaff, B.E. and Moore, D., 2008. No mountains to snow on: major post-Eocene uplift of the East Victoria Highlands; evidence from Cenozoic deposits. Australian Journal of Earth Sciences, 55(2): 211-234. Horton, R.E., 1945. Erosional development of streams and their drainage basins; hydrophysical approach to quantitative morphology. Bulletin of the Geological Society of America, 56: 275-370. Houston, B.R., 1967. The post-Palaeozoic sediments and volcanics, Geology of the City of Brisbane, Geological Survey of Queensland, Brisbane, pp. 7-36. Howard, A.D., 1967. Drainage analysis in geologic interpretation: a summary. American Association of Petroleum Geology, 51: 2246-2259. Humphries, D., 2003. An analysis of post-Triassic faulting in southeast Queensland. BSc (Hons) Thesis, University of Queensland, Brisbane, 134 pp. Husen, S., Kissling, E., Flueh, E. and Asch, G., 1999. Accurate hypocentre determination in the seismogenic zone of the subducting Nazca Plate in northern Chile using a combined on-/offshore network. Geophysical Journal International, 138(3): 687-701. Hutjes, R.W.A., Kabat, P., Running, S.W., Shuttleworth, W.J., Field, C., Bass, B., da Silva Dias, M.F., Avissar, R., Becker, A. and Claussen, M., 1998. Biospheric Aspects of the Hydrological Cycle. Journal of Hydrology, 212-213(1-4): 1-21. Jensen, H.I., 1903. The geology of the Glass House Mountains and district. Proceedings of the Linnean Society of New South Wales, 28: 842-875. Jensen, H.I., 1906. Geology of the volcanic area of the east Moreton and Wide Bay districts, Queensland. Proceedings of the Linnean Society of New South Wales, 31: 73-173. Jones, M.R., 2006a. Cenozoic landscape evolution in central Queensland. Australian Journal of Earth Sciences, 53(3): 433-444. Jones, O.A., 1927. The Tertiary deposits of the Moreton District. Proceedings of the Royal Society of Queensland, 38: 23-46. Jones, T., 2006b. MATLAB functions to analyse directional (azimuthal) data - I: Single-sample inference. Computes and Geoscience, 32: 166-175. Jones, T.A., 1968. Statistical analysis of orientation data. Journal of Sedimentary Petrology, 38(1): 61-67.

122

Jorgensen, D.W., 1990. Adjustment of alluvial river morphology and process to localised active tectonics, Unpublished Ph.D. Dissertation. Colorado State University, Fort Collins, pp. 240 (cited by Holbrook and Schumm, 1999). Julian, M. and Anthony, E., 1996. Aspects of landslide activity in the Mercantour Massif and the French Riviera, southeastern France. Geomorphology, 15(3-4): 275-289. Kanamori, H. and Stewart, G.S., 1979. A slow earthquake. Physics of the Earth and Planetary Interiors, 18(3): 167-175. Kershaw, A.P. and Nanson, G.C., 1993. The last full glacial cycle in the Australian region. Global and Planetary Change, 7: 1-9. King, L.C., 1956. A geomorpfologia do Brasil oriental. Revista Geografica Brasileira, 18: 147-265. Kirby, E., Whipple, K.X. and Harkins, N., 2008. Topography reveals seismic hazard. Nature Geoscience, 1(8): 485-487. Kirchner, J.W., 1993. Statistical inevitability of Horton Laws and the apparent randomness of stream channel networks. Geology, 21(7): 591-594. Kitagawa, Y., Koizumi, N., Ohtani, R., Watanabe, K. and Itaba, S., 2006. Detection of aseismic slip on an inland fault by crustal movement and groundwater observations: A case study on the Yamasaki Fault, Japan. Pure and Applied Geophysics, 163(4): 657-673. Knighton, A.D., 1981. Asymmetry of river channel cross-sections: Part 1. Quanitative indices. Earth Surface Processes and Landforms, 6: 581-588. Kohlbeck, F.K. and Scheidegger, A.E., 1985. The power of parametric orientation statistics in the Earth Sciences. Mitteilungen der Österreichischen Geologischen Gesellschaft, 78: 251-265. Kohonen, T., 2001. Self-Organizing Maps, Third extended edition. Springer Series in Information Sciences, Vol. 30. Springer, Berlin, Heidelberg, New York. Kondorskaya, N.V., Slavina, L.B. and Bruk, M.G., 1989. Near earthquake hypocentre determination: the accuracy problem. Tectonophysics, 166(1-3): 247-253. Koss, J.E., Ethridge, F.G. and Schumm, S.A., 1994. An experimental study of the effects of base-level change on fluvial, coastal plain and shelf systems. Journal of Sedimentary Research, B64(2): 90-98. Krieger Lassen, N.C., Juul Jenson, D. and Conradsen, K., 1994. On the statistical analysis of orientation data. Acta Crystallographica, A50: 741-748. Krumbein, W.C., 1939. Preferred orientation of pebbles in sedimentary deposits. Journal of Geology, 47: 673-706. Lawley, R., 1997. qik-orientate-345. Geo Trek Corp, [email protected]. Levshin, A.L. and Ritzwoller, M.H., 2002. Application of a global 3D model to improve regional event locations. Studia Geophysica et Geodaetica, 46(2): 283-292. Lim, K.J., Sagong, M., Engel, B.A., Tang, Z., Choi, J. and Kim, K., 2005. GIS-Based sediment assessment tool. Catena, 64: 61-80. Linde, A.T., Gladwin, M.T., Johnston, M.J.S., Gwyther, R.L. and Bilham, R.G., 1996. A slow earthquake sequence on the San Andreas fault. Nature, 383(6595): 65- 68. Little, T.A., Holcombe, R.J., Gibson, G.M., Offler, R., Gans, P.B. and McWilliams, M.O., 1992. Exhumation of late Paleozoic blueschists in Queensland, Australia, by extensional faulting. Geology, 20(3): 231-234.

123

Lucas, K.G., 1987. Landsat lineaments from Childers to Caboolture. In: C.G. Murray and J.B. Waterhouse (Editors), Field Conference: Gympie District. Geological Society of Australia, Queensland Division, Queensland, Australia, pp. 86-87. Lüdmann, T. and Wong, H.K., 1999. Neotectonic regime on the passive continental margin of the northern South China Sea. Tectonophysics, 311(1-4): 113-138. Malik, J.N. and Mohanty, C., 2006. Active tectonic influence on the evolution of drainage and landscape: geomorphic signatures from frontal and hinterland areas along the Northwestern Himilaya, India. Journal of Asian Earth Sciences, 29: 604-618. Marinoni, O., 2005. An Example for a GIS-Integrated Use of Stochastic Multi- Criteria Evaluations, Proceedings of the IAMG Conference, Toronto, Canada, pp. 517-522. Marks, E.O., 1933. Some observations on the physiography of the Brisbane River and neighbouring watersheds. Proceedings of the Royal Society of Queensland, 44(9): 132-150. Mather, A.E., 1993. Basin inversion: some consequences for drainage evolution and alluvial architecture. Sedimentology, 40(6): 1069-1089. McEwen, L.J. and Matthews, J.A., 1998. Channel form, bed material and sediment sources of the Sprongdøla, southern Norway: evidence for a distinct periglacio-fluvial system. Geografiska Annaler, 80(A): 17-36. McKellar, J.L., In press. Late Early to Late Jurassic palynology, biostratigraphy and palaeogeography of the Roma Shelf area, northwestern Surat Basin, Queensland, Australia. Memoir of the Association of Australasian Palaeontologists. Association of Australasian Palaeontologists, Canberra. McKenzie, D., 1978. Some remarks on the development of sedimentary basins. Earth and Planetary Science Letters, 40: 25-32. Meijer, X.D., 2002. Modelling the drainage evolution of a river-shelf system forced by Quaternary glacio-eustasy. Basin Research, 14: 361-377. Miall, A.D., 1991. Stratigraphic sequences and their chronostratigraphic correlation. Journal of Sedimentary Petrology, 61: 497-505. Michetti, A.M., Audemard, F.A. and Marco, S., 2005. Future trends in paleoseismology: Integrated study of the seismic landscape as a vital tool in siesmic hazard analyses. Tectonophysics, 408: 3-21. Mock, S.J., 1971. A classification of channel links in stream networks. Cold Regions Research, 7: 1558-1566. Mock, S.J., 1976. Topological properties of some trellis pattern channel networks. Cold Regions Research and Engineering Laboratory Report: 46-76. Mohammadioun, B. and Serva, L., 2001. Stress drop, slip type, earthquake magnitude and seismic hazard. Bulletin of Seismological Society of America, 91: 694- 707. Montgomery, D.R., 2003. Predicting landscape-scale erosion rates using digital elevation models. Comptes Rendus Geoscience, 335: 1121-1130. Mora, P., 1996. Earthquake monitoring and hazard studies, QUAKES Report #1. University of Queensland, Brisbane, pp. 105-109. Müller, R.D., Lim, V.S.L. and Isern, A.R., 2000. Late Tertiary tectonic subsidence on the northeast Australian passive margin: response to dynamic topography? Marine Geology, 162(2-4): 337-352.

124

Murphy, P.R., Schwarzbock, H., Cranfield, L.C., Withnall, I.W. and Murray, C.G., 1976. Geology of the Gympie 1:250 000 Sheet Areas. Geological Survey of Queensland Report 96, Geological Survey of Queensland. Murphy, P.R., Trezise, D.L., Hutton, L.J. and Cranfield, L.C., 1979. Kurwongbah Beds - A new stratigraphic unit of the D'Aguilar Block. Queensland Government Mining Journal, 80: 65 - 70. Murray, C.G., McClung, G.R. and Whitaker, W.G., 1979. Early Permian fossils from the Amamoor Beds and the age of the Booloumba Beds - a new stratigraphic unit of the North D'Aguilar Block. Queensland Government Mining Journal, 80: 71-78. Murray, C.G. and Whitaker, W.G., 1982. A review of the stratigraphy, stucture and regional tectonic setting of the Brisbane metamorphics. In: P.G. Flood and B. Runnegar (Editors), New England Geology. Department of Geology, University of New England and AHV Club, Armidale, pp. 79-94. Murray, C.G., Fergusson, C.L., Flood, P.G., Whitaker, W.G. and Korsch, R.J., 1987. Plate tectonic model for the Carboniferous evolution of the New-England Fold Belt. Australian Journal of Earth Sciences, 34(2): 213-236. Murray, C.G., 1988. Tectonostratigraphic terranes of southeast Queensland. In: L. Hamilton (Editor), Field Excursion Handbook for the Ninth Australian Geological Convention. Geological Society of Australia Inc. Queensland Division, Brisbane, pp. 19-51. Nanson, G.C., Cohen, T.J., Doyle, C.J. and Price, D.M., 2003. Alluvial evidence of major Late-Quaternary climate and flow changes on the coastal rivers of NSW, Australia. In: K.J. Gregory and G. Benito (Editors), Palaeohydrology - Understanding Global Change. Wiley, Chichester, pp. 233-258. NASA, 2004. Shuttle Radar Topography Mission. Jet Propulsion Laboratory, California Institute of Technology, www2.jpl.nasa.gov/srtm/. Nelson, E.J. and Hillis, R.R., 2005. In situ stresses of the West Tuna area, Gippsland Basin. Australian Journal of Earth Sciences, 52(2): 299-313. O'Brien, P.E. and Wells, A.T., 1994. Sedimentology of the Bundamba Group. In: A.T. Wells and P.E. O'Brien (Editors), Geology and Petroleum Potential of the Clarence-Moreton Basin, New South Wales and Queensland: Bulletin 241. AGSO Canberra, pp. 72-137. Ogata, Y., Kobayashi, A., Mikami, N., Murata, Y. and Katsura, K., 1998. Correction of earthquake location estimation in a small-seismic-array system. Bernoulli, 4(2): 167-184. Oguchi, T., Aoki, T. and Matsuta, N., 2003. Identification of an active fault in the Japanese alps from DEM-based hill shading. Computers & Geosciences, 29(885-891). Olley, J.M. and Wasson, R.J., 2003. Changes in the flux of sediment in the Upper Murrumbidgee catchment, Southeastern Australia, since European settlement. Hydrological Processes, 17(16): 3307-3320. Olley, J.M., Wilkinson, S., Caitcheon, G. and Read, A., 2006. Protecting Moreton Bay: How can we reduce sediment an d nutrients loads by 50%?, River Symposium, Brisbane. Ollier, C.D., 1982. The Great Escarpment of eastern Australia: tectonic and geomorphic significance. Journal Geological Society of Australia, 29(1-2): 13- 23.

125

Ollier, C.D., 1985. Geomorphology and volcanism in eastern Australia. In: F.L. Sutherland, B.J. Franklin and A.E. Waltho (Editors), Volcanism in eastern Australia with case histories from New South Wales. Geological Society of Australia, New South Wales Division, Sydney, pp. 1-12. Ollier, C.D. and Stevens, N.C., 1989. The Great Escarpment in Queensland. In: R.W. Le Maitre (Editor), Pathways in Geology: Essays in Honour of Edwin Sherbon Hills. Blackwell, Melbourne, Australia, pp. 140-152. Ollier, C.D. and Haworth, R.J., 1994. Geomorphology of the Clarence-Moreton Basin. In: A.T. Wells and P.E. O'Brien (Editors), Geology and petroleum potential of the Clarence-Moreton Basin, New South Wales and Queensland: Bulletin 241. AGSO, Canberra. Ollier, C.D., 1995. Tectonics and landscape evolution in southeast Australia. Geomorphology, 12(1): 37-44. Ollier, C.D. and Pain, C.F., 1997. Equating the basal unconformity with the palaeoplain: a model for passive margins. Geomorphology, 19(1-2): 1-15. Ollier, C.D., 2001. Evolution of the Australian landscape. Marine and Freshwater Research, 52(1): 13-23. Orme, A.R., 2002. Shifting paradigms in geomorphology: the fate of research ideas in an educational context. Geomorphology, 47(2-4): 325-342. Ouchi, S., 1985. Response of alluvial rivers to slow active tectonic movement. Geological Society of America Bulletin, 96(4): 504-515. Ouyang, D. and Bartholic, J., 2001. Web-Based GIS Application for Soil Erosion Prediction, Proceedings of an International Symposium - Soil Erosion Research for the 21st Century, Honolulu. Pain, C.F., (no date). Size does matter: relationships between image pixel size and landscape process scales, Canberra ACT. Palyvos, N., Bantekas, I. and Kranis, H., 2006. Transverse fault zones of subtle geomorphic signature in northern Evia island (central Greece extensional province): An introduction to the Quaternary Nileas graben. Geomorphology, 76(3-4): 363-374. Parkhurst, D.L. and Appelo, C.A.J., 1999. User's guide to PHREEQC Ver. 2 - A computer program for speciation, batch-reaction, one-dimensional transport, and inverse geochemical calculations. U. S. Geological Survey Water- Resources Investigations Report, 99-4259, 310 pp. Peacock, D.C.P. and Sanderson, D.J., 1991. Displacement, segment linkage and relay ramps n normal fault zones. Journal of Structural Geology, 13: 721-733. Pettinga, J.R., 2004. Three-stage massive gravitational collapse of the emergent imbricate frontal wedge, Hikurangi Subduction Zone, New Zealand. New Zealand Journal of Geology and Geophysics, 47: 399-414. Philip, G., 2007. Remote sensing data analysis for mapping active faults in the northwestern part of Kangra valley, NW Himilaya, India. International Journal of Remote Sensing, 28(21): 4745-4761. Phinn, S. and Stanford, M., 2001. Monitoring Land-Cover and Land-Use Change in a Rapidly Urbanising Coastal Environment: the Maroochy and Mooloolah Rivers Catchments, Southeast Queensland, 1988 - 1997. Australian Geographical Studies, 39(2): 217 - 232. Pine Rivers Shire Council, 2004. http://www.healthywaterways.env.qld.gov.au/PAGE170624PMM3HMZ1.htm l.

126

Plumb, K.A., 1979. The tectonic evolution of Australia. Earth-Science Reviews, 14: 205-249. Posamentier, H.W., Jervey, M.T. and Vail, P.R., 1988. Eustatic controls on clastic deposition. I Conceptual Framework. In: C.K. Wilgus et al. (Editors), Sea- level Changes: An Integrated Approach. Society of Economic Paleontologists and Mineralogists, Special Publication 42, pp. 109-124. Posamentier, H.W. and Vail, P.R., 1988. Eustatic controls on clastic deposition. II Sequence and Systems Tract Models. In: C.K. Wilgus et al. (Editors), Society of Economic Paleontologists and Mineralogists, Special Publication 42, pp. 125-154. Powell, C.M., Edgecombe, D.R., Henry, N.M. and Jones, J.G., 1976. Timing of regional deformation at Hill End Trough: a reassessment. Journal of the Geological Society of Australia, 23: 407-421. Powell, J.W., 1875. Exploration of the Colorado River of the West and its Tributaries. Preda, M. and Cox, M.E., 2002. Trace metal occurrence and distribution in sediments and mangroves, Pumicestone region, southeast Queensland, Australia. Environment International, 28: 433-449. Prosser, I.P., Wilkinson, L.J., Hughes, A.O. and Caitcheon, G., 2003. Patterns of erosion and sediment and nutrient transport in the west Brisbane River catchment, Queensland, CSIRO Land and Water Report. Queensland Government, D.N.R.M., 2005. South East Queensland 25 metre Digital Elevation Model - SEQ_DEM_100K. Queensland Government, D.N.R.M.W., 2003. Queensland Geological Digital Data, CD ROM. Rayburg, S.C. and Neave, M., 2008. Assessing morphologic complexity and diversity in river systems using three-dimensional asymmetry indices for bed elements, bedforms and bar units. River Research and Applications: DOI: 10.1002/rra.1096. Read, J.F., Osleger, D. and Elrick, M., 1991. Two-dimensional modeling of carbonate ramp sequences and component cycles. In: E.K. Fransen (Editor), Sedimentary Modelling: Computer Simulations and Methods for Improved Parameter Definition. Bulletin of Kansas Geological Survey, pp. 472-488. Renard, K.G., Foster, G.R., Weesies, G.A., McCool, D.K. and Yoder, D.C., 1997. Predicting Soil Erosion by Water. A Guide to Conservation Planning with the Revised Universal Soil Loss Equation (RUSLE), Agricultural Handbook, vol 703. US Department of Agriculture, Washington, DC, pp. 1-251. Riley, J., Greenberg, R. and Sarid, A., 2006. Europa's South Pole Region: A sequential reconstruction of surface modification processes. Earth and Planetary Science Letters, 248(3-4): 808-821. Roberts, G.P., 1996a. Variation in fault-slip directions along active and segmented normal fault systems. Journal of Structural Geology, 18(6): 835-845. Roberts, G.P., 1996b. Noncharacteristic normal faulting surface ruptures from the Gulf of Corinth, Greece. Journal of Geophysical Research - Solid Earth, 101(B11): 25255-25267. Roberts, G.P., Cowie, P.A., Papanikolaou, I.D. and Michetti, A.M., 2004. Fault scaling relationships, deformation rates and seismic hazards: an example from Lazio-Abruzzo Appennines, central Italy. Journal of Structural Geology, 26: 377-398.

127

Roberts, G.P. and Michetti, A.M., 2004. Spatial and temporal variations in growth rates along active normal fault systems: an example from the Lazio-Abruzzo Appennines, central Italy. Journal of Structural Geology, 26: 339-376. Robertson, A.D., 1985. Upper mantle xenoliths and megacrysts and the origin of the Brigooda basalt and breccia, near Proston, Queensland. Papers - Department of Geology, University of Queensland, 11(3): 58-71. Roy, A.G., Jarvis, R.S. and Arnett, R.R., 1980. Soil-Slope Relationships within a drainage basin. Annals of the Association of American Geographers, 70(3): 397-412. Russ, D.R., 1982. Style and significance of surface deformation in the vicinity of New Madrid, Missouri. U.S. Geological Survey Professional Papers, 1236_H: 95- 114. Rynn, J.M.W., 1987. Queensland seismic risk study, Queensland Department of Mines, Brisbane. Sandiford, M., 2003. Neotectonics of southeastern Australia: linking the Quaternary faulting record with seismicity and in situ stress. Geological Society of Australia Special Publication, 22: 101-113. Santos, M.L.M., Guenat, C., Bouzelboudjen, M. and Golay, F., 2000. Three- dimensional GIS cartography applied to the study of the spatial variation of soil horizons in a Swiss floodplain. Geoderma, 97(3-4): 351-366. Sarup, J., Muthukumaran, M., Mathur, N. and Peshwa, V., 2006. Study of tectonics in relation to the seismic activity of the Dalvat area, Nasik District, Maharashtra, India using remote sensing and GIS techniques. International Journal of Remote Sensing, 27(12): 2371-2387. Scheidegger, A.E., 1965. The algebra of stream order numbers. United States Geological Survey Professional Paper, 525-B: 187-189. Scheidegger, A.E., 1979a. The principle of antagonism in the earth's evolution. Tectonophysics, 55(3-4): T7-T10. Scheidegger, A.E., 1979b. Orientationsstruktur der Talanlagen in der Schweiz. Geographica Helvetica, 34(1): 9-17. Scheidegger, A.E. and Ajakaiye, D.E., 1994. Mass Movements in Hilly Areas (with Examples from Nigeria). Natural Hazards, 9(1-2): 191-196. Scheidegger, A.E., 1998. Tectonic predesign of mass movements, with examples from the Chinese Himalaya. Geomorphology, 26(1-3): 37-46. Scheidegger, A.E., 2001. Surface joint systems, tectonic stresses and geomorphology: a reconciliation of conflicting observations. Geomorphology, 38(3-4): 213- 219. Scheidegger, A.E., 2002. Morphometric analysis and its relation to tectonics in Macaronesia. Geomorphology, 46(1-2): 95-115. Schlunegger, F. and Hinderer, M., 2001. Crustal uplift in the Alps: why the drainage pattern matters. Terra Nova, 13(6): 425-432. Scholz, C.H., 2002. The Mechanics of Earthquakes and Faulting. University Press, Cambridge, 471 pp. Schulz, W.H., 2007. Landslide susceptibility revealed by LIDAR imagery and historical records, Seattle, Washington. Engineering Geology, 89(1-2): 67-87. Schumm, S.A. and Khan, H.R., 1972. Experimental study of channel patterns. Geological Society of America Bulletin, 83(6): 1755-1770. Schumm, S.A., Rutherfurd, I.D. and Brooks, J., 1994. Pre-cutoff morphology of the lower Mississippi River. In: S.A. Schumm and B.R. Winkley (Editors), The

128

variability of large alluvial rivers. American Society of Civil Engineers Press, New York, pp. 13-44. Seidl, M.A., Weissel, J.K. and Pratson, L.F., 1996. The kinematics and pattern of escarpment retreat across the rifted continental margin of SE Australia. Basin Research, 12: 301-316. Selvaradjou, S., Montanarella, L., Carre, F., Jones, A., Panagos, P., Ragunath, K., Kumaraperumal, R. and Natarajan, S., 2007. An Innovative Approach for Updating Soil Information based on Digital Soil mapping Techniques, European Commission Directorate-General, Joint Research Centre, Luxembourg. Seppala, M., 1999. Geomorphological aspects of road construction in a cold environment, Finland. Geomorphology, 31(1-4): 65-91. Séranne, M., 1999. Early oligocene stratigraphic turnover on the West Africa continental margin: a signature of the Tertiary greenhouse-to-icehouse transition? Terra Nova, 11: 135-140. Serva, L., Blumetti, A.M., Guerrieri, L. and Michetti, A.M., 2002. The Appenine intermountain basins: the result of repeated strong earthquakes over a geological time interval. Bollettino della Societa Geologica Italiana., Volume Speciale 1: 939-946. Shanley, K.W. and McCabe, P.J., 1991. Predicting facies architecture through sequence stratigraphy - an example from the Kaiparowits Plateau, Utah. Geology, 19: 742-745. Sharp, J.M. and Hill, D.W., 1995. Land subsidence along the northeastern Texas Gulf coast: effects of deep hydrocarbon production. Environmental Geology and Water Sciences, 25: 181-191. Shreve, R.L., 1966. Statistical law of stream numbers. Journal of Geology, 74: 17-37. Shreve, R.L., 1967. Infinite topologically random channel networks. Journal of Geology, 75: 178-186. Singh, R. and Phadke, V.S., 2006. Assessing soil loss by water erosion in Jamni River Basin, Bundelkhand region, India, adopting universal soil loss equation using GIS. Current Science, 90(10): 1431-1435. Snyder, N.P., Whipple, K.X., Tucker, G.E. and D.J., M., 2002. Interactions between onshore bedrock-channel incision and nearshore wave-base erosion forced by eustasy and tectonics. Basin Research, 14: 105-127. Spotila, J.A., Bank, G.C., Reiners, P.W., Naeser, C.W., Naeser, N.D. and Henika, B.S., 2004. Origin of the Blue Ridge escarpment along the passive margin of Eastern North America. Basin Research, 16: 41-63. Srivastava, P., Sharma, M. and Singhvi, A.K., 2003. Luminescence chronology of incision and channel pattern changes in the River Ganga, India. Geomorphology, 51: 259-268. Steckler, M., Reynolds, D.J., Coakley, B.J., Swift, B.A. and Jarrard, R., 1993. Modelling passive margin sequence stratigraphy. In: H.W. Posamentier (Editor), Sequence Stratigraphy and Facies Associations. Special Publication International Association of Sedimentology, pp. 19-41. Stevens, N.C., 1965. The volcanic rocks of the southern part of the Main Range, south-east Queensland. Proceedings of the Royal Society of Queensland, 77(4): 37-52. Stevens, N.C., 1966. Tertiary volcanic rocks, In: Southern Moreton Basin: field conference. Geological Society of Australia. Queensland Division, pp. 11-14.

