Modelling the Tidal and Sediment Dynamics in Darwin Harbour, Northern Territory, Australia

Home , Darwin Harbour

Li Li

School of Physical, Environmental and Mathematical Sciences

The University of New South Wales

Canberra, ACT, 2600, Australia

A thesis submitted in fulfillment of the requirements for the degree of Doctor of Philosophy

August 2013

ABSTRACT

The suspended-sediment dynamics in Darwin Harbour, Northern Territory Australia, were investigated using a combination of field measurements and numerical modelling. After analysing the harbour’s geophysical characteristics from the field data and an extensive literature review, a hydrodynamic model for the harbour was built using the Finite Volume Coastal Ocean Model (FVCOM), a model suited to coastal ocean simulation. This model was then coupled in this study to the estuarine suspended- sediment model (ESSed) of Wang (2002) to produce the FVCOM-ESSed model.

Both the hydrodynamic and sediment-transport components were calibrated using the field data on sea-surface level, current velocity and suspended-sediment concentration. The sediment-transport model focuses on suspended sediments, with improvements that allow wetting-drying processes, different bathymetry types and a variable thickness of the fine-sediment layer on the harbour bed to be included. The combined hydrodynamic and sediment model provides a reasonable simulation of the tidal and suspended-sediment dynamics in the harbour. Numerical experiments using this model were then designed to determine the effect of the mangrove areas and tidal flats in Darwin Harbour on the tides and tidal asymmetry, and subsequently, on the suspended-sediment dynamics.

This study shows that the hydrodynamics of Darwin Harbour are driven mainly by tides, with the effects of wind and rivers small. The M2 tidal component is dominant, with amplitude near Darwin City about 1.9m and phase 249 degrees. Current flow is dominated by the tides. The current, which reaches a maximum speed of about 2.0 ms-1 at the surface of Middle Arm and decreases gradually from the surface to the bottom, is -1 dominated by the M2 tidal current, with an average vertical speed of about 0.4 ms near East Arm Wharf. The water-flow patterns are in accordance with the shoreline, and rectilinear in all three arms. The energetic hydrodynamic regime, together with the availability of erodible sediment on the seabed, determines the variation in suspended- sediment concentration (SSC) in the harbour. Stronger currents induced by the larger tidal ranges during spring tides generate higher bottom SSC values than those during neap tides, as observed in the harbour in November 2012. A turbid zone appears in the

iii

outer harbor, with bottom SSC values about 0.1kgm-3 during the spring tides and 0.07kgm-3 during neap tides. The water is less turbid during neap tides than during spring tides; vertical gradients of SSC are formed in the channel during neap tides due to weaker currents. Net sediment transport is seaward in the channel and landward at the entrance to East Arm, dominated by Eulerian advection.

Mangrove areas and tidal flats play key roles in modulating the tides and water- flow dynamics of an estuary. Suspended sediments are redistributed as a result of these modulated water dynamics. Removal of the mangrove areas and tidal flats from Darwin

Harbour, a possibility due to planned harbour development, would dampen the M2 amplitude because of decreased shoaling effects in the inner harbour, but would generate a greater M4 amplitude. Removal of the mangroves and tidal flats would also lessen the tidal choking effect, thereby, reducing the current speed; as a result, the water would be less turbid in the harbour, for example, the bottom SSC values in East Arm would be reduced by about 70% during spring tides.

Mangrove areas and tidal flats affect tidal asymmetry via their influence on the amplitudes and phases of the tides, and therefore affect net sediment transport. In Darwin Harbour, these areas significantly reduce tidal asymmetry: for example, the tidal elevation skewness would increase by about 120% in Middle Arm if the mangrove areas and tidal flats were removed. Due to the increased flood dominance, the tidal pumping effect would overtake the Eulerian residual to dominate sediment transport both in the channel and at the entrance to East Arm. This would reverse the transport of suspended sediment in the channel to landward. The landward net sediment flux would be decreased at the entrance to East Arm as a result of reduced currents in the arms because of the reduction in the tidal choking effect, if the mangrove areas and tidal flats were removed.

ORIGINALITY STATEMENT

Li Li

August 2013

AUTHENTICITY STATEMENT

vii

viii

ACKNOWLEDGEMENTS

This study was carried out as a PhD student of Oceanography in the School of Physical, Environmental and Mathematical Sciences (PEMS), the University of New South Wales (UNSW) at Canberra. There are many people I would like to thank who helped me at various stages throughout the task of researching and writing this thesis.

My supervisors deserve special thanks and deepest gratitude. Associate Professors Xiaohua Wang and Harvinder Sidhu, PEMS, UNSW at Canberra, always made time in their busy schedule to help me with various problems during my study. Their scientific ideas, many discussions, proof reading, support and encouragement over the years have been greatly appreciated. Thank you for introducing me to Oceanography and for your encouragement along the way. Thank you for keeping me on track throughout the hard grind of writing a thesis. Dr David Williams, Australian Institute of Marine Science (AIMS) provided this research with field assistance and field data. He has been enormously helpful, tirelessly answering many questions regarding data analysis and the use of field equipment. Professor Peter Ridd and the Marine Geophysics Laboratory at the James Cook University are highly appreciated for the generous loans of nephelometers and their efforts during field work. Dr Fernando Andutta assisted me with his useful physical background and constructive discussions. In addition, I would like to thank Donghui Jiang, Younjong Sun and Dehai Song for their generous help in answering my questions and giving me the benefit of their physical and numerical experience during this study.

Conducting a PhD study can be expensive, and both the Chinese Scholarship Council and the University of New South Wales at Canberra must be acknowledged for their scholarship support. The University of New South Wales funded travel expenses incurred during fieldwork, and PEMS provided equipment, charts and library resources. In particular, I would like to thank Australian Research Council (ARC) and INPEX for their generous financial support all through this study via an ARC Linkage Project (LP110100652), and Dr David Williams for his financial support during field trips and data observation. Without this financial support, this work would never have been possible.

Heartfelt thanks go to PEMS staff: Deborah Bator and Dianne Ferguson for their patient help with all kinds of travel; and Tessa Hodson, Nadia Seselja, Annabelle Boag, Steve James and Peter Scott for their warm introduction and assistance, when I first arrived and during my study. Particular thanks go to Colin Symons and Paul Mckie for answering many questions regarding the use of field equipment. Thanks to Dr Peter McIntyre for his editorial efforts and help with proof reading and constructive criticism. Sincere thanks also go to Julie Kesby, Research Officer, PEMS, for her fantastic help in providing advice on the PhD confirmation during my early research phase, for her patient assistance with my academic writing and chart reading, and for help with reference advice and proof reading.

Dr Kate Wilson and Geoff Millar from the Academic Language and Learning Unit, UNSW Canberra, helped me to improve my grammar in English writing. Sincere thanks to colleagues from the library for helping me with book reservations, database search methods and online literature retrieval. Many thanks go to Information Communication and Technology Service for patiently helping me with computer and internet problems.

As any researcher and writer can attest, no such work would be possible without the support of family and friends. In particular I would like to pay special tribute to the incredible support of my family, my father, mother, brother, and my aunt and her family. Heartfelt thanks and sincere appreciation go to Ms J. Xu and her family for bringing me the warmest feeling of home in Australia. Thank you for giving me so much love and encouragement. I also would like to thank all my friends for their warm care, lovely company and sincere friendship during this study.

TABLE OF CONTENTS

ABSTRACT ...... III

ORIGINALITY STATEMENT ...... V

AUTHENTICITY STATEMENT ...... VII

ACKNOWLEDGEMENTS ...... IX

TABLE OF CONTENTS ...... XI

LIST OF FIGURES ...... XV

LIST OF TABLES ...... XXI

CHAPTER 1 INTRODUCTION ...... 1

1.1 SIGNIFICANCE OF SUSPENDED-SEDIMENT DYNAMICS IN HARBOURS ...... 1

1.2 THE PROBLEM ...... 2

1.3 THE STUDY AREA ...... 5 1.3.1 Topography of Darwin Harbour ...... 5 1.3.2 Oceanography in Darwin Harbour ...... 7 1.3.3 Sediment characteristics of Darwin Harbour ...... 8 1.3.4 Meteorology of Darwin Harbour ...... 10 1.3.5 Ecology characteristics of Darwin Harbour ...... 11 1.3.6 Socio-economy of Darwin Harbour ...... 11

1.4 RESEARCH AIMS ...... 12

1.5 RESEARCH INNOVATIONS ...... 13

1.6 ORGANISATION OF THE THESIS ...... 13

CHAPTER 2 LITERATURE REVIEW ...... 15

2.1 INTRODUCTION ...... 15

2.2 HYDRODYNAMICS RESEARCH IN ESTUARIES ...... 16 2.2.1 Factors controlling hydrodynamics in estuaries ...... 16 2.2.2 Hydrodynamics in tidal flats and mangrove areas ...... 17 2.2.3 Tidal asymmetry in estuaries ...... 18

2.3 SEDIMENT DYNAMICS IN ESTUARIES ...... 18 2.3.1 Sediment re-suspension and the bottom boundary layer ...... 19 2.3.2 Settling velocity and flocculation ...... 20 2.3.3 Estuary Turbidity Maxima ...... 21

2.4 NUMERICAL MODELLING OF HYDRODYNAMICS AND SEDIMENT

DYNAMICS IN ESTUARIES ...... 22 2.4.1 Modelling studies of hydrodynamics in estuaries ...... 22 2.4.2 Modelling studies of suspended-sediment dynamics in estuaries ...... 23

2.5 HYDRODYNAMICS AND SEDIMENT TRANSPORT IN DARWIN HARBOUR ...... 24

CHAPTER 3 NUMERICAL STUDY OF HYDRODYNAMICS IN DARWIN HARBOUR ...... 27

3.1 INTRODUCTION ...... 27

3.2 METHODOLOGY ...... 28 3.2.1 Data collection and field measurements ...... 28 3.2.2 Model description ...... 30 3.2.3 Model configuration ...... 31 3.2.4 Calculating tidal asymmetry ...... 36

3.3 MODEL CALIBRATION ...... 38 3.3.1 Sea-surface elevation ...... 38 3.3.2 Current velocity ...... 41

3.4 MODEL VERIFICATION ...... 44 3.4.1 Experiment 1A ...... 45 3.4.2 Experiment 1B ...... 45 3.4.3 Experiment 1C...... 47

3.5 MODEL RESULTS AND DISCUSSION ...... 49 3.5.1 Tides in Darwin Harbour ...... 49 3.5.2 Effects of mangrove areas and tidal flats on tides ...... 52 3.5.3 Effects of the mangrove areas and tidal flats on tidal asymmetry ...... 57 3.5.4 Effects of mangrove areas and tidal flats on tidal energy and bottom dissipation ...... 59

3.6 CONCLUSIONS ...... 61

xii

CHAPTER 4 SUSPENDED-SEDIMENT DYNAMICS IN DARWIN HARBOUR ...... 63

4.1 INTRODUCTION ...... 63

4.2 METHODOLOGY ...... 64 4.2.1 Analysis of field data ...... 64 4.2.2 Model development ...... 68 4.2.3 Model configuration ...... 71 4.2.4 Sediment flux decomposition ...... 75

4.3 MODEL CALIBRATION ...... 76

4.4 MODEL RESULTS ...... 80 4.4.1 Suspended-sediment distribution in Darwin Harbour ...... 80 4.4.2 Vertical suspended-sediment concentration profile in Darwin Harbour ...... 85 4.4.3 Net sediment transport in Darwin Harbour ...... 87 4.4.4 Time variation of suspended-sediment concentration and seabed thickness ...... 88

4.5 DISCUSSION ...... 90 4.5.1 Effect of mangrove areas and tidal flats on suspended-sediment distribution ...... 90 4.5.2 Effect of the mangrove areas and tidal flats on net sediment flux ...... 92 4.5.3 Effect of dredging in East Arm on the suspended-sediment dynamics ...... 96 4.5.4 Effect of the disposed material on suspended-sediment dynamics.... 102

4.6 CONCLUSIONS ...... 103

CHAPTER 5 CONCLUSIONS AND RECOMMENDATIONS FOR FUTURE WORK ...... 107

5.1 CONCLUSIONS ...... 107 5.1.1 Hydrodynamics of Darwin Harbour ...... 107 5.1.2 Effect of mangrove areas and tidal flats on tidal asymmetry ...... 108 5.1.3 Suspended-sediment model development ...... 108 5.1.4 Suspended-sediment dynamics of Darwin Harbour ...... 109 5.1.5 Effect of mangrove areas and tidal flats on sediment transport ...... 110

xiii

5.2 RECOMMENDATIONS ...... 110

5.3 FUTURE WORK ...... 111

5.4 WIDER APPLICATIONS OF THIS STUDY ...... 113

REFERENCES ...... 115

xiv

LIST OF FIGURES

Figure 1.1: Map of Darwin Harbour showing the extent of intertidal mudflats and mangroves (Smith & Haese 2009)...... 4

Figure 1.2: Monthly mean meteorology data from the Bureau of Meteorology: (a) mean maximum/minimum air temperature; (b) mean rainfall; (c) mean wind speed...... 11

Figure 3.1: Darwin Harbour and the locations for field measurements. The ten locations are labelled on the map: A BoM station; B Location Hudson; C Location Blay; Locations MA1 and MA2 are in Middle Arm, EA1 and EA2 in East Arm, WA1 in West Arm, and CL and CR on the west and east coast of the outer harbour, respectively. Cross-section D is at the entrance to East Arm...... 28

Figure 3.2: Grids for the model domain (left-hand figure) and near the wharfs (right-hand figure). The contour heights in meters are relative to Lowest Astronomical Tides (LAT), with positive upward...... 32

Figure 3.3: Comparison of model and observed sea-surface levels at the ten observation locations: (a) and (b) Experiment 1 in 2009; (c) and (d) Experiment 1B in 2009; and (e) to (k) Experiment 1C in 2012...... 39

Figure 3.4: Comparison of model (Experiment 1) and observed current speeds at Location Blay in East Arm...... 42

Figure 3.5: Comparison of model (Experiment 1) and observed vertical profiles -1 of current speed (ms ) along its major axis for tidal constituents M2 and S2: (a) Location Blay, (c) Location Hudson, and (e) Location MA2; and deviations of the model values from the observed values (%): (b) Location Blay, (d) Location Hudson, and (f) Location MA2...... 43

Figure 3.6: Comparison of model (Experiment 1) and observed near-surface

current ellipses for tidal constituent M2 near locations: (a) Blay; (b) Hudson; and (c) MA2...... 44

Figure 3.7: Current speeds near peak ebb tide (GMT about 2:10am 30th September 2007) along Cross-section D near East Arm Wharf: (a) observation; and (b) model (Experiment 1A). The blue line in panel (a) shows the actual depth of the seabed...... 45

Figure 3.8: Comparison of model (Experiment 1B) and observed along-channel current speeds at Location Hudson in East Arm...... 46

Figure 3.9: Comparison of model (Experiment 1C) and observed along-channel currents at the surface, middle and bottom levels, and the average vertical current speeds at Location MA2...... 48

Figure 3.10: The M2 tidal amplitudes and phases at the surface...... 49

Figure 3.11: Vertically averaged velocities vectors of the spring (a) flood and (b) ebb currents at GMT about 1900 hours and 1300 hours, respectively, on 25th June 2009...... 50

Figure 3.12: Vertical profile of the M2 tidal-current major axis at four locations shown in Figure 3.1...... 51

Figure 3.13: M2 tidal-current ellipses at the four locations: (a) current-ellipse orientations; and (b) surface current ellipse distribution...... 52

Figure 3.14: Changes in M2 amplitude and phase when the mangrove areas are removed (Experiment 2 – Experiment 1)...... 53

Figure 3.15: Changes in M2 amplitude and phase when the mangrove areas and tidal flats are removed (Experiment 3 – Experiment 1)...... 53

Figure 3.16: Changes in M4 amplitude and phase when the mangrove areas are removed (Experiment 2 – Experiment 1)...... 54

Figure 3.17: Changes in M4 amplitude and phase when mangrove areas and tidal flats are removed (Experiment 3 – Experiment 1)...... 55

Figure 3.18: Changes in tidal asymmetry skewness  MM2/ 4 : (a) when all the mangrove areas are removed (Experiment 2–Experiment 1); (b) when all the mangrove areas and tidal flats are removed (Experiment 3 – Experiment 1)...... 58

xvi

Figure 3.19: Tidal asymmetry as a function of the percentage of mangrove areas removed in East Arm near Location Blay...... 58

Figure 3.20: Changes in tidal energy density: (a) when all the mangrove areas are removed (Experiment 2 – Experiment 1); (b) when all the mangrove areas and tidal flats are removed (Experiment 3 – Experiment 1)...... 59

Figure 3.21: Changes in bottom energy dissipation: (a) when all the mangrove areas are removed (Experiment 2 – Experiment 1); (b) when all the mangrove areas and tidal flats are removed (Experiment 3 – Experiment 1)...... 60

Figure 4.1: Research domain. Cross-sections: 1 in the channel; 2 at the entrance to East Arm; 3 from the outer harbour to Middle Arm. Locations: A – Charles Point; B – Lee Point; C – Mandorah Point; D – East Point; E - Darwin City and old wharfs; F – East Arm Wharf; G – Nightcliff; H - outer harbour; I – Inner harbour; J – Channel Island; K – Location to examine SSC time variation. The contour depths in meters are relative to mean surface level, with positive downward. Seven observation locations are labelled on the map: MA1 and MA2 in Middle Arm; EA1 and EA2 in East Arm; WA1 in West Arm; CL and CR on the west and east coast, respectively, of the outer harbour...... 65

Figure 4.2: Observed suspended-sediment concentrations at the seven locations in the harbour...... 66

Figure 4.3: Observed bottom-current speeds in East Arm (EA2) and Middle Arm (MA2)...... 67

Figure 4.4: The initial thickness of the fine-sediment layer used for Case 2...... 74

Figure 4.5: Comparison of model and observed time series of bottom suspended- sediment concentrations at the seven locations...... 79

Figure 4.6: Error statistics: (a) RMSE for the model suspended-sediment concentrations at the seven locations; (b) temporal (summation over time in Equation (4.7)) ACC at the seven locations; and (c) time variation (summation over locations in Equation (4.7)) in ACC during the simulation period...... 80

xvii

Figure 4.7: Suspended-sediment concentration at (a) the bottom level and (b) the surface level in the spring tidal cycle on 16th November 2012...... 82

Figure 4.8: Suspended-sediment concentration at (a) the bottom level and (b) the surface level in the neap tidal cycle on 7th December 2012...... 84

Figure 4.9: Vertical profiles of suspended-sediment concentration along Cross- section 3 in the spring tidal cycle on 16th November 2012. ‘A’ indicates the entrance to Middle Arm...... 85

Figure 4.10: Vertical profiles of current speed along Cross-section 3 during: (a) spring peak flood; (b) spring peak ebb; (c) neap peak flood; and (d) neap peak ebb...... 86

Figure 4.11: Vertical profiles of suspended-sediment concentration along Cross- section 3 in the neap tidal cycle on 7th December 2012. ‘A’ indicates the entrance to Middle Arm...... 86

Figure 4.12: Residual current and sediment-flux distributions along Cross-section 1, averaged over one lunar month: (a) residual-current distribution; (b) water volume flux time series; (c) sediment-flux distribution; and (d) sediment-flux time series (positive is seaward)...... 87

Figure 4.13: Time series of (a) surface elevation; (b) surface suspended-sediment concentration (SSC); (c) bottom SSC; and (d) fine-sediment layer thickness at location K near East Arm Wharf...... 88

Figure 4.14: Erosion and deposition after one lunar month (positive values indicate erosion)...... 89

Figure 4.15: Suspended-sediment concentration at the bottom level in Experiment B during the spring tidal cycle on 16th November 2012...... 90

Figure 4.16: Suspended-sediment concentration at the bottom level in Experiment B during the neap tidal cycle on 7th December 2012...... 91

Figure 4.17: Erosion and deposition accumulated in one lunar month in Experiment B (positive values indicate erosion)...... 92

xviii

Figure 4.18: Components of net sediment flux through Cross-section 1 from: (a)

Experiment A; and (b) Experiment B. T1 Eulerian velocity; T2 Stokes drift;

T3, T4 and T5 tidal pumping; T6 gravitation circulation; and T7 changing forms of the vertical profiles of velocity and concentration with the tide. (c) Depth: positive indicates seaward direction...... 95

Figure 4.19: As for in Figure 4.18, but along Cross-section 2...... 96

Figure 4.20: Location of dredging and dumping zones, and pipelines (black doted lines). A dredging zones in the shipping channel, approach area, berthing area and turning basin near East Arm Wharf; B dredging zones for offloading facilities near East Arm Wharf; C offshore disposal area; F pipeline. Inset: D location used in the model to simulate dumping; E the dredging zone in the model domain...... 97

Figure 4.21: Comparison of the time series for bottom SSC values from the dredging simulation (blue lines), material-dumping simulation (purple lines) and the observed values (black dots) at the seven locations...... 99

Figure 4.22: Suspended-sediment concentration values in the spring tidal cycle on 1st December 2012 from Experiment A1 at: (a) the surface level and (b) the bottom level; and from Experiment A2 at: (c) the surface level and (d) the bottom level...... 101

xix

LIST OF TABLES

Table 3.1: Geographic locations of the field-measurement sites in the harbour...... 29

Table 3.2: Configuration of key model parameters...... 34

Table 3.3: Experiment descriptions...... 35

Table 3.4: Skewness  c of the observed current velocities at Location Blay...... 37

Table 3.6: Model and observed sea-surface-level amplitudes and phases of the M2

and M4 tides near Location Blay...... 56

Table 3.7: Predicted and observed vertically averaged tidal current major axes of

the M2 and M4 constituents near Location Blay...... 56

Table 4.1: Model initial conditions and constants...... 73

Table 4.2: Test Cases and Experiments descriptions...... 75

Table 4.3: Comparison of net sediment transport...... 93

Table 4.4: Sediment-transport components (kg/s) through Cross-sections 1 and 2 (rounded to one decimal place)...... 94

xxi

xxii

CHAPTER 1 INTRODUCTION

1.1 Significance of suspended-sediment dynamics in harbours

Modelling the tidal and suspended sediment dynamics is fundamental for the development in marine industry and coastal management, in particularly, the wharf planning and other coastal construction (Toorman 2001). Recently, increasing attention has been paid to sediment transport in harbours and coastal areas due to some negative effects caused by increased suspended-sediment levels and in some cases the suspension of polluted sediments (Den Besten et al. 2003). Suspended-sediment dynamics in estuaries determines the fate of particle-bound nutrients and harmful materials, the rate of erosion or accretion of mudflats and sand beaches, the speed of siltation of navigation channels and the generation of zones of high turbidity (Van Leussen 2011). Turbid water leads to reduction in incomes from harbour activities such as fishing, recreation and entertainment, e.g. in Providence, RI, USA (Grigalunas et al. 2001). Also, as marine transportation in harbours is affected by siltation caused by the deposition of suspended sediments, wharf companies and the government may incur dredging costs and economic loss (Liu et al. 2011; Wang et al. 2011a; Wu et al. 2009).

Sediment suspension and re-suspension can be harmful to a harbour’s ecological system (Corbett et al. 2004) by degrading water quality. A high concentration of suspended sediments can cause reduction in phytoplankton biomass especially in estuaries because of their low levels of available light, leading to low growth rates, and smaller areal coverage and depth of colonies of macrophytes and macroalgae, e.g. in Venice Lagoon, Italy (Facca et al. 2002). Zooplankton can also be affected by turbidity in estuaries. In addition, suspended sediments can affect benthic communities, such as fish populations, and habitat complexities by restricting the growth of or removing alltogether macroalgae and other aquatic plants on which the fish feed. Birds and sea mammals may also be affected by changes in food supply as turbidity caused by the re- suspension of sediments results in anoxia and the release of potentially toxic substances (Zhang et al. 2010).

1.2 The problem

Suspended sediment in estuaries can affect estuarine morphology by its erosion and deposition. Highly concentrated suspended sediments in estuaries can induce stratification, which in turn can trap the sediments (Xing & Davies 2003). The trapped sediment makes the estuaries shallower and narrower. Sediment deposition can occur in river deltas and offshore (Holland et al. 2009; Wolanski et al. 1980). In tropical estuaries with large areas of mangrove forests, sediment from erosion of agricultural soil accumulates in the mangrove areas, e.g. Jiulongjiang Estuary in China (Alongi et al. 2005).

A comprehensive understanding of suspended-sediment dynamics in harbours is essential for harbour management and development because, if the locations of high- turbidity zones and the mechanisms of their formation can be predicted, appropriate policies and strategies can be implemented. Siltation and erosion due to the construction of wharfs and dykes, e.g. Changjiang Estuary (Jiang et al. 2012) and Scheldt River Estuary (Van Maren et al. 2009), can be minimised by properly selecting their locations and predicting accurately the consequent long-term deposition and erosion rates.

1.2 The problem

Darwin Harbour (12°28'S, 130°51'E), Northern Territory Australia, is the location for this study (Figure 1.1). It is a rapidly growing economic harbour in Northern Australia as it is located at the intersection of the Asia region and the tropical region, which are the two great regions of global economic and population growth (Coalition 2013). Since 2000-2001, the Northern Territory has had a high rate of economic growth, an average of 3.6% per annum. In 2008-2009, the growth rate is estimated to increase to 4.1% (Northern Territory Government 2009). In 2013, the Coalition (2013) proposed to develop the potential agriculture, tourism and energy export potential of Northern Australia. Darwin Harbour will be increasingly busy as one of the main transport harbours in the Northern Territory and is also an increasingly important center for tourism. With economic development, several types of marine-related activities, such as recreation, fishing and transportation, have been flourishing near and in the harbour. For example, in 2007-2008, total vessel calls to the Port of Darwin increased by 13.2% to a total of 5,340 (Northern Territory Government 2009).

