Physical Processes of a Shallow Subtropical Estuarine System : Coombabah Lake, Gold Coast, Australia
Author Ali, Ayub
Published 2010
Thesis Type Thesis (PhD Doctorate)
School School of Engineering
DOI https://doi.org/10.25904/1912/1067
Copyright Statement The author owns the copyright in this thesis, unless stated otherwise.
Downloaded from http://hdl.handle.net/10072/366810
Griffith Research Online https://research-repository.griffith.edu.au
Physical Processes of a Shallow Subtropical
Estuarine System:
Coombabah Lake, Gold Coast, Australia
Ayub Ali
BSc in Civil Engineering, MSc in Hydroinformatics (Netherlands)
Griffith School of Engineering
Science, Environment, Engineering and Technology
Griffith University
Submitted in fulfilment of the requirements of the degree of
Doctor of Philosophy
February 2010
Declaration
This work has not previously been submitted for a degree or diploma in any university. To the best of my knowledge and belief, the thesis contains no material previously published or written by another person except where due reference is made in the thesis itself.
______
Ayub Ali
February 2010
i
ii
Acknowledgements
After four years of study, which was full of exciting new learning experiences, albeit far away from my beloved Country, Bangladesh, I have succeeded to complete my PhD study in Griffith University, Gold Coast, Queensland, Australia. Here, I would like to take the opportunity to express my sincere appreciation and deep gratitude to those who assisted me in my work and supported me in one way or another during this particular period.
I like to express my great respect and gratitude to my principal supervisor Charles Lemckert for his kind supervision, excellent guidance, continuous encouragement and valuable advice and suggestions during all stages of my study. He was always very helpful for discussion about research and career development. I was constantly inspired by your ability to manage your amazing academic career, adoring wife and five children – thank you so much for all of my PhD experiences.
I like to express my sincere thanks and appreciation to my associate supervisor Dr Hong Zhang for her guidance and cooperation during this study especially with numerical modelling. I like to thank all the research students and my friends forever, at Griffith University, who helped me during my Candidature: Ryan J K Dunn, Nathan Benfer, Peta Williams and Johann Gustafson for helping me in data collection. Thanks to the lab technicians who helped me in preparation and installation of the data collection devices. Thanks to other PhD students whom I shared my office with especially Hassan Jabur and Mainul Islam for their daily accompany.
The author is grateful to DHI Water and Environment for making the MIKE modelling software available for this study. Thanks to the Centre for Infrastructure and Environmental Management (CIEM), Griffith School of Engineering and Griffith University for providing financial support and necessary facilities to make this research possible.
And most importantly, I wish to extend my profound gratitude to my beloved wife and kids for their continuous sacrifice and inspiration to make this study successful. I would
iii
like to extend my sincere thanks to my brothers, sisters and relatives who all supported me whole-heartedly for their moral support during my long stay away from them.
Ayub Ali
Gold Coast
QLD, Australia
iv
Dedicated to my deceased parents
Who could be extremely happy with this achievement
If they were alive today.
v
vi
List of Publications
Below is a list of articles that were written and, either published or submitted for publication during the thesis candidature:
Publications directly related to this thesis: Ali, A. and Lemckert, C.J., 2009. A traversing system to measure bottom boundary layer hydraulic properties, Estuarine Coastal and Shelf Science. doi:10.1016/j.ecss.2009.04.017 Ali, A., Lemckert, C.J. and Dunn, R.J.K., 2010. Salt fluxes in a very shallow sub- tropical estuary. Journal of Coastal Research., 26(3), 436-443. Ali, A., Lemckert, C.J., Zhang, H. and Dunn, R.J.K., 2009b. Sediment dynamics of a very shallow sub-tropical estuarine system – Coombabah Lake, Gold Coast, Australia. Estuarine, Coastal and Shelf Science (under revision). Ali, A., Zhang, H. and Lemckert, C.J., 2009c. Numerical study of the hydrodynamics of a very shallow estuarine system - Coombabah Lake, Gold Coast, Australia. Journal of Coastal Research, SI 56, 922-926. Dunn, R.J.K., Ali, A., Lemckert, C.J., Teasdale, P.R., and Welsh, D.T., 2007. Short- term variability of physio-chemical parameters and the estimated transport of filterable nutrients and chlorophyll-a in the urbanised Coombabah Lake and Coombabah Creek system, southern Moreton Bay, Australia. Journal of Coastal Research, SI 50, 1099-1105.
Additional publications to that which makes up this thesis: Ali, A., Mynett, A.E. and Azam, M.H., 2007. Sediment Dynamics in the Meghna Estuary, Bangladesh: A Model Study. Journal of Waterway, Port, Coastal and Ocean Engineering, 133 (4), 255-263. Ali, A., Maraqa, M.A., Imran, H.D., Hamza, W. and Al Awadi, S. (2009). Effects of discharge characteristics on pollutant concentration at Jebel Ali Harbor, Dubai- UAE: A case study. Journal of Aquatic Ecosystems Health & Management (submitted).
vii
Maraqa, M.A., Ali, A., Imran, H.D., Hamza, W., and Al Awadi, S., 2008. Simulation of the hydrodynamic regime of Jebel Ali Harbor, Dubai-UAE. Journal of Aquatic Ecosystems Health & Management, 11 (1), 105-115. Maraqa, M.A., Ali, A. and Khan, N., 2007. Modelling Selected Water Quality Parameters at Jebel Ali Harbour, Dubai-UAE. Journal of Coastal Research, SI 50, 819-824.
viii
Abstract
Estuaries are of immense importance to many communities. It has been estimated that 60 to 80 % of commercial marine fishery resources depend on estuaries for part, or all of, their life cycle. The characteristics of estuarine flow, water quality and sediment conditions are important as they play a critical role in the functionality and health of these systems. This study utilised both field data and numerical modelling technique to help enhance our understanding of the physical processes of a very shallow subtropical estuarine system.
This study first quantified various salt flux components within the shallow subtropical estuarine system Coombabah Lake in Gold Coast, Queensland, Australia to better understand the system’s physical processes for assisting future management decisions in this ecologically and economically significant region. Residual water transport was identified as the dominant factor influencing residual salt transport, which alternates direction frequently.
This study then developed a new simple and robust traversing system to measure flow properties within estuarine bottom boundary layers to estimate two important parameters used in numerical modelling of aquatic systems: bed shear stress and bed roughness height. Four commonly-employed techniques: (1) Log-Profile; (2) Reynolds stress; (3) Turbulent Kinetic Energy; and (4) Inertial Dissipation used to estimate bed shear stresses from velocity measurements were also compared. Bed shear stresses estimated with these four methods agreed reasonably well; of these, the Log Profile method was found to be most useful and reliable for the unstratified conditions studied.
A three-dimensional hydrodynamic model with unstructured meshes utilising the MIKE3 FM modelling system and simulated the hydrodynamic regime was set up for Coombabah Lake to assist with enhancing our understanding of the hydraulic properties within this shallow sub-tropical estuarine system. The sensitivity of calibration parameters of a very shallow estuarine model was also investigated. The model utilised the hydraulic data collected by the newly developed traversing system and that collected
ix
during the first part of the study used to examine salt flux dynamics. The hydrodynamic regime of the lake was found to be favourable for settlement of suspended sediments. The results also revealed that the correct bathymetry is the most important parameter for accurate modelling, followed by appropriate bed roughness in the numerical scheme for very shallow environments.
This study finally provided an understanding of the sediment dynamics within Coombabah Lake and the surrounding waters. It utilised ten days of observed hydrodynamic and sediment data and employed the three-dimensional model with unstructured meshes utilising the MIKE3 FM modelling system. Sediment dynamics of the lake were found to be dominated by advection process driven by tides with wave and wind playing minor roles – even though the system was shallow. Simulation results agreed well with field data and supported the aforementioned findings. Correlation between TSS and turbidity was very poor; therefore, the employed automatic data logging system (turbidity meters) was determined inappropriate for the estimation of TSS concentration in the very shallow subtropical estuarine system.
x
Table of Contents
Title ...... page
Declaration ...... i
Acknowledgements...... iii
List of Publications ...... vii
Abstract ...... ix
Table of Contents...... xi
List of Figures ...... xvii
List of Tables...... xxi
List of symbols and acronyms ...... xxiii
Preface ...... xxvii
CHAPTER I INTRODUCTION AND LITERATURE REVIEW ...... 1
1.1 Introduction ...... 1
1.2 Literature Review...... 8 1.2.1 Salt Fluxes...... 8 1.2.2 Classification of Open Channel Flow...... 9 1.2.3 Bottom Boundary Layer ...... 12 1.2.4 Velocity Distributions in the Bottom Boundary Layer ...... 15 1.2.5 Techniques for Estimating Bed Shear Stress...... 19 1.2.6 Technique for Estimating Roughness Height and Drag Coefficient ...... 22
1.3 Thesis Objectives and Scopes ...... 23
1.4 Organisation of Thesis ...... 23
xi
1.5 References ...... 25
CHAPTER II SALT FLUXES WITHIN A VERY SHALLOW SUBTROPICAL ESTUARY ...... 35 Abstract...... 35
2.1 Introduction ...... 35
2.2 Methods ...... 40 2.2.1 Study site...... 40 2.2.2 Field measurement and data processing ...... 42 2.2.3 Quantification of salt fluxes ...... 43
2.3 Results and discussion...... 44 2.3.1 Hydrometeorological properties...... 44 2.3.2 Salt fluxes...... 46
2.4 Conclusions ...... 51
2.5 Acknowledgement...... 51
2.6 References ...... 51
CHAPTER III A TRAVERSING SYSTEM TO MEASURE BOTTOM BOUNDARY LAYER HYDRAULIC PROPERTIES ...... 55 Abstract...... 55
3.1 Introduction ...... 56
3.2 Theoretical Background ...... 57
3.3 Methods and Materials ...... 60 3.3.1 Techniques for estimating bed shear stress ...... 60 3.3.2 Technique for estimating roughness height and drag coefficient...... 63 3.3.3 New traversing system ...... 63 3.3.4 Data processing...... 68
xii
3.4 Results and Discussion ...... 72
3.5 Conclusions ...... 76
3.6 Acknowledgement...... 77
3.7 References ...... 77
CHAPTER IV NUMERICAL STUDY OF THE HYDRODYNAMICS OF A VERY SHALLOW ESTUARINE SYSTEM - COOMBABAH LAKE, GOLD COAST, AUSTRALIA ...... 83 Abstract...... 83
4.1 Introduction ...... 84
4.2 Study site ...... 87
4.3 Methods ...... 87 4.3.1 Field Measurement ...... 87 4.3.2 Numerical Model...... 88
4.4 Results and Discussion ...... 89 4.4.1 Field Data...... 89 4.4.2 Model Calibration...... 90 4.4.3 Simulation Results...... 93
4.5 Conclusions ...... 96
4.6 Acknowledgement...... 96
4.7 References ...... 96
CHAPTER V SEDIMENT DYNAMICS OF A VERY SHALLOW SUBTROPICAL ESTUARINE SYSTEM – COOMBABAH LAKE, GOLD COAST, AUSTRALIA ...... 99 Abstract...... 99
xiii
5.1 Introduction ...... 100
5.2 Coombabah Lake...... 102
5.3 Methods ...... 104 5.3.1 Field and Laboratory Measurements ...... 104 5.3.2 Numerical Model...... 106 5.3.3 Data Analysis...... 109
5.4 Results and Discussion ...... 109 5.4.1 Field Data...... 109 5.4.2 Model Calibration...... 115 5.4.3 Simulation Results...... 117
5.5 Conclusions ...... 121
5.6 Acknowledgments...... 122
5.7 References ...... 122
CHAPTER VI CONCLUSIONS AND RECOMMENDATIONS ...... 127
APPENDIX A SHORT-TERM VARIABILITY OF PHYSIO-CHEMICAL PARAMETERS AND THE ESTIMATED TRANSPORT OF FILTERABLE NUTRIENTS AND CHLOROPHYLL-A IN THE URBANISED COOMBABAH LAKE AND COOMBABAH CREEK SYSTEM, SOUTHERN MORETON BAY, AUSTRALIA ...... A-1 Abstract...... A-1
A.1 Introduction ...... A-2
A.2 Methodology...... A-4 A.2.1 Site description...... A-4 A.2.2 Experimental design...... A-5 A.2.3 Water sample collection and analysis...... A-6 A.2.4 In situ water quality parameters ...... A-7
xiv
A.2.5 Field instrumentation...... A-7 A.2.6 Transport estimations ...... A-8 A.2.7 Meteorological measurements...... A-8
A.3 Results and discussion...... A-9 A.3.1 Hydrological data...... A-9 A.3.2 Variability within Coombabah Lake and Creek...... A-10 A.3.3 Filterable nutrient and chl-a transport estimations ...... A-16
A.4 Conclusion...... A-17
A.5 Acknowledgements...... A-17
A.6 References ...... A-18
xv
xvi
List of Figures
Figure page
Figure 1.1: Coombabah Lake and surrounding water courses. 4
Figure 1.2: Coombabah Lake viewed from north along the main channel during low tide. 5
Figure 1.3: (a) Laminar and (b) turbulent flow. 11
Figure 1.4: Scientific classification of flow region. 13
Figure 1.5: Engineering classification of flow region (adapted from Liu, 2001). 14
Figure 1.6: Prandtl’s mixing length theory (adapted from Liu, 2001). 16
Figure 1.7: Illustration of the velocity profile in hydraulically smooth and rough flows. 19
Figure 2.1: Location map of the study site, Coombabah Lake (southern Moreton Bay, Australia). 42
Figure 2.2: Meteorological data collected on site: (a) wind speed; (b) wind direction; (c) rainfall; (d) evaporation; and (e) air pressure. 46
Figure 2.3: Hydraulic data collected at Station A: (a) water depth; (b) water velocity; (c) salinity; and (d) computed salt fluxes. 48
Figure 2.4: Hydraulic data collected at Station B: (a) water depth, (b) water velocity, (c) salinity and (d) computed salt fluxes. 49
Figure 2.5: Average hydraulic data at both Stations A and B: (a) water depth; (b) salinity; and (c) water velocity. 50
xvii
Figure 3.1: Typical velocity and shear stress distribution within different flow regions (layer thickness is not to scale) of a turbulent bottom boundary layer. 58
Figure 3.2: New traversing system. 64
Figure 3.3: Schematic set-up of Altimeter and ADV probe. 65
Figure 3.4: Location map of the study site, Coombabah Creek and its adjacent estuaries (adapted from Benfer et al., 2007). 68
Figure 3.5: Sample of measured and fitted velocity profiles: (a) stationary ADV; and (b) moving ADV. 69
Figure 3.6: Time series of measured and estimated parameters: (a) tidal level; (b) mean velocity; (c) bed shear stress; (d) roughness height; and (e) drag coefficient. 72
Figure 3.7: Sample profiles of shear stress: (a) turbulent shear stress estimated by Reynolds stress and TKE methods; and (b) bed shear stress estimated by ID method using dissipation rates at different heights and the same by LP method. 73
Figure 4.1: Location map of study site (left) and bathymetry of Coombabah Lake including measurement stations (right). 86
Figure 4.2: Correlation coefficients between observed and simulated water levels and velocities for various bed level lowering. 91
Figure 4.3: Comparison of tide level (top) and velocity (bottom) at Station 1. 92
Figure 4.4: Simulated flow field at the onsets of flood tide. 93
xviii
Figure 4.5: Simulated flow through bifurcation channels at Lake entrance (positive seaward). 94
Figure 4.6: Simulated bed shear stress at ebb tide. 95
Figure 4.7: Observed and simulated turbulent kinetic energy. 95
Figure 5.1: (a) Location map of the study site; and (b) Lake bathymetry with sampling (circle) and model boundary (triangle) locations (AHD means the Australian Height Datum). 103
Figure 5.2: Measured hydraulic and water quality data: (a) water depth; (b) flow velocity; (c) salinity; and (d) turbidity. 110
Figure 5.3: Tidally averaged hydraulic and water quality data: (a) water depth; (c) salinity; and (d) turbidity. 111