129

Stevens, N.C., 1969. The volcanism of southern Queensland. Geological Society of Australia Special Publication, 2: 193-202. Stevens, N.C., 1971. The Glasshouses. In: G. Playford (Editor), Geological Excursions Handbook, ANZAAS, 43rd Congress. ANZAAS and Geological Society of Australia, Queensland Division, Brisbane, pp. 153-158. Strahler, A.N., 1954. Quantitative geomorphology of erosional landscapes, C.-R. 19th International Geol. Cong., Algiers, 1952, pp. 341-354. Strahler, A.N., 1957. Quantitative analysis of watershed geomorphology. Transactions, American Geophysical Union, 38(6): 913-920. Strahler, A.N., 1960. Physical Geography. John Wiley and Sons, New York. Strahler, A.N., 1966. The Earth Sciences. Harper International, 681 pp. Sussmilch, C.A., 1933. The geomorphology of the Moreton District, Queensland; an interpretation. Proceedings of the Royal Society of Queensland, 44: 104-131. Sutherland, F.L., 1985. Regional controls in eastern Australian volcanism, Publication - New South Wales Division. Geological Society of Australia, New South Wales Division : Sydney, N.S.W., Australia, Australia, pp. 13-32. Sutherland, F.L., 2003. 'Boomerang' migratory intraplate Cenozoic volcanism, eastern Australian rift margins and the Indian-Pacific mantle boundary. Special Paper - Geological Society of America, 372: 203-221. Taylor, R.G. and Howard, K.W.F., 1999. The influence of tectonic setting on the hydrological characteristics of deeply weathered terrains: evidence from Uganda. Journal of Hydrology, 218: 44-71. Taylor, T.G., 1911. Physiography of Eastern Australia: Bulletin 8, Commonwealth Bureau of Meteorology., Melbourne. Tebbens, L.A., Veldkamp, A., Van Dijke, J.J. and Schoorl, J.M., 2000. Modeling longitudinal-profile development in response to Late Quaternary tectonics, climate and sea-level changes: the River Meuse. Global and Planetary Change, 27(1-4): 165-186. Tejero, R., Gonzalez-Casado, J.M., Gomez-Ortiz, D. and Sanchez-Serrano, F., 2006. Insights into the "tectonic topography" of the present-day landscape of the central Iberian Peninsula (Spain). Geomorphology, 76(3-4): 280-294. Thomas, M.F., Nott, J., Murray, A.S. and Price, D.M., 2007. Fluvial response to late Quaternary climate change in NE Queensland, Australia. Palaeogeography, Palaeoclimatology, Palaeoecology, 251: 119-136. Thoms, M.C. and Sheldon, F., 2002. An ecosystem approach for determining environmental water allocations in Australian dryland river systems: the role of geomorphology. Geomorphology, 47(2-4): 153-168. Thornes, J.B. and Alcantara-Ayala, I., 1998. Modelling mass failure in a Mediterranean mountain environment: climatic, geological, topographical and erosional controls. Geomorphology, 24(1): 87-100. Tucker, R.M., 1967. Geology of the City of Brisbane. Part 1. The pre-Mesozoic sediments and volcanics and the related intrusives. Publications of the Geological Survey of Queensland, 324(1-6). Twidale, C.R., 1966. Late Cenozoic activity of the Selwyn up-warp, northwest Queensland. Journal of the Geological Society of Australia, 13: 491-494. Twidale, C.R., 1972. The neglected third dimension. Zeitschrift für Geomorphologie, 16: 283-300. Twidale, C.R., 1980. Geomorphology. Thomas Nelson, 406 pp.

130

Twidale, C.R., 2004. River patterns and their meaning. Earth-Science Reviews, 67(3- 4): 159-218. Twidale, C.R. and Bourne, J.A., 2004. Neotectonism in Australia; its expressions and implications. Geomorphologie, 2004(3): 179-193. USDA, 2008. Soil and Water Assessment Tool/ArcSWAT. USDA Agricultural Research Service at the Grassland, Soil and Water Research Laboratory, Temple, Texas, USA. Vail, P., 1987. Seismic stratigraphy interpretation using sequence stratigraphy. Part 1: Seismic stratigraphy interpretation procedure. In: A.W. Bally (Editor), Atlas of Seismic Stratigraphy. American Association of Petroleum Geologists Studies in Geology, 27, pp. 1-10. Vail, P.R., Mitchum, R.M. and Thompson, S., 1977. Seismic stratigraphic and global changes in sea level. Part 4: Global cycles of relative changes in sea level. In: C.E. Payton (Editor), Seismic stratigraphy - applications to hydrocarbon exploration. American Association of Petroleum Geologists, Memoir 26, pp. 83-97. Vannoli, P., Basili, R. and Valensise, G., 2004. New geomorphic evidence for anticlinal growth driven by blind-thrust faulting along the northern Marche coastal belt (central Italy). Journal of Seismology, 8(3): 297-312. Veevers, J.J. and Evans, P.R., 1975. Late Palaeozoic and Mesozoic history of Australia. In: K.S.W. Campbell (Editor), Gondwana Geology: papers from the 3rd Gondwana symposium. ANU Press, Canberra, Canberra, ACT, pp. 579- 607. Veevers, J.J., 2001. Atlas of Billion-year earth history of Australia and neighbours in Gondwanaland. GEMOC Press, Sydney, 76 pp. Vétel, W., Le Gall, B. and Johnson, T.C., 2004. Recent tectonics in the Turkana Rift (North Kenya): an integrated approach from drainage network, satellite imagery and reflection seismic analyses. Basin Research, 16: 165-181. Walsh, F., 1972. Application of stream order numbers to the Merrimack River basin. Water Resources Research, 8: 141-144. Ward, W.T. and Hacker, J.L.F., 2006. Brisbane Airport: an alluvial landscape veiled by marine sediments. Australian Journal of Earth Sciences, 53(6): 1001-1012. Watkins, J.R., 1967. The relationship between climate and the development of landforms in the Cainozoic rocks of Queensland. Journal of the Geological Society of Australia, 14(1): 153-168. Webb, A.W. and McDougall, I., 1967. Isotopic dating evidence on the age of the Upper Permian and Middle Triassic. Earth and Planetary Science Letters, 2: 483-488. Webb, A.W., Stevens, N.C. and McDougall, I., 1967. Isotopic age determinations on Tertiary volcanic rocks and intrusives of southeastern Queensland. Proceedings of the Royal Society of Queensland, 79: 79-92. Welch, D.M., 1973. Channel form and bank erosion, Red River, Manitoba, Fluvial Processes and Sedimentation. Proceedings of the Hydrology Symposium, University of Alberta, pp. 284-293. Wellman, P., Williams, J.W. and Maher, A.R., 1994. Interpretation of gravity and magnetic anomalies in the Clarence-Moreton Basin region. In: A.T. Wells and P.E. O'Brien (Editors), Geology and petroleum potential of the Clarence- Moreton Basin, New South Wales and Queensland: Bulletin 241. AGSO, Canberra, pp. 217-229.

131

Wells, D.L. and Coppersmith, K.J., 1994. New empirical relationships among magnitude, rupture length, rupture width, rupture area and surface displacement. Bulletin of the Seismological Society of America, 84(4): 974- 1002. Westaway, R., 2002a. Geomorphological consequences of weak lower continental crust, and its significance for studies of uplift, landscape evolution and the interpretation of river terrace sequences. Netherlands Journal of Geosciences, 81(3-4): 283-303. Westaway, R., 2002b. Long-term river terrace sequences: Evidence for global increases in surface uplift rates in the Late Pliocene and early Middle Pleistocene caused by flow in the lower continental crust induced by surface processes. Netherlands Journal of Geosciences, 81(3-4): 305-328. Whipple, K.X., 2004. Bedrock rivers and the geomorphology of active orogens. Annual Review of Earth and Planetary Sciences, 32: 151-185. Whitaker, W.G. and Green, P.M., 1980. Moreton Geology 1:500 000 map. Department of Mines, Queensland. Whitehouse, F.W., 1940. Studies in the Late Geological History of Queensland. Papers of the Department of Geology, University of Queensland, 2, 74 pp. Whitlow, R., 2000. The Geology of the Gold Coast Region, TriMap Pty Ltd, Prepared for Strategic and Environmental Branch, Planning Directorate, Gold Coast City Council. Williams, M., Dunkerley, D., De Decker, P., Kershaw, P. and Chappell, J., 1998. Quaternary Environments. Hodder Headline Group, London. Willmott, W., 2004. Rocks and Landscapes of the National Parks of Southern Queensland. Geological Society of Australia Incorporated, Queensland Division, Brisbane, 162 pp. Willmott, W.F., 1987. Landslides and Land-use planning on the Malaney-Mapleton Plateau. In: C.G. Murray and J.B. Waterhouse (Editors), Field Conference: Gympie District. Geological Society of Australia, Queensland Division, Queensland, Australia, pp. 114-118. Willmott, W.F. and Suwitadiredja, D.A., 2003. Slope stability and its constraints on closer settlement - Mouont Mee Plateau, south-east Queensland. Queensland Geological Record 2003/03, Department of Mines and Energy, Queensland Government. Wilson, M.M., 1973. Geological notes on the Rocksberg Greenstones and overlying rock sequences, Field Conference Guidebook, Brisbane-Gympie Area. Geological Society of Australia, Queensland Division, pp. 19-29. Winkler, H.G.F., 1967. Protogenesis of metamorphic rocks. Springer-Verlag, New York, 237 pp. Woldenberg, M.J., 1967. Geography and properties of surfaces. Harvard Papers in Theoretical Geography, 1: 95-189. Wood, L.J., Ethridge, F.G. and Schumm, S.A., 1993. The effects of rate of base-level fluctuation on coastal plain, shelf and slope depositional systems: an experimental approach. In: H.W. Posamentier, C.P. Summerhayes, B.U. Haq and G.P. Allen (Editors), Sequence Stratigraphy and Facies Associations. International Association of Sedimentologists Special Publication, 18, pp. 43- 54.

132

Woolfe, K.J., Larcombe, P., Naish, T. and Purdon, R.G., 1998. Lowstand rivers need not incise the shelf: an example from the Great Barrier Reef, Australia, with implications for sequence stratigraphic models. Geology, 26: 75-78. Woolfe, K.J., Larcombe, P., Orpin, A.R. and Purdon, R.G., 2000. Spatial variability in fluvial style and likely response to sea-level change, Herbert River, Queensland. Australian Journal of Earth Sciences, 47: 689-694. Woolnough, W.G., 1927. The Chemical Criteria of Peneplanation. Journal of the Proceedings of the Royal Society of New South Wales, 61, 153 pp. Zelilidis, A., 2000. Drainage evolution in a rifted basin, Corinth Graben, Greece. Geomorphology, 35: 69-85. Zernitz, E.R., 1932. Drainage patterns and their significance. Journal of Geology, 40: 498-521. Zoback, M.D. and Zoback, M.L., 1991. Tectonic stress field of North America and relative plate motion. In: D.B. Slemmons, E.R. Engdahl, M.D. Zoback and D.D. Blackwell (Editors), Neotectonics of North America. Geological Society of America, pp. 339-366. Zoback, M.L., 1992. First- and second- order patterns of stress in the lithosphere: the World stress map project. Journal of Geophysical Research(97): 11703-11728.

133

134

PAPER 1

TITLE The influence of geological fabric and scale on drainage pattern analysis in a catchment of metamorphic terrain: Laceys Creek, southeast Queensland, Australia

AUTHORS Jane Helen Hodgkinson, Stephen McLoughlin, Malcolm Cox

School of Natural Resource Sciences Queensland University of Technology

Published in Geomorphology November 2006

135

STATEMENT OF ORIGINAL AUTHORSHIP

Jane Helen Hodgkinson (PhD Candidate): reviewed previous work and literature; planned and conducted field work, collated datasets and conducted analysis, devised ordering system and interpreted data; wrote paper

Stephen McLoughlin (Principal PhD Supervisor): reviewed research program; discussed methods and ordering system; reviewed, discussed and edited paper

Malcolm Cox (Associate PhD Supervisor): reviewed, discussed and edited paper

136 Abstract

The relationship between geological fabric and drainage patterns in the 81.8 km2 Laceys Creek sub-catchment of the North Pine River catchment, southeast Queensland, Australia, is analysed using a new channel-ordination system. The Laceys Creek catchment is situated on the South D'Aguilar Block, which underwent metamorphism, faulting and uplift from the Late Carboniferous to Late Triassic. The catchment drains exposures of two main rock units, the Neranleigh-Fernvale Beds and the Bunya Phyllite. Both units are composed of metamorphosed deep-sea sediments that accumulated as an accretionary wedge during late Palaeozoic subduction of the palaeo-Pacific plate under the eastern margin of the Australian craton. The new channel ordination system used in this study allows improved classification of stream segments of equal prominence or rank in comparison to previous schemes. A 10 m contour digital elevation model (DEM) was produced within which drainage channels were digitised. Planar geological features, including bedding, faults, joints and cleavage, were mapped in the field and collated with data from previous geological mapping programs. Regional and local trends of geological fabric are reflected in the variable orientation of channels of different rank in the catchment. Cleavage and fractures are the dominant planar features of the Bunya Phyllite and these correlate most closely with the orientation of middle-order incised stream segments. In contrast, middle- order channels on the Neranleigh-Fernvale Beds most closely correlate with bedding, which dominates the fabric of this unit. Although anthropogenic factors exert local influence and climatic processes exert broad influence on the catchment, this study focuses on structural and lithological fabrics, which are the apparent dominant controls on middle-order channel orientations. Identification of congruent patterns between bedrock fabric and channel ranks is variable, depending on the scale and number of channels included in the analysis. Many low-rank channels correspond closely to the orientation of fine-scale bedding and foliation and these influences may not be detected by coarse-scale mapping. Understanding the extent of geological controls on the morphology of a catchment may assist geo-hazard identification, land use planning and civil-engineering projects.

137 Key words: Metamorphic terrain; Catchment; Channel orientation; Drainage pattern; Channel ordination; Queensland.

1. Introduction

1.1 Background The relationship between geology and landforms has long been established (e.g. Hobbs, 1904, 1911; Zernitz, 1932; Twidale, 1980). Structure and slope are accepted to be the primary controls on the spatial arrangement of channels. Some rivers show correlation and repeating alignment with straight or gently arcuate structural features such as faults, cleavage and fold axes to produce drainage systems with a recognisable pattern (e.g., trellis, rectangular and parallel drainage networks). In areas of poor outcrop, channel patterns can be a valuable guide to underlying geological structures (Twidale, 1980). Initially following slope, rivers will later adjust to structure as they incise into bedrock (Twidale, 2004). However, slope itself may be controlled by active tectonism, implying that even low order channels, prior to incision, may be geologically controlled. Rock fabric has been shown to influence drainage patterns at very fine scales (e.g. Twidale, 1972; Scheidegger, 1979a, b; Ackermann et al., 1997; Eyles et al., 1997; Eyles and Scheidegger, 1999; Beneduce et al., 2004). It has long been assumed the relationship occurs as a result of zones of weakness in the bedrock becoming enhanced by weathering and erosion processes, although data mostly have been insufficient to confirm this (Ericson et al., 2005). Where exposures are abundant and regolith cover does not obscure structural details, it has been possible to accurately map and measure large numbers of fracture traces but this method is less suitable for assessing the structure-drainage relationship in catchments where regolith, vegetation and land uses conceal natural bedrock features. Differential weathering of various metamorphic rocks will form belts of hills and ridges depending on the resistance of each rock unit (Strahler, 1966). Differential weathering of internal fabric, such as cleavage, will be expressed by finer-scale controls on the landscape (Twidale, 1972). Some previous studies have assessed diverse structural influences on drainage pattern. High-quality aerial photographs have been used to demonstrate the relationship between joints and channels in part of a granite-dominated catchment in the Sierra Nevada, USA, where the high degree of bedrock exposure assisted

138 correlation. In places, this relationship was evident irrespective of regional slope (Ericson et al., 2005). At a fine scale, despite the presence of regolith and ground cover, studies have shown a correlation between bedrock joints and gully orientation and, based on investigations in New South Wales, the strength of this correlation is strongly dependent on the bedrock lithology (Beavis, 2000). In the Macaronesian Islands, stream orientations and ridge trends are similarly oriented, which suggests a common tectonic control on the landforms (Scheidegger, 2002). On a coarser scale, tectonic fabric was also found to exert a first-order control on landscape patterns in the Chinese Himalayas (Scheidegger, 1998). Similarly, slopes and valley trends were found to strongly correspond with Appenine and anti-Appenine lineaments in central Italy (Alexander and Formichi, 1993). Tectonic controls appear to be the main influences on many other large-scale landforms, particularly valleys and gorges (Scheidegger, 2001). Although suspected for many decades, a clear relationship between geological controls and trellis style channel networks was found to be weak by some workers (Mock, 1971, 1976). More recently such a relationship has been quantitatively demonstrated (Abrahams and Flint, 1983). Bedrock fractures and metamorphic cleavage have been suggested as the likely controls on the incised meanders of the Shenandoah River, USA, and the channel and gorge systems of the meandering River Torrens, Mt Lofty Ranges, Australia, respectively (Twidale, 2004). Such previous studies have clearly identified relationships among fracture patterns, cleavage direction and stream orientation, but this does not imply that all channels or valleys are endogenically controlled. The arrangement of non-random channel orientations may be controlled by endogenic influences such as geological structure or by ancient stress regimes (Scheidegger, 1979b). If channel orientations prove to be random, exogenic forces such as climate are most likely the cause (Scheidegger, 1998).

1.2 Purpose of this study

This study aims to identify the extent to which structural grain and lithological fabric, collectively described herein as geological fabric, controls channel alignment. The relationship between the drainage system morphology and geological fabric in a high-energy, sub-tropical catchment dominated by low-grade

139 metamorphic rocks in southeast Queensland, Australia is investigated (Fig. 1a, b). The dominant rock units are phyllitic and are transected by a range of structural discontinuities (faults and joints) that are variably orientated with respect to cleavage. Geological fabric elements in the study area include: bedding, cleavage, faults, and joints , hereafter cumulatively referred to as lineaments (their expression in outcrop) or planar features (their two-dimensional representation). Geological structures occur at a variety of scales, and consequently their effects on channel form might be expected at varying magnitudes. The study area is the Laceys Creek catchment (Fig. 1c), which has two primary rock units of similar sedimentary origin but of different, albeit low, metamorphic grade. Both rock types have been subjected to almost the same climate across the catchment so differences in endogenic controls on channel form should be quantifiable within each lithological unit. Despite the complex nature of the planar geological features caused by multiple deformation events and the nature of the metamorphic lithology, any relationship between drainage and structure should be determinable by detailed mapping and analysis of preferred orientations of channel segments. This study aims to identify whether stream pattern varies across the catchment and whether the lithological fabric of each rock unit affects stream orientation and the pattern of channels. It also aims to demonstrate the extent to which channels are controlled endogenically and exogenically, and finally illustrate how portrayal of network orientation is affected by varying the scale of the mapped channels. Many civil engineering and land management projects employ evaluation of only coarse-scale lithological and structural properties of the bedrock but greater detail may be required for specific local-scale projects. Changing the resolution of the data may significantly alter their interpretation. Therefore, fine-scale channel mapping and the effects of the underlying geology upon the position of those channels may also be of assistance to local land-use and construction planning in the target catchment.

2. The study area

The Laceys Creek catchment is the largest subcatchment of the North Pine River system and is situated approximately 50 km north of Brisbane, southeast Queensland (Fig 1a, b). The catchment is

140 located upstream of the artificial Lake Samsonvale, an important water resource for southeast Queensland. The hilly to mountainous terrain of the catchment is sparsely populated, with fewer than 1000 residents. Although the catchment has not been widely developed, forest was cleared in 60% of the catchment during the early 1900s and at present this area is used for grazing and rural-residential land; the remaining 40% is forest reserve and state forest. Natural vegetation cover consists mostly of Eucalyptus-dominated open sclerophyll forest with minor riparian closed-vine forest. The region experiences predominantly summer rainfall, but this may be erratic and of variable intensity. The catchment is located within the South D’Aguilar Block of the New England orogenic belt, which is characterised by north-northwesterly trending geological assemblages (Fig. 2a) that represent volcanic arc, forearc/backarc basins and subduction complexes (Murray et al., 1987; Coney, 1992; Little et al., 1992; Holcombe et al., 1997a,b; Betts et al., 2002). The South D’Aguilar Block consists of steeply dipping, north-northwesterly to south-southeasterly striking, pre-Permian meta-sediments and meta- volcanics (Denmead, 1928; Belford, 1950; Bryan and Jones, 1962; Tucker, 1967; Wilson, 1973; Cranfield et al., 1976). The Bunya Phyllite and the Neranleigh-Fernvale Beds are the two main geological units in the Laceys Creek catchment (Fig. 2). They are of similar sedimentary origin but have different fabrics. The Bunya Phyllite crops out in a north-northwest-trending belt on the southwest flank of the Rocksberg Greenstone in the Pine Rivers catchment (Fig. 2).

Fig 1. Study area catchment (a) Location in relation to major Australian cities. (b) Location of North Pine River catchment in relation to other local geological provinces of southeast Queensland. (c) Map showing drainage channels of the North Pine catchment detailing Laceys Creek catchment, shaded grey.

141 Fig 2. Geological setting. (a) Geology of the North Pine catchment and boundaries of the North Pine River and Laceys Creek catchments (Queensland Government, 2003). (b) Geology of the Laceys Creek catchment. Stream reaches have been straightened from source to outlet (Queensland Government, 2003).

The Bunya Phyllite overlies the Rocksberg Greenstone, and the transition is represented by the intercalation of meta-sediments and meta-volcanics. The unit is strongly foliated and consists of lutites and arenites metamorphosed to greenschist facies. It is a dominantly pelitic suite of metasediments with some intercalated metabasic volcanic rocks. Alternating micaceous and quartzose bands are transected

142 by veins of white quartz with accessory graphite and calcite. Grain size is generally less than 0.01 mm and the mineral assemblage is primarily quartz, muscovite and chlorite and is assigned to the lowest Barrovian-type greenschist subfacies (Winkler, 1967). The original sediments are interpreted as having been deposited in a relatively deep-water marine environment and may represent oceanward facies equivalents of the Neranleigh-Fernvale Beds. Cranfield et al., (1976) suggested the age of the Bunya Phyllite is not younger than Carboniferous. Within the Laceys Creek catchment, the Bunya Phyllite covers 12.9 km2 and forms rugged slopes with elevations of 80 to 420 a.s.l. The Neranleigh-Fernvale Beds are exposed against the southwestern margin of the Bunya Phyllite and cover an area of 68.9 km2 in the Laceys Creek catchment (Fig. 2). Exposures form rugged hills with elevations of 100 to 660 m a.s.l. The unit consists of argillaceous rocks, conglomerates, radiolarian cherts, arenites, basic volcanics and minor limestone (Cranfield et al., 1976). Arenites are well developed within the Pine Rivers area, and thinly bedded shale and siltstone are widespread, in some places sheared and phyllitic. Original bedding and cross-bedding are well preserved. The representative lithologies and bedforms suggest deposition took place in a relatively deepwater marine environment, experiencing periodic turbidity flows. The unit has been regionally metamorphosed to the lowest Barrovian-type greenschist subfacies but to a grade less than that of the Bunya Phyllite (Winkler, 1967). The mineral assemblage consists of quartz, sericite and chlorite. Triassic intrusions, including the Mount Samson and the Samford granodiorites, have thermally metamorphosed the Neranleigh-Fernvale Beds in the Pine Rivers area (Fig. 2a). An accurate age for the Neranleigh-Fernvale Beds has not been identified but is placed between the mid-Devonian (Fleming et al., 1974) and mid-Carboniferous (Green, 1973). The unit was subject to low-grade metamorphism towards the end of the Carboniferous (Cranfield et al., 1976). The Neranleigh-Fernvale Beds are considered to have been thrust over the Bunya Phyllite although the exact stratigraphic relationship between these units has not been clearly established. A major shear zone is located along their contact (Cranfield et al., 1976). The beds have been folded and several north-northwest trending fold axes have been mapped (Denmead, 1928; Belford, 1950; Cranfield et al., 1976). However, due to a paucity of detailed work in this area, interpreting precise structural relationships is difficult (Cranfield et al., 1976). During the Late

143 Carboniferous New England orogenic event, the pre-Permian rocks were folded along north-northwesterly trending axes. Further deformation occurred during the Late Permian and Early Triassic, followed by block faulting and associated emplacement of post-tectonic granitic intrusions several of which are within the North Pine catchment. During the late Middle Triassic further folding took place along north and northwest trending axes (Cranfield et al., 1976; Holcombe et al., 1993, 1997a). The multiple episodes of deformation experienced by the pre-Permian sediments and volcanics of the South D’Aguilar Block produced diverse bedding, cleavage, fault and joint orientations within the units. The Paleogene saw development of small intermontane sedimentary basins through eastern Australia, including the Petrie Basin within the North Pine catchment, situated at the southeastern edge of Lake Samsonvale (Fig. 2a). The Petrie Basin received 300 m of sediment and is considered to be partly erosional and partly tectonically controlled (Cranfield et al., 1976). The Paleogene sediments are largely undisturbed except along rejuvenated basement faults such as the North Pine Fault (Fig. 2a). Three of the main drainage pattern categories (e.g. Zernitz, 1932; Twidale, 2004) can be used to describe the drainage pattern of the Laceys Creek catchment. Drainage on the Bunya Phyllite is generally dendritic. The drainage pattern on the Neranleigh-Fernvale Beds is generally dendritic with a strong parallel influence and angular junctions; a centrifugal drainage pattern is observed in the most proximal part of the network. Hill slopes are typically 20-25° in the upper portion of the study area, becoming more gentle closer to the main perennial channel. The trunk channel of the network has a mean slope gradient of ca. 1°. Overall, the incision of the channels forms ‘V’ shaped valleys. Shallow rudosols and tenosols (Isbell, 1996) dominate the catchment on the steep slopes, but thicker profiles are seen on very gently inclined lower slopes and in depressions. Laterite occurs on some of the lower slopes and flats.

3. Methods

Existing digital maps available for the area provide only coarse resolution and greatly under-represent small channels. Analysing only the channels from those maps

144 would lead to results biased to the orientation of larger channels only. To reduce this effect, an alternative map of channel networks was compiled (Figs. 1c and 2). NASA Shuttle Radar Topography Mission (SRTM) data (NASA, 2004) and the MapInfo Professional 7.8 software were used to extract a 10 m contour digital elevation model (DEM) of the catchment. Where contour deflections indicated concave-downhill patterns, channels were digitised over the base map as a separate GIS layer. A proportion of the channels digitised were randomly selected to verify the accuracy of their positions, using aerial photographs, existing maps and field observations using a compass and GPS. Catchment and subcatchment boundaries were also digitised following topographic highs around each drainage basin. Digital geological maps of lithological and structural data collated previously (Queensland Government, 2003) were also included as additional GIS layers. To quantitatively measure the channels and their orientations, channels were ‘straightened’ following the method of Scheidegger (1979a) by digitising axes from node to node throughout the network (Fig. 2). The orientation of each channel reach was then measured. Because identification of first-order channel head locations can be problematical (Heine et al., 2004) and lead to spurious orientation measurements for these reaches, the orientation data of first order channels were treated with caution. Standard statistical procedures cannot confidently be applied to directional data due to the very nature of circularly derived measurements where 0° = 360° (Jones, 1968). Statistical analysis of directional data has been explored to some extent in a variety of scientific disciplines where orientation data naturally occur (Krieger Lassen et al., 1994). Parametric orientation statistics in relation to earth sciences have been discussed by Kohlbeck and Scheidegger (1985). Statistical methods commonly seek to describe a mean value, deviation from that mean, or closeness of fit between data sets. However, for the purposes of this study, it is more important to seek correlations between the various datasets, which themselves may display multiple clustering of data causing further complications to the analysis. In an attempt to avoid the problems associated with circular data in standard statistical procedures, simple statistical analysis was performed on selected segments of the circular data where the selection did not incorporate orientations towards or either side of 360°. For example, a portion of the dataset ranging from 90-180° was analysed and subsequently narrowed to incorporate only those data within the 140-

145 170° range. Normally it would be inappropriate not to apply the same method to an entire dataset, but in this situation, it was not possible to do so as the problem of analysing circular data returns when the portion of the selected dataset includes 360/0°. Additionally, this practice may intentionally exclude pertinent peripheral data that may relate to minor trends present. Although the results of selective statistical analysis are discussed, the current methodologies are not deemed to be reliable as a quantitative analytical method for this field of study. A review of current literature suggests that there is not yet a suitably rigorous technique to statistically analyse circular orientation data. Therefore, visual comparison of channel orientation patterns has been essential throughout this study. Channel orders with similar trends were grouped together and a new group was formed where the dominant trend showed significant change. Channel orders were grouped to both simplify diagrams and to better indicate visually where trends change as orders progress through the drainage network.