1.2 The problem

To meet this economic potential, Darwin Harbour facilities are being expanded. One significant example is the liquid natural gas plant (LNGP) located on Wickham Point, close to which a new wharf, East Arm Wharf, is currently under construction. The LNGP will process gas from Browse Basin, the Ichthys Project managed by INPEX Browse Ltd (INPEX). The Ichthys Project, began in May 2012, will have an initial capacity to produce 8.4 million tonnes of liquid natural gas (LNG), 1.6 million tonnes of liquefied petroleum gas (LPG) annually, and approximately 100,000 barrels of condensate per day at peak (Kelly 2012). An amount of US$34 billion has been committed by the Australian Government to the Ichthys Project in Darwin (INPEX 2012). INPEX is bringing in the largest dredge ever seen in Australia to carry out an associated dredging operation in Darwin Harbour (Sinclair Knight Merz Pty Ltd 2011). On completion, the harbour will have a deeper channel for LNG carriers to load gas at East Arm Wharf. Territorians and visitors to the Northern Territory place a high value on Darwin Harbour, which is used for fishing, boating and other recreational pursuits, and also holds significant heritage and cultural value (Fortune & Drewry 2009).

1.2 The problem

Figure 1.1: Map of Darwin Harbour showing the extent of intertidal mudflats and mangroves (Smith & Haese 2009).

All of these economic activities are likely to have an impact on the harbour, and potentially lead to further pressure on the harbour’s aquatic environments. For example, silting of the harbour could lead to a threat to shipping and to fishing which could threaten economic development. In order to sustain the economic development, understanding the harbour’s hydrodynamics and sediment-transport patterns has become increasingly important so as to conserve the economic, environmental and cultural value of the harbour.

A quantitative capability to model the hydrodynamics and sediment dynamics in the harbour is essential to properly address harbour management issues. Mathematical modelling is now a useful approach, due to substantial advances in numerical

1.3 The Study area techniques and computer capabilities in recent years. Also, the development of appropriate instruments has made sediment measurements feasible and available, with suspended-sediment concentrations able to be measured by nephelometers, acoustic backscatter profiling sensors (ABS) and optical backscatter profiling sensors (OBS). The data from these can be used to calibrate the numerical models and thereby display the sediment dynamics of the entire harbour.

This study uses the three-dimensional (3-D) Finite Volume Coastal Ocean Model (FVCOM) to explore the hydrodynamics in Darwin Harbour. A sediment model developed by Wang (2002) is incorporated into the FVCOM hydrodynamic model. This sediment model focuses on sediment suspension and re-suspension, and the physical mechanisms controlling these processes. The study will help to determine the distribution and transportation patterns of sediments and their sources. An understanding and prediction of sediment deposition will also allow siltation and the dredging of navigational channels to be managed. Therefore, this research will be of importance for Darwin Harbour’s management and development, thereby, delivering significant socio-economic benefits to the Darwin community.

1.3 The study area

1.3.1 Topography of Darwin Harbour

Darwin Harbour is on the north coast of Australia. It opens to the north along a line from Charles Point to Lee Point into the Beagle Gulf and the Timor Sea. Darwin Harbour has complex shoreline, characterised by dense mangrove forests, rivers and headlands. The 50 km length from its outer boundary to its uppermost estuarine limit in Middle Arm comprises three components: the outer harbour; the inner harbour; and three arms, East, West and Middle Arms. The coastline geometry, which provides an essential spatial reference for the land boundary in the hydrodynamic model, is defined by Charts AUS 24 (2007), 26 (1995), 28 (2008) and satellite images. Data from ETOPO11 are used as a reference to double check the map used in the model. Darwin city, the old wharfs, East Arm Wharf and the LNGP are all located on headlands near East Arm. These headlands make the hydrodynamics in the harbour complicated.

1 www.ngdc.noaa.gov/mgg/global/global.html 5

1.3 The Study area

The bathymetry of Darwin Harbour changes gradually, with an average depth of about 15 meters. The channel is the deepest area. A navigation channel from the outer harbour to the entrance of East Arm is being dredged as part of East Arm Wharf construction. The harbour bathymetry was obtained from the official navigation charts. In addition, the Australian Institute of Marine Science (AIMS) has high-resolution bathymetry data for Darwin Harbour, covering all the mangrove areas, tidal flats and areas outside the harbour mouth.

There are three main rivers flowing into Darwin Harbour. The largest is Blackmore River, followed by Elizabeth River, running into Middle Arm and East Arm, respectively. The runoff catchment area of the harbour is very small. The catchment of the two main rivers is about 120 km2, which is only ~18% of the harbour surface area (660 km2). According to runoff records, if both the Elizabeth and Blackmore Rivers peaked at the same time, the combined instantaneous runoff would be about 1,000 m3s-1 (Williams et al. 2006). This peak event would only last for about an hour before falling to a base level of less than 100 m3s-1. The probability of the two rivers peaking at the same time is quite low, a one-in-a-hundred-year event. In contrast, the peak spring flood tidal flow measured along a line from East Point to Mandorah is 120,000 m3s-1 (Williams et al. 2006): the flood tidal prism is three orders of magnitude greater than the river input. Both the Elizabeth and Blackmore Rivers cease to flow during the dry season. Berry Creek supplies freshwater all year round to Blackmore Estuary, but at a low rate (< 1 m3s-1). According to the assessment by Asia-Pacific Applied Science Associates (2010), the flow of the Elizabeth River during the wet season has some impact in terms of current speed only at the upstream extent of East Arm, where the speed may increase by as much as 0.2 ms-1 after rain. Any influence would however be hardly traceable by midway along East Arm, and the East Arm port area would not be affected.

In recent years, coastal facilities have been and will continue to be constructed to cater for the fast-developing coastal economy of Darwin Harbour. The infrastructure has grown steadily, especially since the mid 1990s, e.g. wharfs, marinas and the LNGP (Williams et al. 2006); almost all these facilities have being constructed near East Arm. This construction has changed the morphology of the harbour, e.g. the width of East Arm during high tide has changed from 3.5 km before to 2.2 km after the East Arm

1.3 The Study area

Wharf construction.

1.3.2 Oceanography in Darwin Harbour

Darwin Harbour is semi-diurnal macro-tidal, with a maximum tidal range of 7.8 m, mean spring and neap tidal ranges of 5.5 m and 1.9 m, respectively (Woodroffe et al. 1988), and a mean tidal range of 3.7 m (Michie 1987). The large tidal range dominates most of the characteristics of the harbour, e.g. the concentration of suspended sediment varies over the tidal cycle (McKinnon et al. 2006). There is a 1.5 h lag in the tides between the harbor mouth and its upper reaches. The tides become asymmetric as they propagate into the upper reaches, as indicated by ebb tides about one hour longer than flood tides (Williams et al. 2006).

Because the harbour is macro-tidal, the prevailing currents are of tidal origin. Current speeds vary with bathymetry and proximity to the shore. Maximum flood and ebb current speeds near the entrance to the harbour are 0.3 ms-1 and 1 ms-1 during neap tides, and 0.5 ms-1 and 1.6 ms-1 during spring tides (Asia-Pacific Applied Science Associates 2010; HR Wallingford 2010a). Occasionally, maximum currents speeds in the harbour can reach 2 ms-1 during spring tides (Li et al. 2011; Williams et al. 2006). Peak tidal currents are about 25% faster at flood tide than at ebb tide in the upper reaches of the harbour (Williams et al. 2006).

The strong tides mix the freshwater runoff and ocean salt water, so there is no marked river plume exiting the harbour during the wet season. For most of the year, Darwin Harbour is vertically well mixed with respect to salinity. The oceanic salinity at the harbour mouth remains almost vertically constant throughout the year, because the freshwater is strongly mixed with salt water before it reaches the mouth. There is only a very gradual salinity gradient between the upper Blackmore River and the outer harbour (Andutta et al. 2013).

Darwin Harbour is well sheltered from long-period tsunamis and ocean-swell waves because of its geography. Waves within the harbour are generally of short mean period (3-5 s), with heights well below one meter. Waves with a significant wave height of 4.5 m and mean period of 7.5 s are observed at the harbour entrance, but the wave heights are reduced down to 0.7 m inside the harbour (Asia-Pacific Applied Science Associates 2010).

1.3 The Study area

Water temperature in the harbour varies seasonally, with higher temperatures during the wet season and lower temperatures during the dry season. During 2003-2004, the water temperature ranged between about 24°C in the dry season (June 2003) to 33°C in the wet season (December 2003). However, the spatial variation in the harbour at any time is very small (Andutta et al. 2013; McKinnon et al. 2006).

1.3.3 Sediment characteristics of Darwin Harbour

Sediments in the harbour are predominantly fine-grained, mainly clay and silt and have been investigated by Williams et al. (2006). The sediment grain-size distribution over the entire harbour was sampled by Fortune (2006). East Arm has mainly mud beds, with a large calcareous sand deposit upstream of East Arm Wharf. Silt and clay (<63 µm) comprise about 18% on average of the sediment in East Arm. More than 80% of the sediment near Darwin City is fine sand (< 250 µm). The silt and clay component varies in the range 1-28%, depending on the sample site.

In the channel and part of the inner harbour, fine sand (63-250 µm), coarse sand (0.25-2 mm) and gravels (>0.25 mm) comprise about 95% of the sediment, with only 5% silt and clay. The sediment in West Arm is dominated by fine sand (~32%) and coarse sand (~43%), with granular material (~32%) throughout the area. Middle Arm has mainly mud beds, but also some significant shoals of siliceous sands that become fine seaward, indicating a terrestrial origin. Fine sand comprises more than 50% here, and the clay fraction is low (0-5.3%). The silt fraction ranges between 17% (near Channel Island) and 28% (at the entrance to Middle Arm). Suspended sediment flocculates, with floc size typically 50-200 µm, as shown by microscope pictures (Williams et al. 2006). Near the mouth of each arm, in waters with suspended-sediment concentrations less than 0.04 kgm-3, there exists a plume zone where the floc size is in the range 200-1,000 µm (Williams et al. 2006). Fieldwork shows that the west coast of the outer harbour is sandy, whereas the east coast is muddy, with large areas of tidal flats. The seabed in the inner harbour and along the first few kilometers of Middle Arm contains mostly coarse sediment because strong currents at these places inhibit deposition of fine sediment.

Zones of maximum turbidity lie inside the mouth of the three arms in both the dry and wet seasons, with suspended-sediment concentration reaching 0.25 kgm-3 (Williams et al. (2006). This is controlled by a combination of the complex circulation near

1.3 The Study area headlands and embayments, and the asymmetry of the tidal currents. The net sediment flux in the wet season is landward, at rates of 4.8, 13.4 and 8.5 tonnes m-1d-1 near Wickham Point, Channel Island and the tip of Middle Arm, respectively (Williams et al. (2006). However, in the dry season, the net sediment flux near Wickham Point is seaward, at a rate of 1.1 tonnes m-1d-1. Upstream in Middle Arm, the net sediment flux is still landward, at a rate of 3.7 tonnes m-1d-1.

Tidal currents in Darwin Harbour determine which kind of sediment and how much remains in suspension or is deposited. Strong currents in the channels erode fine sediments from the substrate, leaving coarser material behind, and redepositing the finer suspended sediments in areas of weaker currents such as the subtidal and intertidal mud flats and the mangrove areas. The concentration of suspended sediment varies over a tidal cycle (Padovan 1997; Wilson et al. 2004); this being most pronounced in the upper reaches of the harbour. The turbidity originates from a number of sources including catchment inflow, bottom-sediment re-suspension, mud transported from mangrove areas, flocculation of organic material in the water column and stocks of phytoplankton.

As shown by McKinnon (2006) derived from CTD (Seabird SEB 19plus) casts, in June 2003, there is little vertical stratification in turbidity, and only a very gradual pattern of decreased turbidity with distance from upper Blackmore River to Middle Arm and the outer harbour. Turbidity decreases at the seaward end along the channel of the harbour as a result of the dilution of the highly turbid water by low-turbidity water from the Timor Sea (McKinnon 2006).

INPEX has modelled the coarse- and suspended-sediment dynamics induced by dredging work for the construction of East Arm Wharf and the LNPG (HR Wallingford 2010b). The model was a two-dimensional model (TELEMAC-2D) coupled with the DELFT-3D water-quality module. Dredging is being conducted in the shipping channel, the approach area, the turning basin, the berthing area and the offloading facility area in East Arm, and along the inshore pipe route from the outer harbour to Wickham Point. The dredging disposal area is located at outside the harbor mouth. The INPEX model predicted that the dredging could affect almost all the East Arm channel and the entrance to Middle Arm.

1.3 The Study area

1.3.4 Meteorology of Darwin Harbour

Darwin Harbour has a tropical monsoon climate, with two distinct seasons, dry and wet. The dry season is the six months between April and September, the wet season October to March. Monthly climate statistics from the Bureau of Meteorology (BoM), including mean temperatures, rainfall and wind speeds, are shown in Figure 1.2. These data were taken at Darwin Airport (12.42 °S, 130.89 °E).

As can be seen from Figure 1.2(a), the mean maximum and mean minimum air temperatures from 1981 to 2010 were about 33°C and 25°C, respectively. There is little variation in heat flux at the sea surface. During the same period, there was negligible rainfall during the dry seasons, whereas the wet seasons had an average monthly rainfall of 275mm. The bulk of the rainfall is between December and March (Figure 1.2(b)). Wind speed was negligible, with averaged speeds in the years 1995 to 2010 less than 4.5 ms-1 (Figure 1.2(c)).

Figure 1.2: Monthly mean meteorology data from the Bureau of Meteorology: (a) mean maximum/minimum air temperature; (b) mean rainfall; (c) mean wind speed.

1.3 The Study area

1.3.5 Ecology characteristics of Darwin Harbour

The Darwin Harbour region includes large water areas, tidal flats, rivers, bare rocky outcrops, coral reefs, subtidal rocky reefs and dense mangrove forests. The total area of mangroves is about 238 km2, which is 22% of the total area of the harbour to the high water mark (Water Monitoring Branch 2005). They form valuable ecosystems in sheltered tropical and sub-tropical coastal environments that are periodically inundated by tidal waters. Extensive mangrove forests around the harbour fringe provide important habitat, generate food supply for other species and maintain the productivity of the harbour. These habitats are home to a diverse range of species of flora and fauna. Estimates of the number of marine invertebrate species exceed 3,000. The checklist of marine fish has 415 species (Fortune & Drewry 2011). Some of these marine species are the foundation for marine tourism in Darwin. All these species depend on high-quality water and habitats in the harbour to ensure their preservation. One of the most valuable marine organisms in Darwin Harbour is coral: the harbour has 55 species of hard coral (Fortune & Drewry 2011). Coral reefs are considered to be among the most sensitive ecosystems to long-term climate change, pollution, dredging and other environmental and anthropogenic factors. Well-developed coral reefs in the harbour are confined to Channel Island, Wickham Point, Weed Reef and South Shell Island, where relatively high water clarity, strong currents and therefore low sedimentation rates, create a habitat favorable to corals reefs (Fortune & Drewry 2011). The extensive dredging proposed for the East Arm Wharf development puts the coral at high risk because of elevated amounts of suspended sediment that can settle on the reefs.

Mangrove forest occupies at least two-thirds of the harbour foreshore. A total of 36 mangrove species occur in the harbour, constituting about half the world’s mangrove species. About 2% of the mangrove areas in the harbour have been reclaimed, with nearly 96% of the remainder conserved; any change to this requires the approval of the Northern Territory Government (McKinnon et al. 2006).

1.3.6 Socio-economy of Darwin Harbour

The Darwin Harbour economy is expected to grow in the near future in order to cater for the expected boost in activities related to the increased exploitation of petroleum in the nearby Timor Sea, to the completion of the Adelaide-Darwin railway link and to the

1.4 Research aims continued expansion in trade with Asia. The Darwin Port Corporation has an AUD$150 million infrastructure program, including the East Arm Wharf Facilities Masterplan 2030, to upgrade the harbour’s existing facilities. These investments guide the development of East Arm Wharf and provide direction and certainty to stakeholders, industry and the economic growth of the Territory. In addition, between 2000 and 2001 three exploratory wells resulted in the discovery of an extremely promising gas and condensate field, now known as the Ichthys Field (around 13°44' S 123°15' E). The gas from the Ichthys Project will be transported to the onshore LNG processing plant proposed for Blaydin Point on the Middle Arm Peninsula in Darwin Harbour. The Ichthys Project has begun and it is expecting to bring large economic profits to Darwin

Darwin Harbour also provides a venue for commercial, recreational and subsistence fishing. Recreational fishing in the harbour is substantial: in 2000, 37% of the harbour’s residents spent some of their time fishing, and one in every five resident households owned a pleasure boat used at least in part for recreational fishing. The tourism industry is also very important because 21% of the recreational fishing in and near the harbour is by visitors (Coleman 2004). Aquaculture has also been developed in the harbor, with the first project a pearl oyster farm near Wickham Point. Several prawn farms have also been established in Middle Arm (Williams et al. 2006).

This attractive future of a growing economy results in an increasing demand on facilities and resources. East Arm Wharf and the LNGP are two new large coastal facilities in the harbour. There will be an inevitable increase in the stresses on the harbour’s terrestrial and aquatic environments with this increase in economic development.

1.4 Research aims

This study examines the 3-D hydrodynamics and sediment-transport dynamics of Darwin Harbour, based on field measurements and numerical modelling. The effects of mangrove and tidal-flat reclamation on the tides and sediment transport are also explored.

More specifically, this thesis aims to achieve the following objectives:

(1) Collect and analyse hydrographic and sediment data from Darwin Harbour;

1.5 Research innovations

(2) Conduct fieldwork to quantify the bottom boundary layer and sediment dynamics in the harbour;

(3) Use the data collected to build a hydrodynamics model to simulate tides in the harbour using the FVCOM, a 3-D finite-volume model with an unstructured mesh;

(4) Couple the sediment model developed by Wang (2002) to the hydrodynamic model to simulate sediment transport in the harbour.

1.5 Research innovations

This study makes the following novel contributions to this research area.

Firstly, it examines the 3-D hydrodynamics of the entire Darwin Harbour, which has not been previously investigated.

Secondly, it couples a hydrodynamics model with a sediment-dynamics model to reveal the 3-D sediment dynamics of the entire Darwin Harbour.

Finally, the effects of mangrove areas and tidal flats on tides and sediment- transport dynamics in the harbour are studied; the results from this study can provide scientific data for both harbour management and development.

1.6 Organisation of the thesis

This thesis is organized as follows.

Chapter 1 introduces the study area, including its topography, meteorology, biological diversity, socio-economy, oceanography and sediment dynamics. A literature review is presented in Chapter 2.

Chapter 3 presents the hydrodynamics model of Darwin Harbour; its results are used to examine the harbour’s 3-D tidal characteristics, including the effects of mangrove areas and tidal flats on the tides.

In Chapter 4, a sediment-dynamics model of Darwin Harbour is constructed. The model results are used to describe and quantify sediment distribution and flux in the harbour. The effects of mangrove areas and tidal flats on sediment-transport dynamics are also discussed.

1.6 Organisation of the thesis

Chapter 5 presents a summary and conclusions drawn from this study. Recommendations for future research work to assist Darwin Harbour’s ongoing development and management are also offered.

CHAPTER 2 LITERATURE REVIEW

2.1 Introduction

Suspended sediment in estuaries and harbours has been observed to accumulate slowly on the bottom, on the tidal flats and in the inner regions where low hydrodynamic energy conditions prevail (Xu 2009). Suspended-sediment deposition and re-suspension constitute a significant loop for sediment dispersion in estuaries and harbours. Suspended-sediment dynamics in estuaries is continuously changed by local hydrodynamics, such as tides, freshwater incursion, wind and waves. The tides determine how suspended sediments are distributed and transported (Nelson 2001), e.g., in Changjiang Estuary (Shi 2010). The location of the zone of maximum turbidity is related to gravitational circulation and salinity stratification within estuaries (Dyer 1997), together with the freshwater incursion, which itself can generate offshore net sediment transport in tropical estuaries, e.g. Daly Estuary (Wolanski et al. 2006c). On tidal flats, suspended-sediment transport is influenced by tides and waves (Christie et al. 1999; Holland et al. 2009; Pejrup 1988), which also determine their physical characteristics (Le Hir et al. 2000), nutrient and metal exchange between the seabed and water column (Christiansen et al. 2004), and the formation of the seabed (Whitehouse et al. 2011). Wind has often been considered a predominant driving force for sediment re- suspension in many estuaries, especially shallow ones (Ven Beusekom & Jonge 1995). The net sediment transport over tidal flats increases with wind speed (Ridderinkhof et al. 2000), and its direction is dependent on wind direction and the location of source material (Yang & Khangaonkar 2009). Therefore, an understanding of the hydrodynamics in estuaries is fundamental to the study of suspended-sediment dynamics. In this chapter, the literature on hydrodynamics in estuaries is first reviewed in Section 2.2 to help in building a hydrodynamic model for Darwin Harbour. The suspended-sediment dynamics is then investigated in Section 2.3. A review of the numerical modelling studies of hydrodynamics and sediment dynamics is presented in Section 2.4. Finally, the hydrodynamics and sediment dynamics in Darwin Harbour are reviewed in Section 2.5.

2.2 Hydrodynamics research in estuaries

Estuaries, which are conjunctions of open oceans and lands, often have complex bathymetry, high biological diversity and flourishing marine economies (Kingsford et al. 2005). This complexity determines the estuarine hydrodynamic characteristics and properties. As these are important for sediment dynamics, accurate predictions of the hydrodynamics has attracted much attention worldwide. Special attention has also been paid to tides and tidal asymmetry in estuaries, as they are closely linked with sediment transport.

2.2.1 Factors controlling hydrodynamics in estuaries

The spatial and temporal variabilities of estuarine circulation are caused by its complicated bathymetry and three main driving forces: tides; winds; and freshwater discharges.

According to Mehta (1988), an estuary can be classified into one of three different hydrodynamic regimes, based on its tidal range,  . This study shows the dominant hydrodynamic forces are tidal currents, wind waves and wind currents when 1.0m; tidal currents and wind waves when 1.0   3.0 m; and tidal currents only when 3.0 m. Based on this classification, as the maximum and mean tidal ranges in Darwin Harbour are 7.8 m and 3.7 m, respectively, sediment re-suspension in the estuary is expected to be driven only by tidal currents.

The hydrodynamics can be significantly affected by wind events in estuaries, especially flow structure in shallow areas (Collins et al. 1998), e.g. current speeds measured in the upper part of a water column may deviate from the standard assumption of a logarithmic profile. Wind-induced surges cause significant changes to tidal elevation, currents and the wetting-drying processes in shallow coastal regions (Jones & Davies 2008), to which the hydrodynamics and the suspended sediment-transport process are very sensitive (Dyer et al. 2000). This is because sediments can be eroded in the shallows, even by small waves, when onshore winds are present (Lumborg et al. 2006). A complete reversal of the sediment-transport processes, from ebb dominance to flood dominance, can then occur. This would change an estuary from an erosion type to a deposition type.

2.2 Hydrodynamics research in estuaries

River inflow is a common characteristic of an estuary. High river inflow generates a strong freshwater front and high seaward residual currents which affect the water circulation (e.g. Ría de Muros in Spain (Carballo et al. 2009)). The process of fresh water mixing with salt water affects sediment transport and stratification, forming zones of high turbidity, which can then impact the biogeochemical processes that occur there, e.g. Changjiang Estuary in China (Shi & Zhang 2011), Columbia River Estuary in the United States (MacCready et al. 2009), and Tana Estuary in Kenya (Kitheka et al. 2005).

2.2.2 Hydrodynamics in tidal flats and mangrove areas

The hydrodynamics in tidal flats are complex, characterised by extremely shallow and rapidly varying water depths, both spatial and temporal, with additional air exposure and special bottom conditions. The main forces that tidal flats experience are tides, wind-induced circulation, waves, density-driven circulation and drainage processes (Le Hir et al. 2000). Among these factors, tide is naturally the most important process in macro-tidal estuaries. Mangrove-forested wetlands occupy the upper intertidal zone between terrestrial and marine ecosystems (Woodroffe et al. 1988). To enable a better understanding of the roles of tidal flats and mangroves in the exchange of sediment between mangrove-swamp forests and adjacent coastal waters, considerable effort has recently been focused on studying the hydrodynamic processes in tidal flats and mangrove swamps.

It is generally understood that intertidal wetland characteristics are determined by the cumulative and complex interactions of many factors, e.g. hydrology, sediment dynamics and sea-level changes (Knight et al. 2008). Vegetation density, evapotranspiration, rainfall and, in some cases, groundwater, can also be important for mangrove tidal flooding patterns (Mazda et al. 1995; Wolanski 1992). Of these factors, the hydrodynamics of the adjacent estuary is crucial for the ecosystems of mangrove forests, because it changes their physical properties and determines their exchanges, e.g. dispersion of sediment, nutrients and mangrove seeds (Varnell et al. 2003).

In turn, mangroves play a key role in modulating the hydrodynamics, e.g. the tides (Mazda et al. 2005). Mangroves create their own ecosystems, with a high bottom drag due to their dense roots, and trap sediment that forms their substrate (Furukawa &

2.3 Sediment dynamics in estuaries

Wolanski 2004; Mazda et al. 1997b). A mangrove tidal creek experiences an asymmetry in its tidal currents due to vegetation-induced friction (Wolanski et al. 1980).

2.2.3 Tidal asymmetry in estuaries

According to Walton (2002), if the duration of the falling tide is longer/shorter than that of the rising tide, leading to a larger peak flood/ebb current, the system is referred to as flood/ebb dominant or flood/ebb asymmetric. Tidal asymmetry can promote residual currents and net mass transport (tide-induced residual circulation). As asymmetry plays an important role in sediment transport dynamics and channel morphologies (Aldridge 1997; Dronkers 1986) in various tidal embayments and estuaries, and understanding of its mechanisms is essential for suspended-sediment studies.