Figure 5.4: On-site meteorological conditions: (a) wind speed; (b) wind direction; (c) rainfall; and (d) air pressure. 112
Figure 5.5: Observed TSS concentrations (two data almost at the same time represent two samples) and total water depth at Station 4. 113
Figure 5.6: Correlation between TSS and turbidity measured at Stations 1-8 during the study period. 114
Figure 5.7: Comparison of simulated and observed TSS: (a) entire study period; and (b) one tidal cycle. 116
Figure 5.8: Simulated TSS concentrations: (a, b) with no waves and winds; (c, d) with waves; and (e, f) with winds (left panel during flood tides and right panel during ebb tides). 118
Figure 5.9: Simulated TSS concentrations at Station 4 without and with wave and wind activity included. 119
xix
Figure 5.10: Simulated TSS concentration for various bed roughness values. 120
Figure A.1: Study area of Coombabah Lake and Coombabah Creek in southern Moreton Bay, Australia (inset). A-5
Figure A.2: Water depth (left) and sample velocity profiles (right) of the sampled Coombabah Lake entrance channel. A-9
Figure A.3: Coombabah Lake-Creek system filterable nutrient concentration box-plot representation. A-14
Figure A.4: Coombabah Lake filterable reactive PO43- mean concentrations during summer ebb and flood tides. A-16
xx
List of Tables
Table page
Table 2.1: Previously reported salt fluxes per unit width (positive seaward and negative landward) divided by mean water depth 36
Table 3.1: Bed shear stresses (N/m2) estimated by various methods 74
Table 4.1: Observed bed shear stress and bed roughness length 89
Table A.1: Minimum (min), maximum (max), mean, standard deviation (sd), and percent relative standard deviation of physio-chemical parameters, chl-a and suspended solids within the Coombabah Lake-Creek system (n = 42) A-11
xxi
xxii
List of symbols and acronyms
ADCP Acoustic Doppler Current Profiler
ADV Acoustic Doppler Velocimeter
BBL Bottom Boundary Layer
DHI Danish Hydraulic Institute
FM Flexible Mesh
TSS Total suspended solids
LP Logarithmic Profile
RS Reynolds Stress
TKE Turbulent Kinetic Energy
ID Initial Dissipation km kilometre ppt parts per thousand psu practical salinity unit
SNR Signal Noise Ration
3D three dimensional
CTD Conductivity, Temperature and Depth
GHD Gutteeridge, Haskins and Davey
GPS Global Positioning System
DEM Digital Elevation Model
xxiii
Hz Hertz
RTK Real Time Kinematic
PLC Procedural Language Code
A cross sectional area
CD Drag coefficient
Cuu autocovariance of u
Cww autocovariance of w
Cuw covariance of u and w
d50 median diameter of sediment err errors related to shear stress in LP method
euw errors of shear stress in RS method
F Flux
fs sampling frequency h water depth
ks roughness height
L characteristic length l Prandtl’s mixing length
P shear production
P wetted perimeter
N degrees of freedom
xxiv
T sampling period
R hydraulic radius
Re Reynolds number s salinity s depth-averaged mean salinity s deviation from the depth-averaged mean salinity
sT time-averaged mean salinity
s time-averaged mean of depth-averaged mean salinity
U instantaneous velocity in x direction
U mean speed u flow velocity in x direction u depth-averaged mean velocity in x direction u deviation from averaged velocity in x direction
uT time-averaged mean velocity in x direction
u time-averaged mean of depth-averaged mean velocities in x
direction
u* shear velocity
V instantaneous velocity in y direction v flow velocity in y direction v deviation from averaged velocity in x direction
xxv
W instantaneous velocity in z direction w flow velocity in z direction w deviation from averaged velocity in x direction z vertical coordinate
z0 elevation where velocity is zero
δ thickness of boundary layer
δv thickness of viscus sublayer
ε energy dissipation
η dynamic viscosity of fluid
κ von-Kármán constant
μm micrometre
ν kinematic viscosity of fluid
ρ density of fluid
b bed shear stress
t turbulent shear stress
v viscus shear stress
z total shear stress
xxvi
Preface
This thesis is based around a series of preprints (Chapters 2, 3 and 4) whilst Chapter 5 was submitted to Estuarine, Coastal and Shelf Science for review and is currently under revision. The preprints are published works in fully refereed journals and are reproduced here word by word except for figure, table and equation numbers. The first author (i.e. the PhD candidate) was responsible for data collection, analysis and interpretation of the results with the guidance of second the author, whilst the third and fourth authors (where named) helped in data collection process and physical understanding of the results. All of these works are related to the physical processes of a shallow subtropical estuarine system – Coombabah Lake in Gold Coast, Australia.
Chapter 1 contains the thesis introduction and a literature review followed by brief descriptions about the scope of the thesis and organisation of the thesis report. Chapter 6 summarises the research conclusions and recommendations of future studies. All chapters including Appendix A are self contained with figure, table and equation numbering unique to each. Each chapter contains the references cited within its texts.
xxvii
xxviii Chapter I Introduction and literature review
CHAPTER I
Introduction and Literature Review
1.1 Introduction
Estuaries are of immense importance to many communities. It has been estimated that 60 to 80 % of the commercial marine fishery resources depend on estuaries for part of or all of their life cycle (Klen, 2006). An estuary is a semi-enclosed coastal body of water, which has a free connection with the sea, and is thus strongly affected by tidal action. The characteristics of estuarine flow and sediment (and/or pollutant) transport patterns are important as they play a critical role in the functionality and health of these systems. When bottom sediment is resuspended, trace metals, nutrients and organic contaminants are released into the water column, which in turn can limit the amount of light entering the water and reduce water quality (Morris and Howarth, 1998). Sediment settling can inhibit channel continuity by deposition in navigational areas. If any of these issues creates a significant problem, management strategies must be developed and implemented in order to rectify the situation and/or preserve the environment in a healthy state. These strategies usually involve collection of field data and development of numerical models. The numerical models must be developed using sound scientific principles. However, many knowledge gaps still exist, which results in most models relying on the use of approximations when determining boundary conditions and sediment (or pollutant) transport dynamics.
The mechanisms that control the transport, resuspension and deposition of the fine suspended sediments (or pollutants) in tidal estuaries are extremely complex. They are directly influenced by highly variable hydrodynamic conditions near the bed and by sediment type, size and composition. There have been many investigations of near-bed flows, resuspension and transport of sediments under natural field conditions, with the majority of these having non-cohesive sediments in estuaries [e.g. (Soulsby et al., 1984,
Ayub Ali 1 Chapter I Introduction and literature review
1994; Brown and Davies, 2009) offshore sites (Soulsby and Dyer, 1981; Williams et al., 1999; Nikora et al., 2002, Ma at al., 2008) or coastlines dominated by wind wave effects (Davies, 1985; Butt and Russell, 2000; Conley and Griffin, 2004); Verney et al, 2007]. However, studies of near-bed flows, resuspension and transport of cohesive sediments within shallow estuarine systems are limited (Uncles and Stephens, 2009). Compared to sand dynamics, cohesive sediment dynamics are significantly more complicated due to the complexity of relevant physical, chemical, and biological processes and their spatial and temporal variability.
Cohesive sediments and other suspended (or dissolved) water quality parameters are primarily transported by advection-dispersion processes. The advection process transports the materials by fluid velocity and the dispersion process mixes them with the transporting medium by velocity gradients. On the other hand, resuspension and deposition depends primarily on the hydrodynamic properties (e.g. shear stress) of the Bottom Boundary Layers (BBL) and the bed properties (e.g. bed materials and bed forms) rather than the properties of fluid and transporting substances. The effects of grain size and bed forms are among the bed properties that can be expressed by bed roughness height.
It is relatively difficult to determine the bed shear stress directly in the field because it requires measurements very near the bottom (within a centimetre). Previously, point source devices, such as S4 or Acoustic Doppler Velocimeters (ADVs) current meters (Jing and Ridd, 1996; Osborne and Boak, 1999; Stips et al., 1998, Gross et al., 1994; Black, 1998; Yuan et al., 2008) have been used to derive boundary layer properties. However, in traditional fixed mooring arrangements they cannot fully resolve the boundary layer properties. Additionally, if a detailed boundary layer profile is to be determined, a number of devices must be deployed at one time, which is usually beyond the scope of most researchers. More recently, Acoustic Doppler Current Profilers (ADCP) have been used to record velocity data near the bed (Thomsen, 1999; Osalusi et al., 2009) as they can provide near instantaneous three-dimensional velocity profile data that can be used to estimate the flow structure and provide rough estimates of shear stress. However, they have limitations in that they cannot sample close to the bed
Ayub Ali 2 Chapter I Introduction and literature review
(approximately 10% of the sampling depth), which is the most important part in case of thin boundary layers.
Another important parameter, bed roughness (ks), is an essential parameter in modelling of current circulations, wave height attenuations and sediment transport in estuarine and coastal waters, but it is often unknown and difficult to measure directly in the field (You, 2005). The bed roughness generally consists of three roughness components as grain or Nikuradse roughness; bed form roughness; and sediment saltation roughness (You, 2006). The grain roughness is the smallest form of roughness and is commonly taken as ks = 2.5d50 where d50 is the median diameter of sediment. The bed form roughness is generated by sand ripples, biogenic mounds or benthic seagrasses. Several empirical formulas have been suggested for estimation of ripple length and height in steady flow (van Rijn, 1984; Raudkivi, 1998) and in oscillatory flow (Nielsen, 1981; Grant and Madsen, 1982). The roughness of sediment saltation has been studied by few investigators, e.g. Raudkivi (1989) and You and Nielsen (1996). The roughness of biogenic mounds is often empirically estimated from photographs of the seabed (Grant et al., 1984; Wheatcroft, 1994). No single empirical formula has been made available for estimation of the total roughness of sediment grain, irregular sand ripples, biogenic mounds, benthic seagrasses and sediment saltation. Alternatively, the total bed roughness may be directly estimated by fitting the logarithmic velocity distribution, the von Kármán–Prandtl velocity equation, to individual current profile measured near the bed in the absence of surface waves. This fitting method is called the log-fit method (You, 2005). However, two essential conditions are required by the log-fit method. One is that the current profile must be measured at more than three levels close to the bed, and the other is that the measured current profile must be of the logarithmic distribution. The first condition requires more than three current meters to measure individual current profile. However, the number of the current meters that can be deployed within the logarithmic layer is often limited to four in the field (e.g. Ke et al., 1994). The current research selected Coombabah Lake-Creek system as a case study site.
Coombabah Lake-Creek system situated in Gold Coast, Queensland, Australia (see Figure 1.1) is one of the fastest growing cities in the developed world (Skinner et al., 1998). The lake is the largest estuarine lake in southern Moreton Bay covering ~2 km2
Ayub Ali 3 Chapter I Introduction and literature review
(GHD, 2003) with an urbanised catchment area of 44 km2, characterised by residential, commercial and light industrial developments. Coombabah Creek is a 17 km long, moderately impacted (Cox and Moss, 1999; Lee et al., 2006) sub-tropical tidal creek that flows through Coombabah Lake. The lake is a shallow body of water (see Figure 1.2) characterised by fine sediments (Dunn et al., 2007b) located in the mid-tidal region of the creek, with urban development positioned to the east and along the southern and western shorelines. The creek enters into the lake at south-west side and leaves the same from the north-east side. Ultimately, Coombabah Creek discharges into the Gold Coast Broadwater, within southern Moreton Bay.
Figure 1.1: Coombabah Lake and surrounding water courses.
As a consequence of the ecological significance and the potential for anthropogenic disturbances and inputs into the semi-urbanised Coombabah Lake, the lake and surrounding wetlands have been the focus of recent scientific effort (e.g. Frank and Fielding, 2004; Hollingsworth and Connolly, 2006; Burton et al., 2008; Dunn et al., 2008; Knight et al., 2008; Benfer et al., 2007). Variations in surface sediments nutrient
Ayub Ali 4 Chapter I Introduction and literature review concentrations, and observed cyclic variations in intratidal variability of physio- chemical and biological parameters have previously been attributed to the variability of sediment sources, hydrodynamic regime and increased freshwater input following rainfall events, and diurnal cycles, respectively (Dunn et al., 2007a, 2007b).
Figure 1.2: Coombabah Lake viewed from north along the main channel during low tide.
Dunn et al. (2007b) determined the physio-chemical properties of the bed sediments within Coombabah Lake and attributed variations of nutrient concentrations to the variations in sediment source, hydrodynamic regime of the lake and inputs during periods of freshwater flow after rainfall events. Dunn et al. (2007a) explored the intratidal variability of water quality parameters within the Coombabah Lake-Creek waters and estimated the transport of filterable nutrients and chlorophyll-a at the Coombabah Lake entrance (Appendix A). Cyclic variations of physio-chemical parameters (suspended solids and chlorophyll-a concentrations) were observed, supposedly as a result of tidal and diurnal cycles.
Ayub Ali 5 Chapter I Introduction and literature review
Observations are the most realistic source of data for understanding the physical processes, in particular the mixing and transport of sediments/contaminants within estuarine systems. Additionally, a well calibrated and validated numerical model can provide spatially and temporally variable data using information only at boundaries. The numerical model is very useful for assessing impacts of any intervention within the modelling domain and it is very important for selecting a suitable management strategy for aquatic systems. This study adopted both techniques for better understanding of the physical processes within a very shallow subtropical estuarine system.
Since, water circulation and mixing processes are the primary component to understand the transport processes within estuarine systems; this study first estimated various components of salt fluxes and determined the dominant transport processes for initial understanding of the sediment/contaminant transport processes within a very shallow subtropical estuarine system – Coombabah Lake-Creek system in Gold Coast, Queensland, Australia (Ali et al., 2009a). The salt fluxes were estimated utilising intensively collected velocity and salinity data.