3.1 Ordering the channels

In order to partition the channels into similar channel-order ‘sets’, the catchment channels were ordered, using a hierarchical numbering system. As this study seeks to determine if different structural features typically control stream segments of different magnitude, a stream ordering system is required that groups channel segments on the basis of similarity of scale and position within the drainage network. It is generally accepted that all finger-tip tributaries or channels in a network, are designated as 1st order. This suggests the 1st order channels have something in common and are comparable with one another and, therefore, may be analysed as a set. To analyse the whole network, all channel sets, or orders, should be similarly comparable. The ordering scheme used for this work also needs to ensure that every channel order increases after each node, so that every successive reach is identified and separable from the rest of the network for analysis. Channel-order numbers of some schemes, such as the Horton (1945) and Strahler (1954) methods, may be manipulated in various ways. Similarly derived dimensionless numbers might be treated in a similar way. Shreve (Shreve, 1966, 1967) noted that Horton’s and Strahler’s Laws should be expected from any topologically random distribution. A later review of the relationships confirmed this

146 argument, establishing that, from the properties the laws describe, no conclusion can be drawn to explain the structure or origin of the stream network (Kirchner, 1993). However, for practical purposes, ordering systems continue to be used as ranking systems (Orme, 2002). Some mathematically derived ordering systems, such as those proposed by Scheidegger (1965), Woldenberg (1967) and Walsh (1972), provide complex results that lead to operational difficulties (Gardiner, 1975). Therefore, these methods have not been incorporated for this project. The three most commonly used ordering systems are those of Horton (1945), Strahler (1957) (now commonly referred to as the Horton-Strahler method) and Shreve (1967) and each was assessed for suitability of channel ranking within this project. In summary, the Horton (1945) method was unsuitable as a high order may be designated to any scale of channel from fingertip tributaries through to major rivers, and comparison between orders grouped in this way cannot be justified for this study (Fig. 3a). The Strahler (1957) ordering system, a derivation of the Horton method, now referred to as the Horton-Strahler method) was inappropriate as it does not allocate a new order at every node, disregards some tributaries when orders are designated and effectively ignores the presence and influence (discharge and capacity) of some channels in the system (Fig. 3b). The Shreve (1967) numbering system,

Fig 3. Differences among stream ordering systems: a) Horton (1945); b) Strahler (1957); c) Shreve (1967); d) new system described in this paper. to some extent provides order numbers that relate to channel magnitude or position and is more suitable for the present study than the Horton and Strahler methods.

147 However, it leads to situations where similarly positioned channels are assigned to substantially different orders and the relationship between channels in each order is inadequate to allow comparison in this study (Fig. 3c). Since none of the schemes was fully suited to the requirements of this study, a new numbering method was devised. As with the Horton, Strahler and Shreve methods, fingertip tributaries are designated as 1st order (Fig. 4a), and the ordering system is then ‘counted’ in a downstream direction. Once designated to an order, a channel is then hidden, temporarily pruned or ‘greyed out’ depending on the technique available on the GIS package. The next channels that appear to be fingertip tributaries on the remaining network are designated the 2nd order (Fig. 4b). These can then be hidden, removed or greyed out revealing the next order of channels that again appear to be fingertip tributaries and are designated 3rd order (Fig. 4c). This method is carried out systematically throughout the entire catchment network, always treating the visible channels that appear to be fingertips as the new order (Fig. 4d-f).

Fig 4. Portion of the Laceys Creek catchment showing the methodology of the new channel ordering system. a) The fingertip tributaries are designated as order 1 (or 1st order); b) By greying out or removing the 1st order channels, those channels that now appear to be the fingertip tributaries are designated the 2nd order; c) By greying out or removing the 2nd order channels, those channels that would appear to be fingertip are designated the 3rd order; d) Similarly, 3rd order channels are removed or greyed out to reveal channels now designated as 4th order; e) The next channel that appears to be fingertip is designated as 5th order; f) The final order in this sub-catchment is designated as 6th order.

148 This ordination system need not be numerical. Each order may be given a letter, or for visual aids, a colour and be equally effective as the order descriptor remains dimensionless. However, numbers provide a limitless supply of sequential descriptors and are, therefore, possibly the most useful, coupled with varying channel colours in a GIS. The order designated to each channel reach will be a relative indication of influence upon that reach from others upstream (like the Shreve method but unlike the Horton and Strahler methods). Although working on a similar principle to the Shreve method, it allows for all orders to be accounted for in the system, keeps the final order number low, and provides a more suitable method for comparative studies across networks and between catchments. All orders can be sequentially removed and remaining ‘fingertip’ channels are of equal rank and therefore comparable. No other scheme achieves this while assigning every successive reach to a consecutive separate order. The number designated to each order is dimensionless, yet since it is relative to others in the network, it is useful as a comparative tool regarding upstream influence upon that particular order. It also allows groups to be removed from the network in order to analyse the remaining network. Although this method creates a scale-dependant system, at present a method for scale independency has not be established. This method can be used to show the level to which analytical results are scale dependant by sequentially removing orders of the same weighting, and viewing the remaining catchment.

3.2 Rock fabric Field work was carried out to measure the orientation of planar features to augment data supplied by Queensland Government DNR (Queensland Government, 2003). Given the paucity of outcrop within the Laceys Creek area, particularly on the Bunya Phyllite, data was collected from additional sites (primarily road cuttings) close to, but outside of the study area. Strike and dip of planar-features were measured using the ‘right-hand-thumb along strike’ rule, where dip is recorded 90° clockwise from strike. Strikes range from 1° to 360° requiring a circular rose diagram for visual representation. Each planar-feature strike is a double-headed vector, causing the opposite direction of each measurement to be equally relevant to this analysis. Both orientations are incorporated in the portrayed data giving the rose diagrams 180° symmetry (Krumbein, 1939; Jones, 1968; Kohlbeck and Scheidegger, 1985). The

149 strike of each planar feature was sorted by both feature type and lithological unit in which it appeared and rose diagrams were produced to visually express the datasets as a circular distribution (Krumbein, 1939; Jones, 1968; Reyment, 1971).

All fingertip Sequential Every reach of the Numbers treated tributaries ordering (orders network accounted as designations designated as 1st present increase for by increase of rather than values Reference order in units of 1) order number Horton NO YES1 NO YES (1945) Strahler YES YES NO YES (1957) Shreve YES NO YES NO (1967)

This paper YES YES YES YES

1Although numbers increase downstream on the whole, the Horton system reassigns entire streams with a higher order and as such, does not entirely fit these criteria (see also Fig. 3a). Table 1 Comparisons among the Horton, Strahler and Shreve schemes and proposed new method.

Channel-reach measurements were taken in the direction of flow, therefore, each datum was within the range of 1° to 360° and these are represented on circular rose-diagrams as unidirectional vectors. Rose diagram peaks are then used to identify the mean orientations of each dataset (Müller-Salzburg, 1963). Orientations of all channels were first grouped depending on underlying lithology. Rose diagrams were constructed to show the orientation trends for each channel order, as well as for clusters of channel orders. Rose diagrams were produced using all channels in the network on each lithology and further diagrams were produced after removing low- order and then low- and middle-order channels to assess the effects of scale on network orientation analysis. The symmetrical rose diagrams of planar features were compared with asymmetrical diagrams of channel orientations for each rock unit.

4. Results Well preserved, thin (3-10 cm), turbidite beds, some containing planar- and cross-lamination, are evident in the Neranleigh-Fernvale Beds. Moderate weathering has occurred, including mechanical weathering by root wedging. Some preferential weathering of the coarser fractions within the turbidites has formed corrugated exposures. Slatey cleavage is locally evident but the planar surfaces are poorly

150 defined. Crenulation cleavage, along which quartz is present, was observed. Fractures are commonly filled with quartz. Two main trends of bedding planes are evident across the catchment, striking approximately 100° and 138° from north and dipping approximately 50° and 80° towards the southwest, respectively (Fig. 5a). Fractures typically have northeast strikes and dip 40° to 90° to the southeast or similar angles to the northwest (Fig. 5b). Cleavage is well developed in the Bunya Phyllite, within the Laceys Creek catchment. Fractures are common and bedding is seen rarely. Quartz veins are located along some cleavage, fracture and bedding planes. Deformation of primary foliation and of some quartz veins is evident. Cleavage planes strike approximately 140° and typically dip 65° towards the southwest (Fig. 6a). Fractures are generally more vertically orientated than cleavage planes, striking predominantly northeast and southwest (Fig. 6a). Much of the outcrop visible in road cuts is weathered where water has penetrated along cleavage planes, altering minerals and mobilizing the clay fraction.

Observations with respect to channels traversing the Neranleigh-Fernvale Beds (Fig. 5b) include The 1st to 3rd order channels trend mainly northeast, east and west with a few to the north; The 4th to 7th order channels trend mainly east to south-southeast and also to the northwest with a few to the west; The 8th to 15th order channels trend mainly east to southeast and northwest to north-northwest with some channels to the northeast; Channels of or above 16th order trend mainly north-northwest through north to east.

The 1st to 3rd order channels display a strong trend, but do not clearly display a correlation with the planar features measured (Fig. 5a, b). However, there is a possible relationship with bedding orientation. Channels from 4th to 15th order show a good correlation with bedding and a less-strong correlation with cleavage. Channels of or above 16th order display a strong trend that may correlate with fractures and possibly bedding although this relationship is weak.

151 Examination of channel orientations on the Bunya Phyllite reveals the following (Fig. 6b): • The 1st order channels trend mainly northeast; • The 2nd to 4th order channels trend mainly north-northwest, north-northeast and southwest; • The 5th to 9th order channels trend mainly to the southeast with some towards the north-northeast, north and north-northwest; • The 10th to 13th order channels trend towards the northeast, east and southeast; • Channels of or above 14th order trend strongly northeast with some towards the east.

The 1st order channels on the Bunya Phyllite display trends that do not strongly correlate with measured planar features, although some correlation with cleavage may be present (Fig. 6a, b). Channels from 2nd to 9th orders have trends that correlate with cleavage and bedding. Channels from 10th to 13th orders show some correlation with cleavage. Channels from the 14th order do not strongly correlate with orientations of planar features although a weak correlation with fractures may be present and a strong trend is evident. Cleavage and bedding plane orientations correlate better with lower-order channels on the Bunya Phyllite than on the Neranleigh-Fernvale Beds. These planar features correlate with the orientation of higher-order channels on the Neranleigh- Fernvale Beds. On both units geological fabric exerts the least control on the lowest order channels and the most control on low- to middle and middle (around 3rd to 15th) order channels. High (greater than 16th) order channels showed an average orientation of northeast through north to north-northwest on the Neranleigh-Fernvale Beds and approximately southeast on the Bunya Phyllite. The highest order channels may be influenced by geological fabric to a limited extent but their strong disparate trend suggests alternative controls. Rose diagrams including all channels in the network on each lithology revealed an average orientation mainly east-west with some north-south trends on the Neranleigh-Fernvale Beds and approximately northeast through to south on the Bunya Phyllite (Fig. 7). The rose diagrams show the average orientation of the remaining network to alter direction after progressively more ranks of low and then

152 middle order channels were deducted from the dataset, although this was most apparent on the Neranleigh-Fernvale Beds. Results from the trials of selective statistical analysis provided correlation coefficients varying from strong through to weak depending upon the range and position of the dataset selected. For example, a southeasterly channel orientation trend is apparent in the 4th to 7th channel set on the Neranleigh-Fernvale Beds (Fig. 5b). Data ranging from 90-180° were analysed incorporating both cleavage and channel orientations and resulted in a moderate statistical correlation. The dataset was then narrowed to include only those data within the 140-170° range, which statistically resulted in a very strong correlation. It is clear that the results are biased by the data selection process and, therefore, provide no greater insights than visual analysis.

6 Discussion The strong trends displayed by 1st to 3rd order channels on the Neranleigh- Fernvale Beds and first order channels on the Bunya Phyllite suggests there may be additional endogenic controls on channel courses in the catchment. The poor correlation between low order channels and planar features suggests these channels may not yet have sufficiently down-cut into bedrock for fractures and foliation to be influential. Planar features may be the main control on 4th to 15th order channel orientation. The strong trends of channel orientation from the 16th order on the Neranleigh-Fernvale Beds and from the 14th order on the Bunya Phyllite, do not correlate with measured planar features, but may indicate control by larger scale structural features or by processes or structures no longer evident in the environment. The variation in channel orientation evident where the network involves fewer data and less detail (Fig. 7), strongly implies that network analysis is scale- dependant and that better accuracy is achieved where more channels are measured. Analysis of each channel order under the new ordination scheme reveals correlations with different geological or other environmental influences at different scales. Although existing ordination systems have their own merits, the new method devised for this project was found to be an effective alternative and may be employed by both traditional manual methods and with GIS. The sequential removal or addition of orders ensures that the scale of the drainage system is reduced or enhanced equally across the network where scaling is required. Every sequential order is accounted for

153 allowing direct comparisons to be made in the event that more than one catchment is being studied. Escalating order numbers are avoided unlike the systems where orders are derived by the addition of previous order numbers. Whilst still assigning every reach to a new order, this method may be of benefit when working with extremely complex or large catchments.

Fig. 5a Rose diagrams of channels grouped in similar orientations in Laceys Creek catchment on the Neranleigh-Fernvale Beds

1st to 3rd order 4th to 7th order 8th to 15th order 16th order

channels channels channels and over

Fig. 5b Rose diagrams of planar feature orientations in Laceys Creek catchment in the Neranleigh-Fernvale Beds

n=100 n=38 n=18

Strike of bedding planes Strike of cleavage planes Strike of fault and on Neranleigh-Fernvale on Neranleigh-Fernvale fracture planes on Beds Beds Neranleigh-Fernvale

Beds

154

Fig. 6a Rose diagrams of channels grouped in similar orientations in Laceys Creek catchment on the Bunya Phyllite

nd th th th th th th 1st order channels 2 to 4 5 to 9 order 10 to 13 14 order order channels order channels and channels channels over

Fig. 6b Rose diagrams of planar feature orientations in Laceys Creek catchment in the Bunya Phyllite

n = 9 n = 70 n = 26

Strike of bedding planes Strike of cleavage planes Strike of fault and on Bunya Phyllite on Bunya Phyllite fracture planes on Bunya Phyllite

155

Fig. 7. Comparisons among the Horton, Strahler and Shreve schemes and proposed new method.

The new channel ordination system allows channels of equal rank to be removed and analysed sequentially for investigating the relative influence of structures and fabric on channel orientation at different scales. On both lithologies in the Laceys Creek catchment, combinations of planar-feature orientations correspond to the orientation of low to middle-order channels but relationships are less well resolved in the lowest and highest-order parts of the network. As noted, lowest order channels show a strong trend, especially on the Neranleigh-Fernvale Beds, but do not strongly correlate with planar bedrock features, suggesting another influence such as neotectonic tilt. Shallow, low-order channels forming at the surface and not yet incised to bedrock may be more susceptible to recent geological changes than are older, higher order or incised channels. The Australian continent is currently under a compressional stress regime that may cause warping and reverse faulting of the crust. Following the construction of the Australian stress map (Hillis and Reynolds, 2000), modelling has revealed a roughly north-northeast south-southwest trending stress field in southeast

156 Queensland (Reynolds et al., 2002, 2003). Local work on recent, shallow earthquake activity indicates that the principal compressional stress field in this region is currently acting in a northeast-southwest direction (Cuthbertson, 1990), and this may have influenced the orientation of the smallest and youngest channels. The highest order channels also do not correlate with the measured planar features and may be controlled by other geological structures such as deep-seated faults and fractures or by conditions that are no longer prevalent. Previous analyses of an area of uniform rock-type such as granite, with high exposure, have revealed strong stream-fracture orientation relationships (Scheidegger, 1979a; Ericson et al., 2005). A similar relationship (albeit on very fine-scale channels) has also been shown to exist in areas of variable rock type where regolith cover is substantial (Beavis, 2000). This study shows that, at multiple scales within a whole catchment, there is correspondence between the orientation of planar bedrock features and channels despite rock types of two metamorphic grades each with broadly varying degrees of soil and vegetation cover. The correlations vary between the two rock units investigated. This study also reveals evidence of the evolution of the Laceys Creek drainage network and the extent to which geological fabric is controlling the drainage pattern. Large-scale geological structures and palaeo-controls are likely to be the dominant influences on highest order streams. The middle orders are mainly controlled by the structural grain and lithological fabric and the lowest orders, not yet incised to bedrock, may be influenced initially by neotectonism and exogenic controls. In summary, an assessment of the influence of rock architecture on drainage patterns is strongly affected by the scale of analysis. Given the limitations of current statistical methods when dealing with circular orientation data, visual analysis is better suited to this study. However, the types of information derived from catchment segment and geological fabric orientations would offer useful datasets for rigorous analysis should an appropriate statistical procedure become available in future. Identification of endogenic controls on channel orientation and scale by mapping and analysis using the methods outlined above, may be of use for local scale land-use planning and prediction of geohazards such as mass wasting, substrate stability and stream avulsion.

157

Acknowledgements We are grateful for the constructive input of both Dr Mike Daniels and an anonymous reviewer. We are also grateful to Dr. Micaela Preda for assisting with GIS and for helpful discussions, Dr. Andrew Hammond for useful input and advice and Jonathan Hodgkinson for his field assistance and valued support.

References

Abrahams, A.D., Flint, J.-J., 1983. Geological controls on the topological properties of some trellis channel networks. Geological Society of America Bulletin 94, 80-91. Ackermann, R.V., Seidl, M.A., Schlische, R.W., 1997. First-order regional structural controls on modern drainage networks within the Newark Rift Basin, New Jersey. Geological Society of America Abstracts with Programs 29, 25. Alexander, D., Formichi, R., 1993. Tectonic causes of landslides. Earth Surface Processes and Landforms 18, 311-338. Beavis, S.G., 2000. Structural controls on the orientation of erosion gullies in mid- western New South Wales, Australia. Geomorphology 33, 59-72. Belford, D.J., 1950. The Neranleigh Series. Unpublished Honours Thesis, University of Queensland, Brisbane. Beneduce, P., Festa, V., Francioso, R., Schiattarella, M., Tropeano, M., 2004. Conflicting drainage patterns in the Matera Horst Area, southern Italy. Physics and Chemistry of the Earth, Parts A/B/C 29, 717-724. Betts, P.G., Giles, D., Lister, G.S., Frick, L.R., 2002. Evolution of the Australian lithosphere. Australian Journal of Earth Sciences 49, 661-695. Bryan, W.H., Jones, O.A., 1962. Contributions to the geology of Brisbane No. 3. The bedded cherts of the Neranleigh-Fernvale Group of south-east Queensland. Proceedings of the Royal Society of Queensland 73, 17-36. Coney, P.J., 1992. The Lachlan Belt of eastern Australia and circum-Pacific tectonic evolution. Tectonophysics 214, 1-25. Cranfield, L.C., Schwarzbock, H., Day, R.W., 1976. Geology of the Ipswich and Brisbane 1:250 000 Sheet Areas; Report No 95, Geological Survey of Queensland, Queensland Department of Mines. Cuthbertson, R.J., 1990. The seismo-tectonics of southeast Queensland. Bulletin - Australia, Bureau of Mineral Resources, Geology and Geophysics 232, 67- 81. Denmead, A.K., 1928. A survey of the Brisbane schists. Proceedings of the Royal Society of Queensland 39, 71-106. Ericson, K., Migon, P., Olvmo, M., 2005. Fractures and drainage in the granite mountainous area: A study from Sierra Nevada, USA. Geomorphology 64, 97-116. Eyles, N., Scheidegger, A.E., 1999. Neotectonic jointing control on Lake Ontario shoreline orientation at Scarborough, Ontario. Geoscience Canada 26, 27-30.

158 Eyles, N., Arnaud, E., Scheidegger, A.E., Eyles, C.H., 1997. Bedrock jointing and geomorphology in southwestern Ontario, Canada: an example of tectonic predesign. Geomorphology 19, 17-34. Fleming, P.J.G., Day, R.W., Murray, C.G., Whitaker, W.G., 1974. Late Palaeozoic invertebrate macrofossils in the Neranleigh-Fernvale Beds. Queensland Government Mining Journal 75, 104-107. Gardiner, V., 1975. Drainage Basin Morphometry. Department of Geography, Wolverhampton Polytechnic, 48 p. Green, D.C., 1973. Radiometric evidence of the Kanimblan orogeny in southeast Queensland and the age of the Neranleigh-Fernvale Group. Journal of the Geological Society of Australia 20, 153-160. Heine, R.A., Lant, C.L., Sengupta, R.R., 2004. Development and comparison of approaches for automated mapping of stream channel networks. Annals of the Association of American Geographers 94, 477-490. Hillis, R.R., Reynolds, S.D., 2000. The Australian Stress Map. Journal of the Geological Society 157, 915-921. Hobbs, W.H., 1904. Lineaments of the Atlantic border region. Geological Society of America Bulletin 15, 483-506. Hobbs, W.H., 1911. Repeating patterns in the relief and structure of the land. Geological Society of America Bulletin 22, 123-176. Holcombe, R.J., Little, T.A., Sliwa, R., Fielding, C.R., 1993. Tectonic elements and evolution of SE Queensland during the Late Palaeozoic to Early Mesozoic. In: Flood, P.G., Aitchison, J.C. (Eds.), New England Orogen, Eastern Australia. Department of Geology and Geophysics, University of New England, Armidale, pp 517-527. Holcombe, R.J., Stephens, C.J., Fielding, C.R., Gust, D., Little, T.A., Sliwa, R., Kassan, J., McPhie, J., Ewart, A., 1997a. Tectonic evolution of the northern New England Fold Belt: Carboniferous to Early Permian transition from active accretion to extension. In: Ashley, P.M., Flood, P.G. (Eds.), Tectonics and Metallogenesis of the New England Orogen. Geological Society of Australia. Special Publication 19, pp 66-79. Holcombe, R.J., Stephens, C.J., Fielding, C.R., Gust, D., Little, T.A., Sliwa, R., Kassan, J., McPhie, J., Ewart, A., 1997b. Tectonic evolution of the northern New England Fold Belt: the Permian-Triassic Hunter Bowen event. In: Ashley, P.M., Flood, P.G. (Eds.), Tectonics and Metallogenesis of the New England Orogen. Geological Society of Australia. Special Publication 19, pp 52-65. Horton, R.E., 1945. Erosional development of streams and their drainage basins; hydrophysical approach to quantitative morphology. Bulletin of the Geological Society of America 56, 275-370. Isbell, R.F., 1996. The Australian Soil Classification. CSIRO Publishing, Melbourne. Jones, T.A., 1968. Statistical analysis of orientation data. Journal of Sedimentary Petrology 38, 61-67. Kirchner, J.W., 1993. Statistical inevitability of Horton Laws and the apparent randomness of stream channel networks. Geology 21, 591-594. Kohlbeck, F.K., Scheidegger, A.E., 1985. The power of parametric orientation statistics in the Earth Sciences. Mitteilungen der Österreichischen Geologischen Gesellschaft 78, 251-265. Krieger Lassen, N.C., Juul Jenson, D., Conradsen, K., 1994. On the statistical analysis of orientation data. Acta Crystallographica A50, 741-748.

159 Krumbein, W.C., 1939. Preferred orientation of pebbles in sedimentary deposits. Journal of Geology 47, 673-706. Little, T.A., Holcombe, R.J., Gibson, G.M., Offler, R., Gans, P.B., McWilliams, M.O., 1992. Exhumation of late Paleozoic blueschists in Queensland, Australia, by extensional faulting. Geology 20, 231-234. Mock, S.J., 1971. A classification of channel links in stream networks. Cold Regions Research 7, 1558-1566. Mock, S.J., 1976. Topological properties of some trellis pattern channel networks. Cold Regions Research and Engineering Laboratory Report, 46-76. Müller-Salzburg, L., 1963. Der Felsbau. F. Enke, Stuttgart, 624 p. Murray, C.G., Fergusson, C.L., Flood, P.G., Whitaker, W.G., Korsch, R.J., 1987. Plate Tectonic model for the Carboniferous evolution of the New-England Fold Belt. Australian Journal of Earth Sciences 34, 213-236. NASA, 2004. Shuttle Radar Topography Mission. Jet Propulsion Laboratory, California Institute of Technology, http://www2.jpl.nasa.gov/srtm/. Orme, A.R., 2002. Shifting paradigms in geomorphology: the fate of research ideas in an educational context. Geomorphology 47, 325-342. Queensland Government, D.N.R.M., 2003. Queensland Geological Digital Data. Reyment, R.A., 1971. Introduction to Quantitative Palaeoecology. Elsevier, Amsterdam, 226 p. Reynolds, S.D., Coblentz, D.D., Hillis, R.R., 2002. Tectonic forces controlling the regional intraplate stress field in continental Australia: Results from new finite element modelling. Journal of Geophysical Research 107, 10.1029/2001JB000408. Reynolds, S.D., Coblentz, D.D., Hillis, R.R., 2003. Influences of plate-boundary forces on the regional intraplate stress field of continental Australia. Geological Society of Australia Special Publication 22, 53-64. Scheidegger, A.E., 1965. The algebra of stream order numbers. United States Geological Survey Professional Paper 525-B, 187-189. Scheidegger, A.E., 1979a. Orientationsstruktur der Talanlagen in der Schweiz. Geographica Helvetica 34, 9-17. Scheidegger, A.E., 1979b. The principle of antagonism in the earth's evolution. Tectonophysics 55, T7-T10. Scheidegger, A.E., 1998. Tectonic predesign of mass movements, with examples from the Chinese Himalaya. Geomorphology 26, 37-46. Scheidegger, A.E., 2001. Surface joint systems, tectonic stresses and geomorphology: a reconciliation of conflicting observations. Geomorphology 38, 213-219. Scheidegger, A.E., 2002. Morphometric analysis and its relation to tectonics in Macaronesia. Geomorphology 46, 95-115. Shreve, R.L., 1966. Statistical law of stream numbers. Journal of Geology 74, 17-37. Shreve, R.L., 1967. Infinite topologically random channel networks. Journal of Geology 75, 178-186. Strahler, A.N., 1954. Quantitative geomorphology of erosional landscapes, C.-R. 19th International Geol. Cong., Algiers, 1952, pp. 341-354. Strahler, A.N., 1957. Quantitative analysis of watershed geomorphology. Transactions, American Geophysical Union 38, 913-920. Strahler, A.N., 1966. The Earth Sciences. Harper International, 681 p.