Astronomical tides (e.g., M2) that propagate into estuaries can generate shallow- water and compound constituents (e.g., M4, M6 and others) (Huang et al. 2008; Nidzieko 2010). Tidal asymmetry may be produced by the interactions of these tidal constituents, including astronomical tides, shallow water tides and compound tides (Blanton et al. 2002; Song et al. 2011). Two metrics have been used to describe tidal asymmetry: the amplitude ratio, aa/ , which measures the degree of asymmetry; DD12 and the phase difference, 2 , which indicates its direction. The subscripts D DD12 1 and D indicate diurnal and semidiurnal tidal constituents, respectively (Hoitink et al. 2 2003; Speer & Aubrey 1985). In recent years, several other ways of defining tidal asymmetry have been developed. The tidal-duration asymmetry,  1 , is the normalized sample skewness of the time derivative of the tidal elevation (Nidzieko 2010). The ebb- tide duration is shorter for g 1 < 0, the flood-tide duration shorter for g 1 > 0. This theory has been further improved to include all the tidal constituent groups which generate tidal asymmetry, and to calculate their contributions to the total asymmetry created by tidal interactions using the amplitudes, frequencies and phases of all astronomical tides (Song et al. 2011).

2.3 Sediment dynamics in estuaries

There are essentially two physical processes that control sediment transport in estuaries: one, sediment re-suspension, occurs at the seabed and produces sediment in the water

2.3 Sediment dynamics in estuaries column; the other, advection, transports sediment suspended in the water column away from its source, such as rivers (Kombiadou & Krestenitis 2013). These are important processes in Darwin Harbour; in modelling the harbour, more attention needs to be paid to sediment re-suspension and the bottom boundary layer, and to sediment transport, including settling velocity, flocculation and Estuary Turbidity Maxima.

2.3.1 Sediment re-suspension and the bottom boundary layer

Although sediments in estuaries come from various sources, including the land, ocean and bottom erosion (Escapa et al. 2008; Wolanski & Simon 2000), the sediment dynamics in the dry season in Darwin Harbour consists mainly of the suspension and re- suspension of local sediments. Sediment is lifted from the seabed into the water column by erosion due to the large bottom shear stress induced by strong currents. Deposition is the opposite process to erosion. It governs the removal of sediment particles from water column. The bottom shear stress,  b , and critical shear stress,  c , are important variables that govern and distinguish erosion and deposition processes (Ariathurai & Krone 1976). One pioneering work associating the erosion rate with hydrodynamic forcing is Partheniades (1962), as cited by Sun (2001), which described the erosion using , and the erosion rate E0. Subsequently, a general form of erosion formula was introduced by Gailani et al. (1991), as cited by Sun (2001), to account for the seabed armouring effect by considering the time after deposition.

The parameter characterising bottom shear stress has been estimated by several methods and various values obtained for different bathymetry situations, e.g., Le Hir et al. (2000). It can be derived from measured current-velocity profiles on the basis of the Kármán-Prandtl model (Dyer 1986), which is classically referred to as the ‘law of the wall’ method as cited in Wang (2002).

The critical shear stress parameter can be estimated in the laboratory using sediments from the particular site under investigation (Wang et al. 2011b). Values of the critical shear stress are mostly determined by bottom conditions (Peterson 1999): in a clay zone, a stress of 0.01–0.03 Nm-2 is expected to move non-cohesive clay-sized particles, although it is acknowledged that the cohesive properties of clays may be much more durable. In a silt zone, non-cohesive particles in the silt size range begin scouring

2.3 Sediment dynamics in estuaries between 0.03 and 0.1 Nm-2. As silts finer than 10 or 20 μm are mobilised more easily, they are the most common constituent in dead-zone sludge; in a sand zone, sand begins to move with a shear stress of 0.1 Nm-2. However, the characteristics of the bottom boundary layer (BBL) are expected to be modified in a flow with suspended sediment. High concentrations of re-suspended sediment lead to an increase in flow velocity and ‘drag reduction’, which is perceived mainly as a reduction in bed shear stress (Gust 1976). Therefore, to deal with the BBL, a stability function has been added to the calculation of the drag coefficient (Wang 2002).

Near-shore waves have both positive asymmetry and skewness, which generate a thinner boundary layer and, consequently, a larger bed shear stress (Gonzalez-Rodriguez & Madsen 2007). Consequently, a large sediment transport often occurs in a wave boundary layer (Malarkey & Davies 2012).

2.3.2 Settling velocity and flocculation

Sediment settling and the vertical motion of the water particles control the vertical sediment transport. The additional effects of river discharge, waves and biota make this process quite complex (Wright & Schoellhamer 2005). An increase in suspended- sediment concentration (SSC) increases the settling velocity by increasing the probability of flocculation, as is shown in the settling velocity equation of Stokes’ Law of Sedimentation. However, when the flocs are too large, the upward friction increases and the settling velocity decreases due to the increased interaction between the flocs and between the flocs and the water flow (Sun 2001). The concentration at which the settling velocity starts to decrease is known as hindering settling concentration. The suspended-sediment concentration is included in the calculation of the settling velocity; a critical SSC value separates the flocculation and hindering settling processes (Sun 2001).

Turbulent shear in water can significantly affect the flocculation process, and hence, the settling velocity of flocculated sediment in a turbulent flow (Milligan & Hill 1998). The impact of turbulent shear in water on the median settling velocity has been addressed by Pejrup and Mikkelsen (2010). More-comprehensive formulas for settling velocity have also been presented. For example, Strom and Keyvani (2011) derived an

2.3 Sediment dynamics in estuaries explicit formula for the floc settling velocity over the viscous, transitional and inertial ranges. This formula considers parameters like particle shape and floc diameter.

2.3.3 Estuary Turbidity Maxima

An estuarine turbidity maximum (ETM) is usually defined as a region where the SSC is a maximum relative to both upstream and downstream in an estuary (Nichols & Biggs 1985; Wai et al. 2004). It generally occurs in the upper parts of estuaries (Mitchell et al. 1999). The formation mechanisms of an ETM vary according to the characteristics of the different estuaries. Tides, bathymetry, river inflow and particle properties can influence the locations and SSC of ETM zones (Brenon & Le Hir 1999).

In tidally controlled estuaries, ETMs are, to a large extent, governed by tides, especially in macro-tidal estuaries (Brenon & Le Hir 1999). The variations in the tidal conditions along an estuary affect the magnitude of tidal pumping which, typically, controls the location of the ETM (Manning et al. 2010; Uncles & Stephens 2010). Tidally averaged gravitational circulation is a mechanism by which sediment is trapped in an ETM in an estuarine flow. In estuaries characterised by medium-to-low tidal ranges (<4 m) and strong river inflow, the resulting residual seaward flow near the surface must be balanced by an equivalent residual landward flow near the seabed, which has the effect of maintaining the ETM in a location near the fresh-salt-water interface (Uncles & Stephens 1993). An alternative mechanism put forward is that tidal asymmetry causes landward movement of sediment over successive tides. This is balanced by a seaward transport influenced by the freshwater flow (Brenon & Le Hir 1999; Dronkers 1986). The predominance of either of these two mechanisms appears to be related to the magnitudes of the tidal range and the freshwater flow (Dyer 1997). In many cases, a combination of the two processes has been proposed, such as in Tamar Estuary (Uncles & Stephens 1989) and Gironde Estuary (Doxaran et al. 2009).

The effects of turbulence, saline-induced flocculation, wind waves, stratification and sediment deposition on intertidal zones are also potentially important mechanisms for ETM dynamics (e.g. Nunes Vaz et al. 1989; Weir & McManus 1987).

2.4 Numerical modelling of hydrodynamics and sediment dynamics in estuaries

Numerical models simulating hydrodynamics in estuaries vary in the number of dimensions, the numerical techniques adopted, the mathematical description of the hydrodynamic processes and the forces included (e.g. Bourgault & Kelley 2004; Chen et al. 2003; King 2009; Son & Hsu 2011). Based on hydrodynamic models, numerical models for sediment dynamics deal with different kinds of sediment, e.g. sand, silt or clay. They focus on the different sediment properties and processes, including cohesive or non-cohesive, erosion or deposition, settling or flocculation, and bottom boundary layers. Conventionally, 3-D advective-diffusion equation is split into horizontal (2-D) and vertical (1-D) equations to deal with the different scales in the horizontal and vertical planes. A similar split algorithm is used to separate the 3-D sediment-transport equation. Different methods are used to improve the accuracy of the sediment transport models, e.g. a refined layer-integrated algorithm has been applied to solve the transport equation in the cohesive-sediment simulation in Humber Estuary (Wu et al. 1999).

Selection of a model depends on the availability of input and calibration data, the physical reliability, the problem scale, the required accuracy and the available budget (Van Rijn 1993). For specific problems in certain estuaries, model selection largely depends on estuarine characteristics and the study purpose. For example, fresh-water inflow cannot be neglected in Changjiang Estuary (Ma et al. 2011), so that a 3-D model is required to display vertical salinity variation.

2.4.1 Modelling studies of hydrodynamics in estuaries

One-, two- and three-dimensional models have been developed for different purposes and used in various aspects of hydrodynamics simulation. One-dimensional models are usually used in tidal channels and rivers due to their special morphology, or used to investigate certain vertical properties (Chang et al. 2011). Two-dimensional models are used, for example in the Lister Dyb tidal area in Denmark (Lumborg & Pejrup 2005), as they are suited to the specific research requirements and take less time to run. Three- dimensional models have mostly been developed in recent years, for example the FVCOM (Chen et al. 2003) and Princeton Ocean Model (POM) (Blumberg & Mellor

2.4 Numerical modelling of hydrodynamics and sediment dynamics in estuaries

1987); transport of suspended sediment, salinity and heat needs 3-D hydrodynamic information to show variations in the vertical direction.

The numerical techniques adopted in estuarine hydrodynamics models allow both horizontal and vertical variations. Horizontally, the complicated bathymetry of estuaries is usually decomposed with structured or unstructured grids. Vertically, a depth- averaged algorithm is used in 2-D models. For 3-D models, the s -coordinates have been developed and are widely used, for example in FVCOM, as they can represent the irregular variable topography and reproduce realistic bottom boundary layers in estuaries better than average distributed layers (Mellor & Blumberg 1985). Another method to compensate for the disadvantages of average distributed layers is to add additional bottom layers to display better the bottom properties. For example, Resource Management Associates (RMA) model uses bottom layers at fixed heights just above the bed.

2.4.2 Modelling studies of suspended-sediment dynamics in estuaries

Three-dimensional models are preferred for the simulation of suspended-sediment dynamics, as large horizontal and vertical variations in suspended-sediment concentration (SSC) need to be included. For example, sediment dynamics in Humber Estuary (Wu et al. 1999) and Pearl Estuary (Hu et al. 2011) have been simulated using 3-D sediment models. Two-dimensional sediment models are used in many studies, due to easier parameter settings and shorter computing time, when they can fulfill the objectives of the research, e.g. sediment dynamics in Darwin Harbour by HR Wallingford (2010b) and Williams (2009).

Sediment models for estuaries have been developed for either cohesive or non- cohesive (e.g. Van Rijn 1993) sediments, as the mechanisms of cohesive and non- cohesive sediment dynamics are quite different. For cohesive sediments, a large number of processes may affect its dynamics, as discussed in Section 2.3, including the hydrodynamics, consolidation of the sediment bed and biological processes. For non- cohesive sediments, the grain size, shape and density of the sediments are more important (Sun 2001). Therefore, it is hard to simulate cohesive and non-cohesive sediments together, because defining the erosion and deposition of a cohesive and non- cohesive sediment mixture is quite a challenge. A sediment dynamics model currently

2.5 Hydrodynamics and sediment transport in Darwin Harbour adopted by FVCOM can simulate both the cohesive and non-cohesive sediments, but their properties are still calculated separately using different settling-velocity algorithms (Warner et al. 2008). Cohesive sediments have attracted much more attention than non- cohesive sediments in estuarine numerical modelling because of the specific seabed properties: sediment size is always considered as smaller than fine sand (<0.25 mm). Some sediment models have been developed specifically for suspended sediments, e.g. Wang (2002).

2.5 Hydrodynamics and sediment transport in Darwin Harbour

Of all the research conducted in coastal zones, only a small number of studies have been undertaken of Darwin Harbour. These includes studies of its biology (Burford et al. 2008; McKinnon et al. 2006; Metcalfe & Glasby 2008; Noske 1996; Peerzada & Kozlik 1992; Wolanski et al. 2006a; Woodroffe et al. 1988) and of the pollution in it (Esslemont 1999; Peerzada & Dickinson 1988; Peerzada & Ryan 1987).

As Darwin Harbour is fringed by large areas of mangroves, research has been conducted into the relationship between its ecology and the mangrove systems (Metcalfe et al. 2011; Metcalfe & Glasby 2008; Woodroffe et al. 1988); these show that the mangrove areas can affect the hydrodynamics and act as a sink for suspended sediments.

The coastal water in East Arm is turbid, with a high concentration of suspended sediments from the mangroves and surrounding coastal land of which, in an average wet season, about 36,000 t flow into the harbour (Fortune & Drewry 2009). During the dry season, there is no input from the river into the harbour: sediments are from the open sea. The suspension and re-suspension of sediments has been demonstrated to change the water quality (Wolanski et al. 2006b), and the factors controlling the hydrodynamics and sedimentation in the harbour have been studied (Williams et al. 2006). East Arm, where the central business district and the wharfs are located, is subject to deposition of silt and clay (Wolanski et al. 2006b), although it too has large areas of mangroves that act as efficient sediment traps (Furukawa & Wolanski 2004).

Since 1993, the Water Research Laboratory (WRL), in association with the Northern Territory Government, has conducted hydrodynamic and sediment-transport

2.5 Hydrodynamics and sediment transport in Darwin Harbour modelling of Darwin Harbour, including modelling of its tidal dynamics and the sedimentation of dredge plumes, using the RMA Modelling Suite (Water Research Laboratory 2000). Williams et al. (2006) discussed the 2-D hydrodynamics of Darwin Harbour. Williams (2009) used a 2-D sediment model to assess the sediment dynamics if the sandbar in East Arm were removed. More recently, to guide their Ichthys Project, INPEX estimated the impact of dredging on the harbour environment using a 2-D numerical model and published an Environmental Impact Statement (HR Wallingford 2010b).

However, to date, no study has been conducted into Darwin Harbour’s 3-D sediment dynamics, and its 3-D hydrodynamics and sediment-transport mechanisms are still largely unknown. The sediment bottom boundary layer has not been explored, nor have the effects of mangrove areas and tidal flats on tides and sediment transport been examined.

2.5 Hydrodynamics and sediment transport in Darwin Harbour

CHAPTER 3 NUMERICAL STUDY OF HYDRODYNAMICS IN DARWIN HARBOUR

The work in this chapter has been published (Li et al. 2011; Li et al. 2012).

3.1 Introduction

Understanding the hydrodynamics in a harbour is fundamental to the analysis of its transport properties, such as sediment transport, and to planning harbour development, such as wharf construction. In particular, the pattern of sediment transport determines the fate of a harbour’s morphology, siltation of the navigation channels and generation of turbid zones (Van Leussen 2011). Harbours have the potential to be degraded when their hydrodynamic environments are changed by new construction, e.g. in Alhama de Granada, southern Spain (Viseras et al. 2009) and on the Dutch coast (Wijnberg 2002). In harbours dominated by tides, such as Darwin Harbour, strong interactions between the hydrodynamics and sediment dynamics originate from the cyclical processes of suspension, mixing and deposition (Baird et al. 1987; Ven Beusekom & Jonge 1995). In the long run, harbours may silt up even when there are only small quantities of new sediment. Therefore, detailed consideration of the hydrodynamics of Darwin Harbour is necessary to prepare for future construction and development. The hydrodynamics in Darwin Harbour are complex because of its complicated shorelines and bathymetry.

The primary aims of this chapter are to: firstly build and calibrate a hydrodynamic model of Darwin Harbour against field data collected by the Australian Institute of Marine Science (AIMS); and secondly, use this model to understand the effects of mangrove areas and tidal flats on the harbour hydrodynamics, including tidal dynamics and tidal asymmetry.

Section 3.2 analyses the field data using Harmonic Analysis. An introduction to the methodology of the hydrodynamics model and the tidal-asymmetry calculations are also presented in this section. The calibration and verification of the hydrodynamics model are discussed in Sections 3.3 and 3.4. Model results and discussion are presented in Section 3.5. Conclusions are provided in Section 3.6.

3.2 Methodology

3.2.1 Data collection and field measurements

In order to obtain the basic hydrodynamic characteristics and to calibrate the hydrodynamics model of the harbour, data from various sources were collected, and field measurements then conducted to supplement these data. The observation locations at which these data were measured are labelled on the map of Darwin Harbour in Figure 3.1; their latitudes and longitudes are listed in Table 3.1.

3.2 Methodology

Table 3.1: Geographic locations of the field-measurement sites in the harbour. Location Latitude (S) Longitude (E) BoM* station 12°28' 130°51' Blay 12°30' 130°54' Hudson 12°30' 130°55' MA1 12°35' 130°52' MA2 12°36 130°55' EA1 12°30' 130°54' EA2 12°32' 130°58' WA1 12°32' 130°47' CL 12°24' 130°43' CR 12°20' 130°50' *Australian Bureau of Meteorology

The bathymetry, sea-surface level and current data, obtained from the AIMS, cover the entire harbour, including the water areas, tidal flats and mangrove area. These data are almost uniformly distributed, with resolution in most areas of about 500 m by 500 m. The resolution in East Arm and along the navigational channel is higher, about 30 m by 30 m. These data were the primary source of the bathymetry for this study. Isobaths in the water area were taken from charts (AUS 24 (2007), AUS 26 (1995), AUS 28 (2008)) to improve the accuracy of the geometry. Hourly sea-surface level data were obtained from the Bureau of Meteorology (BoM) for the years 1992 to 2009 to study the principle tidal characteristics in the harbour: their observation station is near Darwin City. The AIMS sea-surface level and current data were measured at 10-minute intervals at Blaydin Point (Location Blay) and Location Hudson near East Arm Wharf. These data were measured by a bottom-mounted upward-looking Nortek ADCP. The sea-surface level data from Location Blay is from 20th June to 10th July 2009. The current data, from the same time period at this location, were measured in ten evenly distributed vertical layers, with a bin size of 1 m. The sea-surface level data from Location Hudson are from 1st to 30th August 2009. The current data, from the same time period at this location, were taken from more vertical layers (20 evenly distributed vertical layers), with a smaller bin size (0.5 m), than those at Location Blay. To get a full

3.2 Methodology picture of the vertical current profile, current transect data along Cross-section D (Figure 3.1), taken by a boat-mounted, downward-looking Nortek ADCP on 30th September 2009, were obtained from AIMS.

The AIMS data are limited to three locations and along one cross-section, near East Arm Wharf and Darwin City. For this reason, more data are required to reveal the overall hydrodynamic situation and to calibrate the model results over the entire harbour. Therefore, seven more locations in the harbour were selected at which to conduct hydrodynamic measurements (see Figure 3.1 for these locations). These locations were chosen to sample the geophysical characteristics of the entire harbour; equipment deployment was also technically feasible there, with reasonable water depth and moderate current velocities. The field measurements cover the time period from early November 2012 to early February 2013, the transient time of the dry and wet season in preparation for the sediment study described in Chapter 4. Sea-surface level was then measured at all seven locations at 10-minute intervals. Current data were measured at Location MA2 at 10-minute intervals, with more vertical layers (59 evenly distributed vertical layers) and finer resolution (bin size of 0.25 m) than those from Locations Blay and Hudson. The data were measured by a bottom-mounted upward-facing ADCP on a tripod 0.4 m above the seabed. The measurements between early November and early December are used to calibrate the hydrodynamic model because this period is in accordance with the sediment model calibration as the suspended-sediment concentration measurements are compromised by dredging after mid-December. As this period has not much rain, river discharge is not considered.

Overall, sea-surface level data were taken from ten locations and current data from three locations. These data were filtered to remove equipment noise and bad measurements. Harmonic Analysis was then conducted to resolve the tidal constituents to obtain the tidal characteristics of the harbour and prepare for calibration of the hydrodynamic model.

3.2.2 Model description

FVCOM (Chen et al. 2003) is a 3-D hydrodynamic model which uses an unstructured, finite-element grid. Compared with structured-grid models, such as POM (Blumberg & Mellor 1987) and the Environmental Fluid Dynamics Code (Hamrick 1992), the

3.2 Methodology unstructured grid used by FVCOM is especially suited to Darwin Harbour, which has complex shoreline geometries and dynamic physical processes. As FVCOM solves the 3-D momentum, continuity and density equations using a finite-volume method, it allows mass conservation to be strictly maintained.

A σ-stretched coordinate system is applied in the vertical direction to improve the representation of the complicated bathymetry and obtain a more accurate representation of the irregular variable bottom topography. The σ-coordinate transformation is defined as zz  , HD where σ varies from −1 at the bottom to 0 at the surface. The total water-column depth is D = H + ζ , where H is the bottom depth and ζ is the height of the free surface, both relative to z = 0 (mean sea-surface level). The continuity equation is:  ()()Du  Dv  w     0, (3.1) t  x  y  where x, y and σ are the east, north and vertical coordinates, respectively, and u, v and w the corresponding velocity components. The model employs the Mellor-Yamada Level 2.5 turbulence closure scheme (Mellor & Yamada 1982) for vertical mixing and the Smagorinsky scheme for horizontal mixing (Smagorinsky 1963). The drag coefficient

Cd is determined by matching a logarithmic bottom layer to the model at a height zab above the bottom,

 2 Cd  max2 , 0.0025 , (3.2) (ln(zzab /0 )) where κ = 0.4 is the von Karman constant and z0 the bottom roughness parameter.

3.2.3 Model configuration

3.2.3.1 Domain

This study focuses on Darwin Harbour from Charles Point to Lee Point (Figure 3.1), with the model domain expanded to include areas outside the mouth of the harbour to keep all open-boundary nodes in the open ocean and the mangrove areas in order to check the hydrodynamics there. The water areas, tidal flats and mangrove areas are

3.2 Methodology shown in Figure 3.1. The locations of Darwin City and the wharfs are also indicated on the map.

Construction of the unstructured triangular model grid, which consists of 9,666 elements and 5,205 nodes in the horizontal plane, is based on all the available bathymetry and shoreline data from AIMS. Figure 3.2 shows the model grid for the entire domain; the resolution around the wharfs and in the three arms is especially high. The cell sizes of the domain range from 18 m near the wharfs to 3,300 m at the open boundary. To simulate the vertical profiles of the currents accurately, 20 uniform vertical layers are specified in the water column using the σ-coordinate system. Four locations, 1, 2, 3 and 4, in the outer harbour, near the channel, in the inner harbour and in Middle Arm, respectively (Figure 3.2), are selected to display the tidal current vertical profiles.

Figure 3.2: Grids for the model domain (left-hand figure) and near the wharfs (right- hand figure). The contour heights in meters are relative to Lowest Astronomical Tides (LAT), with positive upward.

3.2.3.2 Boundary conditions and forcing

The oceanic open boundary is located outside the harbour mouth, extending into the open ocean from Charles Point and Lee Point (Figure 3.1). The open-boundary

3.2 Methodology conditions for the water level were specified using tidal elevations from the TPXO7.2 global model of ocean tides1. Hourly tidal elevations, constructed using four diurnal components (K1, O1, P1, Q1), four semidiurnal components (M2, S2, N2, K2), three shallow-water components (M4, MS4, MN4) and two long-period components (Mf, Mm), were applied at the open-ocean boundary.

River discharge is small according to the discussion in Section 1.3.1. Therefore, there is no stratification in Darwin Harbour even during the wet season (summer), as the strong tidal velocities cause thorough mixing. Although the upper arms become less saline during flood events, they do not stratify. However, the flood flows do play a role in flushing, and may well affect the re-suspension of sediments from the mangrove zones and those that have been moved upstream and deposited during the long dry season. The salinity of the water column varies with horizontal position, increasing in the downstream direction until it reaches the background salinity value. This gradual horizontal change in salinity/density can generate density-driven currents, but these are estimated to be less than 0.09 ms-1, about 3.0% of the maximum current observed in the study area.

This chapter concentrates on modelling the harbour hydrodynamics in the dry season, when no thermal stratification from the arms to the outer harbour has been observed (Williams 2009). There is a small along-channel salinity gradient created by evapotranspiration; the density currents this generates are again negligible compared with the tidal currents.

Wind effects at the free-surface boundary can be neglected, according to the sensitivity model test of Li et al. (2011): the east and north current differences caused by wind are less than 0.05 ms-1. Given that the maximum and mean tidal ranges in Darwin Harbour are 7.8 m and 3.7 m, respectively, the hydrodynamics in the harbour are expected to be driven only by tidal currents (Mehta 1988). Therefore, both wind and heat fluxes at the free-surface boundary are neglected in our model.

The bottom roughness used in the model is calculated according to the water depth. If a node depth is less than 3.0 m, its depth is set to 3.0 m; the bottom drag

1 http://volkov.oce.orst.edu/tides/TPXO7.2.html 33

3.2 Methodology

coefficient Cd is calculated from Equation (3.2). The mangrove zones are treated differently from the water areas and tidal flats because of the larger drag forces caused by mangrove trees and their roots (Mazda et al. 1997a). According to field experience and Mazda et al. (1997a), ranges from about 1.0 to 10.0, depending on the observation site, mangrove species and tidal conditions. In this study, the median value

Cd  5.0 is set for the mangrove areas if they are completely covered by water.

3.2.3.3 Initial conditions

The model was initialized with a constant value for salinity of 33psu and for temperature of 25°C, typical of the harbour’s mean salinity and temperature during the dry season. Together with the assumptions of no river inflow and zero heat flux at the surface, these values result in barotropic conditions in the model. The model was run for 31 days, from 20 June 2009 to 21 July 2009; its key parameters are summarised in Table 3.2.

Table 3.2: Configuration of key model parameters. Model parameter Value Model time setup 1.0 s Bottom friction coefficient Minimum 0.0025; 5.0 for mangrove areas Horizontal diffusion Smagorinsky scheme Vertical eddy viscosity M-Y 2.5 turbulent closure Node, element, vertical layers 5205, 9666, 20 uniform σ layers Open boundary condition Tidal time series from TPXO7.2

3.2.3.4 Model runs and sensitivity tests

In order to check the effects of the mangrove areas and tidal flats on tides in the harbour, three numerical experiments using the model were designed: in Experiment 1, the model was run with all the mangrove areas and tidal flats included; in Experiments 2 and 3, the mangrove areas and the mangrove areas plus tidal flats, respectively, were removed from the model (converted to land). Experiment 1A, using the observed vertical current profile data from 30th September 2007, and Experiments 1B and 1C,

3.2 Methodology using observed elevation and current data from August 2009 and November 2012, respectively, were used to further verify the model in Experiment 1.