It is well recognised that the reliability of numerical models depends extensively on the accuracy of input parameters (Lasance, 2002; Vreugdenhil, 2002; Owais et al., 2008; Mitchell et al., 2008). Further, knowledge of the hydrodynamic properties in the bottom boundary layers is necessary to establish relationships between bottom currents, shear stress, and near-bed sediment response to flows. These relationships are essential to understand sediment transport pathways. This study developed a new simple and robust traversing system to measure flow properties within estuarine bottom boundary layers to estimate the bed shear stress and the bed roughness height (Ali and Lemckert, 2009). The system was successfully applied to Coombabah Lake-Creek system.
This study then set up a three-dimensional hydrodynamic model with unstructured meshes utilising the MIKE3 FM (DHI, 2008) modelling system and simulated the hydrodynamic regime of Coombabah Lake for better understanding of hydraulic properties within shallow subtropical estuarine systems (Ali et al., 2009b). The sensitivity of calibration parameters of a very shallow estuarine model is also investigated. The model utilised the hydraulic data collected by the newly developed traversing system.
Ayub Ali 6 Chapter I Introduction and literature review
This study finally provided an understanding of the sediment dynamics within Coombabah Lake and the surrounding waters. It utilised ten days of hydrodynamic and sediment data and employed a three-dimensional model with unstructured meshes utilising the MIKE3 FM (DHI, 2008) modelling system. The model simulated the sediment dynamics for a better understanding of the sediment/contaminant transport processes within the shallow intertidal lake from a well calibrated numerical model. Furthermore, the sensitivity of bed roughness on TSS concentration in this shallow system was also examined utilizing the numerical model.
The following section summarises the literature on estuarine mixing, bottom boundary layer hydraulic properties and sediment dynamics that have been extensively used in this study.
Ayub Ali 7 Chapter I Introduction and literature review
1.2 Literature Review
This thesis investigated the mixing and hydraulic properties of bottom boundary layers because a fully developed bottom boundary layer occupies the entire water column of a shallow estuarine system. Therefore, some important literature related to mixing and bottom boundary layer is illustrated in this section. This section summarised the classification of open channel flow and its hydraulic properties (e.g. bed shear stress, velocity) based on the effect of active forces. Characteristics of a bottom boundary layer and various techniques of measuring the bottom boundary layer hydraulic properties are also briefly described in this section.
1.2.1 Salt Fluxes
To quantify the water circulation and mixing processes within estuarine systems, it is often convenient to study salinity distribution patterns because salinity is typically considered as a conservative tracer (Dyer, 1997). The behaviour of salts within the water column provides a basis for predicting transport of other soluble conservative substances. Total salt flux within an estuarine system is typically estimated using velocity and salinity concentrations at several water depths along a vertical section measured at different locations across the section (e.g., Bowden, 1963; Fischer, 1972; Sylaios et al., 2006). However, observations of currents and salinity collected at one location along the cross-section of a main channel within an estuarine system may also provide a valid experimental approach when assessing the relative importance of different transport processes (e.g., Lewis and Lewis, 1983; Pritchard, 1954; Restrepo and Kjerfve, 2002; Simpson et al., 2001; Uncles and Jordan, 1979; Uncles et al., 1985). In general, the net salt flux per unit width perpendicular to the main flow (F) can be calculated as follows (Bowden, 1963; Dyer, 1997; Kjerfve, 1986; Restrepo and Kjerfve, 2002):
h F u(z)s(z)dz (1.1) 0 where F represents the net flux (ppt m2/s); u(z) represents the observed velocity (m/s); s(z) represents the observed salinity (ppt); z represents the vertical coordinate; and h
Ayub Ali 8 Chapter I Introduction and literature review represents water depth (m). The instantaneous velocity and salinity can be decomposed as u u uand s s s , where the primed quantities represent deviations from the depth-averaged means, u and s . The depth-averaged means may be decomposed into time-averaged means and time varying components: u u uT and s s sT such that:
u u uT u (1.2)
s s sT s (1.3) where angle brackets are net, or time-averaged, over at least one complete tidal cycle. By substituting Equations (1.2) and (1.3) into Equation (1.1), the net salt flux can be decomposed into the following terms (Restrepo and Kjerfve, 2002):
F h u s huT sT u hT sT s hT uT hus (1.4) which, for simplicity, can be written as:
Net flux = Flux 1 + Flux 2 + Flux 3 + Flux 4 + Flux 5 (1.5) where Flux 1 represents the advective salt flux due to water discharge and change in storage volume during the tidal cycle; Flux 2 represents the sloshing effect – the tidal dispersion via triple correlation between tidal depth change, tidal current, and tidal salinity, usually directed upstream; Flux 3 represents the cross correlation between tide and salinity; maximum (and positive) when tide and salinity are in phase, and minimum (and negative) when they are out of phase; Flux 4 represents the Stokes’ drift dispersion; and Flux 5 represents the salt dispersion due to mean shear produced by gravitational circulation.
1.2.2 Classification of Open Channel Flow
Mixing in estuarine systems primarily depends on hydraulic properties of the system and hydraulic properties again depend on the type of flow. Therefore, this section briefly discusses flow types and its key characteristics. Open channel flow is governed
Ayub Ali 9 Chapter I Introduction and literature review to a large degree by viscosity, gravity and inertial forces. Based on the effect of viscosity relative to inertia, the flow is classified as laminar, turbulent, or transitional.
Laminar flow
The flow is laminar if the viscous forces are so strong relative to the inertia forces that viscosity plays a significant part in determining the flow behaviour. The Reynolds number (Re), defined by Equation (1.6) is the ratio of inertial to viscous forces:
uL Re (1.6) where u is the mean velocity of flow, L is the characteristic length and ν is the kinematic viscosity of the fluid. For laminar flow, Re is small, usually less than 500 in open channels.
The characteristic length L is chosen according to the characteristics of the flow. With pipe flow, the diameter D or the hydraulic radius R of the pipe is the length L. In free surface flow, the choice of L is difficult, particularly if the shape of the cross section is complex. It is customary to choose the hydraulic radius, R as the characteristic length which is defined as:
A R (1.7) P where A is the flow area and P is the wetted perimeter. However, in practice water depth h is used as hydraulic radius for open channel flow.
In laminar flow, fluid particles appear to move in definite smooth, continuous paths in straight channels and there is no significant transverse mixing as the fluid flows from point-to-point (Figure 1.3 (a)). However, in natural flows the disturbances are so great that laminar conditions rarely exist.
For the analysis of laminar flow, the shear stress ( v ) is defined as:
du (1.8) dz
Ayub Ali 10 Chapter I Introduction and literature review where is the density of the fluid.
The shear stress is the internal stress within a fluid which resists deformation or change of shape of fluid mass during motion and it exists only under dynamic conditions.
velocity, u velocity, u
τ
dz
du height, z height, z (a) Laminar flow (b) Turbulent flow
Figure 1.3: (a) Laminar and (b) turbulent flow.
Turbulent flow
The flow is turbulent if the viscous forces are weak relative to the inertial forces, the Reynolds number is large, usually greater than 2000. With turbulent flow, the fluid particles move in irregular paths which are neither smooth nor fixed but still represent the forward motion of the fluid (Figure 1.3 (b)). For the analysis of turbulent flow, the shear stress ( t ) is defined by the relation:
du t (1.9) dz where η is a turbulent mixing coefficient, corresponding to the dynamic viscosity μ in laminar flow and is, therefore, often called “dynamic,” or “eddy” viscosity. The eddy viscosity η is not a property of the fluid like ρ and ν, but depends on the velocity u.
Most flows in nature are turbulent. Turbulence is generated by instability in the flow, which triggers vortices. A typical phenomenon of turbulent flow is the fluctuation of velocity:
Ayub Ali 11 Chapter I Introduction and literature review
U u u', V v v' and W w w' (1.10) where U, V and W are instantaneous velocity in x, y and z directions respectively; u, v and w are time-averaged velocity in x, y and z directions respectively; u΄, v´ and ẃ are instantaneous velocity fluctuation in x, y and z directions respectively.
In turbulent flow, the water particles move in very irregular paths, causing an exchange of momentum from one portion of fluid to another, and hence, the turbulent shear stress (Reynolds stress) which is given by (time-averaging of the Navier-Stokes equation):
t u'w' (1.11)
In turbulent flow both viscosity and turbulence contribute to shear stress. The total shear
( z ) stress is:
z v t (1.12)
Transitional flow
Between laminar flow and turbulent flow, there is a mixed or transitional state. The value of the Reynolds number is usually between 500 and 2000 in the case of pipe flow. With free surface flows, the limits defining laminar and turbulent flow are slightly different. Laminar flow ends when the Reynolds number is around 103 and the fully 4 turbulent flow begins when it is about 10 . But these limits are also a function of R/ks (R is the hydraulic radius and ks is the Nikuradse roughness height) and form of the channel in which the flow occurs.
1.2.3 Bottom Boundary Layer
Near the bottom of an open channel, the flow has a distinct structure, called a boundary layer. The most important aspect of a boundary layer is that the velocity of the fluid goes to zero at the boundary. This is called the "no-slip" condition, i.e. the fluid velocity matches the boundary velocity. Figure 1.4 depicts a typical mean velocity profile, u(z), above a solid boundary where z denotes the distance above the boundary. With u(0)=0, at some distance above the boundary the velocity reaches a constant value, U∞, called the free stream velocity. Between the bed and the free stream the velocity varies over the
Ayub Ali 12 Chapter I Introduction and literature review vertical coordinate. The height of the boundary layer, δ, is typically defined as the distance above the bed at which u = 0.99U∞.
τz Total shear stress Velocity profile Flow layer classification τv Viscous shear stress velocity, u(z) τt Turbulent shear stress Free stream
Turbulent outer layer
τz = τt
z δ height,
τt Turbulent logarithmic layer τz = τt = const.
Transition layer τz = τt + τv = const. τ δv Viscous sublayer τz = τv = const. v
u(δ) = 0.99u∞ Bottom shear stress, τb Figure 1.4: Scientific classification of flow region (reproduced from Ali & Lemckert, 2009).
When the velocity is comparatively low, the heights of the protrusions forming the irregularities of the bottom of a channel are small compared to the thickness of the laminar flow. Small eddies appear at the top of the protrusions but are rapidly damped in the laminar flow. These eddies continue developing as the flow velocity increases. When eddies are so strong that they regenerate turbulence within the boundary layer, the depth of the laminar flow decreases. The small laminar flow region is called the viscous sub layer and its thickness is denoted as δν. The velocity distribution is parabolic in the laminar flow and logarithmic in the turbulent flow region.
For the viscous sub layer, the effect of the bottom (or wall) roughness on the velocity distribution was first investigated for pipe flows by Nikuradse (1933). He introduced the concept of equivalent grain roughness ks (Nikuradse roughness or bed roughness). The only situation where we can directly obtain the bed roughness is a flat bed consisting of uniform spheres, where ks = diameter of sphere. However, in nature, the bed is composed of grains with different sizes. Moreover, the bed is not flat and various bed forms, e.g. sand ripples or dunes, will appear depending on grain size and current. In
Ayub Ali 13 Chapter I Introduction and literature review that case, the bed roughness can be obtained indirectly by the velocity measurement. Based on experimental data, it was found (Figure 1.5):
u k (1) Hydraulically smooth flow for * s 5 , where u is the shear velocity. Bed * roughness is much smaller than the thickness of viscous sublayer. Therefore, the bed roughness will not affect the velocity distribution.
u k (2) Hydraulically rough flow for * s 70 . Bed roughness is so large that it produces eddies close to the bottom. A viscous sub layer does not exist and the flow velocity is not dependent on viscosity.
u k (3) Hydraulically transitional flow for 5 * s 70 . The velocity distribution is affected by bed roughness and viscosity.
Figure 1.5: Engineering classification of flow region (adapted from Liu, 2001).
The Bottom Boundary Layer (BBL) of a turbulent flow can be subdivided into four regions (see Figure 1.4). Starting from the bottom, we have:
(1) Viscous sub layer: a thin layer just above the bottom. In this layer there is almost no turbulence and the viscous shear stress in this layer is constant (Granger, 1985). The flow is laminar. Above this layer the flow is turbulent.
(2) Transition layer: also called buffer layer. Viscosity and turbulence are equally important.
Ayub Ali 14 Chapter I Introduction and literature review
(3) Turbulent logarithmic layer: viscous shear stress can be neglected in this layer. Based on measurement, it is assumed that the turbulent shear stress is constant and equal to bottom shear stress. It is in this layer where Prandtl (1926) introduced the mixing length concept and derived the logarithmic velocity profile.
(4) Turbulent outer layer: velocities are almost constant because of the presence of large eddies which produce strong mixing of the flow.
In the turbulent logarithmic layer, the measurements show that the turbulent shear stress is constant and equal to the bottom shear stress. By assuming that the mixing length is proportional to the distance to the bottom (l = κz, where κ is the proportionally constant), Prandtl (1926) obtained the logarithmic velocity profile.
Various expressions have been proposed for the velocity distribution in the transitional layer and the turbulent outer layer. None of them are widely accepted. However, by the modification of the mixing length assumption, the logarithmic velocity profile applies also to the transitional layer and the turbulent outer layer. Measurement and computed velocities show reasonable agreement (Liu, 2001). Therefore, from an engineering point of view, a turbulent layer with the logarithmic velocity profile covers the transitional layer, the turbulent logarithmic layer and the turbulent outer layer (Figure 1.5).
1.2.4 Velocity Distributions in the Bottom Boundary Layer
Turbulent layer
In the turbulent layer the total shear stress contains only the turbulent shear stress. The total shear stress increases linearly with depth such that:
z t z b 1 (1.13) h
where b is the bed shear stress; h is the total depth; z is the elevation above bed and t is the shear stress at elevation z.
Ayub Ali 15 Chapter I Introduction and literature review
Prandtl (1926) introduced the mixing length concept in order to calculate the turbulent shear stress. He assumed that a fluid parcel travels over a length l (Figure 1.6) before its momentum is transferred.
Figure 1.6 shows the time-averaged velocity profile. A fluid parcel, located in layer 1 and having the velocityu1 , moves to layer 2 due to eddy motion. There is no momentum transfer during this movement, i.e. the velocity of the fluid parcel is still u1 when it just arrives at layer 2, and decreases to u2 some time later by the momentum exchange with other fluid in layer 2. This action will speed up the fluid in layer 2, which can be seen as a turbulent shear stress t acting on layer 2 trying to accelerate layer 2.
Layer 1
Layer 2
Figure 1.6: Prandtl’s mixing length theory (adapted from Liu, 2001).