160 Tucker, R.M., 1967. Geology of the City of Brisbane. Part 1. The pre-Mesozoic sediments and volcanics and the related intrusives. Publications of the Geological Survey of Queensland 324. Twidale, C.R., 1972. The neglected third dimension. Zeitschrift für Geomorphologie 16, 283-300. Twidale, C.R., 1980. Geomorphology. Thomas Nelson, 406 p. Twidale, C.R., 2004. River patterns and their meaning. Earth-Science Reviews 67, 159-218. Walsh, F., 1972. Application of stream order numbers to the Merrimack River basin. Water Resources Research 8, 141-144. Wilson, M.M., 1973. Geological notes on the Rocksberg Greenstones and overlying rock sequences, Field Conference Guidebook, Brisbane-Gympie Area. Geological Society of Australia, Queensland Division, pp 19-29. Winkler, H.G.F., 1967. Protogenesis of metamorphic rocks. Springer-Verlag, New York, 237 p. Woldenberg, M.J., 1967. Geography and properties of surfaces. Harvard Papers in Theoretical Geography 1, 95-189. Zernitz, E.R., 1932. Drainage patterns and their significance. Journal of Geology 40, 498 - 521.

161

162 INTRODUCTION AND BACKGROUND TO PAPER 2

As the results from the first paper showed a correlation between faulting and drainage of the higher order channels in the Laceys Creek catchment, this raised the question whether there were further correlations between drainage channels and mapped faults at a broader scale. Regional drainage and topography were compared with the mapped faults and joints, and some correlation was evident, particularly in the upper Brisbane River valley where major ancient fault systems exist and where two major artificial reservoirs are presently located (Lakes Somerset and Wivenhoe). As the reservoirs are built on known faulted valleys, a seismic monitoring network is located at both Wivenhoe and Somerset dams to detect earthquakes that may document premature releases of stress in the crust caused by the crustal loading and pore-pressure changes induced by the water bodies (reservoir-induced earthquakes). The relationship between the physical features and structure in the region suggests the landscape is ancient in origin. For example, the upper Brisbane Valley is located in a narrow downfaulted zone that may constitute a late Palaeozoic to early Mesozoic remnant foreland basin or intra-cratonic graben (Campbell et al., 1999). However, in order to identify whether those original primary processes are continuing to impose controls on the landscape, the correlation between some known physical features and recent earthquake epicentres was analysed. The datasets, as previously mentioned, provided locations of epicentres that may have errors of up to 30 km in radius and this was taken into consideration during analysis, by allowing a ‘buffer’ around the location of each earthquake. Thus, where more than 3 ‘buffered’ earthquake epicentres revealed some clustering or linear patterns, these are described in the following paper as ‘corridors’. The corridors do not necessarily imply the positions of major fault planes, as the majority of earthquakes represented are of low magnitude, but the linear alignments suggest that such patterns may be related to narrow zones where stress is released periodically. The earthquakes in the databases covering the past 100 years are mainly less than magnitude 2 but these should not be considered representative of the earthquakes in the region throughout geological time. The low magnitude earthquakes experienced most frequently in this region over the past century are unlikely to cause large, if any, surface displacement. However, they may be responsible for continued weakening of the rocks, indicating the preferred locations

163 of less-common larger magnitude earthquakes, and they may define zones of weakness that could eventually lead to preferred fluid drainage, weathering and possibly erosion in these locations. The resulting corridors suggest that groups of earthquakes may be causing small weaknesses in preferred locations across the region. Paper 2 considers whether the earthquake corridors indeed show correspondence with drainage features, and whether ongoing small-scale seismic activity may be responsible for enhancing the topographic expression above existing structural features. The database limitations are well known to the earthquake engineering community, hence a thorough discussion of seismic data and its limitations was not outlined at the conference at which the following paper was presented. Limitations and further references relating to earthquake data are provided in the literature review.

CAMPBELL, L. M., HOLCOMBE, R. J. and FIELDING, C. R. 1999. The Esk Basin - a Triassic foreland basin within the northern New England Orogen. In: Flood, P. G., ed. Regional Geology, Tectonics and metallogenesis, New England Orogen, NEO '99. pp. 275-284. Armidale Dept of Earth Sciences, University of New England.

164

PAPER 2

TITLE The correlation between physiography and neotectonism in southeast Queensland

AUTHORS Jane Helen Hodgkinson. Stephen McLoughlin, Malcolm Cox

School of Natural Resource Sciences Queensland University of Technology

Reviewed for DEST purposes and presented at the Australian Earthquake Engineering Society Conference, Canberra, ACT November 2006

165 STATEMENT OF ORIGINAL AUTHORSHIP

Jane Helen Hodgkinson (PhD Candidate): reviewed previous work and literature; planned methodology, collated data and conducted GIS work; analysed and interpreted results, wrote paper and poster, presented both at conference

Stephen McLoughlin (Principal PhD Supervisor): reviewed and discussed methods and results; made suggestions regarding location of earthquake ‘corridors’; reviewed, discussed and edited paper

Malcolm Cox (Associate PhD Supervisor): suggested need for a review of faults in the region; reviewed, discussed and edited paper

166 Abstract We tested for correlation between recent earthquake epicentre data and the distribution of major physiographic features, such as escarpments and river channels, in southeast Queensland. Preliminary results indicate that many of the known earthquake epicentres over the past century are distributed in several broad belts, corresponding in location and orientation to major structural discontinuities or narrow sedimentary basins bounded by faults. Other earthquake clusters show broad correlation with linear segments of major river systems where no major faults have been mapped. Several domains dominated by Palaeozoic-Triassic rocks, such as the North and South D’Aguilar blocks, are represented by high terrain flanked by faults that may have been active back to at least the mid-Mesozoic. Reactivation and subsidence along some of these faults may account for the local accumulation of thick sedimentary piles during the Paleogene. Modern earthquake epicentre distributions along the margins of these blocks suggest that recent and on-going tectonism may be enhancing the escarpments flanking the uplands. Unmapped, concealed or deep-seated geological discontinuities may exist where earthquake epicentres correspond to linear physiographic features but not to currently mapped faults or joints. Identification of such concealed geological structures will be important for developing accurate earthquake hazard maps into the future.

Introduction Tectonism is one of the primary driving forces behind the structural and physiographic modification of land masses. For example, tectonism is largely responsible for terrain uplift and basin subsidence, which allow modification by secondary processes such as weathering and erosion. As tectonics is generally accepted to cause geomorphological change, it is logical to study geomorphological features to identify the influences of tectonics on the landscape (for example Burrato et al., 2003; Vannoli et al., 2004). The Australian continent is situated within the Indo-Australia Plate. Compared to plate boundary earthquakes, intraplate earthquakes are few and shallow. However, the Global Seismic Hazard Map shows that Australian earthquake activity is moderate to high, relative to other intraplate regions (GSHAP, 1992- 1997). The Indo-Australian Plate is presently under compressional stress (Hillis, 1998; Hillis and Reynolds, 2000) and modification of the land will occur in order to

167 accommodate shortening of the continental mass where the stress exceeds the strength of the crust. For example, Neogene-Quaternary reverse faulting and compressional folding has clearly influenced landscape evolution in both central western (for example, Clark, 2005), and southeastern Australia (for example, Sandiford, 2003). Folds, faults, joints, shears and rock fabric alteration, resulting from tectonic movement, have long been known to control the formation of a variety of distinctive land surface features such as scarps and river channels (for example: Hobbs, 1904; Hobbs, 1911; Zernitz, 1932; Strahler, 1960; Strahler, 1966; Twidale, 1980; Scheidegger, 1998; Scheidegger, 2002; Ericson et al., 2005; Hodgkinson et al., in press). The relationship between Australia’s geology and its earthquakes is poorly understood and many earthquakes cannot be assigned to known structures (Clark and McCue, 2003). Physiographic analysis in conjunction with recent seismic data will assist identifying those landscape features that are likely to be tectonically controlled and presently active; such analysis has the potential to refine zones hazardous to the population and infrastructure. For dipping faults, earthquake epicentres will appear more distant from the surface trace with increasing hypocentral depth. As a consequence, epicentres may not be expected to align precisely with surface features, such as mapped faults, scarps and joint systems. Therefore, broad sectors in which earthquake epicentres are located should be identified to determine potentially active fault zones. Earthquake locations may also be inaccurate, particularly those identified from early records and, therefore, care must be taken when relating them to local physiographic features. Since the calculation of actual earthquake depth is typically inaccurate, this parameter should be treated with caution. Using available data, this study aims to provide evidence of tectonic control upon the landscape of southeast Queensland.

Background Southeast Queensland’s geology is complex and derived from several cycles of compressional and extensional tectonic activity since about 370 Ma. Palaeozoic to modern sedimentary and igneous rocks are interspersed with large belts of metamorphic rocks throughout the region. Many of the geological units (Queensland Government, 2003) are bounded by faults. Extensive regolith, vegetation and infrastructure conceals many of southeast Queensland’s geological discontinuities including faults, joints and formation boundaries. Some areas appear to be relatively

168 fault-free and, although this may be the case, the apparent dearth of these features may be attributable to concealment by ground cover or the lack of detailed mapping. Recent core logging in southeast Queensland has revealed discrepancies with the published geological maps (geological map, Queensland Government, 2003; Brisbane City Council, 2006 pers. comm.) and confirmed that faults and joints are common in the region. Fault distributions have been analysed recently (Humphries, 2003) but little has been published regarding age constraints on fault activity. Childs (1991) analysed Landsat images of the northern part of the region and showed that the main ranges and drainage systems are strongly concordant with the bedrock geology, and that faults also correspond with channel orientation. However, Humphries noted that some major faults and a shear zone on geological maps were not identifiable on the Landsat image, and may either lack surface expression or have had unfavourable illumination for Landsat (Childs, 1991). The region’s elevation ranges from sea-level to 1360 m a.s.l. and the area can be divided into three general terrains: highlands (>300 m a.s.l.), hills (30-300 m a.s.l.) and lowlands (<30 m a.s.l.) (Fig. 1). The greatest portion of southeast Queensland is situated in the hilly to highland terrains. However, most of the population presently resides within the coastal lowlands, especially within the expanding cities of Brisbane, Ipswich and the Gold Coast. Three artificial reservoirs, Somerset, Wivenhoe and Samsonvale, provide southeast Queensland’s main water supply: each is situated in known faulted and seismically active zones.

Earthquake monitoring In order to compare neotectonism and its geomorphological effects, detailed earthquake data are needed. Earthquake monitoring in Queensland is generally sparse by international standards and has only operated intermittently. Earthquakes have been recorded since the late 1800’s. In 1937, Queensland’s first international monitoring station was opened in Brisbane, followed by the Charters Towers station in 1957. Subsequently, a seismic monitoring network developed slowly, broadening considerably after 1977, when more detailed instrumental monitoring was implemented around the large dams. Dam-site and other seismographs were integrated into a state-wide network, monitored by The University of Queensland (UQ) from 1993 (the QUAKES Centre) but since 1998, much of the operational instrumentation has been progressively discontinued from service. Monitoring is now

169 restricted to southeast Queensland. Since 2000, 22 Queensland Government seismograph stations continue to collect data under commercial contract to Environmental Systems and Services (ES&S, Victoria,). As well as temporal discontinuities, data completeness is also affected by differences in the resolution of monitoring, spatially.

Detail shown in Fig. 6

Fig. 1 25m DEM of southeast Queensland

170

Fig 2 Main drainage systems and locations of southeast Queensland

171 Fig. 3 Earthquake epicentre corridors superimposed onto slope map. Scarps highlighted in black

172 Methods Digital topographic data at 25 m intervals (Queensland Government, 2005) (Fig. 1), geological and drainage (Fig. 2) maps for southeast Queensland were obtained from the Department of Natural Resources, Mines and Water (DNRM&W)(Queensland Government, 2003). Using the ArcGIS 9 software, a digital elevation map (DEM) was produced and slope maps (Fig. 3) created from which scarp features were extracted. Earthquake data retrieved from Geoscience Australia (2006) provided information for 100 earthquakes recorded in the region since 1872. Further data were supplied by the Earth Systems Science Computational Centre (ESSCC) at The University of Queensland, increasing the total number of earthquakes recorded in the region to 344 (Figs. 3,4,5). The digital maps were combined with seismic data for identification of concordant patterns of geomorphological lineaments and earthquake epicentres. Where 4 or more earthquake epicentres clustered or were well-aligned within a 12-15 kilometre wide corridor, they have been considered to possibly have a common source and be related to similar zones of seismic activity. Such clusters and alignments are referred to here as ‘earthquake corridors’.

Results Highlands and scarps The highland areas are situated mainly in the west and north of the study area, and generally trend in a northwest-southeasterly direction. A discrete area of highlands, situated in the central region, is separated from the west by the Brisbane River valley. Some highland terrain is situated in the southeast, associated with the Mount Warning shield volcano. The physiography of the latter highland area does not correspond to the predominant northwest-southeast geological trends in southeast Queensland. Scarps are common across the region (Fig. 3). In places they coincide with the orientation of highlands, geological units, faults and drainage.

173

Fig. 4 Main drainage, structural features and earthquake epicentres in southeast Queensland

174

Fig. 5 Data from ‘Fig. 4’ superimposed with earthquake corridors

175 Drainage Channel orientation in the region is predominantly northwest-southeasterly and northeast-southwesterly (Fig 2.). These trends are particularly strong in the Brisbane River system, which may be described as a trellis or rectilinear drainage pattern. Secondary trends occur in an east-west orientation and other trends are evident at a finer scale. Drainage in the southeast is radial, away from the centre of the Mount Warning volcanic complex.

Geological discontinuities High angle dip-slip faults, joints, thrusts and shear zones have been mapped throughout the region (Queensland Government, 2003), although a large area in the southwest and on the coastal plains in the east appears to have few faults. Fault orientations in the remainder of the region strongly trend in a northwest-southeasterly orientation although various other trends are also evident (Fig. 4).

Seismicity Earthquakes occur throughout the region on all terrains (Figs. 3,4,5). As depth to focus data are highly uncertain, earthquake epicentres only have been considered in this analysis. Their distribution shows some concurrence with drainage, structural features (Figs. 4,5) and scarp distributions (Fig. 3). The epicentres cluster most prominently within a broad northwest-southeast trend but subsidiary southwest- northeast and roughly east-west trends are also evident (Figs. 3,5).

Discussion Physical relationships There is some concurrence between the location and orientation of faults and rivers, especially the Brisbane River system (Fig. 4). The presence of a scarp may be due to surface displacement by faulting, mass wasting or by other surface processes such as fluvial erosion. Some scarps appear to have no correlation with present drainage or mapped faults and may represent features with historical controls such as retreating coastal escarpments. A relationship between drainage system pattern and highland location is present, although such definition would be expected due to normal down- cutting of rivers between resistant rock units. The lowland areas primarily consist of unconsolidated Cenozoic sediments, which may conceal faults or joints.

176

Seismic and physiographic relationships Earthquake epicentre and drainage patterns commonly concur throughout the region (Figs. 4,5). North of Brisbane/Toowoomba, this association is also closely aligned with faults. However, south of Brisbane/Toowoomba, virtually no mapped faults coincide with earthquake and river-trends. In the north, this alignment suggests that mapped faults may be active and controlling drainage channel orientation and position. In the south however, where some earthquake zones align with drainage but do not coincide with mapped faults, other faults may be concealed and/or not yet mapped. Several scarps coincide spatially with earthquake-prone zones (Fig. 3) and an apparent alignment between some scarps, faults and drainage channels, suggests that these scarps may be fault and/or river controlled. Some earthquake epicentres cluster in close proximity to mapped faults, such as those flanking much of the South D’Aguilar Block (Figs. 4,7). However, some clusters do not appear to align with currently mapped faults despite their linear spatial distribution. The proximity of earthquake activity to channel location suggests that there may be a relationship requiring further investigation.

Fig. 6 DEM detail (northwest of region) showing linear escarpments with mapped faults superimposed

177

Fig. 7 Stable geological blocks and basins of southeast Queensland

Seismic zones lacking physiographic relationships There are several earthquake-prone zones in the region that are commonly located in flat or gently undulating terrain and free from scarps, large river channels and faults. The earthquake epicentres in these areas may be associated with very deep faults that presently have no surface expression, may be in areas that are poorly mapped due to ground cover or may be beyond the resolution of the DEM. Equally, they may be associated with mapped faults with very shallow angles of dip causing the surface expression to be far enough away from the epicentre to appear unrelated. Data may also inaccurately reflect the position of the epicentre.

General Deep earth investigation such as drilling, reflection seismography or GPR, together with more accurate measurements of hypocentre depths may assist in identifying the

178 faults with which recent earthquakes are associated. Such work may also identify unmapped faults where surface features and earthquakes suggest there is potential faulting in the vicinity. For example, topography, river orientation and earthquake activity suggests faulting occurs along the southern edge of the South D’Aguilar Block. Many of the rock units throughout the region are bounded by faults implying tectonism is responsible for their current position. Recent earthquake activity in the vicinity of these faults suggests continued or sporadic movement of these units is occurring.

Conclusions This preliminary study suggests that geomorphological evidence, when combined with geological and earthquake data, may be used to successfully identify zones of current faulting. An important consideration in this study has been the scale of viewing both temporally and spatially. Earthquake corridors suggested in our results may not be apparent if each epicentre was viewed in isolation or at a less broad scale in time and space. Although features such as channels and slopes may be controlled by differential weathering, neotectonism may also be influential and this may pose a greater threat than geohazard maps imply. The most widely used seismic hazard map of southeast Queensland (McCue et al., 1998) is a classic representation of a probabilistic earthquake model (defined by Cornell, 1968). The hazard designations are a product of available data, which may be sparse, temporally and spatially. Equally, the earthquakes may not be probabilistic in nature (Clark and McCue, 2003). Consequently, the map may not fully represent actual hazards in the area. Therefore, more detailed, widespread, long-term monitoring programs would be a valuable addition to future hazard assessment and mitigation, together with deterministic seismic modelling. First motion studies are useful in determining the dip orientation of dip-slip faults and whether normal or reverse. Accurate focal depths would also better constrain fault locations. Broad ‘zones’ surrounding the implied epicentre positions should be used to relate earthquakes to potential, local, physiographic features, unless absolutely certain of the data accuracy. The most densely populated area in southeast Queensland is situated within the lowlands, which hosts a widespread veneer of unconsolidated Cenozoic sediments that may conceal potentially active faults. Better collection and availability of structural data,

179 in combination with high resolution digital terrain models, would enable more thorough landscape analysis to identify potentially active fault zones, areas of concealed faults and deep sediment zones which are conducive to seismic amplification. Ultimately this would provide a better understanding of both neotectonism and localised seismic hazard zones in southeast Queensland.

Acknowledgements We wish to thank Dr Dion Weatherly and Col Lynam of ESSCC and QUAKES Group at the ESSC Centre, The University of Queensland, for providing us with the ESSCC Earthquakes Database and other useful references, and for interesting collaborative discussions. We are also grateful for the valuable suggestions and comments from Dan Clark and an anonymous reviewer.

References Brisbane City Council, 2006. City Design, Ground Engineering, Brisbane, pp. Core logging showing that geology maps incorrect in many parts of Brisbane. Burrato, P., Ciucci, F. and Valensise, G., 2003. An inventory of river anomalies in the Po Plain, Northern Italy: evidence for active blind thrust faulting. Annals of Geophysics, 46(5): 865-882. Campbell, L.M., Holcombe, R.J. and Fielding, C.R., 1999. The Esk Basin - a Triassic foreland basin within the northern New England Orogen. In: P.G. Flood (Editor), Regional Geology, Tectonics and metallogenesis, New England Orogen, NEO '99. Dept of Earth Sciences, University of New England, Armidale, pp. 275-284. Childs, I., 1991. Earthquake risk, Landsat imagery and fault zones in the Bundaberg area of central eastern Queensland. Queensland Geographical Journal, Series 4, 6: 59-70. Clark, D. and McCue, K., 2003. Australian paleoseismology: towards a better basis for seismic hazard estimation. Annals of Geophysics, 46(5): 1087-1105. Clark, D., 2005. Identification of Quaternary faults in southwest and central western Western Australia using DEM-based hill shading. Geoscience Australia Record: 60p. Cornell, C.H., 1968. Engineering seismic risk analysis. Bulletin of the Seismological Society of America, 58: 1583-1606. Ericson, K., Migon, P. and Olvmo, M., 2005. Fractures and drainage in the granite mountainous area: A study from Sierra Nevada, USA. Geomorphology, 64(1-2): 97-116. ESSCC, 2006. Earth Systems Science Computational Centre, Earthquake Database, University of Queensland, Brisbane. Geoscience Australia, 2006. http://www.ga.gov.au/oracle/quake/quake_online.jsp. GSHAP and (Global Seismic Hazard Assessment Program), 1992-1997. Global Seismic Hazard Map, http://www.seismo.ethz.ch/GSHAP/global/. Hillis, R.R., 1998. The Australian stress map. Exploration Geophysics, 29(3-4): 420.

180 Hillis, R.R. and Reynolds, S.D., 2000. The Australian Stress Map. Journal of the Geological Society, 157: 915-921. Hobbs, W.H., 1904. Lineaments of the Atlantic border region. Geological Society of America Bulletin, 15: 483-506. Hobbs, W.H., 1911. Repeating patterns in the relief and structure of the land. Geological Society of America Bulletin, 22: 123-176. Hodgkinson, J.H., McLoughlin, S. and Cox, M., in press. The influence of geological fabric and scale on drainage pattern analysis in a catchment of metamorphic terrain: Laceys Creek, southeast Queensland, Australia. Geomorphology, doi:10.1016/j.geomorph.2006.04.019. Humphries, D., 2003. An analysis of post-Triassic faulting in southeast Queensland. BSc (Hons) Thesis, University of Queensland, Brisbane, 134 pp. McCue, K.F., Somerville, M. and Sinadinovski, C., 1998. The new Australian earthquake hazard map, Proceedings of the Australasian Structural Engineering Conference, Aukland, NZ, pp. 433-438. Queensland Government, D.N.R.M., 2005. South East Queensland 25 metre Digital Elevation Model - SEQ_DEM_100K. Queensland Government, D.N.R.M.W., 2003. Queensland Geological Digital Data, CD ROM. Sandiford, M., 2003. Neotectonics of southeastern Australia: linking the Quaternary faulting record with seismicity and in situ stress. Geological Society of Australia Special Publication, 22: 101-113. Scheidegger, A.E., 1998. Tectonic predesign of mass movements, with examples from the Chinese Himalaya. Geomorphology, 26(1-3): 37-46. Scheidegger, A.E., 2002. Morphometric analysis and its relation to tectonics in Macaronesia. Geomorphology, 46(1-2): 95-115. Strahler, A.N., 1960. Physical Geography. John Wiley and Sons, New York. Strahler, A.N., 1966. The Earth Sciences. Harper International, 681 pp. Twidale, C.R., 1980. Geomorphology. Thomas Nelson, 406 pp. Vannoli, P., Basili, R. and Valensise, G., 2004. New geomorphic evidence for anticlinal growth driven by blind-thrust faulting along the northern Marche coastal belt (central Italy). Journal of Seismology, 8(3): 297-312. Zernitz, E.R., 1932. Drainage patterns and their significance. Journal of Geology, 40: 498-521.

181

182 PAPER 3

TITLE Drainage patterns in southeast Queensland: the key to concealed geological structures?

AUTHORS Jane Helen Hodgkinson, Stephen McLoughlin, Malcolm Cox

School of Natural Resource Sciences Queensland University of Technology

Published in Australian Journal of Earth Sciences December 2007

183 STATEMENT OF ORIGINAL AUTHORSHIP

Jane Helen Hodgkinson (PhD Candidate): reviewed previous work and literature; planned methodology, collated data, conducted GIS work and analysis, interpreted results; wrote paper

Stephen McLoughlin (Principal PhD Supervisor): reviewed and discussed methods, analysis and results; identified some drainage patterns and additional references; reviewed and edited paper

Malcolm Cox (Associate PhD Supervisor): reviewed and discussed results and reviewed and edited paper

184 Abstract

Southeast Queensland’s geomorphology is characterised by northwest-southeast trending trunk drainage channels and highlands that strongly correlate with the distribution of geological units and major faults. Other geomorphological trends strongly coincide with subsidiary faults and geological domains. Australia is presently under compressional stress. Seismicity over the past 130 years records 56 earthquakes of >2 magnitude indicating continuing small-scale earth movements in the Moreton region. Highlands in this region are dominated by Palaeozoic to Triassic metamorphic and igneous rocks, and are generally 20-80 km from the coastline. Coastal lowlands are largely dominated by Mesozoic sedimentary basins and a veneer of surficial sediments. The eastern coast of Australia represents a passive margin; crustal sag along this margin could be expected to produce relatively short, high-energy, eastward flowing drainage systems. We performed a geomorphological analysis to characterise the drainage patterns in southeast Queensland and identify associations with geological features. Anomalous channel, valley and escarpment features were identified, which failed to match the anticipated drainage model and also lacked obvious geological control. Despite their proximity to the coast (base level), these features include areas where drainage channels flow consistently away from, or parallel to, the coastline. Although many channels do coincide with geological structures, the drainage anomalies cannot be directly related to known structural discontinuities. Anomalous drainage patterns are suggested to indicate previously unidentified structural features and in some cases relatively young tectonic control on the landscape. Recent seismicity data have also been analysed to assess spatial correlations between earthquakes and geomorphological features. Our results show that structure largely controls drainage patterns in this region and we suggest that a presently unmapped and potentially active, deep-seated structure may exist parallel to the coast in the northern coastal region. We propose that this structure has been associated with uplift in the coastal region of southeast Queensland since mid-Cenozoic times.

KEYWORDS: Drainage patterns; Anomalous drainage; Geological structure; Geomorphology; Tectonic control

185 INTRODUCTION Fluvial drainage patterns have been studied broadly as a tool for a wide range of theoretical and applied geological investigations (for example Strahler, 1966; Schumm and Khan, 1972; Twidale, 2004). As stated by Hills (1963) and reiterated by Twidale (2004), the greatest assistance that geomorphology can offer structural geology may be derived from the interpretation of drainage patterns. Twidale (2004) concluded that most river patterns are determined by geological structure and slope, and anomalies in, and diversions from, these patterns are generally caused by active faults and folds. Control of drainage patterns by geological structure and tectonics is widely accepted and has been investigated extensively (for example Burnett and Schumm, 1983; Ouchi, 1985; Mather, 1993; Jackson et al., 1998; Zelilidis, 2000; Maynard, 2006). Tectonic control of geomorphological features is broadly accepted (e.g. Ollier, 1981). Recorded earthquakes are widespread in Australia but relatively few can be assigned to known structures (Clark and McCue, 2003). Deep regolith cover across much of the continent conceals many bedrock structural details. Low resolution mapping and sparse seismic profiles have also constrained fault identification in many areas. Additional methodologies are required to identify active and ancient faults for geohazard mapping and geotechnical surveys. Geomorphological analyses offer a means to identify many concealed structural discontinuities. Entrapment of drainage channels along faults is generally a function of preferential weathering and erosion along zones of rock weakness and/or rotation of fault blocks influencing the topography. Geomorphological evidence such as linear trends of major drainage channels and steep slopes may be used to identify zones of faulting. Analysis of drainage patterns and recognition of anomalies in these systems has become a popular method to identify obscure geological structures and neotectonism (for example Jackson et al., 1998; Goldsworthy and Jackson, 2000; Burrato et al., 2003; Vannoli et al., 2004; Delcaillau et al., 2006; Hodgkinson et al., 2006a). This is especially effective when combined with spatial earthquake data. This study aims to characterise the drainage systems of southeast Queensland and identify broad-scale anomalies in stream patterns and their potential geological controls. A multidisciplinary approach is employed in which remote sensing techniques are integrated with structural geology, earthquake records and

186 geomorphological observations within a geographic information system (GIS). The objective of this study is to improve the understanding of geological controls on the landscape and the effects of neotectonism in the region.