Three further sensitivity experiments were run to examine the effect of mangrove areas in more detail by progressively removing them from the model: in Experiment 4, about 30% of the total mangrove area was removed, from around East Arm; in Experiment 5, about 50% was removed from around East Arm and the east side of Middle Arm; and in Experiment 6, about 70% was removed from around East Arm and Middle Arm. Descriptions of all experiments are given in Table 3.3. These mangrove- removal scenarios are in accordance with those which may occur as a result of development, because almost all the social and economic activities of Darwin take place near East Arm.

Table 3.3: Experiment descriptions. Experiment Description 1 Reference experiment; all mangrove areas and tidal flats included. 1A As for Experiment 1, but for 30th September 2007. 1B As for Experiment 1, but for August 2009.

1C As for Experiment 1, but for November 2012.

2 All mangrove areas removed; tidal flats included.

3 All mangrove areas and tidal flats removed. 4 30% of the total mangrove areas removed (around East Arm); tidal flats included. 5 50% of the total mangrove areas removed (around East Arm and east side of Middle Arm); tidal flats included. 6 70% of the total mangrove areas removed (around East Arm and Middle Arm); tidal flats included.

3.2 Methodology

3.2.4 Calculating tidal asymmetry

As discussed in Section 2.2.3, different methods can be used to quantify the tidal asymmetry. The tidal duration asymmetry  1 is the normalized sample skewness of the  time derivative of the tidal elevation   ' (Nidzieko 2010) t

1 T 3 '' t1 t 3 T 1   1 3 3/2 , 2 (3.3)  1 T ''  t1 t T 1  

where z is the sea-surface elevation, 3 is the third sample moment about its mean and  the standard deviation (Emery & Thomson 2001). The summation is from time t = 1 to t = T. The ebb-tide duration is shorter if 1  0 ; the flood-tide duration shorter if

1  0 .

In an estuarine environment with the tidal elevation and tidal currents generally 90°

out of phase, the asymmetry  1 computed from the tidal elevation Equation (3.3) will be

similar to the asymmetry  c calculated from current velocities in the absence of river flow, stratification and bathymetry effects (Nidzieko 2010), as is the case for Darwin

Harbour.

According to Song et al. (2011),  1 can also be calculated using the amplitudes, frequencies and phases of all the components of the astronomical tides by:

3322 aiijjkk a  a sin  ijk     a iijj  a sin  2  ij    i  j   k242  i   j 1  , (3.4) 1 N ( a2 2) 3/2 2 i1 ii

where the ai , i and i are the tidal amplitudes, frequencies and phases of the respective components of the astronomical tides. In this study, the M2/M4 combination makes the most significant contribution. Other constituent combinations play only

3.2 Methodology minor roles and can be neglected. If only these two constituents are considered, the tidal asymmetry  may be approximated by  , where (Song et al. 2011): 1 MM24/

3 aa2 sin 2 MMMM2 4 2 4    2 (3.5) M/M24 3 2 1 22 (aa 4 ) MM24 2

 The approximation MM24/ is used here in analyzing the controlling factors of tidal asymmetry in Darwin Harbour.

The skewness , calculated from the sea-surface elevation (SSE) data near

Location Blay is positive (0.06), which indicates flood dominance (Walton 2002). The skewness  c of each of the observed along-channel current velocities at Location Blay is shown in Table 3.4: these values indicate flood dominance at all depths. In Darwin Harbour, is larger than , because the bathymetry produces local asymmetry in tidal currents that is sometimes not manifested in the free-surface records (Godin 1991).

Table 3.4: Skewness  c of the observed current velocities at Location

Blay. Vertical layer (meters above seabed)

1 (1.8) 0.16 2 (2.8) 0.12 3 (3.8) 0.11 4 (4.8) 0.11 5 (5.8) 0.12 6 (6.8) 0.13 Average Layers 1-6 0.12

3.3 Model calibration

3.3.1 Sea-surface elevation

The hourly SSE data for the years 1992–2009 from the BoM station and Location Blay (AIMS data; Figure 3.1) were analysed to extract the principal tidal characteristics of the harbour. Tidal data from both BoM and AIMS were used to calibrate the model results. Figure 3.3 shows the model sea-surface level time series, which agree well with the field data from the BoM station (Figure 3.3(a)) and Location Blay (Figure 3.3(b)).

Comparison of the model and observed amplitudes and phases of the SSEs of the main tidal constituents, M2, S2, K1, N2, O1 and K2 near Darwin City are shown in Table 3.5. The harmonic constants are averages from the annual harmonic analyses from 1992 to 2009. Of these tidal constituents, M2 is predominant, which indicates that the harbour is a semidiurnal environment. Table 3.5 demonstrates good agreement between the model and observed amplitudes and phases. The deviations of the model M2 and S2 amplitudes/phases from those observed are 8.1%/7.2° and 3.1%/1.5°, respectively. The large discrepancy between the observed and model N2 and K2 phases may be caused by the fact that the tidal analysis is based on a one-month time interval, and it is therefore unable to resolve these constituents accurately.

3.3 Model calibration

Table 3.5: Comparison of model (Experiment 1) and observed tidal harmonic parameters at the ten locations in Figure 3.1. Amp: amplitude in meters. Deviations for amplitude are the percentage differences between the model and observed values; deviations for phase are the differences model – observed. Tidal constituent O1 K1 N2 M2 S2 K2 M4 Location Amp Phase Amp Phase Amp Phase Amp Phase Amp Phase Amp Phase Amp Phase BoM Observed 0.33 190.1 0.58 200.5 0.35 228.4 1.85 249.3 0.96 298.1 0.27 296.0 0.05 107.1 Model 0.30 187.2 0.53 196.2 0.29 212.4 1.70 242.1 0.93 299.6 0.25 322.0 0.04 46.4 Deviation – 9.1 –2.9 –8.6 –4.3 –17.1 –16.0 –8.1 –7.2 –3.1 1.5 –7.4 26.0 –20.0 –60.7 Blay Observed 0.35 189.2 0.54 196.7 0.42 229.3 1.92 249.2 1.07 308.7 0.29 331.1 0.06 113.4 Model 0.31 182.9 0.55 196.8 0.32 220.4 1.74 243.0 0.95 296.2 0.26 318.6 0.04 55.13 Deviation –11.4 –3.3 1.9 0.1 –23.8 –3.9 –9.4 –2.5 –11.2 –4.0 –10.3 –3.8 –33.3 –51.4 Hudson Observed 0.34 192.3 0.63 186.2 0.39 221.7 1.93 248.5 1.08 296.3 0.29 318.7 0.04 115.9 Model 0.32 184.9 0.58 195.5 0.32 223.9 1.74 243.7 0.93 284.0 0.25 306.4 0.02 199.6 Deviation –5.9 –7.4 –7.9 9.3 –17.9 2.2 –9.8 –4.8 –13.9 –12.3 –13.8 –12.3 –50.0 83.7 MA1 Observed 0.35 188.5 0.61 194.7 0.40 219.3 1.84 238.2 0.92 288.9 0.25 311.3 0.06 96.5 Model 0.32 187.9 0.54 200.6 0.32 223.3 1.78 245.1 0.76 293.7 0.21 316.1 0.03 65.0 Deviation –8.6 –0.6 –11.5 5.9 –20.0 4.0 –3.3 6.9 –17.4 4.8 –16.0 4.8 –50.0 –31.5 MA2 Observed 0.35 188.9 0.61 195.2 0.41 220.1 1.87 239.4 0.94 290.1 0.26 312.7 0.06 91.5 Model 0.30 190.3 0.53 202.9 0.32 228.1 1.80 248.4 0.77 298.8 0.21 321.2 0.06 60.6 Deviation –14.3 1.4 –13.1 7.7 –22.0 8.0 –3.7 9.0 –18.1 8.7 –19.2 8.5 0 –30.9 EA1 Observed 0.35 186.9 0.58 193.8 0.39 214.8 1.76 236.9 0.88 286.2 0.24 308.6 0.05 88.2 Model 0.31 185.9 0.54 199.1 0.32 221.0 1.76 243.3 0.75 291.3 0.20 313.7 0.04 50.3 Deviation –11.4 –1.0 –6.9 5.3 –17.9 6.2 0 6.4 –14.8 5.1 –16.7 5.1 –20.0 –37.9 EA2 Observed 0.36 188.2 0.63 194.1 0.42 219.1 1.93 237.7 0.96 288.3 0.26 310.7 0.06 95.9 Model 0.29 186.6 0.52 200.0 0.30 222.4 1.72 244.4 0.73 293.3 0.20 315.7 0.05 72.2 Deviation –19.4 –1.6 –17.5 5.9 –28.6 3.3 –10.9 6.7 –24.0 5.0 –23.1 5.0 –16.7 –23.7 WA1 Observed 0.34 186.7 0.60 193.1 0.39 215.8 1.79 235.2 0.89 284.8 0.24 307.2 0.05 90.1 Model 0.31 186.3 0.54 199.2 0.31 200.4 1.72 242.9 0.73 290.6 0.20 313.0 0.03 38.3 Deviation –8.8 –0.4 –10.0 6.1 –20.5 –15.4 –3.9 7.7 –18.0 5.8 –16.7 5.8 –40.0 –51.8 CL Observed 0.34 183.6 0.59 190.7 0.37 210.3 1.66 230.5 0.82 278.7 0.22 310.1 0.04 70.3 Model 0.31 184.9 0.53 198.0 0.29 217.8 1.58 240.9 0.67 287.9 0.18 310.2 0.03 17.6 Deviation –8.8 1.3 –10.2 7.3 –21.6 7.5 –4.8 10.4 –18.3 9.2 –18.2 0.1 –25.0 –52.7 CR Observed 0.33 183.3 0.58 192.2 0.36 211.7 1.65 231.7 0.81 280.2 0.22 302.6 0.04 52.8 Model 0.31 184.3 0.53 197.5 0.29 216.6 1.58 239.7 0.67 286.8 0.18 309.2 0.03 8.3 Deviation –6.1 1.0 –8.6 5.3 –19.4 4.9 –4.2 8.0 –17.3 6.6 –18.2 6.6 –25.0 –44.5

3.3 Model calibration

3.3.2 Current velocity

Validation of the model current velocities was conducted at the bottom, middle and surface levels; the results are shown in Figure 3.4. Comparison of the model and observed along-channel velocities in East Arm near Location Blay indicates that the model predicts the current speeds and phases very well.

The vertical current profiles of the M2 and S2 tides from the model and the observed data are shown in Figure 3.5. Panel (a) shows the maximum current speeds at Location Blay: the model values are in good agreement with the observed values in both magnitude and trend from the bottom to the surface, with model deviations from the observations being no more than 15%, as shown in panel (b). In addition, the predicted orientations of the surface current ellipses of the M2 tide match well with those from the observed data, as shown in Figure 3.6(a). Although only the orientations near the surface are shown here, those at the other depths have similar directions.

Panels (c) - (f) are discussed in Section 3.4.

3.3 Model calibration

Figure 3.4: Comparison of model (Experiment 1) and observed current speeds at Location Blay in East Arm.

3.3 Model calibration

Figure 3.5: Comparison of model (Experiment 1) and observed vertical profiles of current speed (ms-1) along its major axis for tidal constituents M2 and S2: (a) Location Blay, (c) Location Hudson, and (e) Location MA2; and deviations of the model values from the observed values (%): (b) Location Blay, (d) Location Hudson, and (f) Location MA2.

3.4 Model verification

Figure 3.6: Comparison of model (Experiment 1) and observed near-surface current ellipses for tidal constituent M2 near locations: (a) Blay; (b) Hudson; and (c) MA2.

3.4 Model verification

In order to validate the hydrodynamic model, three additional numerical experiments were conducted to simulate the tides in the harbour in different periods: September 2007; August 2009; and November 2012. These experiments use the same settings as Experiment 1, except that of tidal forcing at the open-ocean boundary: the time series of the tidal forcing was set to be that of the experimental time period at one-hour intervals.

Experiment 1A, September 2007, examined the ability of the model to reproduce the vertical profiles of current speed near East Arm Wharf. Experiment 1B, August 2009, double checked the model accuracy in East Arm. Experiment 1C, November 2012, prepared for the suspended-sediment simulation in this study, as it is in the same modelling period as used in the suspended-sediment model (Chapter 4).

3.4 Model verification

3.4.1 Experiment 1A

Experiment 1A checks the model predictions of the vertical profiles in current speed near East Arm Wharf on 30th September 2007. Figure 3.7 shows a comparison of the observed and model vertical current profiles along Cross-section D (Figure 3.1). Currents are stronger near East Arm Wharf; this is reproduced by the model (right-hand side of Figure 3.7). The discrepancy between the model and the observed values is due to a mismatch of the bathymetry: the depth of the seabed (blue line in panel (a) in Figure 3.7) is not matched by the model (panel (b)). In addition, the model data are saved hourly, whereas the observed data are not exactly hourly, which may also generate discrepancies.

3.4.2 Experiment 1B

Experiment 1B simulates the hydrodynamics in the harbour in August 2009. The observed sea-surface level data from BoM and AIMS, and observed current data at Location Hudson (Figure 3.1) are used to calibrate the model. The model reproduces the sea-surface level well, matching both the amplitudes and phases of the observed data, as

3.4 Model verification shown in Figure 3.3, panels (c) and (d). The model main tidal constituents agree well with the observed values at the BoM station and Location Hudson, in both magnitude and phase. The deviations of the model amplitudes and phases from the observed values are less than 9.8/13.9% and 4.8/12.3° for the M2/S2 tides, respectively (Table 3.5).

A comparison of the observed and model along-channel current speeds at Location Hudson is shown in Figure 3.8, at the near-surface, middle, near-bottom levels and of the vertically averaged. The model reproduces the current speeds and phases well.

A comparison of the vertical current profiles of the M2 and S2 tides is shown in Figure 3.5. Panel (c) indicates a close fit between model and observed data, with deviations of less than 12.5% (panel (d)). The model current ellipses for the M2 tide align reasonably well with those observed; only the surface current orientation is shown in Figure 3.6(b). The current ellipses at other depths have a similar orientation.

Figure 3.8: Comparison of model (Experiment 1B) and observed along-channel current speeds at Location Hudson in East Arm.

3.4 Model verification

3.4.3 Experiment 1C

Experiment 1C, for November 2012, is constructed in preparation for the suspended- sediment simulation described in Chapter 4. The observed sea-surface level data from seven locations and current data from Location MA2 are used to validate the hydrodynamic model. The model time series for the sea-surface level agree well with the field data at all the seven locations, as shown in Figure 3.3, panels (e) to (k). Table 3.5 demonstrates reasonable agreement between the model and observed amplitudes and phases of the sea-surface level. The deviations of the model M2 and S2 amplitudes from those observed are 10.9% and 24%, respectively. The deviations of the phases are all within 10.4°.

A comparison of observed and model along-channel current speeds at Location MA2 is shown in Figure 3.9, for near-surface, middle and near-bottom currents, as well as the vertically averaged currents. The model performs well, matching both the current speeds and phases.

3.4 Model verification

Figure 3.9: Comparison of model (Experiment 1C) and observed along-channel currents at the surface, middle and bottom levels, and the vertically averaged current speeds at Location MA2.

The model and observed vertical profiles of maximum current speed (along the current major axis) for the M2 and S2 tidal constituents are shown in Figure 3.5. The model M2 and S2 tidal-current major axes are close to those from the observed data

(panel (e)). The deviations of the model M2 and S2 current speeds from the observed values are less than 16%, from the near-bottom to the near-surface levels, as shown in panel (f). The model current-ellipse orientations have similar orientations from the bottom to the surface levels, in accordance with those observed; only the surface-current orientations are shown in Figure 3.6 (c).

In conclusion, the model results show reasonably good agreement with the observed values, with acceptable errors, for SSEs, currents, amplitudes and phases of

3.5 Model results and discussion

the M2 and S2 tidal constituents. As a result, we believe that the model can be used to simulate accurately the water-flow dynamics in the harbour.

3.5 Model results and discussion

3.5.1 Tides in Darwin Harbour

The model results from Experiment 1 are used in this section to study the tidal dynamics of the Harbour. The surface amplitudes and phases of the M2 tide are shown in Figure 3.10. The amplitudes increase gradually from about 1.4 m in the outer harbour to about 1.7 m in the arms, then decrease to about 1.0 m in the tidal flats, finally decreasing to almost zero in the mangrove areas. This large decrease in amplitude is due to the dramatic bottom dissipation in the mangrove zones with their high Cd values. The phase increases from the outer harbour to the arms, with the extreme low and high values in the mangrove areas, driven by the wetting-drying process.

Figure 3.10: The M2 tidal amplitudes and phases at the surface.

Figure 3.11 shows the vertically averaged velocities of the spring flood and ebb currents from the model. For clarity, the velocity vectors are interpolated to an averaged grid with horizontal resolution of about 2.0 km. The current velocities in the harbour

3.5 Model results and discussion reach a maximum in Middle Arm, about 2.0 ms-1. In the inner harbour and arms, the water-flow patterns are in accordance with the shoreline. Current speeds fall to zero in the mangrove regions because of the large amount of bottom friction (high Cd value).

Figure 3.11: Vertically averaged velocities vectors of the spring (a) flood and (b) ebb currents at GMT about 1900 hours and 1300 hours, respectively, on 25th June 2009.

Variations in the M2 tidal-current ellipses at the different vertical layers at the four selected locations (Figure 3.1) are shown in Figure 3.12. At the surface, the maximum -1 speed of the M2 tidal current increases from the outer harbour (0.3 ms ) to the channel (0.6 ms-1), then decreases in the inner harbour (0.5 ms-1); it reaches its maximum in -1 Middle Arm, roughly 1.3 ms . The maximum current speeds of the M2 tide decrease from the surface to the bottom at all four locations because of friction at the bottom; Middle Arm experiences the most rapid decrease, 33.5%, as it has the shallowest water, while the location near the channel has the smallest decrease, 24.8%, because of its deep water column.

3.5 Model results and discussion

Figure 3.12: Vertical profile of the M2 tidal-current major axis at four locations shown in Figure 3.1.

The orientations of the M2 tidal-current ellipses at ten vertical layers at the four locations in the harbour are shown in Figure 3.13(a). The current ellipses in all the layers are orientated roughly in the same northwest-southeast direction, from the surface to the bottom, except at Location 1 in the outer harbour. The current ellipse rotates more than 70° in an anticlockwise direction from Location 1 to 4. The current-ellipse orientations are mostly determined by the shoreline and the navigation channel, which has been deepened in a northwest-southeast direction along the channel to East Arm. As Location 1 is in the outer-harbour mouth, away from land, the effect of the shoreline is less there than at locations 2–4. The surface distribution and characteristics of the M2 tidal current ellipses are shown in Figure 3.13(b). The current ellipses in the outer and inner harbours are elliptical, whereas those in the channel and the three arms are rectilinear.

3.5 Model results and discussion

Figure 3.13: M2 tidal-current ellipses at the four locations: (a) current-ellipse orientations; and (b) surface current ellipse distribution.

3.5.2 Effects of mangrove areas and tidal flats on tides

3.5.2.1 Effect on the M2 tide

The results show that if the mangrove areas (Experiment 2: Table 3.3), and mangrove areas and tidal flats (Experiment 3), respectively, are removed (converted to land in the model), the tidal amplitudes in the outer harbour increase by 0.01 m (0.5%) and 0.02 m (1.0%), respectively (compared with the reference, Experiment 1), due to reduced tidal choking (Li et al. 2011); in the inner harbour, the decrease by 0.01 m (0.5%) and 0.02 m (1.0%), respectively, because of the reduced shoaling effect (Figure 3.14).

Compared with Experiment 1, the M2 phase advances (by up to 2° or 4 minutes) when the mangrove areas are removed (Experiment 2), and advances further (by up to 5° or 10 minutes) when the tidal flats are also removed (Experiment 3). Again this is due to a reduced tidal-choking effect. When the inner harbour volume is reduced by converting the mangrove areas and tidal flats to land in the model, the tidal-choking effect is reduced (Figure 3.15).

3.5 Model results and discussion

Figure 3.14: Changes in M2 amplitude and phase when the mangrove areas are removed (Experiment 2 – Experiment 1).

Figure 3.15: Changes in M2 amplitude and phase when the mangrove areas and tidal flats are removed (Experiment 3 – Experiment 1).

3.5 Model results and discussion

3.5.2.2 Effect on the M4 tide

The M4 amplitude differences between Experiment 2 and Experiment 1 (Figure 3.16) increase from the outer to the inner harbour and the arms, with a similar difference between Experiment 3 and Experiment 1 (Figure 3.17). Compared with Experiment 1, the amplitude of M4 in the arms increases by 0.02m (50.0%) when the mangrove areas are removed (Experiment 2), and by 0.03m (75.0%) when the tidal flats are also removed (Experiment 3).

In the inner harbour, when the mangrove areas (Experiment 2) / mangrove areas and tidal flats (Experiment 3) are removed, the M4 phase advances by 20° (18 minutes) / 40° (42 minutes), compared with Experiment 1. However, in the outer harbour, the phases are delayed in both Experiments 2 and 3 compared with Experiment 1.

As the M4 over-tide is generated by the self-interactions of the M2 tide (Egbert et al. 2010), its changes in amplitude and phase result from corresponding changes in M2. In both Experiments 2 and 3, its amplitude increases and its phase advances in the inner harbour compared with the values in Experiment 1.

Figure 3.16: Changes in M4 amplitude and phase when the mangrove areas are removed (Experiment 2 – Experiment 1).

3.5 Model results and discussion

Figure 3.17: Changes in M4 amplitude and phase when mangrove areas and tidal flats are removed (Experiment 3 – Experiment 1).

The data from Location Blay in East Arm were used to examine the variations in the tides in more detail (Tables 3.6 and 3.7). In Experiment 2, the M2 amplitude and phase change only slightly from those of the reference model (Experiment 1); however, the M4 tidal amplitude/phase increases/advances by 50%/16°. Removing the mangrove areas will cause the M2/M4 current speeds to dramatically decrease, by 22.9%/20.0%, compared with those of the reference model; the time at which the maximum M4 current speed occurs is almost 0.5h earlier.

In Experiment 3, small changes occur in the M2 amplitude and phase compared with those of the reference model, whereas the M4 amplitude/phase increases/advances by 75%/28°. The M2/M4 current speeds are reduced by 45.7%/40.0%, compared with those of the reference model; the time at which the maximum M4 current occurs is about 1h earlier.

3.5 Model results and discussion

Table 3.6: Model and observed sea-surface-level amplitudes and phases of the M2 and

M4 tides near Location Blay.

M2 tide M4 tide Experiment/description Amplitude Phase1 Amplitude Phase1 (m) (deg) (m) (deg) Observed BoM Data 1.85 249.3 0.05 107.1 Exp 1 Reference model 1.74 242.5 0.04 55.1 Exp 2 No mangrove areas 1.73 242.3 0.06 39.4 Exp 3 No mangrove areas or tidal flats 1.71 240.3 0.07 27.5 1 Phase angles are relative to GMT

Table 3.7: Predicted and observed vertically averaged tidal current major axes of the M2 and M4 constituents near Location Blay.

M2 tide M4 tide Experiment /description Vmajor1 Phase2 Vmajor1 Phase2 (ms-1) (deg) (ms-1) (deg) Observed Location Blay data 0.35 167.1 0.06 54.1 Exp 1 Reference model 0.35 163.6 0.05 342.6 Exp 2 No mangrove areas 0.27 163.0 0.04 312.9 Exp 3 No mangrove areas 0.19 162.7 0.03 287.7 or tidal flats 1 Vmajor is the tidal current speed along the major axis 2 Phase angles are relative to GMT

3.5 Model results and discussion

According to Equation (3.5), changes in the M2 and M4 amplitudes and phases  lead to changes in tidal asymmetry, measured by MM24/ . Tidal asymmetry in Darwin

Harbour is discussed in the next section.

3.5.3 Effects of the mangrove areas and tidal flats on tidal asymmetry

If the mangrove areas are completely removed in the model (Experiment 2), the tidal  asymmetry skewness MM24/ increases in all water areas, with the maximum increase being about 0.1 (100%) in the arms (Figure 3.18(a)). If the tidal flats are also completely removed (Experiment 3), experiences a todal increase of 0.15 in the arms (Figure

3.18(b)). The same trends occur for  1 , calculated from Equation (3.4).

The increased tidal asymmetry skewness in Experiments 2 and 3 indicates that removal of the mangrove areas or tidal flats can amplify tidal asymmetry, which can lead to greater flood dominance in the harbour.

The mangrove areas are removed incrementally from Experiment 4 to Experiment 6 to further examine their impact on tidal asymmetry in the harbour. Compared with Experiment 1, the asymmetry skewness in all water areas increases when first the mangrove areas around East Arm are removed (Experiment 4: about 30% of the total mangrove area removed). The skewness increases progressively with further removal of the mangrove areas on the east side of Middle Arm (Experiment 5: 50% of the total mangrove area removed) and then all around Middle Arm (Experiment 6: 70% of the total mangrove area removed). The maximum increase in tidal asymmetry occurs if 100% of the mangrove areas are removed, as already shown in Experiment 2.

The variation in tidal asymmetry skewness in East Arm near Location Blay as a function of percentage mangrove removal is shown in Figure 3.19. The relationship between percentage removal and the asymmetry is approximately linear. The tidal asymmetry skewness near Location Blay increases by 0.06, about 60%, compared with that of the reference model, if all the mangrove areas are removed. Therefore, the mangrove areas and tidal flats serve as important buffer zones for dampening tidal asymmetry, which affects the sediment-transport patterns in the harbour.