The horizontal instantaneous velocity fluctuation of the fluid parcel in layer 2 is:
du u' u u l (1.14) 1 2 dz
Assuming the vertical instantaneous velocity fluctuation having the same magnitude:
du w' l (1.15) dz where negative sign is due to the downward movement of the fluid parcel, the turbulent shear stress now becomes:
Ayub Ali 16 Chapter I Introduction and literature review
2 2 du t u'w' l (1.16) dz
By Prandtl’s mixing length theory, assuming the mixing length:
0.5 z l z 1 (1.17) h
The famous logarithmic velocity profile is derived:
u* z uz ln (1.18) z0
where u is the velocity at elevation z, u* is the shear velocity defined as:
u b (1.19) *
and z0 is the elevation where velocity is zero and is the von-Kármán constant equal to 0.4 for pipe flow. Based on experimental data, Nikuradse found that:
u k 0.11 Hydraulically smooth flow * s 5 u * u*k s (1.20) z0 0.033k s Hydraulically rough flow 70 u*k s 0.11 0.033k s Hydraulically transition flow 5 70 u*
It is interesting to note that the friction velocity is the flow velocity at the elevation
z z0e i.e. u z z0e u* . However, the friction velocity is the fluid velocity very close to the bottom.
Viscous sublayer
In the case of hydraulically smooth flow there is a viscous sublayer. Viscous shear stress is constant in this layer and equal to the bottom shear stress:
Ayub Ali 17 Chapter I Introduction and literature review
du (1.21) v dz b
Integrating Equation 1.21 and applying u |z0 0 gives
b u 2 uz z * z (1.22)
Thus, there is a linear velocity distribution in the viscous sublayer. The linear velocity distribution intersect with the logarithmic velocity distribution at elevation z 11.6 u* , yielding a theoretical viscous sublayer thickness 11.6 . The u* velocity profile is illustrated in Figure 1.7, with the detailed description of the fluid velocity near the bottom.
In estuaries the boundary layer is typically rough (Mehta, 1978; Perlin and Kit, 2002; Bricker et al., 2005), hence we need to resolve this to find bed shear stress as this is required in models and design of hydraulic structures. The bed shear stress can be estimated from near-bed velocity data (see also Chapter 3).
Ayub Ali 18 Chapter I Introduction and literature review
Figure 1.7: Illustration of the velocity profile in hydraulically smooth and rough flows (adapted from Liu, 2001).
1.2.5 Techniques for Estimating Bed Shear Stress
There are four commonly-employed techniques to estimate bed shear stress from velocity measurements:
(1) Log-Profile (LP);
(2) Reynolds stress (RS);
(3) Turbulent Kinetic Energy (TKE); and
(4) Inertial Dissipation (ID) methods.
The suitability, assumptions and limitations of these methods have been critically reviewed by Kim et al. (2000) and Pope et al. (2006). These authors concluded that the TKE approach was the most consistent and offered most promise for future development. However, they have suggested simultaneous use of several methods to
Ayub Ali 19 Chapter I Introduction and literature review estimate bed shear stress where possible, as all of these methods have both advantages and disadvantages; in this way, likely sources of errors can be identified.
The LP method fits velocity and height data into the von Kármán–Prandtl equation (Equation 1.18) and estimates shear velocity and roughness height. The shear velocity is used to calculate bed shear stress from:
2 b u* (1.23)
One of the main problems with this law of the wall approach (LP method) is that the theory is strictly valid only for steady flows (Cheng et al., 1999; Pope et al., 2006; Liu et al., 2009). Another fundamental feature of the LP method is that it is critically dependent upon precise knowledge of the elevations above the bed at which the sequence of current velocities are measured (Kabir and Torfs, 1992; Biron et al., 1998). While this may be straightforward for very smooth, fine-grained, abiotic sediments, this can be considerably problematic in the case of natural estuarine systems where grain size variation, bed forms and biota may conspire to increase bed roughness and make precise determination of elevation less certain (Kabir and Torfs, 1992; Wilcock, 1996).
The RS approach (Equation 1.11) may appear to represent a suitable method of estimating bed shear stress for fully turbulent flow with a large Reynolds number (Dyer, 1986), and for cases where measurements close to the bed are available. However, it has been shown that this method may also be largely unsuitable in field or laboratory studies because of errors arising from any tilting of the velocity measuring device or to secondary flows (Kim et al., 2000). Moreover, the measurement must be within the turbulent logarithmic layer (constant stress region), and where density stratification is not important.
Turbulent Kinetic Energy (TKE) is the absolute intensity of velocity fluctuations from the mean velocity, i.e. the variances of the flow within an XYZ co-ordinate system, and is defined as:
1 TKE u'2 v'2 w'2 (1.24) 2
Ayub Ali 20 Chapter I Introduction and literature review
Simple relationships between TKE and shear stress have been formulated in turbulence models (Galperin et al., 1988), while further studies (Soulsby and Dyer, 1981; Stapleton and Huntley, 1995) have shown the ratio of TKE to shear stress is constant, i.e.:
t C1TKE (1.25)
The proportionality constant C1 was found to be 0.20 (Soulsby and Dyer, 1981), while
C1=0.19 has been adopted by others (Soulsby, 1983; Stapleton and Huntley, 1995; Thompson et al., 2003). The main advantage of the TKE method over the LP method is that it does not require accurate knowledge of elevation above the bed, and is therefore less sensitive to conditions, where sediment erosion and deposition can alter sediment levels by several millimetres or more. Furthermore, in inter-tidal field studies some tilting of the acoustic sensor is almost inevitable, and this method is less sensitive to tilting. However, there are some potential disadvantages to the use of the TKE method. Firstly, the exact limits and dimensions of the sampling volume must be known so when measurements are made within the BBL (near the bed) the sampling volume is not mistakenly positioned partially within the bed (Finelli et al., 1999). Secondly, an inherent feature of all Doppler-based backscatter systems is Doppler noise, which is attributable to several sources, including positive and negative buoyancy of particles in the sampling volume; small-scale turbulence (at scales less than that of the sampling volume); and acoustic beam divergence, which in total may lead to high-biased estimates of turbulent energy from Acoustic Doppler devices (Nikora and Goring, 1998). Finally, accelerating and decelerating flows can cause errors in the TKE approach just as in the LP method. However, this may be corrected by detrending the velocity time-series. Similar to the second technique, the measurement must be taken within the turbulent logarithmic layer. Bed shear stress can also be estimated by using spectral analysis of turbulences and energy budgets.
For a log layer, a first-order balance between shear production P and energy dissipation ε is a fair assumption (e.g. Tennekes and Lumley, 1972; Nakagawa and Nezu, 1975)
u P uw 0 (1.26) z
Ayub Ali 21 Chapter I Introduction and literature review
u u Taking uw u 2 from the Reynolds stress method (Equation 1.11) and * from * z z the LP method (Equation 1.18), we have:
1 3 u* z (1.27)
The energy dissipation ε can be estimated from the inertial sub-range of spectral density distribution of the velocity (Grant and Madsen, 1986; Gross et al., 1994) measured at height z. Then the shear velocity can be estimated from Equation (1.27).
Most importantly, all of these methods require the measurement to be made within the constant stress turbulent logarithmic layer.
1.2.6 Technique for Estimating Roughness Height and Drag Coefficient
While fluid flows over a solid surface, it encounters friction termed as bottom friction
(or bed roughness). The roughness height z0 is most often estimated from recorded velocity profiles (Equation 1.18) while bed shear stresses can be computed using velocities at different points in the water column and the heights of those points with reference to the bed. The velocities and corresponding elevations measured from a water column are plotted onto a logarithmic graph, and roughness height z0 and shear velocity are obtained from curve fitting (Wilkinson, 1986; Bergeron and Abrahams, 1992; Ke et al., 1994; Mathisen and Madsen, 1996).
The drag coefficient is also used to represent the bed roughness in numerical models.
The drag coefficient CD (at a referenced height zr) can be calculated using roughness height z0 (Gross et al., 1999; and Bricker et al., 2005) from:
2 (1.28) C D lnz r z0 which depends upon all bed roughness components including bed sediment grain size, bed-form geometry and sediment saltation.
Ayub Ali 22 Chapter I Introduction and literature review
1.3 Thesis Objectives and Scopes
As outlined previously, the main objective of this thesis was to expand our current understanding of the physical processes of shallow estuarine systems by utilising field data and numerical models. In this process, this study first collected a set of hydraulic and water quality data from a shallow subtropical estuarine lake in Gold Coast, Queensland, Australia during November 1-10, 2005. The study also installed a temporary weather station on a houseboat within the lake during this data collection campaign. Then, various components of salt fluxes through main channel were estimated to identify the dominant transport and mixing processes for understanding of the sediment/contaminant transport processes within shallow estuarine systems. Since bed shear stress and bed roughness are two key parameters for numerically modelling any aquatic system, this study then focused on accurately measuring these hydraulic parameters. In this process, this study developed a traversing system to precisely measure the velocity and the height above the bed at various elevations. Bed shear stress and bed roughness height within the lake were estimated utilizing the newly developed traversing system. This study then developed a numerical model for the shallow estuarine system, Coombabah Lake, and simulated the hydrodynamic regime of the lake which is a prerequisite for the sediment/contaminant transport model. The hydrodynamic model was also utilized for sensitivity testing of various important parameters such as accuracy of bathymetric data, bed roughness and eddy viscosity. This study finally developed a sediment transport model for Coombabah Lake and simulated the sediment dynamics of the lake. In addition, this study analysed the field data for calibration and verification of the modelling results.
1.4 Organisation of Thesis
All chapters in this thesis are self-contained with figures, tables and references. The thesis layout is organised in the following manner, following Chapter 1.
In Chapter 2, the transport and mixing processes of a very shallow subtropical estuarine system are investigated. Currents and salinity measured at two locations within Coombabah Lake were decomposed into time-averaged and time-varying components to quantify the salt flux components attributed to various physical processes. The
Ayub Ali 23 Chapter I Introduction and literature review dominant sediment/contaminant transport and mixing processes was identified from the computed salt flux data. This chapter has been published as “Salt Fluxes within a Very Shallow Subtropical Estuary” in the Journal of Coastal Research in year 2009 and is reproduced here with the only change being figure, table and equation numbers.
Chapter 3 describes a new convenient and robust traversing system developed to measure benthic bottom boundary layer hydraulic properties. The traversing system is comprised of a flexible head Acoustic Doppler Velocimeter, a high precision altimeter and a DC underwater motor assembled on a tripod. The traverser was used to measure the velocity along with the elevation at several levels within a water column in the bottom boundary layer. The velocity and elevations were also measured from more than a metre water column moving the traverser up and down. These data were utilized to determine the bed shear stress and bed roughness height, two key parameters for modelling aquatic systems. This chapter has been published as “A traversing system to measure bottom boundary layer hydraulic properties” in Estuarine, Coastal and Shelf Science in year 2009 and is reproduced here with the only change being figure, table and equation numbers.
Chapter 4 describes a numerical study of the hydrodynamics of Coombabah Lake. This study setup a 3D hydrodynamic model utilizing the MIKE3 modelling system. Sensitivity of calibration parameters for a very shallow estuarine system was also investigated. This study found that the accurate bathymetric measurement is the most sensitive parameter followed by bed roughness and eddy viscosity. This chapter has been published as “Numerical Study of the Hydrodynamics of a Very Shallow Estuarine System – Coombabah Lake, Gold Coast, Australia” in the Journal of Coastal Research in year 2009 and is reproduced here with the only change being figure, table and equation numbers.
Chapter 5 describes sediment dynamics of Coombabah Lake. Total suspended solids (TSS) concentrations, turbidity, salinity and tide levels were measured at eight stations within Coombabah Lake and were analysed to determine the dominant transport processes. Sediment resuspension, distribution and transport processes were also simulated using a numerical model to better understand the influence of physical
Ayub Ali 24 Chapter I Introduction and literature review processes. This chapter was submitted to Estuarine, Coastal and Shelf Science for review and is currently under revision.
Finally, Chapter 6 outlines the conclusions of this study and sets recommendations for future research in line with this study.
Appendix-A of this thesis report is also a preprint. It describes intra-tidal variability of some water quality parameters within Coombabah Lake. It is relevant to this thesis and the candidate was a co-author of this article. This appendix has been published as “Short-term Variability of Physio-chemical Parameters and the Estimated Transport of Filterable Nutrients and Chlorophyll-a in the Urbanised Coombabah Lake and Coombabah Creek System, Southern Moreton Bay, Australia” in the Journal of Coastal Research in year 2007. It is reproduced here with the only change being figure, table and equation numbers.