SETTING The study area is on the eastern edge of the Australian continent between 151° 53'E, 26° 11'S to 153° 31'E, 28° 30'S, and covers approximately 41,000 km2. It generally corresponds to the region covered by the Moreton 1:500 000 Geology Map (Whitaker and Green, 1980). The region is currently undergoing extensive urbanisation and development, and represents one of the fastest population growth centres in Australia (Australian Bureau of Statistics, 2006). The region has a subtropical climate, commonly with hot, wet summers (November to February) and warm, dry winters. Rainfall is often heavy and may contribute to extensive areas of mass wasting on escarpments flanking the coastal plain. Water resources are limited by strongly seasonal and highly variable rainfall, and artificial reservoirs provide the majority of the region’s water supply. The largest reservoirs are situated in valleys that are developed on faults or fault zones. The geomorphology of southeast Queensland has been studied extensively (e.g.: Marks, 1933; Watkins, 1967; Arnett, 1969; Arnett, 1971; Donchak, 1976; Beckmann and Stevens, 1978; Murray and Whitaker, 1982 ; Lucas, 1987; Murray, 1987; Murray et al., 1987; Cuthbertson, 1990; Childs, 1991; Little et al., 1992; Holcombe et al., 1993; Little et al., 1993; Holcombe and Little, 1994; Hodgkinson et al., 2006a). In brief, highlands (mostly plateaux over 300 m a.s.l.) fringe the western, southern and northern margins of the study area; an additional isolated, dissected highland area occurs in the centre (Fig. 1). The remainder of the area, principally in the east, can be described as foot hills and coastal plains. Escarpments are common across the region. The main drainage systems of the region commonly display strong northwest-southeasterly and northeast-southwesterly trends. Identified fault orientations also strongly trend in a northwest-southeasterly orientation although various other trends are also evident (Fig. 2). The Australian continent is currently under compressional stress (Hillis, 1998; Hillis and Reynolds, 2000) and may experience warping and reverse faulting of the crust as the landmass shortens.

187

Fig. 1 Digital elevation model showing locations and main rivers in the study area. The Main Range in this area corresponds to the Great Divide and the Great Escarpment.

188 Fig. 2 Faults and joint systems and main drainage in study area.

189 Stress orientations and seismicity in Australia have been modelled and reveal a north-northeast to south-southwest trending stress-field in the region of eastern to southeast Queensland (Cuthbertson, 1990; Hillis, 1998; Hillis et al., 1999; Hillis and Reynolds, 2000; Zhao and Müller, 2001; Reynolds et al., 2002; Hillis and Reynolds, 2003; Reynolds et al., 2003). A strong northwest trend is clearly evident within structure and distribution of the rock units located in southeast Queensland and is related to the late Palaeozoic – Early Mesozoic convergent margin setting. Seismicity monitoring over the past 130 years has recorded 56 earthquakes of >2 magnitude in the region, of which 17 were >3 magnitude and 2 were >5 magnitude (ESSCC 2006). Many epicentres associated with these earthquakes align in discrete zones (Hodgkinson et al., 2006a), some of which correspond to known structural discontinuities.

Geological History The Australian continent is situated wholly within the Indo-Australia Plate. The geology of southeast Queensland (Fig. 3) developed primarily from a complex series of compressional and extensional events from the late Palaeozoic onwards. During the Late Carboniferous, when the Australian continent was part of Gondwana, an Andean-type volcanic chain (the Connors-Auburn Volcanic Arc), a central forearc basin (the Yarrol Basin) and an accretionary prism in the east (Wandilla Slope and Basin) developed in association with a west-dipping subduction zone (Day et al., 1978; Plumb, 1979; Murray and Whitaker, 1982; Day et al., 1983; Fergusson and Leicht, 1993). These structures formed the northern sector of the New England Fold Belt. Shortening and deformation of the accretionary prism and local obduction of oceanic crust occurred (Day et al., 1978; Plumb, 1979), which led to low grade metamorphism and uplift in the southeast Queensland sector of the New England Fold Belt (for example Fleming et al., 1974; Cranfield et al., 1976; Holcombe, 1978; Murphy et al., 1979; Murray et al., 1979). During the Early Permian, andesitic volcanism resumed (Day et al., 1978; Day et al., 1983). The convergent tectonic regime persisted through the remainder of the Permian and most of the Triassic. Associated backarc and forearc subsidence caused widespread shallow seas to form and new sediments were deposited on the earlier Carboniferous metamorphosed terranes. Thick sediments accumulated in the Esk Trough and Brisbane Valley in the Early Permian (Northbrook Beds) and

190 Middle Triassic (Esk Group). The old accretionary wedge was intruded by small intermediate to felsic plutons whilst offshore a new subduction zone became active (Willmott, 2004). During the Middle to Late Triassic an extensional event lead to further volcanics, granite intrusions and initiation of the Ipswich and Tarong basins (Evernden and Richards, 1962; Webb and McDougall, 1967; Cranfield et al., 1976). Uplift and the creation of mountainous terrain followed (Cranfield et al., 1976; Plumb, 1979; Willmott, 2004). From the Late Triassic to Early Cretaceous, extensional epicratonic basins, such as the Clarence-Moreton, Nambour and Maryborough basins formed and accumulated braided river, paludal and deltaic sediments (Day et al., 1983). Approximately 120 million years ago, the eastern fringe of Gondwana began to break up (Veevers and Evans, 1975; Powell et al., 1976; Branson, 1978; Veevers, 2001; Willmott, 2004). More importantly, between 70 and 45 million years ago, crustal doming led to fracturing of the crust and seafloor spreading along the eastern margin of Australia, and the opening of the Tasman Sea. Significant uplift probably occurred along the flanks of this rift system resulting in the inversion of the local Mesozoic basins.

3a) 3b Fig. 3 a) Geology of southeast Queensland (after) Queensland Government) (2003); b) stable blocks and basins of the Moreton district.

191 Passive continental margins such as the east coast of Australia, are commonly characterised by broad, high elevation, low-relief plateaux flanking a dissected coastal belt. Uplift caused by rifting initiates coastal erosion that may create seaward facing escarpments and coastal plains (e.g. Ollier, 1982; Seidl et al., 1996; Ollier and Pain, 1997). The Great Divide is considered to be the result of upwarping of the crust during rifting. To the east of the Great Divide, the Great Escarpment of eastern Australia, as described by Ollier (1982) can be traced almost continuously along the east coast of Australia, although in some places such as immediately north of Brisbane, the escarpment is obscure or absent. Localised subsidence in the Paleogene saw the evolution of the small Petrie, Oxley and Booval basins, which accumulated silt, clay, limestone, oil shale and basalt in lacustrine and paludal environments (Cranfield et al., 1976). Neogene erosion produced the modern, relatively subdued, topography in the Moreton district (Willmott, 2004). However, as Australia migrated northwards during the Cenozoic, the study area is considered to have moved over one or more hotspots causing localised felsic volcanism such as the Glasshouse Mountains (Jensen, 1903; Jensen, 1906; Stevens, 1971; Willmott, 2004) and mafic volcanism such as the Main Range and the Lamington Group volcanics (Stevens, 1965; Stevens, 1966) around 25-20 million years ago. Between approximately 6 million and 400,000 years ago, small basaltic volcanoes erupted in the Gayndah and Bundaberg areas, although the origins of these are not well understood (Willmott, 2004): they do not conform to the southward younging age-trend of other hot spot related volcanism of eastern Australia (Robertson, 1985; Sutherland, 1985; Sutherland, 2003). Since this time, erosion and deposition has continued and siliciclastic deposits have accumulated on flood plains, estuaries, deltas, spits and sandbars. The coastal plain is largely developed on Triassic to Jurassic mudstones and sandstones (for example the Kin-Kin Beds in the north and Landsborough Sandstone in the central coastal region) and Devonian-Carboniferous metamorphics (the Bunya Phyllite and the Neranleigh-Fernvale Beds in the Brisbane region). Neogene deposits are found throughout the coastal area.

PREVIOUS WORK Taylor (1914), related the river patterns of eastern Australia to structural geology, explaining that westward migration (coastal retreat) of the Great Dividing Range caused headwaters of westward-flowing streams to be captured and reversed. The

192 physiography of the Brisbane River and surrounding catchments has been described by several workers who generally concluded that local escarpments are shaped by erosion and not faults; structural lines of faulting, jointing and zones of preferential weathering in the area moderately coincide with drainage patterns (Marks, 1933; Beckmann and Stevens, 1978). An extensive geomorphological review that covered a large portion of our study area (Sussmilch, 1933), considered river channel positions, although not drainage patterns per se. Sussmilch described the general geomorphology and discussed the relationship between the complex series of horsts separating the continuous high western plateau from the eastern coastal plain. In his review, Sussmilch first described the Brisbane Gap, a low-lying area between the D’Aguilar and ‘Tambourine’ (now Beenleigh) Blocks. He observed that the main drainage through the Brisbane Gap does not flow along the central axis of this domain and inferred that the ‘Gap’ was probably not solely an erosional feature. He suggested that the Brisbane Gap is of tectonic origin and that the northern boundary is a fault-scarp along the southern margin of the D’Aguilar Block and its southern margin may also be a line of east-west faulting, previously suggested by Denmead (1928). In his review, Sussmilch also observed that the Stanley River flows in a general southwest direction to join the upper Brisbane River near Esk, despite it starting close to the coast, but he did not provide an explanation for this atypical ‘counter-coastal’ drainage orientation. Beckmann and Stevens (1978) suggested that due to back-cutting during the late Miocene-early Pliocene, drainage was reversed along the majority of the Stanley River. This may have been assisted by slight tilting to the west although explicit evidence was not available to substantiate this hypothesis. They also suggested that, at this time, several coastal rivers, such as the Caboolture and Pine Rivers, may have previously flowed into the Brisbane River and now discharge directly into Moreton Bay. A brief review of drainage in the Pine Rivers catchment defined it as being fault controlled and suggested that the North Pine Fault, which trends approximately northwest, is responsible for the elongate shape of the northwest trending North Pine River catchment (Hofmann, 1980). In the Bundaberg area, Landsat images clearly show a major north-northwest/south-southeast trend in ridges and valleys across the region (Childs, 1991) and the main ranges and drainage systems in that area also display a strong trend corresponding to geological structure. The geomorphological evolution of southeastern Australia has been related to

193 geological events that occurred episodically over a vast time scale. As tectonic movement is part of a dynamic, on-going process, the continued interruption to homogenous and consistent erosion will prevent the mature peneplain stage from being reached and instead cause complex evolution of the landscape (Ollier, 1995). The dynamic process, or ‘evolutionary geomorphology’, was found to be responsible for the anomalous drainage pattern of the Clarence River on Australia’s east coast (Haworth and Ollier, 1992); the ‘barbed’ drainage pattern suggests stream reversal whereby part of the ancient west-flowing Condamine River has been trapped by the east-flowing Clarence River. Drainage character in relation to the Great Divide and Great Escarpment has been described by several workers. In summary, drainage to the west of the Great Divide is relatively simple and commonly dendritic whereas to the east more complex patterns exist (e.g. Ollier and Haworth, 1994; Ollier and Pain, 1997). Drainage between the Great Divide and the Great Escarpment is often highly complicated (Ollier and Stevens, 1989). In some parts of eastern Australia, drainage to the east of the Great Escarpment may rise near the coast and turn inland before flowing out to sea (Ollier and Pain, 1997) and some rivers flow parallel to the coast, possibly following major structural lineaments (Beckmann and Stevens, 1978). However, some drainage east of the Great Escarpment in Queensland, is simple, flowing directly to the Pacific (e.g. at Rockhampton, Ollier and Stevens, 1989). It is clear that geological structure and physiography correlate strongly in southeast Queensland (Jones, 2006). However, some rivers and escarpments in this region show regular patterns that do not correspond to mapped structural features. Preliminary work has shown that zones of low magnitude earthquakes occur throughout the region and commonly correspond to the positions of features such as rivers and faults (Hodgkinson et al., 2006a). Monitoring of seismicity in the region is sparse by international standards, having only operated over short periods of time since the late 1800’s. Earthquake monitoring and prediction modelling is currently in progress at the Earth Systems Science Computational Centre (ESSCC) at the University of Queensland. Few geomorphological analyses have been undertaken to identify the relationships between known faults, landscape features and neotectonism in the region. Similarly, drainage patterns represented in southeast Queensland, and their degree of correlation with known structures, have not previously been described.

194 This study aims to provide a regional-scale review of drainage patterns, and discuss their relationship with known and potentially concealed geological features.

METHODS AND TERMINOLOGY Along the eastern coast of Queensland throughout the Late Cretaceous and Cenozoic, highlands lay to the west of the region and the lowlands and ocean lay to the east. In this paper, we assume that drainage would be sourced from the highlands and flow towards the coast (generally west to east). Deviations from this model are likely to be controlled by processes and features such as zones of differential weathering, faults, synforms or antiforms. Drainage trends that coincide with known geological structures are classified in this paper as ‘predictable’ and drainage that shows a distinctive pattern but is not associated with known geological features is described as ‘anomalous’. River systems of southeast Queensland were analysed and compared to available geological data to identify both predictable (Zernitz, 1932; Twidale, 2004) and anomalous drainage patterns. Digital topographic, geological and drainage maps of southeast Queensland were obtained from the Department of Natural Resources, Mines and Water (Queensland Government, 2003). Using ArcGIS 9 software, a digital slope-map was created. Earthquake epicentre data were obtained from the Earth Systems Science Computational Centre at The University of Queensland (ESSCC, 2006). The digital maps and data were integrated and analysed in ArcGIS 9 to identify anomalous drainage patterns in the region and to identify correspondence between gross drainage patterns and geological units, structures and earthquake epicentres.

RESULTS AND DISCUSSION A large number of rivers in southeast Queensland display patterns that suggest controls by geological features such as lithotypes, faults or folds. To illustrate the diversity and distribution of drainage patterns in the study area, several representative examples have been selected and described here (Fig. 4a).

195 196

Fig. 4 Diverse drainage patterns occur in southeast Queensland: a) map of study area showing locations of; b), dendritic; c), d), e), f), parallel; g), h), i) radial; j) centripedal; k), l) angular channel networks.

‘Predictable drainage patterns’ Examples of ‘predictable’ drainage patterns are expressed by some degree of regularity in their tributary orientations. It is assumed that some aspect of the bedrock geology or regional slope controls the development of these patterns. Several distinct drainage patterns of this type are evident in the study region and examples are presented here to illustrate their diversity. Patterns may vary depending on the scale of analysis (Hodgkinson et al., 2006b). Few drainage systems display a single type of drainage pattern throughout. Dendritic drainage develops where multiple factors influence channel formation or where drainage has evolved on a relatively uniform regional slope. These areas may lack significant structural control (Zernitz, 1932) but the drainage network is, nevertheless, predictable in its form. Dendritic drainage is fairly common in the region although it is not represented in any large-scale river systems and generally occurs in small systems or as sub-units of large drainage systems in which other patterns predominate (Fig. 4b). Parallel drainage patterns are common in broad-scale river systems such as the upper reaches of the Brisbane River where faults commonly correspond to drainage orientation (Fig. 4c). However, in the southwest of the area, where parallel channels drain away from the crest of the Great Escarpment (Fig. 4d), the few known faults are perpendicular to drainage. Many basement faults in the southwest of the area may be concealed by Cenozoic basalts and thick regolith cover. Earthquake epicentre corridors in the upper Brisbane Valley (Fig. 5) are typically parallel to the channels suggesting that these faults remain active and may be influencing modern drainage. Although not closely associated spatially, a strong parallel southwest- northeasterly drainage trend is apparent across the region including (from south to north) Teviot Brook, Logan River, the southern trunk of the Brisbane River, the upper Stanley River, Yabba Creek and Kandanga Creek (Fig. 1). Parallel drainage is also common at a finer scale, such as to the east of the D’Aguilar Range (Fig. 4e), where channel orientation may be largely controlled by rock fabric (Hodgkinson et al., 2006b) and faulting. Parallel drainage is also evident on the major sand barrier islands where Quaternary parabolic dunes are the major influence on topography (Fig. 4f). A radial drainage pattern occurs on a broad scale, on the outer flanks of the Mount Warning shield volcano in the southeast of the region (Fig. 4g) and at a finer

197

scale in several highland sites such as Mt Nebo, Mt Glorious (Fig. 4h) and Mt Perseverance (Fig. 4i). A centripedal drainage pattern occurs within the Mt Warning shield volcano inner escarpment and basin (northern New South Wales) and is also weakly developed at a smaller scale on the deeply weathered Samford batholith in the east (Fig. 4j). Angular and trellis (or rectilinear) drainage patterns occur across the region at various scales. On a broad scale, the Brisbane River system can be described as a combination of both styles, and the Mary River in the north, has similar features (Fig. 2). At an intermediate scale the North Pine River in the east (Fig. 4k) is also rectilinear in style and numerous smaller rivers display this pattern throughout the region. The Brisbane River and North Pine Rivers both closely follow faulted and earthquake-prone zones. However, some linear reaches of the Brisbane River lack mapped faults and these may be the sites of concealed structural features. The numerous meanders of the Mary River (Fig. 4l) may be controlled by the strong northwest-southeasterly and northeast-southwesterly trends of small faults in the area (Fig. 2) but these are not significant enough to control the location and orientation of the river system as a whole. There are relatively few recorded earthquakes along the Mary River axis (ESSCC 2006) although earthquake epicentre corridors correspond to some upper tributaries and a similarly orientated corridor (Fig. 5) is recorded offshore (ESSCC 2006; Hodgkinson et al., 2006a). Other parts of the Mary River system are aligned with larger faults (Fig. 2).

198

Fig. 5 Dominant earthquake corridors in the study area defined by the criteria of Hodgkinson et al. (2006a). As position and depth to earthquake may be inaccurate, broad ‘corridors’ were defined to allow for location errors. The corridors show there is some alignment of earthquakes, several of which correspond to known faults. Earthquake data span past 130 years. Data source: Earth System Science Computational Centre (ESSCC, 2006)

‘Anomalous drainage’ Although the Mary River drainage pattern (Figs. 2, 4l) has been described above as rectilinear, suggesting structural influence, the northward course of the main channel is not coincident with any known major structural feature and is, therefore, anomalous. However, the shore-parallel course of the Mary River may have been incised prior to down-warp, as has been identified in some other rivers, such as the Shoalhaven River in New South Wales (Ollier and Pain, 1994), and may therefore represent antecedent drainage. A strong element of westward drainage (away from the coast) is evident in the eastern part of the Mary River catchment, the control of which is not obvious and, therefore, anomalous. To the south of the Mary River catchment, similar westward drainage persists and is controlled by a drainage divide extending approximately 120 km in length and situated 10 - 30 km from the coast (Fig. 6a). This drainage divide occurs at relatively low elevation (less than 100 m above sea level in some places and not exceeding 200 m above sea level for the majority of its length) from the area of the Mary River catchment in the north, to the North Pine River catchment in the centre of the region (hereafter designated as the ‘coastal drainage divide’: Fig. 6a, b). Although the coastal drainage divide exceeds

199

300 m above sea level in some places, the majority of the divide is less than 150 m above sea level in elevation. The coastal drainage divide crosses several geological units of varying lithology that are of Triassic to Jurassic age in the northern and southern sectors of the divide, and of Permian age in the central part. No known faults or other geological structures correspond to the position of the coastal drainage divide and we, therefore, designate drainage immediately to the west of this feature as anomalous. To the west of the divide, several small river channels display evidence of stream-capture and align with steams to the east of the divide where they may have previously flowed (Fig. 6b). For example, in the upper eastern reaches of the Brisbane River system, the Stanley River is sourced in the Blackall Ranges, flows eastward towards the coastal drainage divide, then turns abruptly southwest over a relatively fault-free area, towards the faulted and earthquake-prone area at the north end of Lake Somerset, after which it joins the Brisbane River. The course of the upper Stanley River, as it approaches the coastal drainage divide, aligns with the course of Coochin Creek to the east of the divide. As well as anomalous drainage and stream-capture, we propose that the upper section of the Stanley River has undergone reverse drainage. Obi Obi Creek also initiates in the Blackall Ranges, flows in an easterly direction then turns north at the coastal drainage divide (Fig. 6b). It flows parallel to the divide for approximately 5 km then turns to the northwest and continues in this direction until it joins the Mary River at Kenilworth. The east- flowing reach of Obi Obi Creek aligns with the Mooloolah River to the east of the divide. Faults or preferential weathering do not appear to control the changes in direction of the creek and we propose this anomalous drainage is controlled by flexure along the coastal drainage divide and capture of the upper Mooloolah River. Similarly, to the north, Six Mile Creek commences by flowing towards the coast but turns northwards and then northwest at the drainage divide and continues in this direction until it reaches the Mary River, south of Gympie (Fig. 6b). To the east of the divide, Ringtail Creek aligns with the east-flowing reach of Six Mile Creek. This may represent a third case of eastward flowing stream headwaters being captured by northward or westward flowing rivers. A linear depression runs from the highland in the northwest, through the area of Kilcoy and towards the Bribie Island region, that may be the pathway of an ancient drainage system (Fig. 7). This valley (hereafter designated the ‘Kilcoy Gap’) begins with an orientation that is conformable to the

200

northwest-southeast trend of faults in the upper Brisbane Valley but turns to the east and descends through foothills towards the modern coastal plain and Bribie Island. Several modern watercourses are situated within this depression including parts of the upper Stanley River draining to the west, and Coochin and Mellum Creeks draining to the southeast and east. Extensive unconsolidated sediments flank these streams.

a) b)

.eps version

Fig. 6 a) Drainage networks in northeastern part of Moreton district showing the ‘coastal drainage divide’ (heavy line) and locations of Cenozoic volcanics; b) 3 examples of potentially reversed and captured drainage along the coastal drainage divide. See text for inferred causal mechanism.

The depression is largely developed on the Triassic Neurum Tonalite and the latest Triassic to Early Jurassic Landsborough Sandstone. Several late Palaeozoic volcanic and metasedimentary units are also exposed within the depression. One site near the northern margin of the valley (grid ref. 152° 50E' 26", 26° 50' 5"S) reveals previously unmapped, poorly consolidated, coarse-grained channel-fill deposits, incisively overlying Palaeozoic andesite along a modern interfluve approximately 260 m above sea level. Sediments in this exposure are oxidized and have not been dated but their relatively unconsolidated state suggests that they are Cenozoic in age. The deposits provide evidence that at least one ancient river channel has been preserved within this broad valley by topographic inversion (Fig. 8). However, the

201

scarcity of such sites suggests that erosion of similar channel fill deposits together with a substantial quantity of volcanic and metamorphic bedrock has occurred in the Kilcoy Gap region.

Fig. 7 Topographic map detailing the broad valley structure between Kilcoy and the Glasshouse Mountains (The ‘Kilcoy Gap’) and the right- angled bend in the Brisbane River near Ipswich.

202

The morphology of the Kilcoy Gap, suggests it may have been a drainage system descending from highlands in the northwest to the coast in the east. The Stanley River, which is situated within this depression, presently flows away from the coast and the coastal drainage divide, and it may have experienced reversal of drainage within a previously major river channel. As mentioned above, Beckmann and Stevens (1978), suggested the reversal of the Stanley River, the mechanism for which they proposed was backcutting of the Stanley headwaters across a previous divide near Kilcoy. However, backcutting alone is unlikely to have caused the reversal of drainage and we suggest the process was combined with one of stream- capture, enhanced by westward tilting. Beckmann and Stevens (1978) also suggested tilting to the west but were unable to substantiate this. From this study, we invoke a mechanism involving doming or tilting along the coastal drainage divide. The cause of doming or flexure along this line may be the late Oligocene to early Miocene emplacement of the Glasshouse Mountains volcanic plugs and a series of more isolated coeval felsic and mafic intrusives along the coastal plain to the north. The changes in direction of Obi Obi Creek and Six Mile Creek, together with the alignment of their eastward flowing reaches with other streams east of the divide, provide further evidence that flexure along the coastal drainage divide has forced stream-capture and potentially stream reversal. Stream-capture is common elsewhere along the eastern Australian highlands such as within the Clarence-Moreton Basin, northern New South Wales (Ollier and Haworth, 1994). Topographically, the coastal drainage divide (Fig. 6) is, for the main part, a low-relief feature throughout its length. If the feature is not currently undergoing uplift, given sufficient time it may eventually erode and retreat westwards allowing coastal streams to re-capture the headwaters of some westward flowing systems. Given that no major streams flow from the Great Divide to the coast, in the area from Brisbane to the northern limit of the study area, we propose that the coastal drainage divide described here, is a relatively young feature generated by Cenozoic uplift or tilting. Two major mechanisms may be responsible for the uplift. Firstly, uplift may have been generated along a deep-seated, unmapped fault, blind thrust, or gentle flexure of the crust, caused by the current compressional tectonic regime. A series of shore-parallel, gently undulating synforms and antiforms or block faulting caused by compression would explain both the drainage divide and the shore-parallel drainage orientation of the Mary River. Similar undulating topography associated with gentle,

203

Late Cretaceous folding and faulting occurs in the central Maryborough Basin, immediately north of the study area (Ellis, 1966). If this model is correct, the processes that generated the coastal drainage divide may currently be inactive as few earthquakes have been recorded in the vicinity of this feature. Although the coastal drainage divide is not perpendicular to the first-order (plate-boundary controlled) northeast-southwest stress orientation in the region (e.g. Cuthbertson, 1990; Hillis et al., 1999), it may be attributable to local, second-order stress variations, which occur throughout Australia, caused by structural, topographic and density variations in the lithosphere (Hillis and Reynolds, 2003). Late Cenozoic stresses have caused extensive flexures and faults in the landscape in diverse parts of the continent such as the Flinders Ranges (Celerier et al., 2005), the Carnarvon Basin in Western Australia (Hocking, 1990) and the Gippsland Basin, Victoria (Nelson and Hillis, 2005). Secondly, some parts of the divide closely coincide with the location of small areas of mid-Cenozoic felsic to mafic igneous intrusions (Fig. 6a). Thermal doming and emplacement of magma may have caused uplift of the crust around these zones, creating a linked meridional zone of flexure. The coastal drainage divide has a topographic expression that is barely perceptible in places and does not precisely coincide with the position of the volcanics for the majority of its length. The gently scalloped or sinuous pattern of the divide over much of its length may be explained by intense subtropical weathering and erosion during the past 20 million years. Dramatic realignment of river segments is evident elsewhere in southeast Queensland. The southern section of the Brisbane River shows an abrupt right- angled change in orientation towards the northeast in the vicinity of Ipswich and then passes in a relatively straight line towards the coast through the Brisbane Gap (Sussmilch, 1933; Fig. 7). No mapped structural feature aligns with this sharp reorientation. Although few faults are mapped in this area (e.g. Queensland Government, 2003), the lower segment of the Brisbane River coincides with a zone of low magnitude earthquakes (Fig. 5) and this may be evidence of a concealed structural feature (Hodgkinson et al., 2006a). Denmead (1928) and Sussmilch (1933) hypothesised that the Brisbane Gap is bounded by faults to the north and south. However, Cranfield et al. (1976) suggested a single major fault [the hypothetical Buranda Fault of Bryan and Jones (1954)] separates the D’Aguilar and Beenleigh Blocks. Some older maps (e.g. Cranfield and Schwarzbock, 1971) include this inferred structure, along the southern margin of the South D’Aguilar Block. A major

204

concealed fault in this location would explain both the linear NE orientation of the lower Brisbane River and the lateral offset of the exposed Palaeozoic basement rocks to the north (South D’Aguilar Block) and south (Beenleigh Block) of the river. Several earthquake epicentres over the last 150 years, plot in this area (ESSCC 2006; Fig. 5), suggesting that a major geological discontinuity indeed exists along this corridor. Teviot Brook and the upper Logan River flow in a similar orientation to the lower Brisbane River (Fig. 1), and may be aligned along parallel structural discontinuities although they both lie in an area where few faults or earthquakes have been recorded. Not all stream-capture features and drainage patterns in southeast Queensland were necessarily generated by differential weathering or neotectonism in the strict sense. Some drainage patterns have clearly been influenced by the emplacement of Quaternary coastal aeolian dune systems (Fig. 4f); other streams are likely to have been diverted or initiated by the emplacement of Cenozoic mafic lava fields (Fig. 6a), and some watercourses may have migrated to their current positions simply due to autogenetic factors. The lower reaches of the Noosa River provide a good example of the last of these processes. Aerial photography reveals the existence of an ancient large meandering channel located around 2 km to the east of the current position of the lower Noosa River (Fig. 9). This feature likely represents the ancient position of the lower Noosa River until an avulsion event in the relatively recent past redirected the main channel into Lake Cooroibah and thence southwards to Tewantin. The lower part of the Noosa, and other major southeast Queensland river valleys, are characterized by broad floodplains (kilometres to 10s of kilometres wide). Hence episodic avulsion events can result in wide-spaced repositioning of the major channel within the lower alluvial valley. Therefore, it is important to consider drainage patterns upstream, if assessing potential controls on river positions in southeast Queensland.