3.5 Model results and discussion

Figure 3.18: Changes in tidal asymmetry skewness  : (a) when all the mangrove MM24/ areas are removed (Experiment 2 – Experiment 1); (b) when all the mangrove areas and tidal flats are removed (Experiment 3 – Experiment 1).

Figure 3.19: Tidal asymmetry as a function of the percentage of mangrove areas removed in East Arm near Location Blay.

3.5 Model results and discussion

3.5.4 Effects of mangrove areas and tidal flats on tidal energy and bottom dissipation

The tidal energy density per unit area, averaged over a tidal cycle, is calculated from the model elevation and current velocity by (Byun et al. 2004):

1 E  g a2  d(), u 2  v 2  (3.6) 4 where  is the water density, u and v the amplitudes of the barotropic tidal currents, a the tidal elevation amplitude and d the mean water depth. The energy density of the barotropic tides indicates the amount of energy available for mixing.

The changes in tidal energy density between Experiment 2, Experiment 3 and the reference Experiment 1 are shown in Figure 3.20. Compared with Experiment 1, in the outer harbour the energy density increases by 500 Wm-2 in Experiment 2 and by 1000Wm-2 in Experiment 3. In the inner harbour and arms, the energy density decreases by more than 2000 Wm-2 in Experiment 2 and more than 4000 Wm-2 in Experiment 3.

3.5 Model results and discussion

The turbulent energy dissipation, averaged over 31 days, is calculated according to Simpson and Hunter (1974) as

1 T  U dt, (3.7) T bb 0 TT 0

where  b is the bottom shear stress, U b the velocity near the bottom, T0 = 0 and T = 31 days. Figure 3.21 shows the bottom energy dissipation changes between Experiments 2 and 1 and between Experiments 3 and 1. Compared with Experiment 1, less energy (~0.05 Wm-2) is dissipated in the inner harbour if the mangrove areas are removed (Experiment 2); bottom energy dissipation is further reduced if the tidal flats are also removed (Experiment 3).

According to the above discussion, the removal of mangrove areas and tidal flats could increase the energy in the outer harbour but reduce it in the channel, in the inner harbour and the arms. Mangrove areas and tidal flats store energy, as well as increase bottom dissipation, by enhancing tidal-choking and shoaling effects, with the energy- storing effect larger than that of energy dissipation.

3.6 Conclusions

An unstructured-grid coastal ocean model (FVCOM) was employed to simulate the hydrodynamics in Darwin Harbour: the observed tides were well reproduced by this model. The model M2 and S2 amplitudes, phases and maximum tidal current speeds at Location Blay are in good agreement with the observed values in both magnitude and trend from the bottom to the surface. The model therefore provides a reasonable simulation of the tidal dynamics in the harbour.

The model results indicate that the hydrodynamics of Darwin Harbour are complex and driven mainly by tides, with the effects of wind and rivers being small.

The harbour is semidiurnal: the M2 tide is predominant, with an amplitude of about 1.7 m, followed by the S2 tide. Current flow is dominated by tides; the preliminary modelling analysis demonstrated that the current speed reaches a maximum of about 2.0 ms-1 at the surface in Middle Arm. The current speed decreases gradually from the surface to the bottom. Middle Arm (shallower than 10 m) and the channel (about 25 m deep) have the largest (33.5%) and smallest (24.8%) decreases, respectively.

A series of numerical experiments was carried out to help understand the overall physical processes in the harbour, especially the effects of the mangrove areas and tidal flats on the tides. These experiments, sensitivity tests, indicated that the mangrove areas and tidal flats play key roles in modulating the tides and currents in Darwin Harbour.

Removal of mangrove areas and tidal flats would decrease the M2 amplitude due to the decreased shoaling effects in the inner harbour, but would increase the M4 amplitude.

With mangrove areas and tidal flats removed, the amplitude of M2 would increase/decrease slightly (< 1%) in the outer harbour/inner harbour and the arms, respectively, and the M2 phase advance by a maximum of 4° over all the water areas.

The M4 amplitude would increase by 75% in the arms and the phase increase/decrease by up to 30°/40° in the outer harbour/inner harbour and the arms, respectively.

Mangrove areas and tidal flats affect tidal asymmetry due to their effect on the amplitudes and phases of the tides. This study has shown that these areas significantly reduce tidal asymmetry: for example, the tidal elevation skewness would increase by 100% and 120% in Middle Arm, if the mangrove areas and the mangroves plus tidal

3.6 Conclusions flats, respectively, were removed. The skewness near Location Blay in East Arm varies roughly linearly with the percentage of mangroves removed. As tidal asymmetry strongly affects sediment transport in the estuaries, care must be taken with any reclamation of the mangrove areas and tidal flats around the harbour watershed. An overall understanding of the hydrodynamics of Darwin Harbour will benefit the study of its sedimentary dynamics, the subject of Chapter 4.

CHAPTER 4 SUSPENDED-SEDIMENT DYNAMICS IN DARWIN HARBOUR

4.1 Introduction

An understanding of suspended-sediment transport in harbours is essential for harbour management and future development, such as wharf planning and coastal construction; the spatial and temporal variability of suspended sediment has hydrographic implications, in particular for maintaining a navigable depth in harbours. Although estuaries like Darwin Harbour contain both non-cohesive and cohesive sediments, it is the dynamics of cohesive fine sediments in a tidal estuary which leads to specific problems for marine industry and the coastal economy. For example, suspended- sediment deposits may sequester heavy metals, pesticides and other pollutants that are readily adsorbed onto surfaces of the particles (Uluturhan et al. 2011). In estuaries, the net transport direction of fine-grained sediment is generally towards the estuary head, caused by estuarine processes including settling and scour lag, tidal pumping and the flocculation process (Dyer 1997; Pejrup 1988). This upstream transport of fine sediment makes the estuaries shallower, e.g. the accumulation of the sandbar in East Arm of Darwin Harbour (Williams 2009), which might reduce access for shipping in the long run. Therefore, suspended-sediment research has received widespread and increasing attention worldwide (Allen et al. 1980; Kessel et al. 2011; Manning et al. 2010; Margvelashvili et al. 2003). An analysis of the suspended-sediment dynamics in Darwin Harbour is of practical significance.

In this chapter, a sediment model (Wang 2002), which focuses on fine-sediment suspension and re-suspension, is coupled to the FVCOM hydrodynamics model of Darwin Harbour (Chapter 3) to study suspended-sediment dynamics in the harbour. The effects of mangrove areas and tidal flats on suspended-sediment dynamics in the harbour are revealed in this chapter. Section 4.2 outlines the methodology; an analysis of the field data, model development and configuration, and a method to explore the factors controlling suspended-sediment transport are provided. Calibration of the model is in Section 4.3. Observed sea-surface level, currents, and suspended-sediment concentration (SSC) data from November 2012 are used to calibrate the sediment model. The model results are described in Section 4.4, including the distributions of horizontal

4.2 Methodology and vertical SSC, the net suspended-sediment transport and SSC time variation. Section 4.5 models the effects of mangrove areas and tidal flats on suspended-sediment dynamics. The impact of dredging near East Arm Wharf on the harbour suspended- sediment dynamics is modelled. The results are discussed in Section 4.5. Conclusions are given in Section 4.6.

4.2 Methodology

4.2.1 Analysis of field data

In order to obtain a comprehensive picture of the suspended-sediment situation in the harbour, and to calibrate and validate the sediment model, suspended-sediment concentration data were measured at seven locations, MA1, MA2, EA1, EA2, WA1, CL and CR, in the outer harbour and the arms (Figure 4.1). These locations are selected as a scientific integral to bracket all of the harbour but with practical feasibility to avoid very strong currents and heavily trafficked shipping lane. Seven bottom nephelometers were deployed about 0.4 m above the seabed. Each nephelometer contained a pressure sensor, a deposition sensor, a turbidity sensor, a light sensor and a temperature sensor. These data were measured at 10-minute intervals from early November 2012 to early February 2013. This time period is chosen to cover the transient time from dry season to wet season to examine the sediment bottom boundary layer. However, after mid-December 2012, the dredging activities for the East Arm Wharf expansion compromised the SSC records and reduced the data quality. For this reason, only the observed data from early November to early December in 2012 are reported herein and used to calibrate the sediment model. As there is not much rain during November, the beginning of the wet season, river discharge is not considered. The locations of instruments deployed are shown in Table 3.1.

4.2 Methodology

The observed suspended-sediment concentrations from 7th November to 7th December 2012 at the seven locations are shown in Figure 4.2. The SSC varies with the spring and neap tidal cycles, with large SSC values during spring tides and small SSC values during neap tides. The SSC values on the harbour bottom during spring tides can be up to 100 times larger than during neap tides, but still only of the order of 10-3 kgm-3 to 10-4 kgm-3. The most turbid location is CL, with SSC values of more than 0.1 kgm-3, followed by locations WA1 and EA1. 65

4.2 Methodology

Abnormally high or low SSC values, for example, at EA2 on 14th November (~0.3 kgm-3, Figure 4.2(b)) and at MA2 on 18th November (0 kgm-3, Figure 4.2(c)), are smoothed using a Gaussian low-pass filter before these data are used in model: these abnormally high or low values only occur over short time periods, and are largely the result of special events, e.g. construction or the failure of instruments, rather than the natural suspended-sediment dynamics.

The SSC values in the second spring tidal cycle are normally no more than 50% of those in the first spring tidal cycle, as shown in Figure 4.2 (a) (c) and (d). (The exception is at EA1 Figure 4.2(b)). This is because of the larger bottom-current speeds induced by the larger tidal range in the first spring tidal cycle compared with the second cycle, for example, at EA2 and MA2 (Figure 4.3).

Figure 4.2: Observed suspended-sediment concentrations at the seven locations in the harbour.

4.2 Methodology

Figure 4.3: Observed bottom-current speeds in East Arm (EA2) and Middle Arm (MA2).

In East Arm, Location EA1 has relatively high turbidity throughout the entire observation period, especially during the second spring tidal cycle (Figure 4.2(b)). The SSC reaches up to 0.09 kgm-3, in spikes in November, with one abnormally large spike on 14th November reaching 0.25 kgm-3. This variation is abnormal compared with the suspended-sediment concentration values at other six locations, where the second spring tidal cycle has lower SSC values than that in the first spring tidal cycle. At Location EA2, spikes of high SSC values appear in the spring tidal cycle.

As observed during field work, this abnormally high SSC value at EA1 was caused by dredging near East Arm Wharf. This dredging had a greater effect in the second spring tidal cycle because at that time more sediment was lifted into water column than in the first spring tidal cycle. These suspended sediments at EA1 were further advected to EA2 by strong flooding currents during the spring tide, and caused large peaks in the SSC there. This effect was reproduced by this study in Section 4.5.3.

In the outer harbour, obvious spikes of high SSC occurred at both locations CL and CR in the spring tidal cycle, shown in Figure 4.2(a). These high SSC spikes were related to the disposal of sediment from the dredging, as the sediment can be transported from the offshore disposal area outside the harbour to locations CL and CR. This effect

4.2 Methodology was modelled, and is discussed in Section 4.5.4, supplementing the 2-D simulation of HR Wallingford (2010b).

Field observation shows that the water is normally free of sediment in the channel and in the main channel of Middle Arm. A group of well-developed coral reefs, which can only survive in water with high clarity, lies near Channel Island in Middle Arm.

These sediment distribution characteristics and observed SSC data are used to set up the sediment model and calibrate the model results.

4.2.2 Model development

4.2.2.1 Sediment dynamics model

Sediment processes are parameterised following the Wang (2002) model to considers the Estuarine Suspended Sediments (ESSed); sediment transport is described by

C ()() uC  vC()w w C   C   s (),KF  (4.1) t  x  y  z  zhc  z where w is the vertical water velocity, ws the settling velocity of suspended sediment and C the suspended-sediment concentration. Kh is the vertical eddy diffusivity for suspended sediment and Fc is the horizontal diffusion term parameterised according to the Smagorinsky diffusion scheme. The Smolarkiewicz iterative anti-diffusive scheme is used for sediment advection to reduce implicit diffusion (Smolarkiewicz 1984).

The density of clear seawater (without sediment) is determined by the equation of state and, when the contribution of the suspended sediments is considered, calculated by a volumetric relationship

w w 1,  C (4.2) s

where w is the density of clear seawater and s the sediment density. The bottom drag coefficient in a sediment-laden bottom boundary layer is given by

2  1 Dz b Cd  ln , (4.3)  /1 AR z  f  0

4.2 Methodology

where zb is depth of the bottom σ – layer (σ = – 0.991 in this study) and z0 is the bottom roughness length. The effect of stratification is specified by a stability function, 1 + ARf, where A is an empirical constant and Rf is the flux Richardson number. Adams and Weatherly (1981) determined A = 5.5 for a sediment-laden oceanic bottom boundary layer.

The vertical sediment flux, E, on the seabed due to erosion/deposition processes is, according to Ariathurai and Krone (1976),

  b E0 1, bc   c  E   , (4.4)   b Cw 1,   b s b d   d where E0 is the erosion coefficient, τc and τd the critical stress for re-suspension and deposition, respectively, and Cb the sediment concentration in the model’s bottom layer. The continuous bottom exchange of sediment between the seabed and the water column through erosion and deposition is a function of spatial shear stress, which varies both in space and time. Refer to Wang (2002) for more details of the model.

4.2.2.2 FVCOM- ESSed coupling

The ESSed model is incorporated here into the FVCOM sediment module to construct the FVCOM-ESSed model. This model, following Wang (2002), focuses on suspended sediment for the following reasons. First, as described in Section 1.3.3, fine sediment is present in the harbour, especially along the east coast and in East Arm. These locations are the center of the Darwin marine economy, with almost all the marine facilities around Darwin City and the wharfs. Second, it is the suspended sediment which is the main threat to water quality in the harbour. For example, turbid water in the harbour is not good for coral reefs: the increase of sediment deposition rate reduces the recruitment rate near southern islands in Singapore (e.g. Dikou & van Woesik 2006). Third, an understanding of sediment deposition, re-suspension and transport is important for harbour development, e.g. the maintaining the navigation channel. Therefore, the ESSed model is well suited to Darwin Harbour. This model uses the basic settings of the FVCOM sediment module, with some structures revised and improved. The main characteristics of ESSed are summarised as follows.

4.2 Methodology

Basic equations

The basic equations of Wang (2002) are incorporated into FVCOM (Bao Min, 2010, personal communication), including calculations of density, bottom drag coefficient and Richardson number. The re-suspending and settling algorithm uses Equation (4.4), focusing on fine sediments. In order to improve the accuracy of the upwind scheme, the multidimensional positive definite advection transport algorithm of Smolarkiewicz (1984) is used to calculate the vertical sediment advection.

Wetting-drying processes

The wetting-drying processes are included in the model as follows: if the water depth is less than a critical depth of 0.5 m, the node is considered as dry by the sediment module and the SSC value in the water column is not updated from the previous step accordingly. The sediment on the seabed remains there to keep mass balance. Advection, erosion and deposition do not occur at dry nodes. When the node becomes wet again, the SSC value is the summation of SSC values at the previous wet time step of this node, the sediment from bottom erosion and deposition, the sediment from advection and diffusion.

Bathymetry localisation

The FVCOM-ESSed model can recognise different bathymetry types, indexed by their respective parameters, which are provided as input to the model by a bathymetry control file. Any number of bathymetry types may be parameterised and assigned to the model grids in the control file according to requirements.

In this study, three bathymetry types are distinguished: water areas; tidal flats; and mangrove areas. Different erosion and deposition parameters are assigned to tidal flats and mangrove areas to characterise their sediment properties.

Seabed thickness

Different initial conditions for seabed thickness, used to characterize its spatial variability, are read into the model from an input seabed-thickness file. If the bed thickness is less than the critical value, set by the user, deposition is allowed, but there is no erosion.

4.2 Methodology

Disposed materials

Disposed materials are included in the model by reading in user-specified time series of disposed-material concentrations. The dumping location, dumping time interval, the disposed-material concentration and discharge occasions are pre-assigned by the user before starting the model. The disposed-material concentration is assigned either to the whole water column or to the sea surface at the dumping location at the dumping times.

4.2.2.3 Sediment model sensitivity to realistic parameters

The sediment model sensitivity is tested to examine the variation of the model results with main realistic parameters of critical erosion and deposition stress, the erosion rate and the bottom fine sediment thickness.

Above all, the bottom fine sediment thickness plays a key role in deciding the SSC values in the water column. The thickness value determines the fine sediment availability on the seabed, and consequently limits the suspended sediment amount in the water column. If the bottom fine sediment thickness is not limited, i.e. infinite fine sediment on the seabed, the SSC values suspended in the water column vary at a similar ratio with that of the erosion rate.

Increase the critical erosion stress value reduces the SSC values in the water column. Double this critical erosion stress reduces less than half the SSC values. Increase the critical deposition stress value also reduces the SSC values in the water column, but the reduced amount is not clear, as the deposition amount is determined by this critical deposition stress, bottom SSC values as well as settling velocity (Equation 4.4).

More information of the sediment model sensitivity to realistic parameters has already been provided in previous research, e.g. Wang (2002), Wang et al. (2005), and Wang & Pinardi (2002).

4.2.3 Model configuration

4.2.3.1 Initial conditions

This study focuses on the sediment dynamics of Darwin Harbour in the dry season from Charles Point to Lee Point, with the model domain expanded to include areas outside the mouth of the harbour. The mangrove areas are all included in the domain in order to

4.2 Methodology investigate the effect of mangrove areas and tidal flats on sediment transport in the harbour. Details of the model bathymetry and grids configuration are referred in Section 3.2.3.1.

The sediment model begins after the hydrodynamic model has run for 12 (model) hours. Only cohesive suspended sediments are included as they constitute the bulk of the suspended sediments and are more easily transported around the harbour (Brenon & Le Hir 1999). An effective approximation is to treat suspended sediments in Darwin Harbour as a single group with a particle size of 0.002 mm (Wang 2002), as this group represents most of the fine sediments found in the harbour (Fortune 2006). Consequently, the settling velocity is represented by an average value of 10-4 ms-1. Using this settling velocity in the model gave a good fit to the time series of the observed bottom SSC values.

A constant settling velocity is suitable for Darwin Harbour, which has low SSC, because flocculation only become significant when the SSC values rise to about 1 kgm-3 (Van Rijn 1993). The critical erosion and deposition stresses are set to 0.1 Nm-2 and 0.08 Nm-2 according to field experience, which has been observed to range between 0.02 – 5.0 Nm-2 for erosion and 0.06 – 0.1 Nm-2 for deposition (HR Wallingford 2010b). The critical erosion and deposition stress values are tested to best fit the overall trend of the field data of SSC time series. Smaller erosion rates (about 1/50 of that in water areas) and larger critical erosion values (about 10 times that in water areas) are assigned to the tidal flats, because of the surface properties of these areas. Mangrove areas are treated as a sink for suspended sediment in the model: only deposition is allowed whereas erosion is prohibited. The model is run for 40 days, from 1st November 2012 to 10th December 2012; its initial conditions and constants are listed in Table 4.1.

4.2 Methodology

Table 4.1: Model initial conditions and constants. Parameters Description Number of seabed layers 1 Sediment type Cohesive

Erosion rate 5 106 kgm-2s-1 in water areas

107 kgm-2s-1 in tidal flats Critical erosion stress 0.1 Nm-2 in water areas 1.0 Nm-2 in tidal flats and mangrove areas Critical deposition stress 0.08 Nm-2

4.2.3.2 Test cases

Two cases were considered for modelling the suspended-sediment transport dynamics. As there were no seabed data such as bed thickness available for this study, in Case 1 we assumed an initial uniform unlimited thickness of the fine-sediment layer for the 40- day model run. The erosion and deposition patterns at the bottom obtained from the results of this run were then input as the thickness of the fine-sediment layer for Case 2. The most eroded grid location from Case 1 was assigned zero thickness because any fine sediment there is eroded in the long run; there is little fine sediment at the bottom. The thickness at the grid location with the greatest deposit was taken to be the sum of the deposition and erosion thicknesses from Case 1. The thickness of the fine-sediment layer at other grid locations in the harbour was set by linear interpolation between the thicknesses of the most eroded and most deposited locations.

In some areas, the thickness was adjusted according to the findings of Fortune (2006) and our own field observations. The thickness of the fine-sediment layer in the channel and at the entrance to Middle Arm was set to zero because the water there has low turbidity and the seabed experienced high erosion in Case 1. The thicknesses in the outer and inner harbour were set to low values, following the variations from Case 1, because the sediment on the bed there is coarser than in the arms. The detailed initial conditions for the thickness of the fine-sediment layer are shown in Figure 4.4.

4.2 Methodology

Figure 4.4: The initial thickness of the fine-sediment layer used for Case 2.

In Case 2, the model with entire model domain discussed above is referred to as Experiment A. In order to examine the effect of the mangrove areas and tidal flats on sediment transport, Experiment B was conducted with the same settings as Experiment A, except that the tidal flats and mangrove areas were removed from the model domain. Experiment A1 and A2 were designed to test the effect of the dredging activity and the disposal materials on SSC values based on Experiment A. Sediment was released at the dredging or dumping location to mimic these activities, while no erosion and re- suspension was allowed in both Experiment A1 and A2. These cases and experiments are described in Table 4.2.

4.2 Methodology

Table 4.2: Test Cases and Experiments descriptions. Cases Experiments Descriptions

1 Uniform bed thickness; input to Case 2.

2 A Reference model; initial seabed condition from Case 1.

2 A1 Experiment A, but including dredging; no erosion and re-suspension. 2 A2 Experiment A, but including disposal of dredging material; no erosion and re-suspension.

2 B Experiment A, but with mangrove areas and tidal flats removed.

4.2.4 Sediment flux decomposition

According to the method of mass transport flux of Dyer (1997), if the short-period turbulence is neglected, the velocity, u , and sediment concentration, c , at any depth can be written as u u uv and c c cv , where uv and cv are the deviations at

1 h 1 h any depth z from the mean values u u dz and c c dz , respectively, h is h 0 h 0 the water depth. As u and c will vary over tidal cycles with tidal fluctuations, they can be expressed by the sum of the tidally averaged value and its deviation as u u0 ut

1 T 1 T and c c0 ct , respectively, where u0  u dt and c0  c dt . T is the tidal T 0 T 0 period, u0 and c0 the mean vertically averaged velocity and sediment concentration over the tidal cycle, respectively, and ut and ct the corresponding deviations of the vertically averaged values from the means.

The instantaneous sediment flux through a unit width of a section perpendicular to

h 1 the mean flow is given by F uc dz huc d , where  is the normalised distance 00 from the seabed ( 1) to the water surface (  0 ). The net sediment flux over a tidal cycle can be partitioned into seven major fluxes as:

4.3 Model calibration

1 T 0 F huc d dt T 01

huc00 0  chu 0tt  uhc 0 t t  huc 0 t t  huc t t t  huc 0 v v  huc t v v (4.5)

TTTTTTT1  2  3  4  5  6  7 , where the brackets, , denote the tidally averaged value of a vertically averaged variable, the overbar denotes a vertically averaged value and h = h0 + ht, where h0 and ht are the tidally averaged water depth and its deviation, respectively. T1 is the flux due to the non-tidal drift, the Eulerian velocity, and T2 the flux due to Stokes drift. T3 + T4 + T5 are the tidal pumping terms produced by the tidal phase differences; T3 is the correlation term between the tidal level and sediment concentration; T4 arises mainly from the consequence of a sediment erosion threshold and lags, which result from sediment re- suspension and deposition; T5 , the correlation term between sea surface level, current velocity and sediment concentration, expresses the role of tidal trapping. T6 is the vertical gravitational circulation, arising from the correlation between the landward (e.g. directly upstream of the estuary) bottom mean flow with a high near-bed sediment concentration and the seaward mean surface flow with a lower concentration. T7 arises from the changing forms of the vertical profiles of current velocity and sediment concentration in the tide due mainly to the lag between scouring and settling activities (Wai et al. 2004, pp. 443-444). This method is used to identify the different sediment- transport mechanisms in Darwin Harbour.

4.3 Model calibration

A calibration of the sediment model was conducted to obtain a set of parameters to set up the model and improve model accuracy; observed suspended-sediment concentrations at the bottom level at seven locations (Figure 4.1) were used. The root mean square error (RMSE) is used to measure the model errors,

T ()fo 2 RMSE  t1 tt, (4.6) T where f and o in this case are the model and observed values of SSC. A value of the RMSE near zero indicates a close match between the observed and model values of SSC.

4.3 Model calibration

In addition to the RMSE statistics, the anomaly correlation coefficient (ACC) is used to quantify the temporal and spatial correlation between model and observed SSC values from the climatology mean based on the method by Krishnamurti et al. (2003):

M ()()fm ct m o m ct m m1 ACC  1 , (4.7) MM 222  fm ct m  o m ct m  mm11 where f, o and ct are the model values, observed values and the climatology mean values of SSC; the summation m is over time or spatial location. The climatology mean of the SSC is the observed mean SSC value in one lunar month of November 2011. The ACC measures the pattern similarity considering the error and bias. A value of 0.6 is the limit below which the prediction is thought to be less valuable.

As shown in Figure 4.5, the sediment model reproduces the SSC reasonably well, matching both the observed concentrations and their variation with the tides at almost all of the seven locations. However, there are large discrepancies at locations EA1 and CR (Figure 4.5(c) and (g)). At EA1, the model SSC values cannot match the high observed values in the second spring tidal cycle, as evidenced by the largest RMSE value at all locations, 0.02 kgm-3, occurring at location EA1 (Figure 4.6(a)).

The lowest ACC value among the seven locations, 0.5, also occurs at EA1 (Figure 4.6(b)). As discussed in Section 4.2.1, the abnormally large SSC values observed at EA1 were caused by dredging during the observation period near East Arm Wharf.

Station CR is located near an offshore dumping zone for dredging materials from the Ichthys Project (HR Wallingford 2010b). The significant mismatch, shown by low ACC value of 0.53, between modelled and observed SSC at CR during the first spring tidal cycle (Figure 4.5(c)) is caused by advection of the dumped sediments by the flood currents. Further discussion on this dumping activity and its effect on SSC distribution in the harbour is presented in Section 4.5.4.