1.5 References
Ali, A. and Lemckert, C.J., 2009. A traversing system to measure bottom boundary layer hydraulic properties. Estuarine, Coastal and Shelf Science, 83, 425-433. Ali, A., Lemckert, C.J. and Dunn, R.J.K., 2009a. Salt Fluxes in a Very Shallow Subtropical Estuary. Journal of Coastal Research. doi: 10.2112/08-1118.1. Ali, A., Zhang, H. and Lemckert, C.J., 2009b. Numerical Study of the Hydrodynamics of a Very Shallow Estuarine System - Coombabah Lake, Gold Coast, Australia. Journal of Coastal Research, SI 56, 922-926. Benfer, N.P., King, B.A. and Lemckert, C.J., 2007. Salinity observations in a subtropical estuarine system on the Gold Coast, Australia. Journal of Coastal Research, SI 50, 646-651. Bergeron, N.E., and Abrahams, A.D., 1992. Estimating Shear Velocity and Roughness Length from Velocity Profiles. Water Resources Research, 28 (8), 2155–2158. Biron, P.M., Lane, S.N., Roy, A.G., Bradbrook, K.F. and Richards, S.K., 1998. Sensitivity of bed shear stress estimated from vertical velocity profiles: the problem of sampling resolution. Earth Surface Processes and Landforms, 23 (2), 133–139. Black, K.S., 1998. Suspended Sediment Dynamics and Bed Erosion in the High Shore
Ayub Ali 25 Chapter I Introduction and literature review
Mudflat Region of the Humber Estuary, UK. Marine Pollution Bulletin, 37 (3-7), 122-133. Bricker, J.D., Inagaki, S. and Monismith, S.G., 2005. Bed drag coefficient variability under wind waves in a tidal estuary. Journal of Hydraulic Engineering, 131 (6), 497–508. Brown, J.M. and Davies, A.G., 2009. Flood/ebb tidal asymmetry in a shallow sandy estuary and the impact on net sand transport. Geomorphology. doi:10.1016/j.geomorph.2009.08.006. Bowden, K.F., 1963. The mixing processes in a tidal estuary. International Journal of Air and Water Pollution, 7, 343-356. Burton, E.D., Sullivan, L.A., Bush, R.T. and Powell, B., 2008. Iron-sulfide and trace element behaviour in sediments of Coombabah Lake, southern Moreton Bay (Australia). Marine Pollution Bulletin, 56, 1353-1376. Butt, T. and Russell, P., 2000. Hydrodynamics and cross-shore sediment transport in the swash-zone of natural beaches: a review. Journal of Coastal Research, 16 (2), 255–268. Cheng, R.T., Ling, C.H., Gartner, J.W. and Wang, P.F., 1999. Estimates of bottom roughness length and bottom shear stress in South San Francisco Bay, California. Journal of Geophysical Research, 104 (C4), 7715-7728. Conley, D.C. and Griffin, J.G., 2004. Direct measurements of bed stress under swash in the field. Journal of Geophysical Research-Oceans, 109 (C3) (art. no.-C03050). Cox, M. and Moss, A., 1999. Nerang River, Tallebudgera, Currumbin and Coombabah Creeks: Water Quality Report 1999. Brisbane, Queensland Environmental Protection Agency, 26p. Davies, A.G., 1985. Field observations of the threshold of sediment motion by wave action. Sedimentology (Oxford), 32 (5), 685–704. DHI (Danish Hydraulic Institute), 2008. Scientific Documentation, MIKE 21 & MIKE 3 Flow Model FM, Hydrodynamic and Transport Module. DHI Water and Environment, Horsholm, Denmark, 52p. Dunn, R.J.K., Ali, A., Lemckert, C.J., Teasdale, P.R., and Welsh, D.T., 2007a. Short- term variability of physio-chemical parameters and the estimated transport of filterable nutrients and chlorophyll-a in the urbanised Coombabah Lake and
Ayub Ali 26 Chapter I Introduction and literature review
Coombabah Creek system, southern Moreton Bay, Australia. Journal of Coastal Research, SI 50, 1062-1068. Dunn, R.J.K., Lemckert, C.J., Teasdale, P.R., and Welsh, D.T., 2007b. Distribution of nutrients in surface and sub-surface sediments of Coombabah Lake, southern Moreton Bay (Australia). Marine Pollution Bulletin, 54, 602-625. Dunn, R.J.K., Welsh, D.T., Lee, S.Y., Lemckert, C.J., Teasdale, P.R. and Meziane, T., 2008. Investigating the distribution and sources of organic matter in surface sediment of Coombabah Lake (Australia) using elemental, isotopic and fatty acid biomarkers. Continental Shelf Research, 28, 2535–2549. Dyer, K.R., 1986. Coastal and Estuarine Sediment Dynamics, vol. xv, Wiley, Chichester, 342p. Dyer, K.R., 1997. Estuaries-A Physical Introduction. John Wiley & Sons, 195p. Finelli, C.M., Hart, D.D. and Fonseca, D.M., 1999. Evaluating the spatial resolution of an Acoustic Doppler Velocimeter and the consequences for measuring near-bed flows. Limnology and Oceanography 44 (7), 1793–1801. Fischer, H.B., 1972. Mass transport mechanisms in partially stratified estuaries. Journal of fluid mechanics, 53 (4), 671-687. Frank, T.D. and Fielding, C.R., 2004. Sedimentology and geochemistry of an urban coastal lake system: Coombabah Lake Nature Reserve, Gold Coast, Queensland. Australian Journal of Earth Sciences, 51, 261-271. Galperin, B., Kantha, L.H., Hassid, S. and Rosati, A., 1988. A quasi-equilibrium turbulent energy—model for geophysical flows. Journal of the Atmospheric Sciences, 45 (1), 55–62. GHD (Gutteridge, Haskins and Davey Pty. Ltd.), 2003. Coombabah Creek Environmental Inventory. Gutteridge, Haskins and Davey Pty Ltd, Brisbane, Australia, 439p. Grant, W.D. and Madsen, O.S., 1982. Movable bed roughness in unsteady oscillatory flow. Journal of Geophysical Research, 87 (C1), 469–481. Grant, W.D. and Madsen, O.S., 1986. The continental-shelf bottom boundary layer. Annual Review of Fluid Mechanics, 18, 265-305. Grant, W.D., Williams, A.J. and Glenn, S.M., 1984. Bottom stress estimates and their prediction on the northern California Continental Shelf during CODE-1: the
Ayub Ali 27 Chapter I Introduction and literature review
importance of wave current interaction. Journal of Physical Oceanography, 14, 506–527. Granger, R.A., 1985. Fluid Mechanics. Holt, Rinehart and Winston, Tokyo, Japan. 884 pp. Gross, E.S., Koseff, J.R. and Manismith, S.G., 1999. Three-dimensional salinity simulations of South San Fransisco Bay. Journal of Hydraulic Engineering, ASCE, 125 (11), 1199-1209. Gross, T.F., Williams, A.J. and Terray, E.A., 1994. Bottom boundary layer spectral dissipation estimates in the presence of wave motions. Continental Shelf Research, 14 (10-11), 1239-1256. Hollingsworth, A. and Connolly, R.M., 2006. Feeding by fish visiting inundated subtropical saltmarsh. Journal of Experimental Marine Biology and Ecology, 336, 88-98. Jing, L. and Ridd, P.V., 1996. Wave-current bottom shear stresses and sediment resuspension in Cleveland Bay, Australia. Coastal Engineering, 29, 169-186. Kabir, M.R. and Torfs, H., 1992. Comparison of different methods to calculate bed shear-stress. Water Science and Technology, 25 (8), 131–140. Ke, X.K., Collin, M.B. and Poulos, S.E., 1994. Velocity structure and sea bed roughness associated with intertidal (sand and mud) flats and saltmarshes of the Wash, UK. Journal of Coastal Research, 10, 702–715. Kim, S.C., Friedrichs, C.T., Maa, J.P.-Y. and Wright, L.D., 2000. Estimating bottom stress in tidal boundary layer from Acoustic Doppler Velocimeter data. Journal of Hydraulic Engineering, ASCE, 126 (6), 399–406. Kjerfve, B., 1986. Circulation and salt-flux in a well-mixed estuary. In: J. van de Kreeke, (Ed), Physics of Shallow Estuaries and Bays, Springer Verlag, Berlin, 22- 29. Klen, T., 2006. Estuaries, An Introduction to Marine Biology and Oceanography,
Ayub Ali 28 Chapter I Introduction and literature review
Thermal Analyses of Electronic Systems. IEEE Transactions on Components and Packaging Technologies, 25 (3), 366.382. Lee, S.Y., Dunn, R.J.K., Young, R.A., Connolly, R.M., Dale, P.E.R., Dehayr, R., Lemckert, C.J., McKinnon, S., Powell, B., Teasdale, P.R. and Welsh, D.T., 2006. Impact of urbanisation on coastal wetland structure and function. Austral Ecology, 31, 149-163. Lewis, R.E. and Lewis, J.O., 1983. The Principal Factors Contributing to the Flux of Salt in a Narrow, partially Stratified Estuary. Estuarine, Coastal and Marine Science, 16, 599-626. Liu, Z., 2001. Sediment transport. Laboratoriet for Hydraulik og Havnebygning, Instituttet for Vand, Jord og Miljøteknik, Aalborg Universitet
Ayub Ali 29 Chapter I Introduction and literature review
Geophysical Research, 86, 6467–6472. Nikora, V.I. and Goring, D.G., 1998. ADV measurements of turbulence: can we improve their interpretation? Journal of Hydraulic Engineering, ASCE, 124 (6), 630–634. Nikora, V., Goring, D. and Ross, A., 2002. The structure and dynamics of the thin near- bed layer in a complex marine environment: a case study in Beatrix Bay, New Zealand. Estuarine, Coastal and Shelf Science, 54 (5), 915–926. Nikuradse, J., 1933. Strömungsgesetze in Rauchen Röohren. Forschungsheft No. 361. Berlin: Verein Deutscher Ingenieure. Osalusi, E., Side, J. and Harris, R., 2009. Reynolds stress and turbulence estimates in bottom boundary layer of Fall of Warness. International Communications in Heat and Mass Transfer, 36, 412–421. Osborne, P.D. and Boak, E.H., 1999. Sediment suspension and morphological response under vessel generated wave groups: Torpedo Bay, Auckland, New Zealand. Journal of Coastal Research, 15 (2), 388-398. Owais, S., Atal, S. and Sreedevi, P.D., 2008. Governing Equations of Groundwater Flow and Aquifer Modelling Using Finite Difference Method, In: S. Ahmed, R. Jayakumar and A. Salih (Ed). Groundwater Dynamics in Hard Rock Aquifers, Springer, Netherlands, pp. 201-218. Perlin, A. and Kit, E., 2002. Apparent roughness in wave-current flow: Implication for coastal studies. Journal of Hydraulic Engineering, ASCE, 128 (8), 729–741. Pope, N.D., Widdows, J. and Brinsley, M.D., 2006. Estimation of bed shear stress using the turbulent kinetic energy approach—A comparison of annular flume and field data. Continental Shelf Research, 26, 959–970. Prandtl, L., 1926. Uber die Ausgebildete Turbulenz. Proceedings of 2nd International Conference on Applied Mechanics, Zurich, 62-75. Pritchard, D.W., 1954. A study of salt balance in a coastal plain estuary. Journal of Marine Research, 13, 133-144. Raudkivi, A.J., 1989. The roughness height under waves. Journal of Hydraulic Research, 26, 569–584. Raudkivi, A.J., 1998. Loose Boundary Hydraulics. Balkema, The Netherlands. Restrepo, J.D. and Kjerfve, B., 2002. The San Juan Delta, Colombia: tides, circulations,
Ayub Ali 30 Chapter I Introduction and literature review
and dispersion. Continental Shelf Research, 22, 1249-1267. Simpson, J. H., Vennel, R. and Souza, A. J, 2001. The salt fluxes in a tidally-energetic estuary. Estuarine, Coastal and Shelf Science, 52, 131-142. Skinner, J.L., Gillam E. and Rohlin, C.J., 1998. The demographic future of the Moreton Region. In: Tibbets, I.R., Hall, N.J. and Dennison, W.C. (Ed.), Moreton Bay and Catchment, School of Marine Science, University of Queensland, Brisbane, pp. 67–78. Soulsby, R.L., 1983. The bottom boundary layer of shelf seas. In: Johns, B. (Ed.), Physical Oceanography of Coastal and Shelf Seas. Elsevier, Amsterdam, pp. 189- 266. Soulsby, R.L., Atkins, R. and Salkield, A.P., 1994. Observations of the turbulent structure of a suspension of sand in a tidal current. Continental Shelf Research, 14 (4), 429–435. Soulsby, R.L. and Dyer, K.R., 1981. The form of the near-bed velocity profile in a tidally accelerating flow. Journal of Geophysical Research—Oceans and Atmospheres, 86 (NC9), 8067–8074. Soulsby, R.L., Salkield, A.P. and Le Good, G.P., 1984. Measurements of the turbulence characteristics of sand suspended by a tidal current. Continental Shelf Research, 3 (4), 439–454. Stapleton, K.R. and Huntley, D.A., 1995. Seabed stress determinations using the inertial dissipation method and the turbulent kinetic energy method. Earth Surface Processes and Landforms, 20 (9), 807–815. Stips, A., Prandke, H. and Neumann, T., 1998. The structure and dynamics of the Bottom Boundary Layer in shallow sea areas without tidal influence: an experimental approach. Progress in Oceanography, 41, 383-453. Sylaios, G.K., Tsihrintzis, V.A., Akratos, C. and Haralambidou, K., 2006. Quantification of water, salt and nutrient exchange processes at the mouth of a Mediterranean coastal lagoon. Environmental Monitoring and Assessment, 119, 275-301. Tennekes, H., and Lumley, J.L., 1972. A first course in turbulence. MIT Press, Cambridge, Massachusetts, USA. Thomsen, L., 1999. Processes in the benthic boundary layer at continental margins and
Ayub Ali 31 Chapter I Introduction and literature review
their implication for the benthic carbon cycle. Journal of Sea Research, 41, 73-86. Thompson, C.E.L., Amos, C.L., Jones, T.E.R. and Chaplin, J., 2003. The manifestation of fluid-transmitted bed shear stress in a smooth annular flume—a comparison of methods. Journal of Coastal Research, 19 (4), 1094–1103. Uncles, R.J., Elliott, R.C.A., and Weston, S.A., 1985. Observed fluxes of water, salt and suspended sediment in a partially mixed estuary. Estuarine, Coastal and Shelf Science, 20, 147-167. Uncles, R.J. and Jordan, M.B., 1979. Residual fluxes of water and salt at two stations in the Severn Estuary. Estuarine, Coastal and Marine Science, 9, 287-302. Uncles, R.J. and Stephens, J.A., 2009. Turbidity and sediment transport in a muddy sub- estuary. Estuarine, Coastal and Shelf Science. doi:10.1016/j.ecss.2009.03.041. You, Z.J., 2005. Estimation of bed roughness from mean velocities measured at two levels near the seabed. Continental Shelf Research, 25, 1043-1051. You, Z.J., 2006. Estimation of mean seabed roughness in a tidal channel with an extended log-fit method. Continental Shelf Research, 26, 283-294. You, Z.J. and Nielsen, P., 1996. Movable bed roughness in the flow of irregular waves and currents over movable beds. In: Proceedings of the 25th International Conference on Coastal Engineering, Orlando, pp. 3495–3506. Yuan, Y., Wei, H., Zhao, L. and Jiang, W., 2008. Observations of sediment resuspension and settling off the mouth of Jiaozhou Bay, Yellow Sea. Continental Shelf Research, 28, 2630–2643. van Rijn, L.C., 1984. Sediment transport, Part III: bed forms and alluvial roughness. Journal of Hydraulic Engineering, ASCE, 110, 1733–1754. Verney, R., Deloffre, J., Brun-Cottan, J.-C. and Lafite, R., 2007. The effect of wave- induced turbulence on intertidal mudflats: Impact of boat traffic and wind. Continental Shelf Research, 27, 594–612. Vreugdenhil, C.B., 2002. Accuracy and Reliability of Numerical River Models. Journal of the American Water Resources Association, 38 (4), 1083-1095. Wheatcroft, R.A., 1994. Temporal variation in bed configuration and one-dimensional bottom roughness at the mid-shelf STRESS site. Continental Shelf Research, 14, 1167–1190. Wilcock, P.R., 1996. Estimating local bed shear stress from velocity observations.
Ayub Ali 32 Chapter I Introduction and literature review
Water Resources Research, 32 (11), 3361–3366. Wilkinson, R.H., 1986. Variation of Roughness Length of a Mobile Sand Bed in a Tidal Flow. Geo-marine Letters, 5, 231-239. Williams, J.J., Rose, C.P., Thorne P.D., O'Connor B.A., Humphery, J.D., Hardcastle, P.J., Moores, S.P., Cooke, J.A. and Wilson, D.J., 1999. Field observations and predictions of bed shear stresses and vertical suspended sediment concentration profiles in wave–current conditions. Continental Shelf Research, 19 (4), 507–536.