205

Fig. 8 Road-cut at Cedarton showing fining-up (conglomerate to medium grained sandstone) channel fill deposits resting erosively on Permian andesites located on a modern interfluve interpreted to represent inverted mid-Cenozoic topography. Thick line represents basal erosion surface. Thin lines represents scour surfaces within the channel fill. Geologist for scale approximately 1.8 m.

a) b)

Fig. 9 Aerial photograph of Noosa River area a) showing current river channel flowing through Lake Cooroibah; b) ancient abandoned channel highlighted to the east of current river position.

CONCLUSIONS The dominant controls on modern drainage patterns in southeast Queensland are differential erosion of lithotypes and entrapment by geological structures, together with late Cenozoic volcanism. These processes are also clearly evident in other parts

206

of eastern Australia (Holdgate et al., 2006). This study has identified a range of drainage patterns at various scales across the region. Although short, high-energy sub-tropical catchments with eastward flowing streams may be expected in a setting such as southeast Queensland, it is evident that far more complex drainage patterns and drainage histories exist in this region. Several anomalies in expected drainage were noted. Some of these are readily explained by differential weathering and erosion of bedrock lithotypes or by alignment with mapped ancient structural features. However, the genesis of other drainage anomalies remains ambiguous. Previous authors recognised that anomalous river orientations and drainage patterns may be caused by active faults or folds or by catastrophic climatic, tectonic or extraterrestrial events (Twidale, 2004) and that dendritic drainage patterns appear to indicate a lack of structural control (e.g. Zernitz, 1932; Ollier and Haworth, 1994). Both the scarcity of dendritic drainage, coupled with the multitude of other drainage patterns identified in this study, suggests that geological structure imposes major controls on drainage in this region. Faults and other geological structures have a profound influence on the geomorphology of the region. The Kosciusko uplift, which may have been responsible for some uplift in this region (Sussmilch, 1933), has received little attention in modern literature, but recent work in NSW has prompted a renewed interest in the nature, timing and extent of this event (Kohn et al., 1999; Sharp, 2004). Further comparisons between the NSW highlands and southeast Queensland may clarify the influence of this phase of tectonism. The lower course of the Brisbane River, running parallel to the southern edge of the D’Aguilar Block, is in close proximity to a belt of recent earthquakes and the drainage in this area may be experiencing ongoing adjustments due to neotectonism. Other areas, such as along the coastal drainage divide, require better structural and seismicity data before the influence of neotectonics can be confirmed or excluded. This study clearly shows that structural features and lithological variations have played a major role in the development of the Moreton District drainage. Therefore, although some antecedent drainage may persist, it is highly likely that anomalous drainage features are also controlled by concealed structures or subtle lithological differences. Future seismic traverses may resolve this issue and further analysis of earthquake data such as focal mechanisms of recent and future earthquakes in the region would better constrain this inference. Although the geology

207

of the area is relatively well-mapped by Australian standards, higher resolution mapping may also provide a better understanding of the genetic controls on channel orientation across the region.

ACKNOWLEDGEMENTS We wish to express our appreciation for the rigorous reviews and valuable suggestions of Dr Colin Pain and Dr Malcolm Jones. We thank Dr Dion Weatherley and Col Lynam of ESSCC, University of Queensland, for important collaborative discussions and for providing earthquake data. We are also grateful to Dr Andrew Hammond for reviewing this paper and for enlightening debate, Bill Ward for informative local field trips, and Jonathan Hodgkinson for invaluable support and assistance in the field.

REFERENCES

Arnett, R.R., 1969. Soil-slope relationships in the Rocksberg Basin, south-east Queensland. M.A. Thesis, University of Queensland, Brisbane, 210 pp. Arnett, R.R., 1971. Slope form and geomorphological process: an Australian example. Slopes, Form and Process. Special Publication, Institute of British Geographers, 3: 81-92. Australian Bureau of Statistics, 2006. Regional Population Growth, Australia, 2004- 05, Cat. no. 3218.0, ABS, Canberra. Beckmann, G.G. and Stevens, N.C., 1978. Geological history of the Brisbane River system. Proceedings of the Royal Society of Queensland, 89: 77-85. Branson, J.C., 1978. Evolution of sedimentary basins from Mesozoic times in Australia's continental slope and shelf. In: E. Scheibner (Editor), The Phanerozoic structure of Australia and variations in tectonic style. Tectonophysics,, pp. 389-412. Bryan, W.H. and Jones, O.A., 1954. Contributions to the geology of Brisbane No. 2. The structural history of the Brisbane Metamorphics. Proceedings of the Royal Society of Queensland, 65: 25-50. Burnett, A.W. and Schumm, S.A., 1983. Alluvial-river response to neotectonic deformation in Louisiana and Mississippi. Science, 222(4619): 49-50. Burrato, P., Ciucci, F. and Valensise, G., 2003. An inventory of river anomalies in the Po Plain, Northern Italy: evidence for active blind thrust faulting. Annals of Geophysics, 46(5): 865-882. Celerier, J., Sandiford, M., Hansen, D.L. and Quigley, M., 2005. Modes of active intraplate deformation, Flinders Ranges, Australia. Tectonics, 24(6): Art. No. TC6006.

208

Childs, I., 1991. Earthquake risk, Landsat imagery and fault zones in the Bundaberg area of central eastern Queensland. Queensland Geographical Journal, Series 4, 6: 59-70. Clark, D. and McCue, K., 2003. Australian paleoseismology: towards a better basis for seismic hazard estimation. Annals of Geophysics, 46(5): 1087-1105. Cranfield, L.C. and Schwarzbock, H., 1971. Geology of Brisbane 1:250 000. Queensland Department of Mines. Cranfield, L.C., Schwarzbock, H. and Day, R.W., 1976. Geology of the Ipswich and Brisbane 1:250 000 Sheet Areas. Geological Survey of Queensland Report 95, Queensland Department of Mines. Cuthbertson, R.J., 1990. The seismo-tectonics of southeast Queensland. Bureau of Mineral Resources, Geology and Geophysics, Australia, Bulletin, 232: 67-81. Day, R.W., Murray, C.G. and Whitaker, W.G., 1978. The eastern part of the Tasman Orogenic Zone. Tectonophysics, 48: 327-364. Day, R.W., Whitaker, W.G., Murray, C.G., Wilson, I.H. and Grimes, K.G. (Editors), 1983. Queensland Geology, A companion volume to the 1:2 500 000 scale geological map (1975). Geological Survey of Queensland, Publication 383. Delcaillau, B., Carozza, J.-M. and Laville, E., 2006. Recent fold growth and drainage development: The Janauri and Chandigarh anticlines in the Siwalik foothills, northwest India. Geomorphology, 76(3-4): 241-256. Denmead, A.K., 1928. A survey of the Brisbane schists. Proceedings of the Royal Society of Queensland, 39: 71-106. Donchak, P.J.T., 1976. A structural study of the greenstone-phyllite sequence of the Petrie-Dayboro area. B.Sc. Honours Thesis, University of Queensland, Brisbane, (Unpublished). Ellis, P.L., 1966. The Maryborough Basin. APEA Journal: 30-36. ESSCC, 2006. Earth Systems Science Computational Centre, Earthquake Database, University of Queensland, Brisbane. Evernden, J.F. and Richards, J.R., 1962. Potassium-argon ages in eastern Australia. Journal of the Geological Society of Australia, 9: 1-50. Fergusson, C.L. and Leicht, E.C., 1993. Late Carboniferous to Early Triassic tectonics of the New England Fold Belt, eastern Australia. In: P.G. Flood and J.C. Aitchison (Editors), New England Orogen, Eastern Australia. Department of Geology and Geophysics, University of New England, Armidale, pp. 53-59. Fleming, P.J.G., Day, R.W., Murray, C.G. and Whitaker, W.G., 1974. Late Palaeozoic invertebrate macrofossils in the Neranleigh-Fernvale Beds. Queensland Government Mining Journal, 75: 104-107. Goldsworthy, M. and Jackson, J., 2000. Active normal fault evolution in Greece revealed by geomorphology and drainage patterns. Journal of the Geological Society, 157: 967-981. Haworth, R.J. and Ollier, C.D., 1992. Continental Rifting and Drainage Reversal - the Clarence River of Eastern Australia. Earth Surface Processes and Landforms, 17(4): 387-397. Hillis, R.R., 1998. The Australian stress map. Exploration Geophysics, 29(3-4): 420. Hillis, R.R., Enever, J.R. and Reynolds, S.D., 1999. In situ stress field of eastern Australia. Australian Journal of Earth Sciences, 46(5): 813-825. Hillis, R.R. and Reynolds, S.D., 2000. The Australian Stress Map. Journal of the Geological Society, 157: 915-921.

209

Hillis, R.R. and Reynolds, S.D., 2003. In situ stress field of Australia. Geological Society of Australia Special Publication, 22: 43 - 52. Hills, E.S., 1963. Elements of Structural Geology. Methuen, London, 483 pp. Hocking, R.M., 1990. Carnarvon Basin, Geology and Mineral Resources of Western Australia: Western Australia Geological Survey, pp. 457-493. Hodgkinson, J.H., McLoughlin, S. and Cox, M., 2006a. The correlation between physiography and neotectonism in southeast Queensland, Proceedings of the Earthquake Engineering Conference AEES2006, Canberra, ACT, Australia. Hodgkinson, J.H., McLoughlin, S. and Cox, M., 2006b. The influence of geological fabric and scale on drainage pattern analysis in a catchment of metamorphic terrain: Laceys Creek, southeast Queensland, Australia. Geomorphology, 81(3-4): 394-407. Hofmann, G.W., 1980. Quaternary sediments and geological history of the Pine Rivers area, Southest Queensland. Queensland Government Mining Journal, 81(948): 502-512. Holcombe, R.J., 1978. Structure and tectonic history of the Brisbane metamorphics in the Brisbane area. Journal of the Geological Society of Australia, 24: 475- 490. Holcombe, R.J., Little, T.A., Sliwa, R. and Fielding, C.R., 1993. Tectonic elements and evolution of SE Queensland during the Late Palaeozoic to Early Mesozoic. In: P.G. Flood and J.C. Aitchison (Editors), New England Orogen, Eastern Australia. Department of Geology and Geophysics, University of New England, Armidale, pp. 517-527. Holcombe, R.J. and Little, T.A., 1994. Blueschists of the New-England Orogen - Structural Development of the Rocksberg Greenstone and Associated Units near Mt-Mee, Southeastern Queensland. Australian Journal of Earth Sciences, 41(2): 115-130. Holdgate, G.R., Wallace, M.W., Gallagher, S.J., Witten, R.B., Stats, B. and Wagstaff, R.E., 2006. Cenozoic fault control on 'deep lead' palaeoriver systems, Central Highlands, Victoria. Australian Journal of Earth Sciences, 53(3): 445-468. Jackson, J., Van Dissen, R. and Berryman, K., 1998. Tilting of active folds and faults in the Manawatu region, New Zealand: Evidence from surface drainage patterns. New Zealand Journal of Geology and Geophysics, 41(4): 377-385. Jensen, H.I., 1903. The geology of the Glass House Mountains and district. Proceedings of the Linnean Society of New South Wales, 28: 842-875. Jensen, H.I., 1906. Geology of the volcanic area of the east Moreton and Wide Bay districts, Queensland. Proceedings of the Linnean Society of New South Wales, 31: 73-173. Jones, M.R., 2006. Cenozoic landscape evolution in central Queensland. Australian Journal of Earth Sciences, 53(3): 433-444. Kohn, B.P., Gleadow, A.J.W. and Cox, S.J.D., 1999. Denudation history of the Snowy Mountains: constraints from apatite fission-track thermochronology. Australian Journal of Earth Sciences, 46(2): 181-198. Little, T.A., Holcombe, R.J., Gibson, G.M., Offler, R., Gans, P.B. and McWilliams, M.O., 1992. Exhumation of late Paleozoic blueschists in Queensland, Australia, by extensional faulting. Geology, 20(3): 231-234. Little, T.A., Holcombe, R.J. and Sliwa, R., 1993. Structural evidence for extensional exhumation of blueschist-bearing serpentinite matrix melange, New-England Orogen, Southeast Queensland, Australia. Tectonics, 12(2): 536-549.

210

Lucas, K.G., 1987. Landsat lineaments from Childers to Caboolture. In: C.G. Murray and J.B. Waterhouse (Editors), Field Conference: Gympie District. Geological Society of Australia, Queensland Division, Queensland, Australia, pp. 86-87. Marks, E.O., 1933. Some observations on the physiography of the Brisbane River and neighbouring watersheds. Proceedings of the Royal Society of Queensland, 44(9): 132-150. Mather, A.E., 1993. Basin inversion: some consequences for drainage evolution and alluvial architecture. Sedimentology, 40(6): 1069-1089. Maynard, J.R., 2006. Fluvial response to active extension: evidence from 3D seismic data from the Frio Formation (Oligo-Miocene) of the Texas Gulf of Mexico Coast, USA. Sedimentology, 53(3): 515-536. Murphy, P.R., Trezise, D.L., Hutton, L.J. and Cranfield, L.C., 1979. Kurwongbah Beds - A new stratigraphic unit of the D'Aguilar Block. Queensland Government Mining Journal, 80: 65 - 70. Murray, C.G., McClung, G.R. and Whitaker, W.G., 1979. Early Permian fossils from the Amamoor Beds and the age of the Booloumba Beds - a new stratigraphic unit of the North D'Aguilar Block. Queensland Government Mining Journal, 80: 71-78. Murray, C.G. and Whitaker, W.G., 1982. A review of the stratigraphy, stucture and regional tectonic setting of the Brisbane metamorphics. In: P.G. Flood and B. Runnegar (Editors), New England Geology. Department of Geology, University of New England and AHV Club, Armidale, pp. 79-94. Murray, C.G., 1987. Review of the Geology of the Gympie District, southeast Queensland, with particular reference to stratigraphic, structural and tectonic problems. In: C.G. Murray and J.B. Waterhouse (Editors), Field Conference: Gympie District. Geological Society of Australia, Queensland Division, Queensland, Australia, pp. 1-17. Murray, C.G., Fergusson, C.L., Flood, P.G., Whitaker, W.G. and Korsch, R.J., 1987. Plate tectonic model for the Carboniferous evolution of the New-England Fold Belt. Australian Journal of Earth Sciences, 34(2): 213-236. Nelson, E.J. and Hillis, R.R., 2005. In situ stresses of the West Tuna area, Gippsland Basin. Australian Journal of Earth Sciences, 52(2): 299-313. Ollier, C.D., 1981. Tectonics and landforms. Longman, Harlow, UK, 324 pp. Ollier, C.D., 1982. The Great Escarpment of eastern Australia: tectonic and geomorphic significance. Journal Geological Society of Australia, 29(1-2): 13-23. Ollier, C.D. and Stevens, N.C., 1989. The Great Escarpment in Queensland. In: R.W. Le Maitre (Editor), Pathways in Geology: Essays in Honour of Edwin Sherbon Hills. Blackwell, Melbourne, Australia, pp. 140-152. Ollier, C.D. and Haworth, R.J., 1994. Geomorphology of the Clarence-Moreton Basin. In: A.T. Wells and P.E. O'Brien (Editors), Geology and petroleum potential of the Clarence-Moreton Basin, New South Wales and Queensland: Bulletin 241. AGSO, Canberra. Ollier, C.D. and Pain, C.F., 1994. Landscape evolution and tectonics in southeastern Australia. AGSO Journal of Australian Geology and Geophysics, 15(3): 335- 345. Ollier, C.D., 1995. Tectonics and landscape evolution in southeast Australia. Geomorphology, 12(1): 37-44.

211

Ollier, C.D. and Pain, C.F., 1997. Equating the basal unconformity with the palaeoplain: a model for passive margins. Geomorphology, 19(1-2): 1-15. Ouchi, S., 1985. Response of alluvial rivers to slow active tectonic movement. Geological Society of America Bulletin, 96(4): 504-515. Plumb, K.A., 1979. The tectonic evolution of Australia. Earth-Science Reviews, 14: 205-249. Powell, C.M., Edgecombe, D.R., Henry, N.M. and Jones, J.G., 1976. Timing of regional deformation at Hill End Trough: a reassessment. Journal of the Geological Society of Australia, 23: 407-421. Queensland Government, D.N.R.M.W., 2003. Queensland Geological Digital Data, CD ROM. Reynolds, S.D., Coblentz, D.D. and Hillis, R.R., 2002. Tectonic forces controlling the regional intraplate stress field in continental Australia: Results from new finite element modelling. Journal of Geophysical Research, 107(B7): 10.1029/2001JB000408. Reynolds, S.D., Coblentz, D.D. and Hillis, R.R., 2003. Influences of plate-boundary forces on the regional intraplate stress field of continental Australia. Geological Society of Australia Special Publication 22 and Geological Society of America Special Paper 372: 53-64. Robertson, A.D., 1985. Upper mantle xenoliths and megacrysts and the origin of the Brigooda basalt and breccia, near Proston, Queensland. Papers - Department of Geology, University of Queensland, 11(3): 58-71. Schumm, S.A. and Khan, H.R., 1972. Experimental study of channel patterns. Geological Society of America Bulletin, 83(6): 1755-1770. Seidl, M.A., Weissel, J.K. and Pratson, L.F., 1996. The kinematics and pattern of escarpment retreat across the rifted continental margin of SE Australia. Basin Research, 12: 301-316. Sharp, K.R., 2004. Cenozoic volcanism, tectonism and stream derangement in the Snowy Mountains and northern Monaro of New South Wales. Australian Journal of Earth Sciences, 51(1): 67-83. Stevens, N.C., 1965. The volcanic rocks of the southern part of the Main Range, south-east Queensland. Proceedings of the Royal Society of Queensland, 77(4): 37-52. Stevens, N.C., 1966. Tertiary volcanic rocks, In: Southern Moreton Basin: field conference. Geological Society of Australia. Queensland Division, pp. 11-14. Stevens, N.C., 1971. The Glasshouses. In: G. Playford (Editor), Geological Excursions Handbook, ANZAAS, 43rd Congress. ANZAAS and Geological Society of Australia, Queensland Division, Brisbane, pp. 153-158. Strahler, A.N., 1966. The Earth Sciences. Harper International, 681 pp. Sussmilch, C.A., 1933. The geomorphology of the Moreton District, Queensland; an interpretation. Proceedings of the Royal Society of Queensland, 44: 104-131. Sutherland, F.L., 1985. Regional controls in eastern Australian volcanism, Publication - New South Wales Division. Geological Society of Australia, New South Wales Division : Sydney, N.S.W., Australia, Australia, pp. 13-32. Sutherland, F.L., 2003. 'Boomerang' migratory intraplate Cenozoic volcanism, eastern Australian rift margins and the Indian-Pacific mantle boundary. Special Paper - Geological Society of America, 372: 203-221. Taylor, T.G., 1914. Physiography of Eastern Australia: Bulletin 8, Commonwealth Bureau of Meteorology, Melbourne.

212

Twidale, C.R., 2004. River patterns and their meaning. Earth-Science Reviews, 67(3- 4): 159-218. Vannoli, P., Basili, R. and Valensise, G., 2004. New geomorphic evidence for anticlinal growth driven by blind-thrust faulting along the northern Marche coastal belt (central Italy). Journal of Seismology, 8(3): 297-312. Veevers, J.J. and Evans, P.R., 1975. Late Palaeozoic and Mesozoic history of Australia. In: K.S.W. Campbell (Editor), Gondwana Geology: papers from the 3rd Gondwana symposium. ANU Press, Canberra, Canberra, ACT, pp. 579-607. Veevers, J.J., 2001. Atlas of Billion-year earth history of Australia and neighbours in Gondwanaland. GEMOC Press, Sydney, 76 pp. Watkins, J.R., 1967. The relationship between climate and the development of landforms in the Cainozoic rocks of Queensland. Journal of the Geological Society of Australia, 14(1): 153-168. Webb, A.W. and McDougall, I., 1967. Isotopic dating evidence on the age of the Upper Permian and Middle Triassic. Earth and Planetary Science Letters, 2: 483-488. Whitaker, W.G. and Green, P.M., 1980. Moreton Geology 1:500 000 map. Department of Mines, Queensland. Willmott, W., 2004. Rocks and Landscapes of the National Parks of Southern Queensland. Geological Society of Australia Incorporated, Queensland Division, Brisbane, 162 pp. Zelilidis, A., 2000. Drainage evolution in a rifted basin, Corinth Graben, Greece. Geomorphology, 35: 69-85. Zernitz, E.R., 1932. Drainage patterns and their significance. Journal of Geology, 40: 498-521. Zhao, S. and Müller, R.D., 2001. The tectonic stress field in eastern Australia. In: K.C. Hill and T. Bernecker (Editors), PESA Eastern Australasian Basins Symposium: a refocussed energy perspective for the future, Melbourne, Victoria, pp. 61-70.

213

214 Conclusions

The multi-criteria analytical approach used in this project has confirmed that underlying geological features control large components of the southeast Queensland landscape. In its entirety, the research covers a far greater area both in extent and scope, has utilized more geological information, and compares more datasets than previous studies. In turn, the work has yielded significant new findings for southeast Queensland, in addition to providing a new stream-ordering system and geomorphological knowledge that may be applied to other regions with similar tectonic settings.

In summary, the significance of this work includes the following: • The analysis presented in paper 1 is the first study of its kind that encompasses a whole sub-catchment that is situated on two differing, juxtaposed meta-sedimentary rock types. Additionally, an entirely new stream-ordering system has been presented in paper 1 that, unlike other ordering systems, importantly retains a topology that is meaningful in terms of stream importance throughout the drainage system. • An assessment of earthquake epicentre locations in relation to geological structure of southeast Queensland is presented in paper 2, which is the first assessment to have been performed utilising the most complete earthquake database currently available. It is also the first analysis of its kind to have been performed for this region in more than 15 years. • The regional-scale geomorphological review of southeast Queensland described in paper 3 has integrated a range of datasets using GIS at a scale and extent that has previously not been achieved. The work has identified some unique findings for the region that have applicability to understanding landscape development in similar tectonic settings globally.

Paper 1 has been cited by Ribohni and Spagnolo (2008) as one of the few studies globally that has dealt with the influence of both the different rate of rock uplift and

215 the selective erosion processes along fault and fracture planes. The three published papers provide a continuous study of multiple aspects of the endogenic controls of geomorphology in southeast Queensland. Together, they provide methods, examples, results and conclusions that may be used for geomorphological studies in other geological and geographic settings globally.

Summary of results and major findings This study has identified that the morphology of southeast Queensland is the cumulative result of multiple tectonic events and the region is not analogous to all other passive margin settings. The study has identified various geological events over the past 350 million years that have influenced, and in some cases continue to influence the evolution of the modern landscape in the region. Integration of the analytical tools used here, with the different datasets and applied at multiple scales, has shown that, although surface processes are on-going, geological controls from various events in the past continue to leave their mark on the landscape.

Hypothesis 1

“Complex geological fabric of metamorphic rocks of southeast Queensland has control over the orientation of streams at the sub-catchment scale”

The first hypothesis that was tested aimed to identify the extent to which streams show alignment with the underlying rock fabric within a catchment that had developed on two juxtaposed metamorphic rock types.

The findings identified that: • the orientation of low order streams are the least strongly controlled by rock fabric and faults; • the orientation of middle order streams are mainly controlled by fine-scale rock fabric features, particularly foliation and joints in phyllitic bedrock; • the orientation of higher order streams are mainly controlled by major faults.

In conclusion from testing this hypothesis, low order streams tend to remain dendritic until they have down-cut to reach bedrock and once this is achieved, the orientation

216 of rock fabric such as cleavage and faults can control the location and orientation of drainage channels. The channel controls and stream ordering system used in this study are discussed below. The control of streams by faulting is a well-known phenomenon and although faults in southeast Queensland are typically ancient, it was important to ascertain whether the present landscape may be controlled by active tectonism or only by geological juxtapositioning generated by very ancient fault movements. Therefore, the second hypothesis was tested.

Hypothesis 2

“The location of recent earthquakes in southeast Queensland align with geomorphological features such as scarps, mountain ranges and valleys.”