More than 75% of the ACC values through the simulation time are larger than 0.6 during the simulation time (Figure 4.6(c)), with values smaller than 0.6 occurring when the model and observed values have different signs relative to the lunar monthly mean. These lower ACC values are mostly caused by the unexpectedly high SSC values at

4.3 Model calibration

EA1 and EA2. If the EA1 and EA2 data are not included, about 80% of the ACC values are larger than 0.6, indicating an acceptable spatial similarity between model and observation.

Therefore, the sediment model is sufficiently accurate for an analysis of the suspended-sediment dynamics in the harbour, including time and spatial variations in the distributions of suspended sediments and net sediment flux.

4.3 Model calibration

Figure 4.5: Comparison of model and observed time series of bottom suspended-sediment concentrations at the seven locations.

4.4 Model results

4.4.1 Suspended-sediment distribution in Darwin Harbour

In Experiment A (Table 4.2), the reference experiment, the SSC at the bottom level is high in the outer harbour, the inner harbour, East Arm and West Arm during the spring tidal cycle (Figure 4.7(a)), as a result of the strong currents coupled with the availability of erodible fine sediment in the harbour. The SSC near the west coast in the outer harbour is higher than that near the east coast. This is due to that the currents are stronger near the west coast, which is near the deep navigation channel, compared with that near the east coast. This phenomenon is also indicated by the field data as shown in

4.4 Model results

Figure 4.2(a). The turbid zone in the harbour moves landward and seaward with the flood and ebb tidal cycles, respectively. The SSC in the turbid zone in the outer harbour is as high as about 0.1 kgm-3 at peak flood current. This turbid zone has higher SSC values at low slack water than at high slack water, because the suspended sediment is diluted by the larger water volume during high slack water. The SSC in East Arm is high throughout the spring tidal cycle, because East Arm has an erodible muddy seabed and relatively strong currents. The suspended sediment in the inner harbour is from both bottom sediment erosion and sediment advection from East Arm. The very high SSC values in the tidal flats and mangrove areas are generated by wetting-drying processes. During low water, these areas are dry, and the model SSC is zero; during high water, these areas are under water, and the SSC is high due to the shallow water and high bottom shear stress. The water in the channel and in Middle Arm is clear due to low sediment availability, in accordance with our field observations. The water is also clear near Channel Island in Middle Arm (Figure 4.1), where the coral reefs are well developed (Fortune & Drewry 2011).

The suspended sediment has a similar distribution at the sea-surface level to that at the bottom level (Figure 4.7(b)). However, the SSC values at the surface are less than 0.07 kgm-3, which is only about 50% of those at the bottom level. The SSC along the navigation channel in the outer harbour is lower than in its vicinity, because the sediment is transported away by strong currents, e.g. in the area outside the channel indicated by A in Figure 4.7(b).

4.4 Model results

Figure 4.7: Suspended-sediment concentration at (a) the bottom level and (b) the surface level in the spring tidal cycle on 16th November 2012.

4.4 Model results

In contrast to the spring tides, the less energetic neap tides do not erode much sediment in the upper part of the harbour, and the bottom SSC in the inner harbour and the arms is mostly less than 0.005 kgm-3 (Figure 4.8(a)). Although the turbid zone in the outer harbour still exists during the neap tidal cycle, its SSC values are less than 0.02 kgm-3 as a result of the relatively strong currents. This is less than 20% of the SSC during the spring tidal cycle. The turbid zone still moves landward and seaward with flood and ebb current cycles, respectively, as a result of current advection. At the surface level the suspended sediment has a similar distribution to that at the bottom level (Figure 4.8(b)), but the SSC is only one third of that at the bottom level. There is net deposition of sediments during a neap tidal cycle, net erosion during a spring tidal cycle.

4.4 Model results

Figure 4.8: Suspended-sediment concentration at (a) the bottom level and (b) the surface level in the neap tidal cycle on 7th December 2012.

4.4 Model results

4.4.2 Vertical suspended-sediment concentration profile in Darwin Harbour

Figure 4.9 shows the vertical SSC profile along Cross-section 3 from the outer harbour to Middle Arm (Figure 4.1) during a spring tidal cycle. The turbid zone in the outer harbour has higher SSC values than near the entrance to Middle Arm, indicated by A in Figure 4.9. The turbid zone in the outer harbour outside the channel (the deepest area in Figure 4.9) moves landward (to the right) and seaward (to the left) with the current cycles, as shown in Figure 4.10(a) and (b), respectively. The turbid zone near the entrance to Middle Arm similarly moves up and down along its channel.

During a spring tidal cycle, sediments tend to be well-mixed throughout the water column due to strong currents. The bottom SSC values at peak flood currents are higher than those at low slack water because of re-suspension.

During a neap tidal cycle, although the water is still turbid in the outer harbour, SSC values are high only at the bottom level, as the current speeds are low, only one third of those during a spring tidal cycle (Figure 4.10 (c) and (d)). As a consequence, the turbid zone at the entrance to Middle Arm disappears and vertical mixing is weaker than during a spring tidal cycle. Therefore, a vertical SSC gradient develops during a neap tidal cycle (Figure 4.11).

Figure 4.9: Vertical profiles of suspended-sediment concentration along Cross-section 3 in the spring tidal cycle on 16th November 2012. ‘A’ indicates the entrance to Middle Arm.

4.4 Model results

Figure 4.10: Vertical profiles of current speed along Cross-section 3 during: (a) spring peak flood; (b) spring peak ebb; (c) neap peak flood; and (d) neap peak ebb.

Figure 4.11: Vertical profiles of suspended-sediment concentration along Cross-section 3 in the neap tidal cycle on 7th December 2012. ‘A’ indicates the entrance to Middle Arm.

4.4 Model results

4.4.3 Net sediment transport in Darwin Harbour

Figure 4.12 shows the vertical profiles and time series of residual currents and net sediment fluxes along Cross-section 1 (Figure 4.1). Averaged over one lunar month, the mean residual current is almost zero (Figure 4.12(a) and (b)). The net sediment flux is approximately 107,263 t/month seaward (Figure 4.12(c) and (d)). The net sediment flux magnitude and direction are in accordance with the observations of Williams et al. (2006). The largest seaward sediment fluxes are near Mandorah Point and the bottom of East Point (Figure 4.1). In the middle of Cross-section 1, the sediment fluxes at the surface are slightly seaward, at the bottom level slightly landward. The net sediment transport through Cross-section 1 is dominated by residual currents, and the distributions of seaward and landward sediment transport are in accordance with those of the residual currents.

In the entrance to East Arm (Cross-section 2, Figure 4.1), the mean residual current averaged over one lunar month is also nearly zero, and the net sediment flux is 18,542 t/month landward. Therefore, in the dry season, most of the outgoing sediment through Cross-section 1 does not come from East Arm.

4.4 Model results

4.4.4 Time variation of suspended-sediment concentration and seabed thickness

Figure 4.13 shows the time series from the model of surface elevation, surface and bottom SSC, and fine-sediment layer thickness at location K (Figure 4.1), selected because of the importance of East Arm Wharf. Here, the SSC has an obvious spring and neap variation, in accordance with the tides, both at the surface and bottom levels. The SSC values during the second spring tidal cycle are much smaller than during the first. The peak SSC only reaches about 25% and 43% of that in the first spring tidal cycle at surface and bottom levels, respectively. This trend is in accordance with the tides, as the second spring tidal cycle is weaker than the first, as indicated by its smaller tidal magnitudes. A similar phenomenon appears in the field data. The peak SSC near East Arm Wharf is 0.08 kgm-3 at the bottom level, about double that at the surface level.

The model time series of the fine-sediment layer thickness at location K is shown in Figure 4.13(d). The layer is eroded throughout a spring-neap tidal cycle. In one month, a total thickness of about 0.003 m of fine sediment is eroded.

Figure 4.13: Time series of (a) surface elevation; (b) surface suspended-sediment concentration (SSC); (c) bottom SSC; and (d) fine-sediment layer thickness at location K near East Arm Wharf.

4.4 Model results

The sediment erosion and deposition in the entire domain after one lunar month are displayed in Figure 4.14. Erosion occurs in the outer harbour, inner harbour, the East Arm and West Arm channels, and in the shallow-water areas, in particular the tidal flats. These erosion zones are the result of strong currents and sediment availability; the surrounding areas are deposition zones.

Along the east coast of the outer harbour, there are mainly deposition areas. Shallow water areas in the mangrove areas are also preferable places for deposition due to current reduction because of high bottom friction generated by the dense mangrove trees. There is no sediment eroded in the channel and in the main channel of Middle Arm, where current is actually very strong, due to the initial null fine sediment thickness there.

Figure 4.14: Erosion and deposition after one lunar month (positive values indicate erosion).

4.5 Discussion

4.5.1 Effect of mangrove areas and tidal flats on suspended-sediment distribution

The removal of the mangrove areas and tidal flats changes the hydrodynamics in the harbour, and consequently, generates a different suspended-sediment transport pattern. During spring tides, with the tidal flats and mangrove areas removed in Experiment B, the turbid zone in the outer harbour still appears (Figure 4.15), but its SSC values at the bottom level are approximately 50% less than those of the turbid zone in Experiment A (mangrove areas and tidal flats present).

Water near Nightcliff in the outer harbour (Figure 4.1) becomes turbid in Experiment B. This turbid zone disappears if the nearby headland, indicated by A in Figure 4.15, is removed from the model domain. As the current speeds in the arms are weakened due to a reduced tidal choking effect, the bottom stress is reduced. Therefore, compared with Experiment A, Experiment B has less suspended sediment in the arms, which leads to reduced SSC values in the inner harbour and arms.

Figure 4.15: Suspended-sediment concentration at the bottom level in Experiment B during the spring tidal cycle on 16th November 2012.

4.5 Discussion

During neap tides, the suspended sediment has similar distributions in Experiment A and B. As the currents during neap tides are weak, except in the outer harbour and the channel, turbidity only appears in the outer harbour, but the SSC values at the bottom level in Experiment B are about 70% less than that in Experiment A (Figure 4.16). Again this is due to reduced currents in the channel caused by a reduced tidal choking when the mangrove areas and tidal flats are removed.

Figure 4.16: Suspended-sediment concentration at the bottom level in Experiment B during the neap tidal cycle on 7th December 2012.

Even with the mangrove areas and tidal flats removed in Experiment B, an erosion zone still appears in the outer harbour, whereas the inner harbour and the arms are mostly deposition zones (Figure 4.17). However, the erosion zones in the inner harbour and the channels of the arms in Experiment B almost disappear compared with Experiment A, as a result of the current reduction caused by reduced tidal choking. An erosion zone near Nightcliff in the outer harbour (Figure 4.1) is formed in Experiment B. Most sediment eroded from coastal areas in the outer harbour in Experiment B is

4.5 Discussion deposited near the channel. Hence, if the mangrove areas and tidal flats are removed, the areas in the outer harbour near the channel will become the main sediment deposition zone.

Figure 4.17: Erosion and deposition accumulated in one lunar month in Experiment B (positive values indicate erosion).

4.5.2 Effect of the mangrove areas and tidal flats on net sediment flux

4.5.2.1 Net sediment fluxes in the channel and the entrance to East Arm

The net sediment flux in the channel and at the entrance to East Arm is examined here to determine the effect of mangrove areas and tidal flats on suspended-sediment transport. In Experiment B, with no mangrove areas or tidal flats, the net sediment flux through Cross-section 1 (Figure 4.1) in the channel is about 5,601 t/month, but landward rather than seaward as in Experiment A (Table 4.3). According to Chapter 3 and Li et al. (2012), flood dominance will be increased in the harbour if the mangrove

4.5 Discussion areas and tidal flats are removed from the model domain. Therefore, over one lunar month, more sediments are transported upstream into the inner harbour through Cross- section 1.

The net sediment flux through the entrance to East Arm (Cross-section 2) is 248 t/month landward, which is in the same direction but with a greatly reduced magnitude (~99% smaller) than in Experiment A. This is due to the reduced currents in the inner harbour and the arms caused by the weakened tidal choking effect when there are no mangrove areas and tidal flats.

Therefore, the reclamation of mangrove areas and tidal flats or coastal construction may cause more sediment to move into the harbour. However, less sediment will be transported upstream into East Arm.

Table 4.3: Comparison of net sediment transport. Cross-section 1 Cross-section 2 Experiments (t/month) (t/month) A 107,263 -18,543 B -5,601 -248 positive values indicate seaward

4.5.2.2 Factors controlling suspended-sediment transport

In order to examine the mechanisms that determine the net sediment transport before and after the removal of the mangrove areas and tidal flats, the factors controlling sediment transport in Experiment A (entire domain) and Experiment B (water areas) were compared. The components of the tidally averaged sediment transport flux through Cross-sections 1 and 2 were extracted from the model using the method of Dyer (1997).

The sediment-transport components through the cross-sections in Experiments A and B are listed in Table 4.4. In the channel (Cross-section 1), the net sediment flux is mostly determined by the Eulerian residual (T1 in Experiment A, but tidal pumping (T4) in Experiment B. The direction of the Eulerian residual changes from seaward in Experiment A to landward in Experiment B, while the second largest flux is caused by tidal pumping (T3 + T4 + T5) in Experiment A which is also seaward. However, in

4.5 Discussion

Experiment B, in which the mangrove areas and tidal flats are moved, the effects of tidal pumping on sediment transport are more important due to the increased flood dominance. At the entrance to East Arm (Cross-section 2), the Eulerian residual and tidal pumping transport sediment landward in both Experiments A and B. The tidal- pumping effect increases to a similar level to the Eulerian residual, when the mangrove areas and tidal flats are removed in Experiment B, because of increased flood dominance.

Table 4.4: Sediment-transport components (kg/s) through Cross-sections 1 and 2 (rounded to one decimal place). Experiment A Experiment B Components Cross-section 1 Cross-section 2 Cross-section 1 Cross-section 2

T1 32.2 -5.3 -0.5 -0.1

T2 0.8 0. 3 0.6 -0.0

T3 0.4 -0.1 -0.0 -0.0

T4 7.5 -2.6 -2.0 -0.1

T5 -0.3 0.2 0.5 -0.0

T6 0.0 -0.0 -0.2 0.0

T7 -0.0 -0.0 0.0 -0.0 Total 40.7 -7.5 -1.5 -0.15 positive values indicate seaward

Figure 4.18 and 4.19 display the variations in the seven sediment-transport components along Cross-sections 1 and 2, respectively. In Experiment A, the tidally induced Eulerian residual dominates the sediment transport in the channel through Cross-section 1. The net sediment transport is landward at the deepest point on Cross- section 1 with no tidal flats and mangrove areas (Experiment B), instead of seaward in Experiment A. The deepest point (Figure 4.18(c)) and the areas near East Point have landward net sediment transport. These net sediment-transport dynamics are in accordance with the residual current distribution shown in Figure 4.12(a).

4.5 Discussion

Figure 4.18: Components of net sediment flux through Cross-section 1 from: (a)

Experiment A; and (b) Experiment B. T1 Eulerian velocity; T2 Stokes drift; T3, T4 and T5 tidal pumping; T6 gravitation circulation; and T7 changing forms of the vertical profiles of velocity and concentration with the tide. (c) Depth: positive indicates seaward direction.

Along Cross-section 2, the Eulerian advection component T1 dominates landward sediment transport in Experiment A, followed by tidal pumping (T3 + T4 + T5) (Figure 4.19(a)). However, the effect of tidal pumping overtakes Eulerian advection near East Arm Wharf (Figure 4.19(b), right-hand side) when the mangrove areas and tidal flats removed, due to the increased flood dominance.

4.5 Discussion

Figure 4.19: As for in Figure 4.18, but along Cross-section 2.

4.5.3 Effect of dredging in East Arm on the suspended-sediment dynamics

As shown by the suspended-sediment concentration (SSC) data in Figure 4.2, the SSC at locations EA1 and EA2 is obviously affected by dredging. This dredging has been modelled by HR Wallingford (2010b) using a 2-D numerical model. To examine further this dredging effect on sediment distribution in the harbour, an additional test (Experiment A1) was conducted to predict the 3-D suspended-sediment dynamics. According to the Ichthys dredging report by HR Wallingford (2010b), the berthing area and the offloading zone (indicated by A and B in Figure 4.20) were dredged during the observation period. A continuous release of sediment at a rate of 0.83 kgs-1 occurred throughout the dredging period. Additional sediment was also released into suspension at a rate of 15 kgs-1 for 52 minutes every 311 minutes. This dredging phase, Phase 3, lasted for three months from mid-September to mid-December 2012.

4.5 Discussion

The Experiment A1 setup is the same as Experiment A in Case 2, but with the dredging process included. An erosion rate of 10-6 kgm-2s-1 throughout the simulation period was assigned to this area, which is indicated by E in the inset to Figure 4.20, to model the sediment release during dredging. The value of 10-6 kgm-2s-1 corresponds approximately to the total sediment release rate reported in HR Wallingford (2010b). The same deposition algorithm as in Experiment A was applied. In the harbour model, deposition was allowed, but no erosion.

4.5 Discussion

A comparison of the model dredging and observed SSC values at the seven locations is shown in Figure 4.21. The model dredging SSC values are higher during neap tides than that during spring tides at EA1 (panel (c)), as the suspended sediment is transported away from EA1 by the strong currents during spring tides. During neap tides, the model dredging SSC values are higher than those observed at EA1 (panel (c). This discrepancy is probably due to the fact that the dredging intensity was reduced during neap tides in practice, whereas the model assumes a constant dredging rate through the entire period. The model spikes in the values at EA2 (panel (d)) have the same phase as the observed values, demonstrating that suspended sediment at EA1 is transported to EA2 during spring tides. Over a month, suspended sediment from the dredging site is transported to all seven locations. In the second spring tidal cycle, more sediment is transported to these locations than in the first.

4.5 Discussion

In order to gain an overall view of the suspended-sediment distribution caused by the dredging activity, the SSC values at the surface and the bottom levels during the second peak spring tide are shown in Figure 4.22. At the surface level, the turbid zone is near East Arm Wharf during peak flood currents and peak ebb currents. It moves further seaward/landward in East Arm at low/high slack water (panel (a)). At the bottom level, the most turbid zone is right at the dredging area. The seaward and landward turbid fronts reach the channel during low slack water and the tip of East Arm during high slack water (panel (b)). Away from the dredging area, the SSC decreases rapidly both at the surface and bottom levels. This model SSC distribution is similar to but of a larger magnitude than that simulated by HR Wallingford (2010b) using a 2-D numerical model, and a 95th percentile SSC criterion. This is probably because the 95th percentile SSC values used in HR Wallingford (2010b) are likely to be exceeded only 5% of the time in each phase of dredging.

100

4.5 Discussion

101

4.5 Discussion

During neap tides, the SSC values at the surface level are lower than during spring tides. At the bottom level, the only turbidity is near the dredging area: the continuously suspended sediment in this area is not transported away because of the weak currents there. Correspondingly less sediment is found in the outer harbour and Middle Arm.

According to the model, dredging causes net suspended-sediment transport seaward through Cross-section 1 in the channel, and landward into East Arm and Middle Arm through Cross-sections 2. However, the model SSC values and magnitudes of the net sediment transport are not necessarily accurate here, as erosion has not been taken into account.

4.5.4 Effect of the disposed material on suspended-sediment dynamics

As shown in Figure 4.5 the model SSC at Location CR does not match the observed data, if only local sediment re-suspension is considered. However, the SSC values at Location CR may also be affected by the dumping of materials in the offshore disposal zone outside the harbour (Figure 4.20). Experiment A2 was designed to test the effect of this dumping on the SSC values at Location CR. The offshore disposal ground is approximately 200 m in diameter and in a water depth of about 20 m (HR Wallingford (2010b). The periods when dumping at the offshore disposal areas occurred match those of the dredging activity in East Arm during the simulation period in November 2012. A 12,000 m3 trailing suction hopper dredger worked for three months in Phase 3 and discharged 367,500 tons of fine sediment at the disposal area. It discharged its load uniformly over a period of 10-15 minute every 311 minutes. A 3,000 m3 barge also worked in Phase 3 discharging 182,100 tons of fine sediment at the disposal areas over 5 minutes every 592 minutes.

Experiment A2 had the same settings as Experiment A in Case 2, but with disposal of sediment at location D included (Figure 4.20). In the model, suspended sediment at a concentration of 50 kgm-3 was released at the surface over ten minutes every six hours at slack water throughout the simulation period to model the actual sediment disposal. This concentration is lower than the average of the actual discharge values because location D is not located right at the disposal area. This approximation is acceptable because the exact values of the SSC were not the concern in this test. In the model, only deposition was allowed, to focus on sediment advection from the disposal

102

4.6 Conclusions area; there was no re-suspension.

The time series of the model SSC values at locations CR and CL at the bottom level are shown in Figure 4.21. The model spikes at Location CR are in phase with the observed values, as shown in panel (g), which demonstrates that the SSC at Location CR is indeed mostly affected by advection rather than local re-suspension. The model SSC spikes also appear at Location CL, as shown in panel (f), and are also in phase with the observed large SSC spikes. The SSC values at the other locations in the inner harbour and the arms are not affected by this disposal activity. Note that a decrease or increase in the SSC values assumed initially at location D has only a very small effect on the extent of the impacted areas in the outer harbour.

The model SSC distribution during spring tides is shown in Figure 4.22. The initial SSC values at the ocean open boundary are set as zero. The radiation condition is used for the SSC open boundary condition. As only the SSC well inside the model domain is concerned (e.g. Station CR and CL), the effects of the open boundary condition on SSC are not considered in this study. The sediment from the disposal area only affects the SSC in the outer harbour during spring tides. An even smaller area is affected during neap tides. The SSC in the inner harbour and the arms is not affected. This SSC distribution in the outer harbour is similar to but of a larger magnitude than that from the 2-D numerical simulation by HR Wallingford (2010b), which showed the 95th percentile in the distribution of the SSC. This difference is probably due to the different sediment release values at the dumping location.

4.6 Conclusions

The suspended-sediment transport dynamics of Darwin Harbour in the dry season were simulated by ESSed, a sediment model (Wang 2002), coupled with FVCOM, a hydrodynamics model. The ESSed model focuses on suspended sediment. Wetting- drying processes were added to determine the effect of tidal flats and mangrove areas on sediment transport. Various bathymetry types were also considered and the initial thickness of the fine-sediment layer varied to model the different geomorphologic conditions.

The observed SSC values from 7th November to 7th December 2012 were used to

103

4.6 Conclusions calibrate and validate the sediment model. The field data show that the suspended- sediment concentration (SSC) at seven locations in the harbour varies with the spring- neap tidal cycle. The SSC values in the second spring tidal cycle are less than 50% of those in the first, due to the stronger currents in the first spring tidal cycle caused by the larger tidal range. The abnormally high SSC values at Location EA1 were caused by dredging for the East Arm Wharf expansion. This dredging also affects the SSC values at EA2, about 7 km upstream, generating large spikes during the spring flood currents. Offshore dumping of the dredging materials also affected SSC values at locations CR and CL around 20 km away in the outer harbour.

The model results are in accordance with the observed SSC at all seven locations in the harbour indicated by RMSE values smaller than 0.02 kgm-3 and ACC values larger than 0.6. There is a mismatch between model and observed SSC values at locations EA1 and CR, but this is due to the dredging for the East Arm Wharf expansion and associated dumping offshore. Therefore, this sediment model is well suited to modelling the suspended-sediment dynamics in the harbour.

According to the model results, there is a turbid zone in the outer harbour with SSC about 0.1 kgm-3 at the bottom level during spring tides. The water is also turbid in the inner harbour, and in the East Arm and West Arm channels during the spring tidal cycle. The turbid zone in the outer harbour still appears during neap tides, but the bottom SSC values only reach about 0.07 kgm-3. Water in the inner harbour and the East Arm and West Arm channels is less turbid during neap tidal cycles due to weaker currents. These turbidity zones are formed by strong currents and sediment availability. Flood and ebb current cycles move the turbid zone in the outer harbour landward and seaward, respectively, by advection. Vertically, sediments are well mixed during spring tides, but vertical gradients of suspended sediment are formed in the channel during neap tides. Net sediment transport is seaward in the channel and landward at the entrance to East Arm, dominated by Eulerian advection. Erosion occurs in the outer harbour and in the East Arm and West Arm channels due to the high bottom stress related to strong currents and erodible sediment on the harbour bed. The mangrove areas are also preferred places for deposition due to the slower currents there, a result of the high bottom friction generated by the dense mangroves. There is no erosion in the channel and in the main channel of Middle Arm, despite strong currents, because of the

104

4.6 Conclusions lack of fine sediment there.

When the tidal flats and mangrove areas are removed from the model domain, the turbid zone in the outer harbour remains during spring tides, but the SSC values at the bottom level are reduced by 50%. The water in the inner harbour, East Arm and West Arm is less turbid as a result of a reduction in the tidal current speeds due to a reduced tidal choking effect. The net sediment flux in the channel is in the landward direction instead of the normal seaward direction due to the increased effect of flood dominance. The landward net sediment flux is reduced at the East Arm entrance as a result of reduced currents in the arms; tidal pumping dominates sediment transport near East Arm Wharf.

The impact of dredging near East Arm Wharf (part of the Ichthys Project) on the suspended-sediment dynamics was modelled. Dredging affects the SSC not only at nearby Location EA1 but also at the other seven locations in the harbour over a period of one month: suspended sediment is transported into the inner harbour, the outer harbour and the three arms. Net sediment transport is seaward in the channel and landward at the entrance to Middle Arm.

The impact of sediment from the offshore disposal of dredging material was also examined using the model. Sediment from the offshore disposal area clearly increases the SSC at Location CR near the disposal area, where the SSC is more affected by advection than local re-suspension. The sediment is also transported to Location CL on the opposite side of the outer harbour, generating large SSC spikes there. However, this sediment does not affect the SSC in the inner harbour and the arms. The model results for the effect of dredging and dumping on harbour suspended-sediment dynamics are similar to those of the 2-D numerical model of HR Wallingford (2010b).