Ayub Ali 33 Chapter I Introduction and literature review
Ayub Ali 34 Chapter II Salt fluxes within a shallow estuarine system
CHAPTER II
Salt Fluxes within a Very Shallow Subtropical Estuary*
Abstract
This paper describes the transport processes and net salt flux within a shallow estuarine system, with particular reference to Coombabah Lake-Creek system in Queensland, Australia. Observations of currents and salinity at two locations within Coombabah Lake provided a basis for assessing the relative importance of various transport processes within a very shallow (water depth < 1 m) subtropical estuary. The instantaneous velocity and salinity data were decomposed into time-averaged means and time-varying components, and were used to quantify the salt flux components attributed to various physical processes. In this study, advection by residual flow, which contributed 65% of total salt flux, was identified as the dominant process in transporting salt. The advective flux also determined the direction of the net salt flux within this shallow estuarine system. This study concludes that the net salt flux varies spatially and temporarily with hydromorphological and meteorological conditions.
2.1 Introduction
Coastal wetlands and estuaries are important environments, providing significant habitats for flora and fauna species, often supporting commercial and recreational fisheries (Stumpf and Haines, 1998). Furthermore, these systems act as filters against contaminants and sediments (Faulkner, 2004), absorb wave energy, and also provide cultural and recreational benefits (Lee et al., 2006b).
* Ali, A., Lemckert, C.J. and Dunn, R.J.K., 2009. Salt fluxes in a very shallow sub- tropical estuary. Journal of Coastal Research. doi:10.2112/08-1118.1
Ayub Ali 35 Chapter II Salt fluxes within a shallow estuarine system
Table 2.1: Previously reported salt fluxes per unit width (positive seaward and negative landward) divided by mean water depth. Study site Spring / Freshwater Mean h u s hu s u h s s h u hus F T T T T T T Reference Neap flow depth Flux 1 Flux 2 Flux 3 Flux 4 Flux 5 Net flux m3/s m ppt cm/s ppt cm/s ppt cm/s ppt cm/s ppt cm/s ppt cm/s Mersey estuary – Neap 41.05 20.4 6.34 -6.95 – – -2.36 -2.97 Bowden (1963) Narrows at St. A Mersey estuary – Spring 18.02 20.4 2.63 -0.44 – – -3.04 -0.85 Bowden (1963) Narrows at St. A Mersey estuary – Neap 66.70 20.4 10.00 -12.27 – – -3.22 -5.49 Bowden (1963) Narrows at St. C Mersey estuary – Spring 42.90 20.4 6.00 -14.40 – – -5.15 -13.65 Bowden (1963) Narrows at St. D Severn Estuary – Pre-neap 154.00 19.4 100.00 3.20 – – -0.24 102.96 Uncles and Jordan (1979) Station A Severn Estuary – Pre- 105.00 14.7 -190.00 -7.40 – – -0.43 -197.83 Uncles and Jordan (1979) Station B spring Hudson Estuary – Calculated from Hunkins 238.00 ~ 24.042.65 8.22 – -19.12 -11.02 20.73 The Narrows I (1981) Hudson Estuary – Calculated from Hunkins – 845.00 ~ 24.078.12 9.25 – -7.47 -57.03 22.87 The Narrows 2 (1981) Hudson Estuary – Calculated from Hunkins – 182.00 ~ 10.036.22 8.78 – -20.18 -8.68 16.14 Yokkers (1981) Tees Estuary – Neap 3.20 ~ 18.0 -79.10 -0.90 0.00 0.00 -0.90 -80.90 Lewis and Lewis (1983) Station 1 Tees Estuary – Spring 4.20 ~ 18.0 -268.40 -0.90 -0.10 0.00 -0.30 -269.40 Lewis and Lewis (1983) Station 1 Tees Estuary – Neap 3.20 ~ 5.0 93.70 -43.90 5.30 2.00 -6.60 50.50 Lewis and Lewis (1983) Station 8
Ayub Ali 36 Chapter II Salt fluxes within a shallow estuarine system
Tees Estuary – Spring 4.20 ~ 5.0 246.20 -98.60 80.10 0.70 -0.80 227.60 Lewis and Lewis (1983) Station 8 North Inlet – – 1.00-5.00 ~ 3.0 -19.20 -315.00 0.20 6.40 -0.10 -327.70 Kjerfve (1986) Flood channel North Inlet – – 1.00-5.00 ~ 3.0 68.40 -234.70 0.30 -24.30 -0.10 -190.70 Kjerfve (1986) Mid channel North Inlet – – 1.00-5.00 ~ 3.0 288.30 -127.20 0.00 -24.20 -0.70 136.20 Kjerfve (1986) Ebb channel Tyne Estuary – – 73.10 4.7 23.70 0.60 -5.20 2.40 0.20 21.70 Park and James (1990) Station 1 Tyne Estuary – – 2.30 11.4 -0.30 -5.10 0.00 2.80 0.40 -2.20 Park and James (1990) Station 4 Laguna de Terminos – – Calculated from David and – 6.4 367.19 26.56 0.00 3.13 0.00 396.88 Carmen Inlet Kjerfve (1998) Laguna de Terminos – – Calculated from David and – 5.0 328.00 -10.00 0.00 156.00 4.00 474.00 Puerto Real Inlet Kjerfve (1998) Bertioga Channel Calculated from Miranda Neap ~10.00 5.5 108.53 -9.64 0.55 -2.91-32.18 64.35 and Kjerfve (1998) Bertioga Channel Calculated from Miranda Spring ~10.00 5.7 66.18 -7.27 0.37 -4.73 -7.27 47.28 and Kjerfve (1998) San Juan River delta – – Restrepo and Kjerfve 2450.00 7.0 135.50 -3.20 3.80 -1.10 -5.30 129.90 San juan (2002) San Juan River delta – – Restrepo and Kjerfve 2450.00 6.0 119.10 -138.20 10.50 -17.50 -2.60 -28.70 Chavica (2002) San Juan River delta – – Restrepo and Kjerfve 2450.00 9.0 635.20 22.90 4.50 18.30 -4.30 676.80 Charambira (2002) San Juan River delta – – Restrepo and Kjerfve 3750.00 7.0 4.50 -111.10 0.00 -10.90 -16.40 -133.80 San juan (2002) San Juan River delta – – Restrepo and Kjerfve 3750.00 6.0 139.70 -250.30 16.50 -10.80 -0.80 -105.80 Chavica (2002)
Ayub Ali 37 Chapter II Salt fluxes within a shallow estuarine system
San Juan River delta – – Restrepo and Kjerfve 3750.00 9.0 111.20 -17.10 0.20 -1.70 -32.60 50.90 Charambira (2002) Vassova lagoon – Spring 0.94 0.75 -0.03 -0.03 – – – -0.06 Sylaios et al. (2006) Entrance canal Vassova lagoon – Neap 0.94 0.72 -0.29 -0.02 – – – -0.31 Sylaios et al. (2006) Entrance canal Coombabah Lake – Spring – 1.12 166.58 -27.86 4.04 -46.09 – 96.67 Present study Station A Coombabah Lake – Neap – 1.30 116.29 -9.38 1.52 -30.81 – 77.62 Present study Station A Coombabah Lake – Spring – 0.62 48.67 -10.64 3.09 22.88 – 64.01 Present study Station B Coombabah Lake – Neap – 0.74 76.75 -9.08 3.60 -5.47 – 65.80 Present study Station B where Flux 1 represents the advective salt flux due to water discharge and change in storage volume during the tidal cycle; Flux 2 represents the sloshing effect – the tidal dispersion via triple correlation between tidal depth change, tidal current, and tidal salinity, usually directed upstream; Flux 3 represents the cross correlation between tide and salinity; maximum (and positive) when tide and salinity are in phase, and minimum (and negative) when they are out of phase; Flux 4 represents the Stokes’ drift dispersion; and Flux 5 represents the salt dispersion due to mean shear produced by gravitational circulation. ppt ≈ psu. – Data not available.
Ayub Ali 38 Chapter II Salt fluxes within a shallow estuarine system
To quantify the water circulation and mixing processes within estuarine systems, it is often convenient to study salinity distribution patterns because salinity is typically considered as a conservative tracer (Dyer, 1997). The behaviour of salts within the water column provides a basis for predicting transport of other soluble conservative substances. Salt is transported downstream by freshwater flow (which enters the estuary through rivers, creeks and groundwater systems) and is mixed back upstream against this seaward flow. The overall landward mixing (termed dispersion rather than diffusion) is produced primarily by the effects of tides, winds, and gravitational circulation (Hunkins, 1981).
To date, a significant number of studies have been undertaken to determine salt fluxes within estuarine systems characterised by very low to high rates of freshwater discharge (2.3 to 3750 m3/s) and varying water depths ranging from 3 to 24 m (e.g., Bowden, 1963; David and Kjerfve, 1998; Hunkins, 1981; Kjerfve, 1986; Lewis and Lewis, 1983; Miranda and Kjerfve, 1998; Restrepo and Kjerfve, 2002). However, the quantification of salt fluxes within very shallow estuaries is limited (e.g., Sylaios et al., 2006).
Previous studies have shown a high degree of variability in salt fluxes within estuarine environments as summarised in Table 2.1. Bowden (1963) calculated the fluxes for four periods (twice at a single station and once at an additional two stations) along and across the Mersey Estuary, with the results indicating a net salt flux consistently directed upstream. Bowden (1963) claimed the discrepancies between salt fluxes were due to variations of velocity across the estuary. Uncles and Jordan (1979) described the residual transport of water as the dominant mechanism for residual transport of salt, observed opposite fluxes seaward during the neap tide phase at one station, and reported fluxes in a landward direction twice the magnitude at another station during spring tides within the deep Severn Estuary (England). However, the latter station was located in a complex topographic region 17.2 km upstream from the previous downstream station. Uncles and Jordan (1979) termed the spring-neap variations in salinity as significant compared with that resulting from time-varying freshwater inputs. Hunkins (1981), Lewis and Lewis (1983), David and Kjerfve (1998), and Miranda and Kjerfve (1998) also observed the residual transport of water as the principal factor in determining the net salt flux. More interestingly, Lewis and Lewis (1983) identified a net salt flux
Ayub Ali 39 Chapter II Salt fluxes within a shallow estuarine system landward at downstream stations and seaward at upstream stations within the Tees Estuary (England). Kjerfve (1986) found seaward transport through ebb channels and landward transport through mid and flood channels within the North Sea Inlet, South Carolina (United States). Park and James (1990) also observed a net seaward transport near the head and landward transport near the mouth of the Tyne Estuary (England). Restrepo and Kjerfve (2002) investigated salt fluxes through three distributaries within the San Juan River Delta (Columbia) and found advection and tidal pumping were the dominant processes in salt transport. Moreover, they observed net seaward flux in one distributary, whereas net landward flux was evident in another distributary in the same estuary. Sylaios et al. (2006) also identified advection as the main mechanism for salt transport, which was consistently landward in direction. Additionally, Sylaios et al. (2006) observed that tidal pumping was an order of magnitude smaller than advection. Although the net flux was landward most of the time, forcing salt into the Vassova Lagoon (Greece), magnitudes were highly variable and changed frequently with meteorological and tidal conditions. Overall, none of the estuaries investigated was observed to exist under a steady condition (i.e., negligible net salt flux). Instead, net salt fluxes varied both spatially and temporarily with changing hydromorphological and meteorological conditions (see Table 2.1), thus warranting further studies in identifying the controlling mechanisms within particular systems.
This study is the first to quantify salt fluxes within the shallow Coombabah Lake–Creek system, providing an initial understanding of the system’s physical processes, assisting future management decisions in this ecologically and economically significant region. Furthermore, this study contributes to a greater understanding of the physical processes occurring within shallow estuarine systems.
2.2 Methods
2.2.1 Study site
Coombabah Creek is a subtropical estuarine system situated in southern Moreton Bay, south-east Queensland (Australia) (see Figure 2.1), one of the fastest growing regions in the developed world (Skinner et al., 1998). The creek is a 17 km long, moderately impacted (Cox and Moss, 1999; Lee et al., 2006a) tidal creek that flows through
Ayub Ali 40 Chapter II Salt fluxes within a shallow estuarine system
Coombabah Lake. The creek enters the lake at the south-west side (hereby referred to as the creek mouth) and leaves the same from the north-east side (hereby referred to as the lake entrance). Ultimately, Coombabah Creek discharges into the Gold Coast Broadwater, within southern Moreton Bay. Thus, the creek mouth has a greater influence on the freshwater flow entering the lake system than the lake entrance, which is dominated by the inflow of seawater.
The lake covers ~2 km2 (GHD, 2003), with an urbanised catchment area of 44 km2, characterised by residential, commercial, and light industrial developments. The lake is a shallow body of water characterised by fine sediments (Dunn et al., 2007b) and is located in the midtidal region of Coombabah Creek, with urban development positioned to the east and along the southern and western shorelines.
With the exception of shallow channels, Coombabah Lake is characterised by a relatively flat bathymetry (Lee et al., 2006a), with depths from 0 to ~1 m. During periods of low water, large portions of the benthic lake sediments become exposed. Episodically large inputs of freshwater occur during periods of heavy rainfall, predominantly during summer periods. Despite its modest dimensions, Coombabah Lake is ecologically significant, being a valuable fish and migratory bird habitat. Additionally, the lake system is unique within southern Moreton Bay because it behaves as an inverse estuary during summer periods, when evaporation exceeds typical freshwater inflows (Benfer et al., 2007).
As a consequence of the ecological significance and the potential for anthropogenic disturbances and inputs into Coombabah Lake, the lake and surrounding wetlands have been the focus of recent scientific efforts (e.g., Burton et al., 2008; Dunn et al., 2007a, 2007b, 2008; Frank and Fielding, 2004; Hollingsworth and Connolly, 2006; Knight et al., 2008).
Ayub Ali 41 Chapter II Salt fluxes within a shallow estuarine system
Gold Coast Broadwater
Lake entrance
Left channel
A Main channel
Bed level (meter) AboveAbove 0.250.25 B 0.000 – - 0.25 0.25 -0.25-0.25 – - 0.00 0 -0.50-0.5 – - -0.25-0.25 -0.75-0.75 – - -0.50 -0.5 Creek mouth -1.00-1 – - -0.75-0.75 -1.25-1.25 – - -1.00 -1 BelowBelow -1.25-1.25
Figure 2.1: Location map of the study site, Coombabah Lake (southern Moreton Bay, Australia).