The second hypothesis aimed to identify whether earthquake epicentres show spatial alignment and if so, whether they correspond with any geological and geomorphological features (including scarps, river valleys and highland lineaments). The information was required in order to consider whether neotectonics currently plays a role in the evolution of the landscape. The findings identified that: • low-magnitude earthquakes in the region cluster in preferred areas; • there are several distinct corridors where earthquake epicentres align with geomorphological and geological structures; • additionally, the location of some earthquake epicentres occur in a linear ‘pattern’ that, although in some places correspond with ancient faults, in other places do not align with known faults or other structural geology. In conclusion from this part of the study, although the geomorphology of some regions clearly corresponds with ancient controls such as the outcrop distribution of Palaeozoic rock units and large, ancient fault systems, other areas may be influenced by some neotectonic control - a function of stress induced by the present compressive regime. Earthquakes presently occurring in the region are not likely to cause surface rupture or displacement. Nevertheless, they may generate subsurface fractures within the rock, which can influence the movement of fluids, enhance local weathering and offer sites for preferential erosion. Additionally, although the low magnitude

217 earthquakes may be releasing stress periodically along ancient faults, some earthquake activity is also related to unknown structures that do not presently have a recognised surface expression and may require further investigation to resolve their character.

Hypothesis 3

“Drainage networks across southeast Queensland show repeating, aligned and anomalous patterns that are controlled by a range of geological structures of varying age.”

The ultimate aim of testing hypothesis 3 was to assess how the regional scale drainage orientations and patterns of southeast Queensland have been influenced by the location of earthquake corridors, and the distribution of geological structure, lithostratigraphy, and igneous intrusions of various ages and origins.

The main findings identified that: • drainage patterns of the majority of large river systems in the region are characteristic of geologically controlled networks; • drainage channel location in the region is commonly controlled by faults and zones of differential weathering on regional and localised scales; • some large drainage systems in the region align with earthquake corridors and large ancient fault systems; • uplifted blocks have caused steep highland drainage proximal to the coast; • some low order streams have been diverted; • there are many incised, meandering streams both in low- and highland areas; • the ‘coastal drainage divide’, first described in this work, has a significant effect on drainage in the region, north of Brisbane; • a large valley is first described in this work (‘the Kilcoy Gap’) that displays evidence of reversed drainage; • despite the eastern Australian coast being a passive tectonic margin, drainage patterns are generally atypical of such a setting.

218 The results of paper 3 provide evidence to conclude that drainage patterns have been influenced by structural and lithological features. These have been emplaced by a long series of tectonic events since the late Palaeozoic that include folding, faulting, metamorphism and volcanism within convergent margin compressional, passive margin extensional, and intra-plate hot-spot volcanism related tectonic regimes. The presence of many incised meandering streams, both in low- and highland areas, suggests there has been relatively recent uplift. The coastal drainage divide is identified by a lack of drainage directly to the coast from the shore-parallel highlands. This feature, previously unidentified, lies to the east of both the Great Divide and the Great Escarpment. It is generally of low-relief, approximately 120 km in length, and is situated 10 to 30 km from the coast in the northeast of the study region, and has a significant influence over drainage in the region. Furthermore, some streams initially flow from the dominant highlands towards the coast, then turn away from the coast where they approach the subtle coastal drainage divide. This is discussed further, below. The divide crosses several geological units of Jurassic, Triassic and Permian ages. No known structure or geological unit corresponds to the position of this feature but various potential structural controls are hypothesized below. A major fault through the Brisbane region that was previously mapped (the Buranda Fault) but later removed from published maps, may be worthy of further investigation as its presence might explain a sharp, northeastward bend in the Brisbane River in the Ipswich area, and may additionally explain the offset of Palaeozoic rock units to the northwest and southeast of Brisbane. The Kilcoy Gap first described in paper 3, may have previously drained directly to the coast and has since experienced reverse drainage due to the mechanism that has generated the coastal drainage divide. Primarily, paper 3 is an example of how a detailed and regional-scale analysis of drainage patterns in relation to geology can identify both the main controls over a landscape, and anomalies that may provide new information towards understanding the evolution of an area’s physiography

Additional findings To assess whether fine-scale geological features control morphology, paper 1 analysed rock fabric in relation to drainage patterns at a sub-catchment scale. This

219 analysis required comparison of topologically similar streams across the network. Testing of the available stream-ordering systems identified that they each insufficiently identify topologically similar stream reaches and do not provide a satisfactory solution for comparative analysis. Therefore, a new stream ordering system was devised, presented and used for the analysis in paper 1. Further to the main findings of paper 2, the results also indicate that there may be a continuous accumulation and release of stress in some regions preferentially to others, such as near water storage reservoirs, although there was no clear indication whether the events were reservoir-induced or whether there is a bias in the dataset due to the clustering of seismographs presently situated around dam sites. As the findings may be significant for the purposes of hazard mapping, it was concluded that further geological and geophysical mapping may identify additional structural and lithological discontinuities where earthquakes cluster along zones with no obvious surface morphological expression. Additionally, it was concluded that more detailed earthquake monitoring would better constrain earthquake locations and the correlation between epicentres and known structures may be improved. Furthermore, small faults (typical of southeast Queensland) can eventually coalesce, forming larger fault traces, along which, rivers are likely to preferentially flow. This process may be slow and difficult to monitor in real-time, but does not preclude it from being a geologically derived landscape control. The work undertaken in paper 3 identified that, despite the eastern Australian coast being a passive tectonic margin, there is a lack of regional subsidence and as previously mentioned, the drainage patterns in the region are generally atypical of such a setting. Some streams run parallel to, or away from the coast, rather than flowing as short streams from the shore-parallel highlands directly to the coast as is typical for such tectonic settings in other parts of the world. In paper 3, the results also indicated that, although some river patterns correspond with and are controlled by known geological structure, not all drainage is controlled in this way. Some drainage patterns, although having linear or angular characteristics of geologically controlled systems, do not correspond with known geological structure, suggesting that they may be controlled by unmapped or concealed structures. Papers 1, 2 and 3 provide multiple forms of evidence at varying scales that suggest there are strong endogenic controls on landscape evolution in the study region. Examples of the multiplicity of geological controls on surface processes can

220 be found within small areas. For example, in the Laceys Creek catchment (paper 1), the phyllitic texture of the rock has a strong influence on the orientation of streams of a certain magnitude. Flow tends to be directed preferentially along the cleavage, later incising deeper into the rock and retaining the same orientation as the rock foliation. In a similar way, fluid preferentially flows along faults, even where the fault produces no initial landscape expression. Fluid flow along such a plane, even beneath a thick regolith can gradually cause weathering and widening, and this has been identified in Laceys Creek (paper 1), and the North Pine and upper Brisbane rivers (paper 3). A corridor of earthquakes occurs along the orientation of the Brisbane River (paper 2), where rock weakening would be expected. These events may lead to fractures, joints, and possibly small faults, causing the rocks to be less resistant to water flow and leading to further erosion and widening by the river along these planar orientations. Paper 3 more fully explored the relationships between drainage patterns and the location of geological units, faults, structure, highlands and valleys. At both coarse and fine scales, many parallel and angular stream patterns exist in southeast Queensland. Although surface water naturally flows away from areas of high relief, specific flow patterns also show evidence of geological control. Radial drainage, for example, is well defined from the remnant crests of the Mount Warning (Tweed) shield volcano, and from the peaks of Mount Perseverance and Mount Glorious. Centripetal drainage has developed from steep highlands into the eroded basin underpinned by the Samford Granodiorite that now forms the Samford Valley. Paper 3 cites many other examples of geologically controlled stream patterns. However, the geological control of streams that flow inland from the coastal drainage divide described in paper 3 is not clear. Some of the streams close to the drainage divide are strongly affected in that they show evidence of diverted drainage. These are low order and therefore, very probably young, so this work concludes that the control of the elevated topography may be relatively recent and is discussed at greater length below. Paper 3 investigated geological control of topography and river patterns in a broader sense, identifying strong correlations between stream and highland distributions and known structure, and more importantly identified some anomalies where geological control is probable but not yet conclusive. This is also discussed at greater length below.

221

Implications for future research The new stream-ordering system that was presented in paper 1 was found to be most beneficial for analysing the correlation between channels and small-scale rock-fabric structures; in itself, the new ordering system is a substantial contribution to this subject. Use of alternative methods in future research will depend on whether topology is of importance to the study. Other methods typically designate a stream order based on a counting system that may not reflect the true topological position of the stream in the network. Furthermore, the stream order designated to a channel reach might not be comparable to other streams in a similar position in the network. The new stream ordering method introduced in paper 1 is better able to compare topologically equivalent segments of a drainage system. This will benefit any future study where comparison of similar order streams is being made across a network, whether this is a comparison of physical properties, orientation or river stage. For example, if evolution of 3rd, 4th and 5th order streams were being studied to identify stream-bed gravel resources or expected flood magnitude of an area, it would be necessary to ensure the orders are equivalent across the network. Otherwise, variability of processes that operate within each stream order would be excessive. Previous stream ordering systems that lack topological similarity would therefore not be practical for this purpose. Comparison of network characteristics becomes more meaningful by using the new ordering system. For example, for a study of flow dynamics in streams during a flood cycle, it would be unlikely that all streams in a catchment could be manually measured. After assigning stream orders across the whole network using the new ordering system, just a small number of the selected order streams could then be monitored, in the knowledge that all other streams of that same order are topologically similar. The similarity allows the results to be used for forecasting or prediction purposes. First ensuring that stream orders are topologically equivalent would enable fewer streams to be monitored. If the alternative (previously established) ordering systems are used, the results may be relatively meaningless. The research presented in paper 1 is also unique as the study was conducted in a sub- catchment on metamorphic rock with substantial ground-cover and little outcrop, in addition to the study incorporating an entire sub-catchment. Previous studies of this type have focussed on granites, in areas with extensive outcrop and covering only portions of a drainage catchment.

222 The results from this study highlight the importance of identifying the orientation of fault and fracture planes, irrespective of earthquake magnitudes typical of an area. This is particularly important because even small faults or joints provide sites for preferential fluid flow that may eventually cause changes in the morphology of the landscape. In locations where small seismic events cluster, this may indicate that rocks are accumulating more stress, or that the rocks there are less able to retain stress and this should also be considered in similarly focussed, future research.

Other observations and general discussion Several scarps in southeast Queensland closely parallel the location of drainage segments. However, some drainage channels, such as the North Pine River and Brisbane River, are aligned with faults, but not situated close to an associated scarp, as the stream has eroded and migrated away from the location of the fault. Differential weathering has also accentuated the scarps where, for example in the North Pine catchment, the fault divides the more resistant Rocksberg Greenstone and the less resistant Bunya Phyllite, the latter of which has eroded more readily. Therefore, although the fault may have been a primary control of the initial location of the river, the river has then exploited the softer rock and differential erosion has widened and deepened the location of flow through this area. Scarps spaced radially around Mount Warning and parallel scarps along the top of the Main Range are closely situated to stream channels and indicate that there has been down-cutting into the hard volcanic rocks with little migration of stream channels. Scarps bounding the edges of the Kilcoy Gap however, are not closely associated with drainage channels or faulting and potentially represent the edges of an ancient drainage corridor (alluvial plain), although to the north, the Maleny Basalt may have provided some protection against erosion in this area. Multiple small, parallel scarps in the northwestern part of the region, are parallel to the upper Brisbane River and the Great Moreton Fault system, and may be fault or fracture controlled. Scarps in the south-central region align in places with the upper Logan River, where differential erosion has caused scarps along the interface between the Jurassic sediments and the Cenozoic basalts. The Como and Glasshouse scarps do not coincide with drainage or fault orientation and represent coastal scarp retreat that occurred at different times, which is discussed further, below.

223 In their review of the Great Escarpment, Ollier and Stevens (1989) stated that complex drainage occurs between the Great Divide and the Great Escarpment and simple drainage occurs to the east of the escarpment. The coastal drainage divide lies to the east of the Great Escarpment. Therefore, according to this model, all drainage to either side of the coastal drainage divide should be simple. Furthermore, as the Great Escarpment is situated to the west of the Brisbane River for the majority of its length, this would suggest that the latter should have a simple drainage pattern. However, ‘simple’ drainage to the east of the Great Escarpment in southeast Queensland, from highlands to shore, is not commonplace. This section of the Great Escarpment is referred to by Ollier and Stevens (1989) as section IX (or the Bellthorpe-Wilsons Peak section). In the very north of the Brisbane River catchment, the river is situated between section IX ending to the river’s west and VIII - Biggenden-Goomeri-Bellthorpe (Ollier and Stevens, 1989) commencing to its east. This alone would suggest some complications should be expected in the drainage patterns in this area. It is evident that drainage patterns in the area flanked by both the Great Escarpment and the coastal drainage divide are very complex. This allows for two possible scenarios: the first may be that the Great Escarpment in this region has been mis-located on maps, and should correspond instead to the position of the coastal drainage divide; or the second, more simple explanation, may be that the coastal drainage divide is a supplementary feature with a separate origin to the Great Escarpment but acting in the same way, with complex drainage to its west and simple drainage to its east. If the former is not the case, and the coastal drainage divide is simply acting in the same way as the Great Escarpment, its origin is probably similarly related to back-tilting generating complex drainage to its west, although the tilting process requires a driving mechanism that is not yet clearly defined. Interpretation of the coastal drainage divide as a simple wave-cut escarpment at a time of higher sea level would not cause back-tilting, so a more complex mechanism for its formation is probable. The coastal drainage divide could be a small, secondary escarpment possibly formed by a second pulse of thermal doming and rifting or localised doming. Alternatively, the back tilting may potentially be caused by a deep seated, unmapped fault, blind thrust or gentle flexure of the crust caused by the current or recent compressional tectonic regime. As there is no evidence of the latter proposed mechanism, thermal doming is more probable. This may be connected with flexure and uplift of the crust related to the mid-Cenozoic emplacement of igneous

224 bodies and hot-spot activity or later underplating. Volcanics related to the mid- Cenozoic events are situated close to both the coastal drainage divide, in the north (Mount Boulder and west of Eumundi) and the Maleny plateau (dated as approximately 25.2 m.y.) in the south. The Glass House Mountains situated on the coastal plain are postulated as having a genetic link with the Maleny basalts (Ewart et al., 1980). The Glass House Mountains erupted approximately 25.4 Ma at a time when the coastal plain was approximately 200 m above the present level prior to erosion. At the time of emplacement, thermal doming would have been central to the location of the volcanic activity, given that it was hot-spot-generated, this centre may have moved over time, generating a roughly meridional elongate domal structure. Coastward erosion, development of a scarp, and landward retreat of the drainage network would have also caused the location of the coastal drainage divide to have shifted to its present location. The Como and Glasshouse scarps that were previously identified by other workers coincide closely, in places within 10s of metres, with the coastal drainage divide but the scarps are situated slightly to the east of the drainage divide proper. In essence, the crests of the coastal scarps appear to be spatially related to the coastal drainage divide but are not fully coincident with it. Scarp retreat through essentially flat to low-angle strata may not yet have had time to “catch” the drainage divide itself. Clearly the origin of the Como and Glasshouse coastal scarps requires further investigation but this study has provided the first viable hypothesis for the origin of these features. The coastal drainage divide has previously not been described as a single continuous feature; only by analysis of drainage channel locations was the full extent of this geomorphological feature identified. A further matter for consideration is that the Great Divide and Great Escarpment are both associated with typical models of a passive margin tectonic setting. The model dictates that the escarpment and divide in such a setting would typically migrate away from the coast line. However, Forsyth and Nott (2003) have suggested that Australia’s eastern highlands do not conform to this model, as divide migration and subsequent stream diversions have not occurred in all places. Furthermore, it is suggested that the continental divide already existed in some areas during the Early to Middle Jurassic, prior to rifting in the Late Cretaceous (Nott and Horton, 2000). From the discussion and research presented here, it is evident that drainage in southeast Queensland is atypical of that expected in a simple passive- margin setting. Forsyth and Nott (2003) cited examples of streams on the Cape York

225 Peninsula, northern Queensland, that previously had been claimed by Pain et al. (1998) as having been ‘captured’ due to formation and migration of the Great Divide. Pain et al. suggested that the Barron and Pasco Rivers, for example, were linked with streams that flow west and are situated west of the Great Divide. However, from drilling surveys and additional field evidence, Forsyth and Nott (2003) confirmed that the Barron and Pascoe Rivers are not related to those that are west of the divide. Furthermore, they concluded that the Barron and Pascoe Rivers are structurally controlled by faults, forcing each to make sharp turns before reaching the coast. This drainage pattern is very similar to that of the Brisbane River. These two areas have similar drainage patterns that starkly differ to the region in between – central Queensland coastal region. In the central region, the Great Divide is situated far inland and the Great Escarpment is close to the coast and approximately shore- parallel. The drainage patterns are very different in that the central Queensland region is dominated by two very large catchments, the Burdekin and Fitzroy, as discussed by Jones (2006 p. 440), who identified that ‘scarp retreat (mostly through relatively soft and gently folded Permian to Jurassic strata) is the most obvious erosional mechanism affecting central Queensland catchments’. The central Queensland catchments act in a very different way to those lying to the north and south, where the Great Divide is closely coincident with the Great Escarpment. This suggests that the location of the Great Divide is a stronger reflection of coastal drainage development than the location of the Great Escarpment. In both Cape York and southeast Queensland, where the escarpment is close to the coast, it is clear that faulting strongly influences drainage. As a comparison, the work of Bezerra et al. (2008) showed that the passive margin of northeastern Brazil, which is now under compressive stress, has experienced subsidence caused by the reactivation of tilted fault blocks. As already stated, a consequence of rifting is back-tilting behind the great escarpment caused by regional doming. However, rifting may also be accommodated by listric and strike slip faulting. In southeast Queensland, it is clear that the Great Moreton Fault system was already an active feature prior to rifting. This fault may have been reactivated during Cretaceous-Cenozoic rifting, and a new perpendicular strike-slip fault (the hypothesized Buranda Fault) may have offset the South D’Aguilar and Beenleigh Blocks. Drainage that was previously captured along the Great Moreton Fault system, now the upper Brisbane River may have been diverted along the transfer

226 fault towards the ocean, through the disjunction between the two blocks of Palaeozoic rocks. Unlike the northeastern Brazil region, southeast Queensland does not appear to be subsiding, perhaps due to magmatic underplating, as discussed further, below. The modern compressional stress regime may be reactivating faults in the region, causing some further tilting and drainage diversion. Jones (2006) showed that fluctuations in sea level can influence sedimentation and drainage pattern development along the Queensland coast. Unlike the rest of the eastern Queensland shelf edge, the central Queensland coast where the Burdekin and Fitzroy Rivers discharge, has a broad shelf whose seaward margin is not coast-parallel and here, large quantities of continental sediment have been deposited from the two rivers. In particular, during late Miocene low sea-levels, continental sediments were transported far out onto the shelf and are now located at 527 m below sea level in accumulations up to 177 m thick (Jones, 2006). At this time, the coastline would have been several hundred kilometres offshore from its present location and a lower base level would have encouraged more stream channel down-cutting and erosion and if head-ward erosion took place, this would have been coupled with enlargement of the catchments on-shore. Although fluvial sediments that are delivered to the sea are commonly deposited close to the shoreline, changes in the location of shoreline as sea-level changed, would have distributed the sediments across the broad shelf. In contrast, southeast Queensland has a narrow continental shelf and the Great Divide is closer to the present-day coast line. Here, the growth of catchments may have been partly limited by the location of the Great Divide. In southeast Queensland, a fall in sea-level would have resulted in less terrestrial erosion than occurred further north, and instead there would have been more erosion of the coastal plain and canyoning of the narrow shelf and slope. Furthermore, the presence of the coastal drainage divide in southeast Queensland highlights that in this area, where highland-parallel coastal drainage is not directly from the highlands to the nearby shore, sedimentation will not be uniform along the coastline. Rather, major depocentres are more likely to be focussed in isolated areas along the shelf. Drainage patterns and sedimentation patterns at passive margins generally, are therefore, likely to be influenced by localised ‘anomalous’ drainage, where drainage is not simply from the coastward side of the great escarpment to the shore and is clearly not ‘simple’ as previously suggested by several other authors. Boyd et al. (2004) identified that the southeast Queensland continental shelf is both

227 narrow and severely sediment deficient in comparison with some other passive margins such as the eastern margin of North America. They concluded that the region has subsided very little since its formation but did not suggest why. The continued buoyancy of the crust in southeast Queensland, perhaps due to crustal underplating (discussed further, below), together with complex, shore-parallel drainage and stream diversion along the coastal drainage divide, may have led to a lack of sediment delivery to some parts of the coast and additionally, sediments may be remaining onshore and stored as sediment veneers on the uplifted highlands and plains or in small Cenozoic basins (e.g., the Pomona, Petrie, Booval and Oxley basins). The Tweed Shield Volcano (Mount Warning), of Miocene-Oligocene age, is situated in the far southeast corner of the region, in alignment with the Great Moreton fault. K-Ar ages of the volcano range from approximately 20.5-22.3 m.y. (Jager 1977). The Tweed Shield Volcano is the largest volcano influencing southeast Queensland. Its eruption would have had major affects on the geomorphology and potentially on the structure of the region. Ewart et al. (1980) concluded that Miocene-Oligocene volcanism of southeast Queensland is the product of magma fractionation and intrusion at the crust-mantle boundary. The process involved ‘underplating’ in which large sill-like intrusions caused thickening of the crust by up to 20 km under the volcanic regions and this has led to a reduction or prevention of subsidence. A similar conclusion was also made by Wellman (1979a; 1979b; 1994) who based studies upon magnetic and gravity anomalies and physiography. Having studied picritic parental magmas, Cox (1980) developed a similar model for lower crustal fractionation and thickening caused by intrusion of sill-like bodies, in the Karoo province of southern Africa and the Parana lavas of South America. Up- doming of the crust during the formation of the Tweed shield volcano is likely to have caused adjustments to the drainage in the region as physiography changed. Drainage from the Tweed shield volcano is radial and the western-most streams in southern Queensland join the Logan River system, which flows around the outskirts of the volcano before reaching the ocean. The eruption and associated tectonics of the volcano may be linked to the location of faulting in the region, where natural planes of weakness provided conduits through which lava could erupt. As stated previously, the volcano is positioned in alignment with a southern extension of the fault system that runs northwest-southeast along the boundary between the Esk

228 Trough and the D’Aguilar Blocks and this plane of weakness may have provided a focus for the eruption of the volcano. O’Brien et al. (1994) described this fault (after Cranfield et al., 1976) as being continuous in a northwest-southeast orientation north of the Ipswich Basin (Eastern Border Fault), and in a north-south orientation south of the basin (West Ipswich Fault), therefore being distant from, and unconnected to the volcano. However, on the Moreton Geology map compiled by the Geological Survey of Queensland (Whitaker and Green, 1980), this is not identified as a continuous fault and there may be a concealed and unmapped fault continuing as far as the shield volcano. Cranfield et al. (1976) and O’Brien et al. (1994) described the Great Moreton Fault System as a network of intersecting and braided faults. This may suggest that, of the multiple associated fault planes in this system, many have yet to be identified. Lavas from the Tweed shield volcano may have concealed any evidence of major bedrock faulting around the volcano itself and would account for the lack of structural detail currently mapped. Hot-spot activity in southeast Queensland is associated with continued underplating, crustal thickening and later eruptions of the Main Range and Mt Warning to the west and south of the Glass House Mountains. Crustal doming associated with hot-spot activity could also account for the back-tilting and drainage anomalies along the coastal drainage divide, but it does not account for anomalously high uplift in the D’Aguilar blocks. Regional uplift or doming that caused tilting and uplift and present-day landward drainage, of the D’Aguilar blocks could be due to a combination of crustal doming and the crust slowly uplifting in order to adjust to isostatic equilibrium, where the thickened, fractionated crust underlies this region. The recent work of Kirby et al. (2008) stated that indices, such as anomalously steep channel profiles and evidence from the topography of long-term uplift were, retrospectively, strong evidence that Sichuan Province of China was seismically active. A surprisingly large earthquake of magnitude 7.9 struck the Sichuan Province in May 2008, in a region where, although previously known to be tectonically active, events were regarded as relatively infrequent and an earthquake of such magnitude was not anticipated. Kirby et al. (2008) stated that indices such as steepened channels and evidence of long-term uplift may be suitable for identifying potential activity on blind or hidden faults or in regions where geodetic networks indicate that little strain is accumulating in the upper crust. Furthermore, beneath the Sichuan Province crustal thickening has occurred, driven by flow and deformation in

229 the lower crust, which may cause seismic activity despite no evidence of shortening in the upper crust. Following the type of evidence and key indicators used by Kirby et al. (2008) in the Sichuan Province, a potentially similar seismic geohazard may exist in the southeast Queensland region. This is evidenced by the anomalous drainage patterns amongst low order streams near the coastal drainage divide, incised meanders in both low and high order streams, numerous terraces and steep headwaters in some catchments draining the D’Aguilar blocks. Palaeodrainage evidence also suggests continued and large-scale uplift of the D’Aguilar blocks. Also similar to the Sichuan Province, crustal thickening has occurred in the southeast Queensland region (for example Ewart et al., 1980) where flow at the crust-mantle boundary may have underplated the crust.

Implications for evaluating the evolution of the landscape Simple geomorphological models are unlikely to fit most landscapes that, like southeast Queensland, have been the subject of a complex, geological history. The results of this study show that where geological control is prevalent, the influence upon the landscape is commonly long-term despite overprinting by numerous subsequent geological events and broad changes in exogenic influences. Although surface features (e.g. surficial sediments) may conceal a geological control such as a fault, the affect of the fault on the landscape is generally still detectable. Surface processes, such as differential weathering, may also enhance evidence of geological control. This work has shown that a multi-criteria approach must be used for landscape evolution analysis. Considering the evidence presented in this thesis, and also discussed further above regarding passive margin settings, it may be necessary to modify the ‘classical’ model of drainage in a passive margin tectonic setting. More work is required in similar geological settings to clarify (1) the range of deformation styles, (2) the character of modern stress regimes, and (3) the drainage architecture, to clarify the range of variation in drainage development along passive continental margins.