105

4.6 Conclusions

106

CHAPTER 5 CONCLUSIONS AND RECOMMENDATIONS FOR FUTURE WORK

5.1 Conclusions

The objective of this study was to understand the dynamics of suspended-sediment transport in Darwin Harbour, Northern Territory Australia. To achieve this objective, two research aims were pursued. The first was to examine the harbour hydrodynamics by constructing a hydrodynamic model using the FVCOM. Special attention was paid to the effects of the mangrove areas and tidal flats on the tides and on tidal asymmetry. The second aim was to understand the suspended-sediment dynamics in the harbour based on the hydrodynamics. A 3-D numerical model, FVCOM-ESSed, was developed which couples the suspended-sediment model of Wang (2002) to the hydrodynamic model of the harbour. Numerical experiments using this model were designed to analyse the factors controlling suspended-sediment transport and the effect of removing the mangrove areas and tidal flats.

To help achieve these aims, hydrographic and sediment data for Darwin Harbour were obtained first to set up and calibrate the FVCOM-ESSed model. These observation data, including sea-surface level, current velocity and suspended-sediment concentration (SSC), were then analysed to obtain a basic understanding of the harbour’s hydrodynamics and suspended-sediment dynamics.

5.1.1 Hydrodynamics of Darwin Harbour

The 3-D hydrodynamic characteristics of Darwin Harbour were determined from both field data and numerical simulations. The hydrodynamics of Darwin Harbour are driven mainly by the tides, with the effects of wind and rivers being small. The main tidal

constituents are M2, S2, N2, K2, K1 and O1. M2 is the dominant component, with

amplitude about 1.9 m and phase 249 degrees near Darwin City. The M2 amplitude increases by about 0.3 m from the outer harbour to the arms, then decreases almost to zero in mangrove areas due to the high bottom stresses there. Current flow is also dominated by tides. The current speed reaches a maximum of about 2.0 ms-1 at the surface of Middle Arm, again falling to zero in the mangrove areas because of the large

107

5.1 Conclusions amount of bottom friction. It decreases gradually from the surface to the bottom: Middle Arm (shallower than 10 m) and the channel (about 25 m deep) have the largest (33.5%) and smallest (24.8%) surface-to-bottom decreases, respectively.

The main tidal current constituent in the harbour is the M2 tidal current, followed -1 by S2, N2 and K1. The maximum speed of the M2 tidal current is about 0.4 ms near East Arm Wharf. In the inner harbour and arms, the water-flow patterns are in accordance with the shoreline. The currents are rectilinear in East Arm, as indicated by near-zero minor axes of the current ellipses, because of the channel morphology.

The tidal system at East Arm is flood dominant according to the field data. This flood dominance is indicated by the tidal asymmetry, calculated via skewness of sea- surface level and current velocities (Nidzieko 2010; Song et al. 2011). The M2 and M4 tidal-constituent combination is the most important contributor to tidal asymmetry, calculated according to Song et al. (2011).

5.1.2 Effect of mangrove areas and tidal flats on tidal asymmetry

Mangrove areas and tidal flats play key roles in modulating the tides and in the water- flow dynamics of the estuary. Removal of the mangrove areas and tidal flats from

Darwin Harbour, would decrease the M2 amplitude due to the decreased shoaling effects in the inner harbour, but would generate a larger M4 amplitude. The phases of both the

M2 and M4 tides would be advanced due to the reduced tidal choking effect. This influence on the amplitudes and phases of the tides is how the mangrove areas and tidal flats affect tidal asymmetry. In Darwin Harbour, these areas significantly reduce tidal asymmetry: for example, the tidal-elevation skewness in Middle Arm would increase by 100% if the mangrove areas were removed, and by 120% if the tidal flats were also removed. The skewness near Location Blay in East Arm varies approximately linearly with the percentage of mangrove removal. As tidal asymmetry strongly affects sediment transport in the estuaries, care must be taken with any reclamation of the mangrove areas and tidal flats around the harbour watershed.

5.1.3 Suspended-sediment model development

The FVCOM-ESSed model was developed in the study to focus on suspended-sediment dynamics, because it is the fine sediment in estuaries which can cause problems like

108

5.1 Conclusions turbidity and siltation. Several features were incorporated into the model to analyses the suspended sediments in Darwin Harbour. Sediment density was coupled to the water- density calculation to consider its impact on the bottom boundary layer dynamics. Wetting-drying processes were included to model the effects of the tidal flats and mangroves areas on suspended sediment. Various bathymetry types were also considered and the initial thickness of the fine-sediment layer varied to model the different geomorphologic conditions. This model could also be used for other harbours that share common characteristics with Darwin Harbour, particularly in terms of extensive tidal flats and mangrove areas.

5.1.4 Suspended-sediment dynamics of Darwin Harbour

New field work was conducted to quantify the bottom boundary layer and suspended sediment dynamics in the harbour. Seven bottom mounted nephelometers were deployed at seven locations in the outer harbour and the three arms, each nephelometer containing a pressure sensor, a deposition sensor, a turbidity sensor, a light sensor, and a temperature sensor. Data were collected at 10-minute intervals from early November 2012 to early February 2013.

In the bottom boundary layer, SSC values vary with the spring-neap tidal cycle, with larger values during spring tides than during neap tides because of the larger spring tidal currents. In Darwin Harbour in November 2012, the bottom SSC values during spring tides were up to 100 times larger than during neap tides, when the SSC was only of the order of 10-3 to 10-4 kgm-3. The near-bottom SSC values measured at the seven locations in the second spring tidal cycle were less than 50% of those in the first.

The 3-D suspended-sediment dynamics were analysed using the FVCOM-ESSed model. According to the model, the water is turbid in the outer harbour, the inner harbour, East Arm and West Arm during spring tides. There is a turbid zone in the outer harbour with SSC values of about 0.1 kgm-3 at the bottom level. This turbid zone still appears during neap tides, but the bottom SSC values are less than about 0.07 kgm-3. Water in the inner harbour and the East Arm and West Arm channels is less turbid during the neap tidal cycle than during the spring tidal cycle, due to weaker currents. These turbid zones are the result of strong currents and sediment availability. Flood and ebb current cycles move the turbid zone in the outer harbour landward and seaward,

109

5.2 Recommendations respectively, by advection. Vertical gradients of SSC develop in the channel during neap tides, but disappear during spring tides. Net sediment transport is seaward in the channel and landward at the entrance to East Arm, dominated by Eulerian advection.

Dredging near East Arm Wharf affects SSC values in other parts of the harbour due to horizontal advection by tidal currents. This was observed during the dredging in November 2012, when large spikes in SSC values were measured at the tip of East Arm. The model indicated that the suspended sediments caused by dredging cin East Arm can be transported into the inner harbour, then the outer harbour, and into the arms.

5.1.5 Effect of mangrove areas and tidal flats on sediment transport

Mangrove areas and tidal flats cause redistribution of suspended sediment by affecting the hydrodynamics in an estuary. In Darwin Harbour, if the tidal flats and mangrove areas were removed, tidal currents would be reduced due to a reduction in the tidal choking effect. Consequently, the SSC in the turbid zone in the outer harbour would be reduced by 50% at the bottom level. The water in the inner harbour, East Arm and West Arm would also be less turbid. Mangrove areas and tidal flats affect net sediment transport by modulating tidal asymmetry. In Darwin Harbour, without the mangrove areas and tidal flats, the net sediment flux in the channel would be landward instead of seaward due to an amplification of flood dominance. Tidal pumping would overtake the Eulerian residual to dominate sediment transport in both the channel and at the entrance to East Arm. The landward net sediment flux would be decreased at the entrance to East Arm as a result of reduced currents in the arms because of a reduction in the tidal choking effect.

5.2 Recommendations

According to the conclusions drawn from this study, more sediment could move landward if the harbour’s construction is not designed and managed correctly. Siltation may then occur in the navigational channel or near the coastal facilities, requiring more dredging to maintain the economic and social functions of the harbour. Therefore, suspended-sediment dynamics should be considered when planning harbour construction and the location of marine industries. Wharfs, dykes and navigational channels should be located away from zones of high erosion or deposition to maintain

110

5.3 Future work their long-term usefulness. Suspended-sediment transport dynamics in the harbour should be considered before the harbour construction, such as mangrove and tidal flats reclamation and port expansion, is undertaken. Water-related activities, such as fishing, diving and snorkelling, should be conducted away from turbid zones.

As indicated by the field SSC data in the bottom boundary layer and the model results, water quality can be degraded if dredging activity is not well planned, and dredged material can be transported back to near shore from disposal areas. Therefore, it is recommended that clear and specific regulations about dredging and dumping activities are needed for any harbour construction. Obviously, the Northern Territory Government has been aware of the importance of environmental protection, as the Northern Territory Environmental Assessment Act (EA Act) was passed in 1984. Projects undertaken in the Northern Territory, e.g. the Ichthys Project, have been assessed under this Act at the level of an Environmental Impact Statement. Projects are also subject to the Environment Protection and Biodiversity Conservation Act 1999 (EPBC Act) and assessed under the bilateral agreement between the Northern Territory and Australian Governments. However, detailed regulations on the management of suspended sediment in Darwin Harbour are limited in these Acts. For example, the assessment and management of suspended sediments must necessarily be part of the Integrated Coastal Zone Management system, which is “the most appropriate process for dealing with long-term challenges, and being a proactive policy process aimed at addressing conflict, interests for coastal space and resources” (Ioppolo et al. 2013). A database containing all data related to sediments is a fundamental requirement for sediment dynamics research and management. Feasible management processes, including emergency and follow-up procedures, for dealing with sediment are necessary to protect marine areas and industries from being damaged or hindered by high water turbidity, sediment erosion or siltation.

5.3 Future work

The field SSC data in the bottom boundary layer were only obtained at seven locations in Darwin Harbour. Furthermore, the data were obtained only at a single depth at the sites. In order to provide a better understanding of the vertical structure of the suspended-sediment dynamics, data are required at more locations and more depths.

111

5.3 Future work

These data can then be used to improve the model accuracy, and increase confidence in the model predictions. As eddies appear where there are headlands in Darwin Harbour, especially near Darwin City and East Arm Wharf, the mechanisms relating to the effects of eddies on sediment transport should be examined; it is believed that eddies can trap sediment (Signell & Geyer 1991). Further measurements are required of relevant parameters, such as eddy diffusivity and entrainment rate, before eddies can be included in the FVCOM-ESSed model.

The suspended-sediment dynamics of Darwin Harbour in the dry season has been numerically modelled here, but the dynamics in the wet season still remain unknown. High rainfall during the wet season gives rise to large river inflows, e.g. the average daily input from the Blackmore River in February 2003 was 2,356 Millions of Liters. A large amount of sediment is then washed into the harbour, generating high turbidity (McKinnon et al. 2006). Therefore, an analysis of the sediment dynamics in the wet season is as important as in the dry season.

Although the effect of waves in the inner harbour on sediment transport is small, wave-current interactions still need to be considered in the shallow-water areas, e.g. the mangrove areas and tidal flats. Here, wave effects become important, especially when there are strong winds. Accordingly, it is suggested that a wave model be incorporated into the current and sediment transport model to allow for wave-current interactions.

Sediments by nature are mixtures of primary particles of various sizes, both cohesive and non-cohesive. However, the interaction between suspended sediment and sediment deposited on the harbour bed is not accounted for in the present study. This interaction requires consideration of different harbour-bed structures and the modelling of consolidation, such as its effect on the bed shear strength.

Theoretically and practically, it is important for further research to be conducted on the impacts of sediment on the marine economy, in order to minimise sediment- related marine economic loss. It is suggested that the relationship between the present of sediment and marine economic loss be examined scientifically. The key factors controlling this interaction should be determined for the harbour and quantified. In this way, sediment research can benefit marine economic development and harbour management.

112

5.4 Wider applications of this study

This study provides contribution to numerical modelling of suspended-sediment dynamics in estuaries. The sediment model proposed in this study provides its capability in studying suspended-sediment dynamics in any estuary by only changing its forcings. It ensures a better understanding of sediment bottom boundary layer, which is important for bottom sediment stratification research in turbid estuaries. It provides an improved ability to characterise different bathymetry types, e.g. mangrove areas, and human activities, e.g. dredging, in sediment simulation.

This study initialises a requirement for assessments of impacts of coastal reclamation on suspended-sediment transport in estuaries. This study highlights the effects of coastal reclamation, e.g. mangrove areas and tidal flats reclamation, on suspended-sediment transport through the mechanism of tidal symmetry. Although the findings in this study are site specific, they are meaningful to coastal reclamation in similar estuaries like Darwin Harbour.

Another contribution of this study is its environmental and socio-economic importance. This study combines suspended-sediment dynamics, environment and human activities together using numerical model to scientifically demonstrate the effect of human activities on suspended-sediment dynamics and environment. The findings of this study are of relevance and importance to estuaries with high frequency and intense of human activities near-shore.

113

5.4 Wider applications of this study

114

REFERENCES

REFERENCES Aldridge, J.N., 1997, Hydrodynamic model predictions of tidal asymmetry and observed sediment transport paths in Morecambe Bay, Estuarine, Coastal and Shelf Science, 44(1), 39-56, doi: 10.1006/ecss.1996.0113. Allen, G.P., Salomon, J.C., Bassoullet, P., Du Penhoat, Y. & de Grandpré, C., 1980, Effects of tides on mixing and suspended sediment transport in macrotidal estuaries, Sedimentary Geology, 26(1-3), 69-90, doi: 10.1016/0037- 0738(80)90006-8. Alongi, D.M., Pfitzner, J., Trott, L.A., Tirendi, F., Dixon, P. & Klumpp, D.W., 2005, Rapid sediment accumulation and microbial mineralization in forests of the mangrove Kandelia candel in the Jiulongjiang Estuary, China, Estuarine, Coastal and Shelf Science, 63(4), 605-618, doi: 10.1016/j.ecss.2005.01.004. Andutta, F.P., Wang, X.H., Li, L. & Williams, D., 2013, Hydrodynamics and sediment transport in a macro-tidal estuary: Darwin Harbour, in Estuaries of Australia in 2050 and Beyond, E. Wolanski (ed.), Springer, Dordrecht, In press. Ariathurai, R. & Krone, R.B., 1976, Finite element model for cohesive sediment transport, Journal of the Hydraulics Division, 102(3), 323-338. Asia-Pacific Applied Science Associates, 2010, Ichthys Gas Field Development Project: Description and validation of hydrodynamic and wave models for discharges, spills, geomorphology and dredge spoil disposal ground selection, Report prepared for INPEX Browse, Ltd., Perth, Western Australia. Baird, D., Winter, P.E.D. & Wendt, G., 1987, The flux of particulate material through a well-mixed estuary, Continental Shelf Research, 7(11-12), 1399-1403, doi: 10.1016/0278-4343(87)90044-6. Blanton, J.O., Lin, G. & Elston, S.A., 2002, Tidal current asymmetry in shallow estuaries and tidal creeks, Continental Shelf Research, 22(11-13), 1731-1743, doi: 10.1016/s0278-4343(02)00035-3. Blumberg, A.F. & Mellor, G.L., 1987, A description of a three dimensional coastal ocean circulation model, in Three-dimensional Coastal Ocean Models, vol. 4, N. Heaps (ed.), American Geophysical Union, Washington, DC, pp. 1-16. Bourgault, D. & Kelley, D.E., 2004, A laterally averaged nonhydrostatic ocean model, Journal of Atmospheric and Oceanic Technology, 21(12), 1910-1924, doi: 10.1175/jtech-1674.1.

115

REFERENCES

Brenon, I. & Le Hir, P., 1999, Modelling the turbidity maximum in the Seine Estuary (France): Identification of formation processes, Estuarine, Coastal and Shelf Science, 49(4), 525-544, doi: 10.1006/ecss.1999.0514. Burford, M.A., Alongi, D.M., McKinnon, A.D. & Trott, L.A., 2008, Primary production and nutrients in a tropical macrotidal estuary, Darwin Harbour, Australia, Estuarine, Coastal and Shelf Science, 79(3), 440-448, doi: 10.1016/j.ecss.2008.04.018. Byun, D.S., Wang, X.H. & Holloway, P.E., 2004, Tidal characteristic adjustment due to dyke and seawall construction in the Mokpo Coastal Zone, Korea, Estuarine, Coastal and Shelf Science, 59(2), 185-196, doi: 10.1016/j.ecss.2003.08.007. Carballo, R., Iglesias, G. & Castro, A., 2009, Residual circulation in the Ría de Muros (NW Spain): A 3D numerical model study, Journal of Marine Systems, 75(1-2), 116-130, doi: 10.1016/j.jmarsys.2008.08.004. Chang, T.-J., Kao, H.-M., Chang, K.-H. & Hsu, M.-H., 2011, Numerical simulation of shallow-water dam break flows in open channels using smoothed particle hydrodynamics, Journal of Hydrology, 408(1-2), 78-90, doi: 10.1016/j.jhydrol.2011.07.023. Chen, C., Liu, H. & Beardsley, R.C., 2003, An unstructured, finite-volume, three- dimensional, primitive equation ocean model: Application to coastal ocean and estuaries, Journal of Atmospheric and Oceanic Technology, 20(1), 159-186, doi: 10.1175/1520-0426(2003)020<0159:augfvt>2.0.co;2. Christiansen, C., Pejrup, M., Kepp, R.R., Nielsen, A., Vølund, G. & Pedersen, J.B.T., 2004, Tidal and meteorological induced nutrient (N, P) dynamics in the micro- tidal Ho Bugt, Danish Wadden Sea, Danish Journal of Geography, 104(1), 87- 96. Christie, M.C., Dyer, K.R. & Turner, P., 1999, Sediment flux and bed level measurements from a macro tidal mudflat, Estuarine, Coastal and Shelf Science, 49(5), 667-688, doi: 10.1006/ecss.1999.0525. Coalition, 2013, The Coalition's 2030 Vision for Developing Northern Australia, the Laberal Party of Australia, 46 pp. Available at: http://www.liberal.org.au/sites/default/files/The%20Coalitions%202030%20Visi on%20for%20Developing%20Northern%20Australia.pdf, accessed on 1st August 2013.

116

REFERENCES

Coleman, A.P.M., 2004, The national recreational fishing survey: The Northern Territory, Northern Territory Department of Business, Industry and Resource Development, Darwin. Collins, M.B., Ke, X. & Gao, S., 1998, Tidally-induced flow structure over intertidal flats, Estuarine, Coastal and Shelf Science, 46(2), 233-250, doi: 10.1006/ecss.1997.0260. Corbett, D.R., McKee, B. & Duncan, D., 2004, An evaluation of mobile mud dynamics in the Mississippi River deltaic region, Marine Geology, 209(1-4), 91-112, doi: 10.1016/j.margeo.2004.05.028. Den Besten, P., Deckere, E.d., Babut, M., Power, B., DelValls, A., Zago, C., Oen, A. & S., H., 2003, Biological effects-based sediment quality in ecological risk assessment for European waters, Journal of Soils and Sediments, 3(3), 144-162. Dikou, A. & van Woesik, R., 2006, Survival under chronic stress from sediment load: Spatial patterns of hard coral communities in the southern islands of Singapore, Marine Pollution Bulletin, 52(1), 7-21, doi: 10.1016/j.marpolbul.2005.07.021. Doxaran, D., Froidefond, J.-M., Castaing, P. & Babin, M., 2009, Dynamics of the turbidity maximum zone in a macrotidal estuary (the Gironde, France): Observations from field and MODIS satellite data, Estuarine, Coastal and Shelf Science, 81(3), 321-332, doi: 10.1016/j.ecss.2008.11.013. Dronkers, J., 1986, Tidal asymmetry and estuarine morphology, Netherlands Journal of Sea Research, 20(2-3), 117-131, doi: 10.1016/0077-7579(86)90036-0. Dyer, K.R., 1986, Coastal and Estuarine Sediment Dynamics, Wiley, Chichester. Dyer, K.R., 1997, Partially mixed and well-mixed estuaries, in Estuaries--A Physical Introduction, 2nd edn, John Wiley and Sons, Chichester, pp. 136-164. Dyer, K.R., Christie, M.C., Feates, N., Fennessy, M.J., Pejrup, M. & van der Lee, W., 2000, An investigation into processes influencing the morphodynamics of an intertidal mudflat, the Dollard estuary, The Netherlands: I. hydrodynamics and suspended sediment, Estuarine, Coastal and Shelf Science, 50(5), 607-625, doi: 10.1006/ecss.1999.0596. Egbert, G.D., Erofeeva, S.Y. & Ray, R.D., 2010, Assimilation of altimetry data for nonlinear shallow-water tides: Quarter-diurnal tides of the Northwest European shelf, Continental Shelf Research, 30(6), 668-679, doi: 10.1016/j.csr.2009.10.011.

117

REFERENCES

Emery, W.J. & Thomson, R.E., 2001, Data Analysis Methods in Physical Oceanography, 2nd edn, Elservier, Amsterdam, The Netherlands, 654 pp. Escapa, M., Perillo, G.M.E. & Iribarne, O., 2008, Sediment dynamics modulated by burrowing crab activities in contrasting SW Atlantic intertidal habitats, Estuarine, Coastal and Shelf Science, 80(3), 365-373, doi: 10.1016/j.ecss.2008.08.020. Esslemont, G., 1999, Heavy metals in corals from Heron Island and Darwin Harbour, Australia, Marine Pollution Bulletin, 38(11), 1051-1054, doi: 10.1016/S0025- 326X(99)00183-6. Facca, C., Sfriso, A. & Socal, G., 2002, Changes in abundance and composition of phytoplankton and microphytobenthos due to increased sediment fluxes in the Venice Lagoon, Italy, Estuarine, Coastal and Shelf Science, 54(5), 773-792, doi: 10.1006/ecss.2001.0848. Fortune, J., 2006, The grainsize and heavy metal content of sediment in Darwin Harbour, Report No.: 14/2006D, Aquatic Health Unit, Environmental Protection Agency, Department of Natural Resources, Environment and the Arts, Darwin, N.T., 74 pp. Available at: http://lrm.nt.gov.au/__data/assets/pdf_file/0007/16882/fortune_grainsize.pdf, accessed on 30th March 2013. Fortune, J. & Drewry, J., 2009, Darwin Harbour Region Report Cards 2009, Aquatic Health Unit, Department of Natural Resources, Environment, The Arts and Sport. Palmerston NT 0831, 62pp. Fortune, J. & Drewry, J., 2011, Darwin Harbour Region Research and Monitoring 2011, 18/2011D, Department of Natural Resources, Environment, The Arts and Sport, Palmerston, NT, Australia. Furukawa, K. & Wolanski, E., 2004, Sedimentation in mangrove forests, Mangroves and Salt Marshes 1(1), 3-10, doi: 10.1023/a:1025973426404. Gailani, J., Ziegler, C.K. & Lick, W., 1991, Transport of suspended solids in the lower Fox River, Journal of Great Lakes Research, 17(4), 479-494, doi: 10.1016/S0380-1330(91)71384-1. Godin, G., 1991, The analysis of tides and currents, in Tidal Hydrodynamics, B.B. Paker (ed.), John Wiley, New York, pp. 675-708. Gonzalez-Rodriguez, D. & Madsen, O.S., 2007, Seabed shear stress and bedload

118

REFERENCES

transport due to asymmetric and skewed waves, Coastal Engineering, 54(12), 914-929, doi: 10.1016/j.coastaleng.2007.06.004. Grigalunas, T., Opaluch, J.J. & Luo, M., 2001, The economic costs to fisheries from marine sediment disposal: case study of Providence, RI, USA, Ecological Economics, 38(1), 47-58, doi: 10.1016/S0921-8009(00)00294-9. Gust, G., 1976, Observations on turbulent-drag reduction in a dilute suspension of clay in sea-water, Journal of Fluid Mechanics, 75(1), 29-47, doi: 10.1017/S0022112076000116. Hamrick, J.M., 1992, A Three-Dimensional Environmental Fluid Dynamics Computer Code: Theoretical and Computational Aspects, Special Report in Applied Marine Science and Ocean Engineering (SPAM_SOE) No.317, The College of William and Mary, Virginia Institute of Marine Science, Gloucester Point, Virginia, 63 pp. Hoitink, A.J.F., Hoekstra, P. & Maren, D.S.v., 2003, Flow asymmetry associated with astronomical tides: Implications for the residual transport of sediment, Journal of Geophysical Research, 108(C10), 13-1-13-8, doi: 10.1029/2002JC001539. Holland, K.T., Vinzon, S.B. & Calliari, L.J., 2009, A field study of coastal dynamics on a muddy coast offshore of Cassino beach, Brazil, Continental Shelf Research, 29(3), 503-514, doi: 10.1016/j.csr.2008.09.023. HR Wallingford, 2010a, Ichthys Gas Field Development Project: Description and validation of hydrodynamic and wave models for dredging and spoil disposal, Report prepared for INPEX Browse, Ltd., Perth, Western Australia. HR Wallingford, 2010b, Ichthys Gas Field Development Project: Dredging and spoil disposal modelling, Report prepared for INPEX Browse, Ltd., Perth, Western Australia. Hu, J., Li, S. & Geng, B., 2011, Modeling the mass flux budgets of water and suspended sediments for the river network and estuary in the Pearl River Delta, China, Journal of Marine Systems, 88(2), 252-266, doi: 10.1016/j.jmarsys.2011.05.002. Huang, H., Chen, C., Blanton, J.O. & Andrade, F.A., 2008, A numerical study of tidal asymmetry in Okatee Creek, South Carolina, Estuarine, Coastal and Shelf Science, 78(1), 190-202, doi: 10.1016/j.ecss.2007.11.027. INPEX, 2012, INPEX and total make final investment decision on ichthys LNG project, Australia, INPEX news, 13 January, pp. 1-4.