2.2.2 Field measurement and data processing
The data collection program was conducted from November 1, 2005, to November 10, 2005. A submersible sensor base was positioned at two individual sample locations characterised by differing salinity concentrations within Coombabah Lake, referred to as Station A (near the lake entrance) and Station B (near the creek mouth) (see Figure 2.1). The stationed instrument packages consisted of a conductivity, temperature, and depth probe (NXIC-CTD; Falmouth Scientific, Inc., Cataumet, Massachusetts) and a high resolution three-dimensional acoustic Doppler velocimeter (Vector velocimeter; Nortek AS, Rud, Norway). High-frequency conductivity, temperature, and depth data were collected from 15 cm above the bed using time-averaged data (3.5 min bursts at frequency of 10 Hz) obtained every 15 minutes. Acoustic Doppler velocimeter data were obtained from 30 cm above the bed at burst intervals of 30 minutes, at a frequency of 32 Hz and 4096 samples per burst.
Ayub Ali 42 Chapter II Salt fluxes within a shallow estuarine system
Meteorological conditions, namely air pressure, solar radiation, humidity, rainfall, and wind speed and direction within the lake environment, were recorded every 15 minutes during the study period using a data logging weather station (WeatherMaster 2000; Environdata, Warwick, Australia). Daily evaporation was estimated from the meteorological data collected using a modified Penman equation (EasiAccess software; Environdata).
Additionally, before the deployment of Stations A and B, water level data from Coombabah Lake were obtained during the period between September 13, 2004, to December 20, 2004, by using a submersible tide gauge (XR-420-TG; Richard Brancker Research Ltd., Ottawa, Canada). The dominant tidal constituents for the area were obtained using this collected data.
2.2.3 Quantification of salt fluxes
Total salt flux within an estuarine system is typically determined using velocity and salinity concentrations at several water depths along a vertical section measured at different locations across the section (e.g., Bowden, 1963; Fischer, 1972; Sylaios et al., 2006). However, observations of currents and salinity collected at one location along the cross-section of a main channel within an estuarine system may also provide a valid experimental approach when assessing the relative importance of different transport processes (e.g., Lewis and Lewis, 1983; Pritchard, 1954; Restrepo and Kjerfve, 2002; Simpson et al., 2001; Uncles and Jordan, 1979; Uncles et al., 1985). The longitudinal salt flux within estuarine systems can be decomposed into components attributed to different physical processes (Table 2.1) (Bowden, 1967; Dyer, 1997; Hunkins, 1981; Kjerfve, 1986; Pritchard, 1954; Restrepo and Kjerfve, 2002). In general, the net salt flux per unit width perpendicular to the main flow (F) may be calculated as the following (Bowden, 1963; Dyer, 1997; Kjerfve, 1986; Restrepo and Kjerfve, 2002):
h F us dz (2.1) 0 where F represents the net flux (ppt m2/s); u(z) represents the observed velocity (m/s); s(z) represents the observed salinity (ppt); z represents the vertical coordinate; and h
Ayub Ali 43 Chapter II Salt fluxes within a shallow estuarine system represents water depth (m). The instantaneous velocity and salinity can be decomposed as u u uand s s s , where the primed quantities represent deviations from the depth-averaged means, u and s . The depth-averaged means may be decomposed into time-averaged means and time varying components: u u uT and s s sT such that
u u uT u (2.2)
s s sT s (2.3) where angle brackets are net, or time-averaged, over at least one complete tidal cycle. By substituting Equations (2.2) and (2.3) into Equation (2.1), the net flux can be decomposed into the following terms (Restrepo and Kjerfve, 2002):
F h u s huT sT u hT sT s hT uT hus (2.4) which, for simplicity, can be written as:
Net flux = Flux 1 + Flux 2 + Flux 3 + Flux 4 + Flux 5 (2.5)
An estuary is considered to be at steady state when the net salt flux over a complete tidal cycle is zero, with the advective and dispersive modes of transport balancing the exchange. Salt fluxes within the lake were calculated for a full tidal cycle (24.5 h) using Equation (2.4). Because of the very shallow and vertically well-mixed nature of the lake (Benfer et al., 2007; Dunn et al., 2007b), gravitational circulation within the lake was considered negligible and therefore flux 5, hus , in Equation (2.4) was considered insignificant.
2.3 Results and discussion
2.3.1 Hydrometeorological properties
Astronomical tidal constituents were obtained from water level data collected in 2004.
The semidiurnal constituents M2, S2, K2, L2 and N2 of Coombabah Lake were 0.37 m,
Ayub Ali 44 Chapter II Salt fluxes within a shallow estuarine system
0.02 m, 0.03 m, 0.09 m and 0.06 m in amplitude with phase lags of 243°, 226°, 121°,
322° and 161°, respectively. The diurnal constituents K1, O1, Q1 and S1 were 0.19 m, 0.11 m, 0.03 m and 0.10 m in amplitude with phase lags of 195°, 99°, 356° and 131°, respectively. Astronomical tidal constituents were assumed to be the same after one year (during this study), because no major changes in geometry and bathymetry of the lake channel system occurred during this period. The lake experiences a mixed tidal regime, mainly of a semidiurnal nature, with a form number of 0.66.
Wind conditions during the 2005 study period were generally moderate but highly variable, ranging from 2 to 20 km/h (see Figure 2.2a), with the average wind speed at 7 km/h and directed from south or south-east (blowing from the creek mouth to the lake entrance, see Figure 2.2b). Light rainfall was recorded every day except on November 7, 2005 (see Figure 2.2c). Rainfall increased gradually from 0.5 mm/d on November 1, 2005 to 18 mm/d (the maximum) on November 6, 2005. Evaporation varied between 5 and 10 mm/d (see Figure 2.2d), which is greater than the annual regional average (3.5 mm/d) and even greater than the monthly average for the same periods in other years (5 mm/d) recorded by the Australian Bureau of Meteorology (2007). Air pressure also dropped by approximately 5 millibars during the study period (Figure 2.2e).
Water depth, velocity, salinity, and the computed salt flux components at Station A are presented in Figures 2.3a–d. A mixed tidal pattern was observed in water depth, velocity, and salinity data, with higher velocity and salinity ranges occurring during spring tides. The opposite trend was observed in the case of the tidal range, where lower ranges occurred during spring tide conditions, compared with the greater ranges observed during neap tide conditions. The spring semidiurnal tide also changed to a mixed type during neap tide conditions. Mean water depth and salinity were greater during neap tide conditions in comparison to the spring tide conditions. Potential explanations for the observed greater water depth during the sampled neap tidal condition were twofold. First, the presence of low air pressure that developed in the region during the neap tidal period resulted in an increase in water level by ~5 cm, and second, a recorded rainfall event within the lake catchment contributed increased freshwater flow into the shallow lake. However, the actual rate of freshwater flow is
Ayub Ali 45 Chapter II Salt fluxes within a shallow estuarine system unknown. The increasing salinity trend was due to the added storage from the seawater, indicating the quantity of freshwater input was relatively small.
30 30 (a) 20 20
10 10
Wind speedWind (km/h) 0 0
360 360 (b)
180 180 Wind direction (°) direction Wind 0 0
20 20 (c)
10 10 Rainfall (mm/day)Rainfall 0 0
) 10 10 (d)
5 5
Evaporation (mm/day 0 0
1025 1025 (e) 1020 1020
1015 1015
1010 1010
Air pressureAir (mbar) 1005 1005 1/11/05 2/11/05 3/11/05 4/11/05 5/11/05 6/11/05 7/11/05 8/11/05 9/11/05 10/11/05 11/11/05 Spring tideDate Neap tide
Figure 2.2: Meteorological data collected on site: (a) wind speed; (b) wind direction; (c) rainfall; (d) evaporation; and (e) air pressure.
2.3.2 Salt fluxes
At Station A, Flux 1 (advective flux) was the dominant flux, which was directed down the estuary (seaward) at a rate of 155 psu cm/s on November 1, 2005, and reached its maximum value of 165 psu cm/s on November 2, 2005. Following this period, the advective flux began to decline gradually until November 5, 2005 (with an abrupt
Ayub Ali 46 Chapter II Salt fluxes within a shallow estuarine system reduction to 30 psu cm/s from 130 psu cm/s), then remained low for 3 days with only minor fluctuations. It then increased to 115 psu cm/s, followed by a declining trend. Flux 2 (tidal dispersion) followed the same trend as Flux 1 but was directed up estuary (landward) (Figure 2.3d). The tidal dispersion value during November 1, 2005, was 40 psu cm/s and gradually reduced to 3 psu cm/s during November 6, 2005. This value continued for 3 days with little variation, followed by an increasing trend. Flux 3 (the cross-correlation between tide and salinity) was the smallest contributor to the net salt flux among the four salt flux components and was always directed seaward following the same trend of Flux 1 (Figure 2.3d). Flux 4 (the Stokes’ drift dispersion) was directed upstream and was approximately 100% greater than Flux 2. Overall, Fluxes 1 and 3 contributed to the seaward (downstream) salt flux, whereas Fluxes 2 and 4 contributed to the landward (upstream) salt flux. Therefore, Stoke’s drift was the main operator in dispersing salt within the water column of Coombabah Lake. However, Flux 1 was the principal contributor, determining the direction of the net salt flux. Overall, Fluxes 1 and 3 contributed 98% and 2%, respectively, to the seaward salt flux, whereas Fluxes 2 and 4 contributed 30% and 70%, respectively, to the landward salt flux.
Water depth, velocity, and salinity data at Station B showed similar characteristics to that of Station A (Figures 2.4a–c). Fluxes 1 and 3 at this station (Figure 2.4d) are also directed seaward (similar to Station A). However, Flux 3 contributed approximately 6% of the total seaward flux (three times that of Station A). In contrast to Station A, a gradual increase of the Flux 1 value from 40 to 70 psu cm/s was observed from the beginning of the study until November 5, 2005, then gradually reduced to a minimum value of 20 psu cm/s in 2 days, where it remained relatively constant for the next day, before a further increase to the same order. Moreover, it was approximately half that of Station A (except on November 6, 2005), although Station B was located near the mouth of the creek where predominantly freshwater enters the lake. Unlike Station A, Flux 2 at Station B was almost constant, with a slightly higher value on November 6, 2005; and Flux 4 was directed seaward during the first half of the study period. The net salt flux at this station was also directed seaward during the study period, with an average value of approximately 50 psu cm/s, with the exception of November 7, 2005, when it was directed landward with a value of 10 psu cm/s.
Ayub Ali 47 Chapter II Salt fluxes within a shallow estuarine system
2.0 2.0 (a) 1.5 1.5
1.0 1.0
0.5 0.5
Water depth (m) depth Water 0.0 0.0
1.0 1.0 (b) 0.5 0.5
0.0 0.0 -0.5 -0.5 Velocity (m/s) -1.0 -1.0
40 40 (c) 30 30
20 20 Salinity (psu) 10 10
200 200 (d) Flux 1 Flux 2 Flux 3 Flux 4 Net flux 100 100
0 0
Salt flux (psuSalt cm/s) -100 -100 1/11/05 2/11/05 3/11/05 4/11/05 5/11/05 6/11/05 7/11/05 8/11/05 9/11/05 10/11/05 Spring tideDate Neap tide
Figure 2.3: Hydraulic data collected at Station A: (a) water depth; (b) water velocity; (c) salinity; and (d) computed salt fluxes.
Overall, salt fluxes at both stations followed very similar trends (net seaward flux), with the highest fluxes occurring during the spring tide period, and the lowest, before the neap tide period. Moreover, the advective flux was always dominant and determined the direction of the net flux at both stations.
No direct relations between salt fluxes and spring-neap tidal variations were observed during this study. Variations of advective, as well as net salt flux at Station B, were, therefore, likely the result of rainfall events because it followed the same trend of rainfall recorded at the study site. Greater advective values correlated with greater recorded rainfall events and, similarly, lower values with reduced rainfall events. However, rainfall within different areas of the catchment might be different from that measured on-site, which caused a time lag/lead between rainfall peaks and salt flux peaks.
Ayub Ali 48 Chapter II Salt fluxes within a shallow estuarine system
Water depth and salinity were lower during spring tide and higher during neap tide conditions (see Figure 2.5). The higher depth during neap tide phase (Figure 2.5a) conditions resulted from higher storage of seawater in addition to the freshwater from the catchments because advective fluxes within the lake were reduced during this period. Furthermore, increasing water depth within the lake aided in diffusion of salts and, thus, reduction in variation of salt concentrations within the lake during the neap tide phase (Figure 2.5b). However, the advective flux was directly related to the net velocity (Figure 2.5c), which controlled the direction of the net salt flux.
2.0 (a) 2.0
1.0 1.0
Water depth (m) 0.0 0.0
1.0 (b) 1.0 0.5 0.5 0.0 0.0 -0.5 -0.5 Velocity (m/s) -1.0 -1.0
40 (c) 40
30 30
20 20 Salinity (psu) 10 10
200 (d) 200 Flux 1 Flux 2 Flux 3 Flux 4 Net flux 100 100
0 0
Salt flux (psuflux cm/s)Salt -100 -100 1/11/05 2/11/05 3/11/05 4/11/05 5/11/05 6/11/05 7/11/05 8/11/05 9/11/05 10/11/05 Spring tide Date Neap tide
Figure 2.4: Hydraulic data collected at Station B: (a) water depth, (b) water velocity, (c) salinity and (d) computed salt fluxes.
The advective flux at Station A was consistently greater in comparison to Station B, with the exception of November 7, 2005, when it was reduced to half that of Station B. An investigation of the data revealed the possibility of the occurrence of horizontal circulation within the lake as a result of a distinct ebb-flood channel system (Ahnert, 1960; Nguyen et al., 2008). Circulation within the lake was generally anticlockwise and
Ayub Ali 49 Chapter II Salt fluxes within a shallow estuarine system opposed by the strong southerly winds that occurred during the study period. The left channel of the lake, on the western side of the large island, (see Figure 2.1), is a flood- dominated channel; on the other hand, the eastern, curved main channel is an ebb- dominated channel. Therefore, net transport of salt through the left channel was landward, whereas that through the main channel was seaward, creating an anticlockwise horizontal circulation (see Figure 2.1). The circulation could be augmented by a wind-induced setup. This circulation increased advective flux at Station A, in addition to increasing the net storage of the lake. There are no data from the left channel because it was believed to be shallow and unimportant at the onset of the study.
1.5 (a) 1.5
1.0 1.0 Average depth (m) 0.5 0.5
35 (b) 35
30 30
25 25
20 20 Average (psu) salinity 15 15
0.10 (c) 0.10 Station A Station B
0.05 0.05
0.00 0.00 1/11/05 2/11/05 3/11/05 4/11/05 5/11/05 6/11/05 7/11/05 8/11/05 9/11/05 10/11/05 11/11/05
Average velocity(m/s) Date -0.05 Spring tide Neap tide -0.05
Figure 2.5: Average hydraulic data at both Stations A and B: (a) water depth; (b) salinity; and (c) water velocity.
The salinity difference between Stations A and B was 7 psu during the spring tide phase, which was reduced to 4 psu during the neap tide phase (see Figure 2.5b). On average, the dilution was observed to be 25% during this study period, which also supports the dominance of advective flux caused by freshwater flow.