Evidence of past geological events in the present landscape The geology of southeast Queensland continues to display some fabric from the Late Carboniferous. During the Late Carboniferous, north-northeast – south-southwest

230 compression caused deformation and uplift of an accretionary prism and the onset of the uplift of the New England Fold Belt. Fabric evident in the present landscape from this time includes the low grade metamorphism seen in the Rocksberg Greenstone, Neranleigh-Fernvale beds and the Bunya Phyllite for example. As stated from paper 1, this fabric controls the orientation of some streams. The geomorphology is also controlled by the structural fabric inherited from late Palaeozoic-early Mesozoic time. At this time, compression in an east-west orientation caused mainly northwest- southeast thrust and strike-slip faulting of the accretionary prism material, during the Hunter Bowen orogenic episode. The structural discontinuities are clearly evident in the present landscape as seen, for example, in the North and South D’Aguilar blocks and the Beenleigh Block, although its ‘offset’ to the South D’Aguilar Block is not explained by this event and as previously discussed, may have been caused by a transfer fault during later rifting. Although further deformation continued through the Middle Triassic, with folding occurring in a north-northwest orientation and igneous emplacements such as the Mount Samson Granodiorite and the Samford Granodiorite, these events did not alter the previously emplaced blocks and units significantly. Extension during the Late Triassic – Early Cretaceous led to the formation of basins such as the Clarence-Moreton, Maryborough and Nambour basins and although these are significant in size, they do not sufficiently conceal or alter the strong north-northwest – south-southeast trending units emplaced during the Palaeozoic to early Mesozoic. During the late Mesozoic, north-northeast – south- southwest extension led to the breakup of the eastern Australian margin, which led to the opening of the Tasman and Coral seas. Strong and obvious morphological evidence of this event is deficient in the region at present; nevertheless, it is possible that during this event, the South D’Aguilar Block and Beenleigh Block and part of the Great Escarpment were off-set. Small basins subsided and filled during the Paleogene in association with local volcanism and possible uplift of adjacent highland blocks. Volcanoes that erupted during the mid-Cenozoic, such as the Tweed Shield Volcano, Main Range Volcanics, the Glass House Mountains and the Maleny basalts, have made an obvious impression on the landscape. The Main Range Volcanics for example, are potentially protecting the highlands in that region from erosion; similarly so the Maleny basalts at the Blackall Range. Nevertheless, the strong northwest-southeast trend of faults and blocks emplaced during the Palaeozoic to early Mesozoic remains a dominant element of the landscape. The modern

231 regional stress field is compressional in a northeast-southwest orientation. No large scale surface ruptures, earthquakes or deformation have been observed in recent times and there are no active volcanoes in the region. However, from the results of paper 2, it is evident that the stress building from the present compressional regime is being released as fairly frequent, low magnitude events, and this is potentially occurring along some ancient faults such as the northwest-southeast trending Great Moreton Fault system. Although the strong northwest-southeast geological trend dominates the orientation of rock units in the area, and some rivers such as the Brisbane and North Pine rivers appear to be strongly controlled by this trend, a variety of the other, aforementioned regimes, presently control other drainage patterns in the region. For example, the volcanic and igneous activity has led to centripetal drainage (for example, on the Samford Granodiorite), radial drainage (for example, around Mount Warning and Mount Glorious) and parallel drainage (for example, flanking the Main Range Volcanics). The various phases of compression, folding and metamorphism have led to parallel and angular drainage patterns, such as that seen in the North Pine River catchment and the streams that feed the upper Brisbane River. However, some drainage such as the anomalous drainage described in paper 3, does not clearly fit with any specific regime described above. Although the north-trending Mary River may be caused by an antecedent drainage regime, the control of some small streams that first flow towards the coastal drainage divide and then turn away, cannot be assigned to a known regime or event and their potential control has been discussed above. In addition to the volcanics that crop out onshore, some exist on the continental shelf, close to and a little way off-shore (Boyd et al., 2004). Their presence, combined with the change in orientation of the coastline, from north-south to northwest-southeast, causes northward longshore drift to diverge from the coast and has created major sand islands in this area (Boyd et al., 2004). Features such as the shoreline, major geological units and large faults (e.g., the Great Moreton and North Pine faults), and the boundaries of geological units discussed above, are all typically orientated northwest-southeast and despite there being several past tectonic events reflected in the landscape of southeast Queensland, this dominant trend has remained since Palaeozoic-early Mesozoic times. In paper 3, evidence of inverted topography and remnant fluvial deposits were presented, as seen in the Kilcoy Gap. Such deposits can provide hard evidence of the location and orientation of palaeochannels. Additionally, it provides

232 information on landscape evolution, including the prevailing conditions during and after deposition. Inverted topography occurs when, sediment initially deposited in a depression, consolidates and becomes more resistant to erosion than the surrounding rock. Once the surrounding rock is eroded, the sediments are left as remnants capping high points in the landscape. Inverted topography may occur as a result of armouring, covering or hardening of the stream fill, the former two, through deposits of large and resistant boulders or basalt flows for example, the latter through diagenetic processes forming ferricrete, silcrete or calcrete for example. Any of these processes may provide resistance to weathering relative to the surrounding rocks that may erode leading to elevation of the previously low-lying stream channel deposits. Inversion of relief is an important contributor to landscape development although, as stated by Pain and Ollier (1995), there is often a reluctance to accept this point of view. Nevertheless, numerous examples exist globally. A particularly large and well- defined example of this can be seen in the southern Pyrenees, Spain, where the Sis conglomerate (Vincent, 2001) represents an enormous palaeovalley fill that now stands elevated by several hundred metres above the older, exposed, underlying sediments. The presence of inverted topography is particularly important for providing evidence of variations in tectonic, climatic and pluvial conditions over time. Inversion of relief has also proven to be a significant method for landscape analysis of Mars (Pain et al., 2007). Several geomorphic features on the Martian surface have been identified as representing inverted relief, including some inverted impact craters. The study of Mars showed that several landscape evolution processes have occurred there, including diagenesis, weathering and the movement of liquid both at the surface and in shallow aquifers. Inverted topography in Australia has been identified in many locations. For example, silcrete derived inversion is evident in the Cape York Peninsular (Pain and Ollier, 1995) and the Kirup Conglomerate, Western Australia (Pain and Ollier, 1995), both of which identify changes in groundwater and overland flow conditions probably linked with climatic changes in the region. The preservation of the inverted Haunted Hill Gravels in southern Victoria has been valuable in identifying the provenance of sediments and pulses of tectonic activity of the area (Bremar, 2002; Kapostasy, 2002). Preservation of palaeochannel deposits may also conserve concentrations of economic deposits such as gold or sapphires within the channel-fill. In Western Australia, ferricrete and nodular iron ore of the Robe River deposit extends tens of kilometres along palaeovalleys (for example

233 McLeod, 1966; Pain and Ollier, 1995). In southeast Queensland, the Oakdale Sandstones (Derrington et al., 1959; Cribb et al., 1960) of Cenozoic age, which consist of sand, shale, conglomerate and coal seams, are capped by rhyolites and basalts and are thought to be of Pliocene age. Similarly the channel fill deposit described in paper 3 that is seen in the Kilcoy Gap is situated close to the Maleny basalts and may have been covered with a thin veneer of basalt providing initial protection from erosion. Basalt from the eruption of the Tweed Shield volcano flowed down pre-existing stream channels and differential weathering has now led to preserved basalt-capped river courses as headlands and ridges such as at Burleigh Heads. The preservation of palaeochannel deposits helps to pinpoint the positions of ancient drainage systems and to evaluate the relative incision and altitude of the landscape in the past. In the case of the deposits preserved in the Kilcoy Gap, the very coarse fluvial sediments indicate that this was the site of a major drainage corridor of probable Paleogene age.

National and international significance Although surface processes, regional tectonic processes and anthropogenic influence may be cited as controlling elements of terrain morphology, the findings in paper 1 have confirmed that the landscape may be controlled at a fine scale and by multiple endogenic controls. This is specifically relevant to understanding landscape evolution in other metamorphic regions both within Australia and internationally. Additionally, it highlights the significance of controls on modern drainage by the foliation and structural discontinuities characterizing ancient accretionary terrains. Understanding these influences has relevance to landscape interpretation in other parts of the eastern Australian coast such as the Narooma accretionary complex in New South Wales and accretionary regions such as western North America and southern Europe. The channel ordering system presented in paper 1 may be applied to any drainage system at any scale. In other similar settings, it would be possible to repeat the methods used in papers 1 and 3 in order to identify the level of geological control that the underlying rocks have on the orientation of the streams. This can be used to establish a base-line before determining if there is any anthropogenic influence on the landscape. More simply though, it can assist in understanding whether rock fabric has a role in drainage orientation or whether the drainage is not strongly influenced by geology but instead by surface processes.

234 Many areas of the world lack well-constrained or long-standing earthquake monitoring systems and paper 2 describes and tests a simple and effective method for a preliminary assessment of the relationship between earthquakes, the structural geology of an area, and its landscape. The findings of paper 2 emphasise the need for a more closely spaced earthquake monitoring system and for more continuous data collection for similar studies to be effectively carried out. However, the paper also identifies that it is possible to associate shallow earthquake events with known fault systems or geomorphological features such as scarps, highland margins or hill-crest alignments, to assess whether structural features are currently active and influencing landscape form. As there is some evidence from extremely well monitored earthquake-prone regions that low magnitude earthquakes may cause surface rupture, it is important that other regions experiencing similar magnitude earthquakes are more closely monitored to examine the cause and effect of activity in the region. The findings of paper 3 provide a regional assessment of how multiple geological influences of ancient and more recent origin have affected the modern form of the landscape. The results will supplement studies of other passive continental margins, such as the coastline surrounding the Red Sea, the east coast of America and the west coast of Europe, to determine the consistency of drainage patterns and evaluate the dominant mechanisms involved in landscape evolution in such settings. Furthermore, as southeast Queensland also contains elements of compressional and accretionary terrains, the patterns identified herein may be particularly applicable to passive margins developed upon older accretionary terrains such as the northern Gulf Coast Basin set against the Palaeozoic rocks of the Ouachita Mountains of the United States. The rifted margin of southern Africa set against the deformed Palaeozoic rocks of the Cape Fold Belt represents another region where similar drainage patterns might be expected.

235 Table 1 Evolutionary model of southeast Queensland and the relationship with knowledge identified and discussed in this thesis

236 Table1- continued

237

Evolutionary model Based on the three components of the study, the geological events that are the controlling factors in the landscape evolution of southeast Queensland are summarised in Table 1. This table also provides a relative time-line for these events and highlights the importance of the multiple processes that have taken place and emphasises why the morphology of southeast Queensland is not typical of a passive margin setting.

Précis of main findings In summary, the main findings of this study include the following: • The study has integrated a range of datasets using GIS at a scale and extent that has previously not been achieved. • An entirely new stream-ordering system has been presented. • The earthquake analysis indicates that low-magnitude earthquakes in the region cluster in preferred areas, some of which align in ‘corridors’ that correspond with known structure and physiographic features, such as highlands, scarps and valleys. • Some epicentres occur in linear patterns that do not align with known faults; some of these align with valleys and scarps. • Low-magnitude earthquakes presently occurring in the region are not likely to cause surface rupture or displacement. • Some large drainage systems in the region align with earthquake corridors and large ancient fault systems suggesting that the ancient faults where river systems have developed may also host recent earthquakes. • The earthquake analysis is the first of its kind to have been performed for this region in more than 15 years. • Low order streams remain dendritic until they have down-cut to reach bedrock. The orientation of cleavage and faults can control the location and orientation of drainage channels after down-cutting has been achieved. • The majority of large river systems in the region display drainage patterns that are characteristic of geologically controlled networks; these patterns are commonly controlled by faults or by zones of differential weathering.

238 • Steep, highland drainage close to coastal zones has been caused by uplifted blocks. • Some low-order streams have been diverted suggesting recent structural control. • Incised meandering streams of both high- and low-orders are common, indicating relatively recent uplift has taken place. • First described in this work, the ‘coastal drainage divide’ has a significant effect on drainage in the region, north of Brisbane, as it causes a distinct lack of drainage directly to the coast from the shore-parallel highlands. • Also first described in this work ‘the Kilcoy Gap’ displays evidence of reversed drainage. • Despite the eastern Australian coast being a passive tectonic margin, drainage patterns are generally atypical of such a setting as they display evidence of past events including convergent margin compressional, passive margin extensional, and intra-plate hot-spot volcanism related tectonic regimes.

Future work The remote-sensing and analytical methods used in this study can be applied to any other location, where similar datasets are available. It would be valuable to carry out similar studies in other passive margin and convergent margin settings in order to clarify the diversity and commonalities of drainage patterns developed in such settings globally. The stream numbering method presented in paper 1, may be used to identify the level of control geology has over other catchments although further work could take into account the different phases of folding and faulting. More detailed geological mapping of southeast Queensland may identify other faults and structures that might correlate with, or better constrain, earthquake corridors described in paper 2; similarly, geological mapping may provide field evidence for the existence of the previously described Buranda Fault through the Brisbane Gap, suggested in this work as having been caused during rifting. Future work may also incorporate further tests of the hypotheses posed in paper 3, regarding the factors controlling the ‘coastal drainage divide’. Such tests may require the use of geophysical methods, provenance studies, bore-hole logging and

239 detailed geological mapping. These studies might also provide additional evidence for the eastward-flowing palaeochannel situated within the Kilcoy Gap. Other future work may include repeating the methods presented here in different areas to test whether geology provides the dominant controls on the landscape or whether exogenic processes are dominant. In particular, the methods presented in paper 1 can provide a useful assessment of which order of streams correspond with various rock fabrics, and may be a guide to the extent to which control is endogenic or exogenic in a small catchment or on multiple or complex combinations of rock types. The methods used in paper 2 can be used to identify whether recent earthquakes are occurring close to, or in association with, known geological structures, although a larger spatio-temporal database from a more extensive monitoring system would be beneficial. It would be important to investigate the potential seismic geohazards in the region, by more extensive earthquake monitoring, detailed examination of gravity anomalies, analysis of upper crustal strain rates and more precise confirmation of the status of the lower crust. The methods used in paper 3 are suitable for remotely studying large areas, particularly inaccessible regions. In isolation, the methods can each provide a limited measure of the geological controls exerted over the landscape, but used together they provide multiple indices to delineate the degree of geological influence and reduce uncertainty in interpretations.

References

Bezerra, F.H.R., Brito Neaves, B.B., Corrêa, A.C.B., Barreto, A.M.F. and Suguio, K., 2008. Late Pleistocene tectonic-geomorphological development within a passive margin - The Cariatá trough, northeast Brazil. Geomorphology, 97(3- 4): 555-582. Boyd, R., Ruming, K. and Roberts, J.J., 2004. Geomorphology and surficial sediments of the southeast Australian continental margin. Australian Journal of Earth Sciences, 51: 743-764. Bremar, K.A., 2002. Reconstructing the paleoenvironment and source of the Haunted Hill Formation (Pliocene-Pleistocene, South Gippsland, Victoria, Australia), 16th Annual Keck Research Symposium of Geology, Carleton College, Minnesota, USA.

240 Campbell, L.M., Holcombe, R.J. and Fielding, C.R., 1999. The Esk Basin - a Triassic foreland basin within the northern New England Orogen. In: P.G. Flood (Editor), Regional Geology, Tectonics and metallogenesis, New England Orogen, NEO '99. Dept of Earth Sciences, University of New England, Armidale, pp. 275-284. Cox, K.G., 1980. A model for flood basalt vulcanism. Journal of Petrology, 21(4): 629-650. Cranfield, L.C., Schwarzbock, H. and Day, R.W., 1976. Geology of the Ipswich and Brisbane 1:250 000 Sheet Areas. Geological Survey of Queensland Report 95, Queensland Department of Mines. Cribb, H.G.S., McTaggart, N.R. and Stained, H.R.E., 1960. Sediments east of the Great Divide. Journal of the Geological Society of Australia (now: Australian Journal of Earth Sciences), 7: 345-355. Derrington, S.S., Glover, J.J.E. and Morgan, K.H., 1959. New names in Queensland stratigraphy. Permian of south-eastern part of the Bowen Syncline, Central Bowen Syncline. Australian Oil and Gas Journal, 5(8): 27-35. Ewart, A., Baxter, K. and Ross, J.A., 1980. The petrology and petrogenesis of the Tertiary anorogenic mafic lavas of southern and central Queensland, Australia - possible implications for crustal thickening. Contributions to Mineralogy and Petrology, 75: 129-152. Fielding, C.R., Sliwa, R., Holcombe, R.J. and Kassan, J., 2000. A new palaeogeographic synthesis of the Bowen Basin of Central Queensland. In: J.W. Beeston (Editor), Bowen Basin Symposium 2000, Proceedings. Geological Society of Australia, Rockhampton, pp. 287-302. Forsyth, A. and Nott, J., 2003. Evolution of drainage patterns on Cape York Peninsula, northeast Queensland. Australian Journal of Earth Sciences, 50: 145-155. Jones, M.R., 2006. Cenozoic landscape evolution in central Queensland. Australian Journal of Earth Sciences, 53(3): 433-444. Kapostasy, D., 2002. Haunted Hill Gravels: deposition and neotectinic history along the southern Australian coast, 16th Annual Keck Research Symposium of Geology, Carleton College, Minnesota, USA. Kirby, E., Whipple, K.X. and Harkins, N., 2008. Topography reveals seismic hazard. Nature Geoscience, 1(8): 485-487. McLeod, W.N., 1966. The geology and iron deposits of the Hamersley Range, Western Australia. Geological Survey of Western Australia Bulletin(61): 240pp. Nott, J.F. and Horton, S., 2000. 180 Ma continental drainage divide in northeast Australia: implications for passive margin tectonics. Geology, 28(8): 763-766. O'Brien, P.E., Korsch, R.J., Wells, A.T., Sexton, M.J. and Wake-Dyster, K.D., 1994. Structure and tectonics of the Clarence-Moreton Basin. In: A.T. Wells and P.E. O'Brien (Editors), Geology and petroleum of the Clarence-Moreton Basin, New South Wales and Queensland. Australian Government Publishing Services, Canberra. Ollier, C.D. and Stevens, N.C., 1989. The Great Escarpment in Queensland. In: R.W. Le Maitre (Editor), Pathways in Geology: Essays in Honour of Edwin Sherbon Hills. Blackwell, Melbourne, Australia, pp. 140-152. Pain, C.F. and Ollier, C.D., 1995. Inversion of relief - a component of landscape evolution. Geomorphology, 12: 151-165.

241 Pain, C.F., Wilford, J.R. and Dohrenwend, J.C., 1998. Regolith of Cape York Peninsula. In: J.H. Bain and J.J. Draper (Editors), North Queensland Geology. Australian Geological Survey Organisation Bulletin 240. Pain, C.F., Clarke, J.D.A. and Thomas, M., 2007. Inversion of relief on Mars. Icarus, 190: 478-491. Ribohni, A. and Spagnolo, M., 2008. Drainage network geometry versus tectonics in the Argentera Massif (French-Italian Alps). Geomorphology, 93(3-4): 253- 266. Vincent, S.J., 2001. The Sis palaeovalley: a record of proximal fluvial sedimentation and drainage basin development in response to Pyrenean mountain building. Sedimentology, 48: 1235-1276. Wellman, P., 1979a. On the Cainozoic uplift of the southeastern Australian highland. Journal Geological Society of Australia, 26: 1-9. Wellman, P., 1979b. On the isostatic compensation of Australian topography. BMR Journal of Australian Geology and Geophysics, 4: 373-382. Wellman, P., Williams, J.W. and Maher, A.R., 1994. Interpretation of gravity and magnetic anomalies in the Clarence-Moreton Basin region. In: A.T. Wells and P.E. O'Brien (Editors), Geology and petroleum potential of the Clarence- Moreton Basin, New South Wales and Queensland: Bulletin 241. AGSO, Canberra, pp. 217-229. Whitaker, W.G. and Green, P.M., 1980. Moreton Geology 1:500 000 map. Department of Mines, Queensland.

242 APPENDICES

243

244 APPENDIX 1 APPENDIX 245

152 246 APPENDIX 2

A GIS and map-analysis deficiency

Fig. 1

Fig. 2

As with other similar geomorphological studies, this work is based on plan view areas and not surface area. A potential deficiency in spatial calculations and morphometric methods with regard to land-surface area was identified at an early stage of this study.

The deficiency is described here for completeness, as surface area calculations were not required and did not affect analyses conducted here.

An area on a map is measured in plan view. For example, in Fig. 1, area A, will have the same measured area as that labelled B: slope is not accounted for. Therefore actual surface area is rarely considered. Similarly shown in Fig. 2, area A and B in plan view would measure the same size (shown by adjoining dotted lines) although the surface area of A is clearly greater than that of B.

247

Bearing this in mind, calculations that consider area may provide differing results if plan area and surface area were compared. For an example, calculating wetness, if two areas each of 1 km2 in plan view were to receive identical amounts of rainfall, they will become wetted to different degrees depending the slope of each; a steeper slope would become less wet than a gently sloped or flat area, because the rain falling on the steeper slope would be more ‘thinly spread’ across the larger surface area than the flatter area.

Drainage density, another common geomorphological consideration, is equally affected by the area measured. For example, Strahler (1966 p. 506-7) uses four ‘1 square mile’ maps, each with different slopes and drainage densities. Although the discussion identifies potential reasons for varying densities, it does not consider the differences in each surface area.

Although surface area currently is not calculable in GIS, the slope for each pixel may be. Therefore, by taking the known size of the pixel and the angle of its slope, the surface-area of each pixel may be computable using cosine (confirmed by Dr. Micaela Preda, pers. comm.). The success of computations for large areas, such as those used in this study, is unlikely to be successful due to the amount of data involved, although average slopes may be utilised if required to decrease data intensity.

Reference

STRAHLER, A. N. 1966. The Earth Sciences, Harper International p. 681.

248 APPENDIX 3

Other software products used

Circular data

In order to present and analyse circular data graphically, data must be exported from

‘qik-orientate-345’ (Lawley 1997) into Microsoft® Office Excel. The data may then be transferred to Stereonet for Windows v1.2 (Allmendinger 2002), to produce 360° rose diagrams for visual analysis.

Images

Adobe Illustrator 10 was used to augment and enhance maps for presentation purposes.

References

ALLMENDINGER, R. 2002. Stereonet for Windows v 1.2, [email protected]. LAWLEY, R. 1997. qik-orientate-345. [email protected], Geo Trek Corp.

249

250 APPENDIX 4

Statistical analysis of planar features in Laceys Creek The orientations of bedding, cleavage, faults and fractures within the Laceys Creek catchment were analysed with respect to channel orientations in an attempt to asses their concurrence with one another. Using the statistical analysis package SPSS, correlation coefficients were produced and are presented here in Tables 1 and 2. High correlation coefficients, and very low significant figures imply a strong relationship. The correlation coefficients revealed by the analysis are not particularly high. Where very low significant figures were calculated by SPSS, the corresponding correlation coefficient is highlighted in bold.

The results suggest that there are some cases of correlation between the orientations of the network and those of the planar features, although these instances are sporadic and difficult to clearly quantify in this way. From these analyses, it is evident that there are more instances of channel correlation with bedding planes than with other planar features on the Neranleigh-Fernvale Beds (Table 1). However, there are more instances of channel correlation with cleavage planes than with other planar features on the Bunya Phyllite (Table 2).

Statistical methods commonly seek to describe a mean value, deviation from that mean, or closeness of fit between data sets. However, for the purposes of this study, it was more important to seek correlations between the various datasets, which themselves may display multiple clustering of data due to the heterogeneous nature of the landscape, causing further complications to the analysis.

The clustering of data for both channel orientations and planar features may explain the low correlation coefficient values. Analysis of the datasets needs to recognize correlations between multiple clusters rather than identify the closeness of fit for the entire dataset. Ordinary statistical procedures cannot confidently be applied to this, or any other directional data due to the very nature of circularly derived measurements where 0° = 360° (Jones, 1968). Statistical analysis of directional data is a relatively new approach but has been explored to some extent in a variety of scientific areas where orientation data naturally occur (Krieger

251 Lassen et al., 1994). Parametric orientation statistics in relation to earth sciences have been discussed by Kohlbeck and Scheidegger (1985).

Table 1

Bedding, Bedding Bedding Cleavage Channels Faults and cleavage Bedding Cleavage and and and excluded fractures and cleavage fractures fractures fractures 1st 0.090 0.046 -0.011 0.115 0.050 0.100 0.096 1-2nd 0.209 -0.060 -0.141 0.163 0.152 -0.052 0.163 1-3rd -0.260 -0.111 0.128 -0.243 -0.238 0.018 -0.243 1-4th -0.033 -0.259 -0.201 -0.033 -0.033 -0.182 -0.033 1-5th 0.048 -0.045 0.172 0.048 0.048 -0.102 0.048 1-6th -0.162 -0.099 -0.046 -0.162 -0.162 -0.164 -0.162 1-7th -0.249 -0.216 0.054 -0.249 -0.249 -0.244 -0.249 1-8th -0.046 0.200 -0.153 -0.046 -0.046 0.046 -0.046 1-9th -0.224 0.214 -0.088 -0.224 -0.224 0.066 -0.224 1-10th -0.177 0.060 -0.084 -0.177 -0.177 0.001 -0.177 1-11th -0.073 0.064 0.012 -0.073 -0.073 0.056 -0.073 1-12th 0.069 -0.050 -0.294 0.069 0.069 -0.034 0.069 1-13th -0.100 0.168 -0.186 -0.100 -0.100 0.147 -0.100 1-14th -0.030 0.170 -0.108 -0.030 -0.030 0.182 -0.030 1-15th 0.058 0.126 0.265 0.058 0.058 0.132 0.058 1-16th 0.216 -0.007 -0.099 0.216 0.216 -0.007 0.216 1-17th 0.144 0.031 -0.090 0.144 0.144 0.031 0.144 1-18th 0.037 0.106 0.372 0.037 0.037 0.106 0.037 1-19th 0.036 -0.212 -0.191 0.036 0.036 -0.212 0.036 1-20th 0.345 -0.064 -0.086 0.345 0.345 -0.064 0.345 1-21st -0.066 0.055 -0.140 -0.066 -0.066 0.055 -0.066 1-22nd 0.140 -0.118 0.001 0.140 0.140 -0.118 0.140 1-23rd -0.106 0.066 0.233 -0.106 -0.106 0.066 -0.106 1-24th 0.018 -0.098 0.379 0.018 0.018 -0.098 0.018 1-25th -0.113 0.231 0.274 -0.113 -0.113 0.231 -0.113

252 Table 2 Bedding Bedding Bedding Cleavage Channels cleavage Bedding Cleavage Fractures and and and excluded and cleavage fractures fractures fractures 1st 0.102 0.382 -0.190 0.470 0.001 0.133 0.155 1-2nd 0.185 0.539 -0.125 0.511 0.058 0.539 0.539 1-3rd 0.068 0.548 -0.229 0.476 0.076 0.548 0.548 1-4th -0.051 0.536 -0.166 0.150 0.048 0.536 0.536 1-5th -0.208 0.552 0.167 -0.227 -0.002 0.552 0.552 1-6th -0.352 0.272 0.114 -0.329 -0.013 0.272 0.272 1-7th -0.587 0.020 0.229 -0.460 -0.242 0.020 0.020 1-8th -0.078 0.100 0.410 -0.397 -0.180 0.100 0.100 1-9th -0.345 -0.059 0.360 -0.530 -0.355 -0.059 -0.059 1-10th -0.347 -0.053 0.389 -0.553 -0.358 -0.053 -0.053 1-11th -0.147 -0.275 0.348 -0.481 -0.440 -0.275 -0.275 1-12th -0.125 -0.250 0.266 -0.485 -0.480 -0.250 -0.250 1-13th -0.363 -0.181 0.214 -0.618 -0.537 -0.181 -0.181 1-14th 0.138 0.189 -0.064 -0.305 -0.256 0.189 0.189 1-15th 0.286 0.421 0.144 -0.164 -0.118 0.421 0.421

References Jones, T.A., 1968. Statistical analysis of orientation data. Journal of Sedimentary Petrology 38, 61-67. Kohlbeck, F.K., Scheidegger, A.E., 1985. The power of parametric orientation statistics in the Earth Sciences. Mitteilungen der Österreichischen Geologischen Gesellschaft 78, 251-265. Krieger Lassen, N.C., Juul Jenson, D., Conradsen, K., 1994. On the statistical analysis of orientation data. Acta Crystallographica A50, 741-748.

253

254