119

REFERENCES

Ioppolo, G., Saija, G. & Salomone, R., 2013, From coastal management to environmental management: The sustainable eco-tourism program for the mid- western coast of Sardinia (Italy), Land Use Policy, 31(0), 460-471, doi: 10.1016/j.landusepol.2012.08.010. Jiang, C., Li, J. & de Swart, H.E., 2012, Effects of navigational works on morphological changes in the bar area of the Yangtze Estuary, Geomorphology, 139-140(0), 205-219, doi: 10.1016/j.geomorph.2011.10.020. Jones, J.E. & Davies, A.M., 2008, On the modification of tides in shallow water regions by wind effects, Journal of Geophysical Research, 113(C5), doi: 10.1029/2007JC004310. Kelly, R., 2012, Inpex, total green light $33bn Ichthys project, The Australian, January 13th 2012, 1p. Print. Kessel, T.V., Vanlede, J. & de Kok, J., 2011, Development of a mud transport model for the Scheldt estuary, Continental Shelf Research, 31(10, Supplement 1), S165- S181, doi: 10.1016/j.csr.2010.12.006. King, l.P., 2009, Chapter2: Advection diffusion equations for transport, in RMA manual: RMA-11 -- A three dimensional finite element model for water quality in estuaries and streams (version 8.3b), Resource Modelling Associates, Sydney, Australia, pp. 1-7. Kingsford, R.T., Dunn, H., Love, D., Nevill, J., Stein, J. & Tait, J., 2005, Protecting Australia's rivers, wetlands and estuaries of high conservation value, Product No.: PR050823., Department of the Environment and Heritage Australia, Canberra. Available at: http://www.environment.gov.au/water/publications/environmental/pubs/protectin g-rivers.pdf, accessed on 30th July 2012. Kitheka, J.U., Obiero, M. & Nthenge, P., 2005, River discharge, sediment transport and exchange in the Tana Estuary, Kenya, Estuarine, Coastal and Shelf Science, 63(3), 455-468, doi: 10.1016/j.ecss.2004.11.011. Knight, J.M., Dale, P.E.R., Dunn, R.J.K., Broadbent, G.J. & Lemckert, C.J., 2008, Patterns of tidal flooding within a mangrove forest: Coombabah Lake, Southeast Queensland, Australia, Estuarine, Coastal and Shelf Science, 76(3), 580-593, doi: 10.1016/j.ecss.2007.07.044. Kombiadou, K. & Krestenitis, Y.N., 2013, Chapter 9: Modelling cohesive sediment

120

REFERENCES

dynamics in the marine environment in Sediment Transport Processes and Their Modelling Applications, A. Manning (ed.), InTech, published March 13, 2013 under CC BY 3.0 license, pp. 213-246. Krishnamurti, T.N., Rajendran, K., Kumar, T.S.V.V., Lord, S., Toth, Z., Zou, X., Navon, I.M. & Ahlquist, J., 2003, Improved skill for the anomaly correlation of geopotential heights at 500 hPa, Monthly Weather Review, 131(1082-1102), 21. Le Hir, P., Roberts, W., Cazaillet, O., Christie, M., Bassoullet, P. & Bacher, C., 2000, Characterization of intertidal flat hydrodynamics, Continental Shelf Research, 20(12-13), 1433-1459, doi: 10.1016/s0278-4343(00)00031-5. Li, L., Wang, X.H., Sidhu, H. & Williams, D., 2011, Modelling of three dimensional tidal dynamics in Darwin Harbour, Australia, in Proceedings of the 15th Biennial Computational Techniques and Applications Conference, CTAC-2010, vol. 52, pp. C103-C123. Li, L., Wang, X.H., Williams, D., Sidhu, H. & Song, D., 2012, Numerical study of the effects of mangrove areas and tidal flats on tides: A case study of Darwin Harbour, Australia, Journal of Geophysical Research–Oceans, 117(C6), C06011, doi: 10.1029/2011jc007494. Liu, G., Zhu, J., Wang, Y., Wu, H. & Wu, J., 2011, Tripod measured residual currents and sediment flux: Impacts on the silting of the deepwater navigation channel in the Changjiang Estuary, Estuarine, Coastal and Shelf Science, 93(3), 192-201, doi: 10.1016/j.ecss.2010.08.008. Lumborg, U., Andersen, T.J. & Pejrup, M., 2006, The effect of Hydrobia ulvae and microphytobenthos on cohesive sediment dynamics on an intertidal mudflat described by means of numerical modelling, Estuarine, Coastal and Shelf Science, 68(1-2), 208-220, doi: 10.1016/j.ecss.2005.11.039. Lumborg, U. & Pejrup, M., 2005, Modelling of cohesive sediment transport in a tidal lagoon–an annual budget, Marine Geology, 218(1-4), 1-16, doi: 10.1016/j.margeo.2005.03.015. Ma, G., Shi, F., Liu, S. & Qi, D., 2011, Hydrodynamic modeling of Changjiang Estuary: Model skill assessment and large-scale structure impacts, Applied Ocean Research, 33(1), 69-78, doi: 10.1016/j.apor.2010.10.004. MacCready, P., Banas, N.S., Hickey, B.M., Dever, E.P. & Liu, Y., 2009, A model study of tide- and wind-induced mixing in the Columbia River Estuary and plume,

121

REFERENCES

Continental Shelf Research, 29(1), 278-291, doi: 10.1016/j.csr.2008.03.015. Malarkey, J. & Davies, A.G., 2012, Free-stream velocity descriptions under waves with skewness and asymmetry, Coastal Engineering, 68(0), 78-95, doi: 10.1016/j.coastaleng.2012.04.009. Manning, A.J., Langston, W.J. & Jonas, P.J.C., 2010, A review of sediment dynamics in the Severn Estuary: Influence of flocculation, Marine Pollution Bulletin, 61(1-3), 37-51, doi: 10.1016/j.marpolbul.2009.12.012. Margvelashvili, N., Robson, B., Sakov, P., Webster, I.T., Parslow, J., Herzfeld, M. & Andrewartha, J., 2003, Numerical models of hydrodynamics, sediment transport and biogeochemistry in the Fitzroy Estuary, Cooperative Research Certer for Coastal Zone Estuary and Waterway Management (Australia), Technical Report 9, Indooroopilly, Qld. Available at: http://www.clw.csiro.au/publications/consultancy/2003/tr9-hydrodynamic- modelling-fitzroy.pdf, accessed on 20th June 2011. Mazda, Y., Kanazawa, N. & Wolanski, E., 1995, Tidal asymmetry in mangrove creeks, Hydrobiologia, 295(1), 51-58, doi: 10.1007/bf00029110. Mazda, Y., Kobashi, D. & Okada, S., 2005, Tidal-scale hydrodynamics within mangrove swamps, Wetlands Ecology and Management, 13(6), 647-655, doi: 10.1007/s11273-005-0613-4. Mazda, Y., Magi, M., Kogo, M. & Hong, P.N., 1997a, Mangroves as a coastal protection from waves in the Tong King delta, Vietnam, Mangroves and Salt Marshes, 1(2), 127-135, doi: 10.1023/a:1009928003700. Mazda, Y., Wolanski, E., King, B., Sase, A., Ohtsuka, D. & Magi, M., 1997b, Drag force due to vegetation in mangrove swamps, Mangroves and Salt Marshes, 1(3), 193- 199, doi: 10.1023/a:1009949411068. McKinnon, A.D., Smit, N., Townsend, S. & Duggan, S., 2006, Darwin Harbour: Water quality and ecosystem structure in a tropical harbour in the early stages of urban development, in The Environment in Asia Pacific Harbour, E. Wolanski (ed.), Springer, Dordrecht, pp. 433-459. Mehta, A.J., 1988, Laboratory studies on cohesive sediment deposition and erosion, in Physical Processes in Estuaries, J. Dronkers & W.V. Leussen (eds), Springer Verlag, Berlin, pp. 427-445. Mellor, G.L. & Blumberg, A.F., 1985, Modeling vertical and horizontal diffusivities

122

REFERENCES

with the sigma coordinate system, Monthly Weather Review, 113(8), 1379-1383, doi: 10.1175/1520-0493(1985)113<1379:MVAHDW>2.0.CO;2. Mellor, G.L. & Yamada, T., 1982, Development of a turbulence closure model for geophysical fluid problems, Reviews of Geophysics and Space Physics, 20(4), 851-875, doi: 10.1029/RG020i004p00851. Metcalfe, K.N., Franklin, D.C. & McGuinness, K.A., 2011, Mangrove litter fall: Extrapolation from traps to a large tropical macrotidal harbour, Estuarine, Coastal and Shelf Science, 95(1), 245-252, doi: 10.1016/j.ecss.2011.09.006. Metcalfe, K.N. & Glasby, C.J., 2008, Diversity of Polychaeta (Annelida) and other worm taxa in mangrove habitats of Darwin Harbour, northern Australia, Journal of Sea Research, 59(1-2), 70-82, doi: 10.1016/j.seares.2007.06.002. Michie, M.G., 1987, Distribution of foraminifera in a macrotidal tropical esturay: Port Darwin, Northern Territory of Australia, Australian Journal of Marine and Freshwater Research, 38(2), 249-259, doi: 10.1071/MF9870249. Milligan, T.G. & Hill, P.S., 1998, A laboratory assessment of the relative importance of turbulence, particle composition, and concentration in limiting maximal floc size and settling behaviour, Journal of Sea Research, 39(3-4), 227-241, doi: 10.1016/S1385-1101(97)00062-2. Mitchell, S.B., West, J.R., Arundale, A.M.W., Guymer, I. & Couperthwaite, J.S., 1999, Dynamics of the turbidity maxima in the Upper Humber estuary system, UK, Marine Pollution Bulletin, 37(3-7), 190-205, doi: 10.1016/s0025- 326x(98)00178-7. Nelson, B.W., 2001, Sediment dynamics in Rangoon River, Myanmar, Science of The Total Environment, 266(1-3), 15-21, doi: 10.1016/S0048-9697(00)00759-2. Nichols, M.N. & Biggs, R.B., 1985, Estuaries, in Coastal Sedimentary Environments, R.A. Davis, Jr. (ed.), Springer-Verlag, New York, pp. 77-186. Nidzieko, N.J., 2010, Tidal asymmetry in estuaries with mixed semidiurnal/diurnal tides, Journal of Geophysical Research:Oceans, 115(C8), C08006-C08006, doi: 10.1029/2009JC005864. Northern Territory Government, 2009, Growing International Trade, Priority Action Plan 2009-2013, Report No.: 978-0-9805709-3-9, Department of the Chief Minister, Darwin, N.T., 1-59 pp. Available at: http://www.dcm.nt.gov.au/__data/assets/pdf_file/0010/45559/DCM_Growing_In

123

REFERENCES

ternational1.pdf, accessed on 20th June 2011. Noske, R.A., 1996, Abundance, zonation and foraging ecology of birds in mangroves of Darwin Harbour, Northern Territory, Wildlife Research, 23(4), 443-474, doi: 10.1071/WR9960443. Nunes Vaz, R.A., Lennon, G.W. & de Silva Samarasinghe, J.R., 1989, The negative role of turbulence in estuarine mass transport, Estuarine, Coastal and Shelf Science, 28(4), 361-377, doi: 10.1016/0272-7714(89)90085-1. Padovan, A., 1997, Darwin Harbour water and sediment quality, Proceedings of the Darwin Harbour Public Presentations, 5-18. Available at: http://lrm.nt.gov.au/__data/assets/pdf_file/0008/16919/sediment.pdf, accessed on 20th June 2011. Partheniades, E., 1962, A Study of Erosion and Deposition of Cohesive Soils in Salt Water, University of California, Berkeley. Peerzada, N. & Dickinson, C., 1988, Heavy metal concentration in oysters from Darwin Harbour, Marine Pollution Bulletin, 19(4), 182-184, doi: 10.1016/0025- 326x(88)90677-7. Peerzada, N. & Kozlik, E., 1992, Seasonal variation of heavy metals in oysters from Darwin Harbor, Northern Territory, Australia, Bulletin of Environmental Contamination and Toxicology, 48(1), 31-36, doi: 10.1007/bf00197480. Peerzada, N. & Ryan, P., 1987, Determination of Pb, Zn, Cu and Cd in Darwin Harbour, Marine Pollution Bulletin, 18(8), 458-461, doi: 10.1016/0025-326x(87)90625-4. Pejrup, M., 1988, Suspended sediment transport across a tidal flat, Marine Geology, 82(3-4), 187-198, doi: 10.1016/0025-3227(88)90140-5. Pejrup, M. & Mikkelsen, O.A., 2010, Factors controlling the field settling velocity of cohesive sediment in estuaries, Estuarine, Coastal and Shelf Science, 87(2), 177-185, doi: 10.1016/j.ecss.2009.09.028. Peterson, E.L., 1999, Benthic shear stress and sediment condition, Aquacultural Engineering, 21(2), 85-111, doi: 10.1016/s0144-8609(99)00025-4. Ridderinkhof, H., van der Ham, R. & van der Lee, W., 2000, Temporal variations in concentration and transport of suspended sediments in a channel-flat system in the Ems-Dollard estuary, Continental Shelf Research, 20(12-13), 1479-1493, doi: 10.1016/s0278-4343(00)00033-9. Shi, J.Z., 2010, Tidal resuspension and transport processes of fine sediment within the

124

REFERENCES

river plume in the partially-mixed Changjiang River estuary, China: A personal perspective, Geomorphology, 121(3-4), 133-151, doi: 10.1016/j.geomorph.2010.04.021. Shi, J.Z. & Zhang, H.-L., 2011, Note on a 2DH finite element model of tidal flow of the North Passage of the partially-mixed Changjiang River estuary, China, Journal of Hydro-environment Research, 5(1), 49-62, doi: 10.1016/j.jher.2010.06.001. Signell, R.P. & Geyer, W.R., 1991, Transient Eddy Formation Around Headlands, Journal of Geophysical Research, 96(C2), 2561-2575, doi: 10.1029/90jc02029. Simpson, J.H. & Hunter, J.R., 1974, Fronts in the Irish Sea, Nature, 250(5465), 404-406, doi: 10.1038/250404a0. Sinclair Knight Merz Pty Ltd, 2011, Ichthys Gas Field Development Project: Assessment of potential impacts on mud crabs in Darwin Harbour, Report prepared by Sinclair Knight Merz Pty Limited, Perth, for INPEX Browse, Ltd., Western Australia, Perth. Smagorinsky, J., 1963, General circulation experiments with the primitive equations, Monthly Weather Review, 91(3), 99-164, doi: 10.1175/1520- 0493(1963)091<0099:gcewtp>2.3.co;2. Smith, J. & Haese, R., 2009, The role of sediments in nutrient cycling in the tidal creeks of Darwin Harbour, AusGeo News, Geoscience Australia, (95), 1-7. Smolarkiewicz, P.K., 1984, A fully multidimensional positive definite advection transport algorithm with small implicit diffusion, Journal of Computational Physics, 54(2), 325-362, doi: 10.1016/0021-9991(84)90121-9. Son, M. & Hsu, T.-J., 2011, Idealized study on cohesive sediment flux by tidal asymmetry, Environmental Fluid Mechanics, 11(2), 183-202, doi: 10.1007/s10652-010-9193-9. Song, D., Wang, X.H., Kiss, A.E. & Bao, X., 2011, The contribution to tidal duration asymmetry by different combinations of tidal constituents, Journal of Geophysical Research--Oceans, 116(C12), C12007, doi: 10.1029/2011jc007270. Speer, P.E. & Aubrey, D.G., 1985, A study of non-linear tidal propagation in shallow inlet/estuarine systems Part II: Theory, Estuarine, Coastal and Shelf Science, 21(2), 207-224, doi: 10.1016/0272-7714(85)90097-6. Strom, K. & Keyvani, A., 2011, An explicit full-range settling velocity equation for mud flocs, Journal of Sedimentary Research, 81(12), 921-934.

125

REFERENCES

Sun, D., 2001, Modeling suspended sediment transport under combined wave current actions in Indian River Lagoon, PhD Thesis, University of Florida,Florida, 251 pp. Toorman, E.A., 2001, Cohesive sediment transport modeling: European perspective, in Coastal and Estuarine Fine Sediment Processes, W.H. McAnally & A.J. Mehta (eds), Elservier, Amsterdam, The Netherlands, pp. 1-18. Uluturhan, E., Kontas, A. & Can, E., 2011, Sediment concentrations of heavy metals in the Homa Lagoon (Eastern Aegean Sea): Assessment of contamination and ecological risks, Marine Pollution Bulletin, 62(9), 1989-1997, doi: 10.1016/j.marpolbul.2011.06.019. Uncles, R.J. & Stephens, J.A., 1989, Distributions of suspended sediment at high water in a macrotidal estuary, Journal of Geophysical Research, 94(C10), 14395- 14405, doi: 10.1029/JC094iC10p14395. Uncles, R.J. & Stephens, J.A., 1993, The freshwater-saltwater interface and its relationship to the turbidity maximum in the Tamar Estuary, United Kingdom, Estuaries, 16(1), 126-141, doi: 10.2307/1352770. Uncles, R.J. & Stephens, J.A., 2010, Turbidity and sediment transport in a muddy sub- estuary, Estuarine, Coastal and Shelf Science, 87(2), 213-224, doi: 10.1016/j.ecss.2009.03.041. Van Leussen, W., 2011, Macroflocs, fine-grained sediment transports, and their longitudinal variations in the Ems Estuary, Ocean Dynamics, 61(2), 387-401, doi: 10.1007/s10236-011-0384-9. Van Maren, D.S., Winterwerp, J.C., Sas, M. & Vanlede, J., 2009, The effect of dock length on harbour siltation, Continental Shelf Research, 29(11-12), 1410-1425, doi: 10.1016/j.csr.2009.03.003. Van Rijn, L.C., 1993, Principles of Sediment Transport in Rivers, Estuaries and Coastal Seas, Aqua Publications, Amsterdam, The Netherlands, 690 pp. Varnell, L.M., Evans, D.A. & Havens, K.J., 2003, A geomorphological model of intertidal cove marshes with application to wetlands management, Ecological Engineering, 19(5), 339-347, doi: 10.1016/s0925-8574(02)00120-9. Ven Beusekom, J. & Jonge, V.N.d., 1995, Wind and tide induced resuspension of sediment and microphytobenthos from tidal flats in the Ems estuary, Limnology and Oceanography, 40(4), 766-778, doi: 10.4319/lo.1995.40.4.0776.

126

REFERENCES

Viseras, C., Fernández, J., García-García, F., Soria, J., Calvache, M. & Jáuregui, P., 2009, Dynamics of sedimentary environments in the accelerated siltation of a reservoir: The case of Alhama de Granada, southern Spain, Environmental Geology, 56(7), 1353-1369, doi: 10.1007/s00254-008-1231-2. Wai, O.W.H., Wang, C.H., Li, Y.S. & Li, X.D., 2004, The formation mechanisms of turbidity maximum in the Pearl River estuary, China, Marine Pollution Bulletin, 48(5-6), 441-448, doi: 10.1016/j.marpolbul.2003.08.019. Walton, T.L.J., 2002, Tidal Velocity Asymmetry at Inlets, ERDC/CHL CHETN, U.S. Army Engineer Research and Development Center. Available at: http://chl.erdc.usace.army.mil, accessed on 10th July 2011. Wang, X.H., 2002, Tide-induced sediment resuspention and the bottom boundary layer in an idealized estuary with a muddy bed, Journal of Physical Oceanography, 32(11), 3113-3131, doi: 10.1175/1520-0485(2002)032<3113:tisrat>2.0.co;2. Wang, X.H., Byun, D.S., Wang, X.L. & Cho, Y.K., 2005, Modelling tidal currents in a sediment stratified idealized estuary, Continental Shelf Research, 25(5-6), 655- 665, doi: 10.1016/j.csr.2004.10.013. Wang, X.H. & Pinardi, N., 2002, Modeling the dynamics of sediment transport and resuspension in the northern Adriatic Sea, Journal of Geophysical Research, 107(C12), 3225, doi: 10.1029/2001jc001303. Wang, X.H., Wang, H., Guan, W. & Guo, Z., 2011a, Dynamics of Chinese muddy coasts and estuaries: An introduction, Estuarine, Coastal and Shelf Sciences, 93(3), 171-172, doi: 10.1016/j.ecss.2011.03.010. Wang, Y., Yu, Q. & Gao, S., 2011b, Relationship between bed shear stress and suspended sediment concentration: Annular flume experiments, International Journal of Sediment Research, 26(4), 513-523, doi: 10.1016/s1001- 6279(12)60009-2. Warner, J.C., Sherwood, C.R., Signell, R.P., Harris, C.K. & Arango, H.G., 2008, Development of a three-dimensional, regional, coupled wave, current, and sediment-transport model, Computers & Geosciences, 34(10), 1284-1306, doi: DOI: 10.1016/j.cageo.2008.02.012. Water Monitoring Branch, 2005, The Health of the Aquatic Environment in the Darwin Harbour Region, 2004, Report No.: 5/2005D, Natural Resource Management Division, Department of Natural Resources, Environment and the Arts, Darwin,

127

REFERENCES

N.T., 70 pp. Available at: http://lrm.nt.gov.au/water/aquatic/ausrivas/darwinharbour#.UefVCdgr9tw, accessed on 10th May 2012. Water Research Laboratory, 2000, Modelling of Darwin Harbour, http://www.wrl.unsw.edu.au/site/wp-content/uploads/darwin-harbour.pdf, accessed on 4th November, 2011. Weir, D.J. & McManus, J., 1987, The role of wind in generating turbidity maxima in the Tay Estuary, Continental Shelf Research, 7(11-12), 1315-1318, doi: 10.1016/0278-4343(87)90035-5. Whitehouse, R.J.S., Harris, J.M., Sutherland, J. & Rees, J., 2011, The nature of scour development and scour protection at offshore windfarm foundations, Marine Pollution Bulletin, 62(1), 73-88, doi: 10.1016/j.marpolbul.2010.09.007. Wijnberg, K.M., 2002, Environmental controls on decadal morphologic behaviour of the Holland coast, Marine Geology, 189(3-4), 227-247, doi: 10.1016/S0025- 3227(02)00480-2. Williams, D., 2009, Part 1:Hydrodynamics and Sediment Transport, Dredging of Sand from Darwin Harbour, Hydrographic and Marine Life, Australian Institute of Marine Science, Arafura Timor Research Facility, Brinkin, Northern Territory, 1- 33 pp. Williams, D., Wolanski, E. & Spagnol, S., 2006, Hydrodynamics of Darwin Harbour, in The Environment in Asia Pacific Harbours, E. Wolanski (ed.), Springer, Dordrecht, The Netherlands, pp. 461-476. Wilson, D., Padovan, A. & Townsend, S., 2004, The water quality of spring and neap tidal cycles in the Middle Arm of Darwin Harbour during dry season, Report No.: 41/2004D, Natural Resource Management Division, Department of Natural Resources, Environment and the Arts, Darwin, N.T., 93 pp. Wolanski, E., 1992, Hydrodynamics of mangrove swamps and their coastal waters, Hydrobiologia, 247(1), 141-161, doi: 10.1007/bf00008214. Wolanski, E., Chicharo, L., Chicharo, M.A. & Morais, P., 2006a, An ecohydrology model of the Guadiana Estuary (South Portugal), Estuarine, Coastal and Shelf Science, 70(1-2), 132-143, doi: 10.1016/j.ecss.2006.05.029. Wolanski, E., Jones, M. & A Bunt, J., 1980, Hydrodynamics of a tidal creek-mangrove swamp system, Marine and Freshwater Research, 31(4), 431-450, doi:

128

REFERENCES

10.1071/MF9800431. Wolanski, E., Mckinnon, A.D., Williams, D. & Alongl, D.M., 2006b, An estuarine ecohydrology model of Darwin Harbour, Australia, in The Environment in Asia Pacific Harbours, E. Wolanski (ed.), Springer, Dordrecht, The Netherlands, pp. 477-488. Wolanski, E. & Simon, S., 2000, Pollution by mud of Great Barrier Reef coastal waters, Journal of Coastal Research, 16(4), 1151-1156, doi: 10.2307/4300132. Wolanski, E., Williams, D. & Hanert, E., 2006c, The sediment trapping efficiency of the macro-tidal Daly Estuary, tropical Australia, Estuarine, Coastal and Shelf Science, 69(1-2), 291-298, doi: 10.1016/j.ecss.2006.04.023. Woodroffe, C.D., Bardsley, K.N., Ward, P.J. & Hanley, J.R., 1988, Production of mangrove litter in a macrotidal embayment, Darwin Harbour, N.T., Australia, Estuarine, Coastal and Shelf Science, 26(6), 581-598, doi: 10.1016/0272- 7714(88)90035-2. Wright, S.A. & Schoellhamer, D.H., 2005, Estimating sediment budgets at the interface between rivers and estuaries with application to the Sacramento-San Joaquin River Delta, Water Resources Research, 41(9), W09428, doi: 10.1029/2004wr003753. Wu, J., Wang, Y. & Cheng, H., 2009, Bedforms and bed material transport pathways in the Changjiang (Yangtze) Estuary, Geomorphology, 104(3-4), 175-184, doi: 10.1016/j.geomorph.2008.08.011. Wu, Y., Falconer, R.A. & Uncles, R.J., 1999, Modelling of water flows and cohesive sediment fluxes in the Humber Estuary, UK, Marine Pollution Bulletin, 37(3-7), 182-189, doi: 10.1016/s0025-326x(99)00103-4. Xing, J. & Davies, A.M., 2003, A model study of tidally induced suspended sediment transport in the Iberian shelf edge region, Estuarine, Coastal and Shelf Science, 58(2), 321-333, doi: 10.1016/s0272-7714(03)00084-2. Xu, F., 2009, Modeling study of flocculation effects on sediment transport in estuaries, PhD Thesis, Stony Brook University, 118 pp. Yang, Z. & Khangaonkar, T., 2009, Modeling tidal circulation and stratification in Skagit River estuary using an unstructured grid ocean model, Ocean Modelling, 28(1-3), 34-49, doi: 10.1016/j.ocemod.2008.07.004. Zhang, S., Zhou, Q., Xu, D., Lin, J., Cheng, S. & Wu, Z., 2010, Effects of sediment

129

REFERENCES dredging on water quality and zooplankton community structure in a shallow of eutrophic lake, Journal of Environmental Sciences, 22(2), 218-224, doi: 10.1016/S1001-0742(09)60096-6.

130