Ayub Ali 50 Chapter II Salt fluxes within a shallow estuarine system
2.4 Conclusions
Similar to previously reported studies, residual water transport within the shallow waters of Coombabah Lake was identified as the dominant factor influencing residual salt transport. Furthermore, this study indicates the net salt flux alternates frequently in contrast to the steady state condition. Future investigations in the form of continuous long-term studies are required to determine when the net salt flux would change in direction at any particular lake location. Additionally, this study has identified that advective flux was of primary importance in the movement of salts within Coombabah Lake (contributing 65% of the total salt flux in this shallow subtropical estuary) and that the lake was characterised by greater average water depth during neap tide phases, which aids in the diffusion of salts, with a 25% dilution occurring within the lake system. The lake was not an inverse type of estuary during this study because of freshwater input. Future modelling studies of the lake will be required to investigate the existence of any horizontal circulation.
2.5 Acknowledgement
The authors would like to acknowledge the financial assistance of the Cooperative Research Centre for Coastal Zone, Estuary and Waterway Management. Acknowledgments are also made to Nathan Benfer and Peta Williams (Griffith University) for their assistance with fieldwork. The authors extend their acknowledgement to anonymous reviewers for their insightful comments.
2.6 References
Ahnert, F., 1960. Estuarine meanders in the Chesapeake Bay area. Geographical Review, 50(3), 390–401. Australian Bureau of Meteorology, 2007. Evaporation. http://www.bom.gov.au/cgi- bin/climate/cgi_bin_scripts/evaporation.cgi (accessed October 15, 2007). Benfer, N.P.; King, B.A., and Lemckert, C.J., 2007. Salinity observations in a subtropical estuarine system on the Gold Coast, Australia. Journal of Coastal Research, Special Issue No. 50, pp. 646–651. Bowden, K.F., 1963. The mixing processes in a tidal estuary. International Journal of
Ayub Ali 51 Chapter II Salt fluxes within a shallow estuarine system
Air and Water Pollution, 7, 343–356. Bowden, K.F., 1967. Stability effects on turbulent mixing in tidal currents. Physics of Fluids, 10(9), S278–S280. Burton, E.D.; Sullivan, L.A.; Bush, R.T., and Powell, B., 2008. Ironsulfide and trace element behaviour in sediments of Coombabah Lake, southern Moreton Bay (Australia). Marine Pollution Bulletin, 56, 1353–1376. Cox, M. and Moss, A., 1999. Nerang River, Tallebudgera, Currumbin and Coombabah Creeks: Water Quality Report 1999. Brisbane, Australia: Queensland Environmental Protection Agency, 26p. David, L.T. and Kjerfve, B., 1998. Tides and currents in a two-inlet coastal lagoon: Laguna de Terminos, Mexico. Continental Shelf Research, 18, 1057–1079. Dunn, R.J.K.; Ali, A.; Lemckert, C.J.; Teasdale, P.R., and Welsh, D.T., 2007a. Short- term variability of physio-chemical parameters and the estimated transport of filterable nutrients and chlorophyll-a in the urbanised Coombabah Lake and Coombabah Creek system, southern Moreton Bay, Australia. Journal of Coastal Research, Special Issue No. 50, pp. 1062–1068. Dunn, R.J.K.; Lemckert, C.J.; Teasdale, P.R., and Welsh, D.T., 2007b. Distribution of nutrients in surface and sub-surface sediments of Coombabah Lake, southern Moreton Bay (Australia). Marine Pollution Bulletin, 54, 606–614. Dunn, R.J.K.; Welsh, D.T.; Lee, S.Y.; Lemckert, C.J.; Teasdale, P.R., and Meziane, T.M., 2008. Investigating the distribution and sources of organic matter in surface sediment of Coombabah Lake (Australia) using elemental, isotopic and fatty acid biomarkers. Continental Shelf Research, 28(18), 2535–2549. Dyer, K.R., 1997. Estuaries—A Physical Introduction. Hoboken, New Jersey: John Wiley & Sons, 195p. Faulkner, S., 2004. Urbanisation impacts on the structure and function of forested wetlands. Urban Ecosystems, 7, 89–106. Fischer, H.B., 1972. Mass transport mechanisms in partially stratified estuaries. Journal of fluid mechanics, 53(4), 671–687. Frank, T.D. and Fielding, C.R., 2004. Sedimentology and geochemistry of an urban coastal lake system: Coombabah Lake Nature Reserve, Gold Coast, Queensland. Australian Journal of Earth Sciences, 51, 261–271.
Ayub Ali 52 Chapter II Salt fluxes within a shallow estuarine system
GHD (Gutteridge, Haskins and Davey Pty. Ltd.), 2003. Coombabah Creek Environmental Inventory. Brisbane, Australia: GHD, 439p. Hollingsworth, A. and Connolly, R.M., 2006. Feeding by fish visiting inundated subtropical saltmarsh. Journal of Experimental Marine Biology and Ecology, 336, 88–98. Hollingsworth, A. and Connolly, R.M, 2006. Feeding by fish visiting inundated subtropical saltmarsh. Journal of Experimental Marine Biology and Ecology, 336, 88-98. Hunkins, K., 1981. Salt dispersion in the Hudson Estuary. Journal of Physical Oceanography, 11, 729–738. Kjerfve, B., 1986. Circulation and salt-flux in a well-mixed estuary. In: van de Kreeke, J. (ed.), Physics of Shallow Estuaries and Bays, Berlin: Springer Verlag, pp. 22– 29. Knight, J.M.; Dale, P.E.R.; Dunn, R.J.K.; Broadbent, G.J., and Lemckert, C.J., 2008. Patterns of tidal flooding within a mangrove forest: Coombabah Lake, Southeast Queensland, Australia. Estuarine Coastal and Shelf Science, 76, 580–593. Lee, S.Y.; Connolly, R.M.; Dale, P.E.R.; Dunn, R.J.K.; Knight, J.M.; Lemckert, C.J.; McKinnon, S.; Powell, B.; Teasdale, P.R.; Welsh, D.T., and Young, R., 2006a. Impact on Urbanisation on Coastal Wetlands: A case study of Coombabah Lake, south-east Queensland. Brisbane, Australia: CRC for Coastal Zone, Estuary and Waterway Management, Technical Report No. 54, 219p. Lee, S.Y.; Dunn, R.J.K.; Young, R.A.; Connolly, R.M.; Dale, P.E.R.; Dehayr, R.; Lemckert, C.J.; McKinnon, S.; Powell, B.; Teasdale, P.R., and Welsh, D.T., 2006b. Impact of urbanization on coastal wetland structure and function. Austral Ecology, 31, 149–163. Lewis, R.E. and Lewis, J.O., 1983. The principal factors contributing to the flux of salt in a narrow, partially stratified estuary. Estuarine, Coastal and Marine Science, 16, 599–626. Miranda, L.B. and Kjerfve, B., 1998. Circulation and mixing due to tidal forcing in the Bertioga Channel, Sao Paulo, Brazil. Estuaries, 21 (2), 204–214. Nguyen, A.D., Savenije, H.H.G., van der Wegen, M., and Roelvink, D., 2008. New analytical equation for dispersion in estuaries with distinct ebb-flood channel system. Estuarine, Coastal and Shelf Science, 79, 7–16.
Ayub Ali 53 Chapter II Salt fluxes within a shallow estuarine system
Park, J.K. and James, A., 1990. Mass flux estimation and mass transport mechanism in estuaries. Limnology and Oceanography, 35 (6), 1301–1313. Pritchard, D.W., 1954. A study of salt balance in a coastal plain estuary. Journal of Marine Research, 13, 133–144. Restrepo, J.D. and Kjerfve, B., 2002. The San Juan Delta, Colombia: tides, circulations and dispersion. Continental Shelf Research, 22, 1249–1267. Simpson, J.H., Vennel, R., and Souza, A. J, 2001. The salt fluxes in a tidally-energetic estuary. Estuarine, Coastal and Shelf Science, 52, 131–142. Skinner, J.L., Gillam, E., and Rohlin, C.-J., 1998. The demographic future of the Moreton region. In: Tibbetts, I.R.; Hall, N.J., and Dennison, W.C. (eds.), Moreton Bay and Catchment. Brisbane, Australia: School of Marine Science, The University of Queensland, pp. 67–80. Stumpf, R.P. and Haines, J.W., 1998. Variations in tidal level in the Gulf of Mexico and implications for tidal wetlands. Estuarine, Coastal and Shelf Science, 46, 165– 173. Sylaios, G.K., Tsihrintzis, V.A., Akratos, C., and Haralambidou, K., 2006. Quantification of water, salt and nutrient exchange processes at the mouth of a Mediterranean coastal lagoon. Environmental Monitoring and Assessment, 119, 275–301. Uncles, R.J., Elliott, R.C.A., and Weston, S.A., 1985. Observed fluxes of water, salt and suspended sediment in a partially mixed estuary. Estuarine, Coastal and Shelf Science, 20, 147–167. Uncles, R.J. and Jordan, M.B., 1979. Residual fluxes of water and salt at two stations in the Severn Estuary. Estuarine, Coastal and Marine Science, 9, 287–302.
Ayub Ali 54 Chapter III A traversing system to measure hydraulic properties
CHAPTER III
A traversing system to measure bottom boundary layer hydraulic properties
Abstract
This study describes a new convenient and robust system developed to measure benthic boundary layer properties, with emphasis placed on the determination of bed shear stress and roughness height distribution within estuarine systems by using velocity measurements. This system consisted of a remotely operated motorised traverser that allowed a single ADV to collect data between 0 and 1 m above the bed. As a case study, we applied the proposed traversing system to investigate bottom boundary layer (BBL) hydraulic properties within Coombabah Creek in Gold Coast, Queensland, Australia. Four commonly-employed techniques: (1) Log-Profile (LP); (2) Reynolds stress (RS); (3) Turbulent Kinetic Energy (TKE); and (4) Inertial Dissipation (ID) used to estimate bed shear stresses from velocity measurements were compared. Bed shear stresses estimated with these four methods agreed reasonably well; of these, the LP method was found to be most useful and reliable. Additionally, the LP method permits the calculation of roughness height, which the other three methods do not. An average value of bed shear stress of 0.46 N/m2, roughness height of 4.3 mm, and drag coefficient of 0.0054 were observed within Coombabah Creek. Results are consistent with that reported for several other silty bed estuaries.
Ali, A. and Lemckert, C.J., 2009. A traversing system to measure bottom boundary layer hydraulic properties, Estuarine Coastal and Shelf Science. doi:10.1016/j.ecss.2009.04.017
Ayub Ali 55 Chapter III A traversing system to measure hydraulic properties
3.1 Introduction
Estuaries are of immense importance to many communities. It has been estimated that 60–80 per cent of commercial marine fishery resources depend on estuaries for part of or all of their life cycle (Klen, 2006). The flow and sediment transport patterns within estuaries are important as they play an important role in the functionality and health of these systems. Due to knowledge gaps, most numerical models used for predicting sediment transport (and related pollutant transport) rely on the use of approximations when determining bottom boundary conditions and sediment transport dynamics.
It is well recognised that the hydrodynamic properties of the bottom boundary layer (BBL) affect sediment resuspension. The shear stress near the bed directly causes sediment erosion, affects vertical mixing, and relates to conditions conducive to sediment deposition. Therefore, to accurately predict and numerically model the flow and sediment transport patterns within estuarine systems, it is important to obtain detailed velocity data near the bed (Soulsby and Dyer, 1981).
It is very difficult to directly determine the bed shear stress in the field as its determination requires the measurement of forces very close to the bed, within the viscous sub-layer (see Figure 3.1) (Ackerman and Hoover, 2001). However, several indirect methods have been developed (see Section 3.3.1) that use more readily measurable velocity data to estimate bed shear stress. Previously, point source current meters, such as the S4 or Acoustic Doppler Velocimeter (ADV) (Gross et al., 1994; Jing and Ridd, 1996; Black, 1998; Stips et al., 1998; Osborne and Boak, 1999) have been used to derive BBL properties. However, in traditional fixed mooring arrangements they cannot usually fully resolve the boundary layer as they are restricted to a single point measurement. Additionally, if a detailed boundary layer profile is to be determined, then a number of devices must be deployed at one location (Gross and Nowell, 1983; Grant and Madsen, 1986; Feddersen et al., 2007), which is usually beyond the scope of most researchers due to the high cost of equipment and installation. More recently, Acoustic Doppler Current Profilers (ADCPs) have been used to record velocity data near the bed (Cheng et al., 1999; Thomsen, 1999), as they can provide near instantaneous three- dimensional velocity profile data that can be used to estimate shear stress. However, ADCPs have limitations in that they have a large (>10 cm) and wide spread (an order of
Ayub Ali 56 Chapter III A traversing system to measure hydraulic properties
1 m) sampling volume, and are unable to sample close to the bed (approximately 10 per cent of the distance from the transducer to the bed), which is the most important region for assessing BBL properties within shallow estuarine systems.
In addition to the bed shear stress, the bed roughness is an essential parameter for modelling current circulations, wave height attenuations and sediment transport within estuarine and coastal waters – but it is often unknown and difficult to measure directly in the field. The majority of modelling software packages (e.g. MIKE21/MIKE3 and ECOMSED) use an estimated roughness height or a drag coefficient as an input parameter for describing the bed shear stress in their sediment transport formulae (e.g. DHI, 2002; HydroQual, 2002). The physical bed roughness generally consists of three roughness components: grain roughness, bed form roughness, and sediment saltation roughness (You, 2005). The total roughness can be measured from the affected velocity profiles using Prandtl’s (1926) law of the wall equation, which would substantially reduce the uncertainties of numerical models.
In this study, a new simple and robust system was developed to measure the flow properties within estuarine BBLs. The system is based around a traversing mechanism used to move an ADV vertically through the water column and, importantly, near the bed, so that hydraulic properties of the BBL could be assessed. Additionally, bed shear stresses measured using four different methods were compared. Results of the successful application of this new system are presented in this paper through a case study of a shallow estuarine system.
3.2 Theoretical Background
The flow of water near a solid boundary has a distinct structure called a boundary layer. An important aspect of a boundary layer is that the velocity of the fluid (u) goes to zero at the boundary. At some distance above the boundary the velocity reaches a constant value (Figure 3.1) called the free stream velocity u∞. Between the bed and the free stream, the velocity varies over the vertical co-ordinate. The height of the boundary layer, δ, is typically defined as the distance above the bed at which u(δ) = 0.99u∞ (see Figure 3.1) (Douglas et al., 1986).
Ayub Ali 57 Chapter III A traversing system to measure hydraulic properties
τ Total shear stress Velocity profile Flow layer classification τv Viscous shear stress velocity, u τt Turbulent shear stress Free stream
Turbulent outer layer
τ = τt
z δ height, height,
τt Turbulent logarithmic layer τ = τt = const.
Transition layer τ = τt + τv = const. τv δv Viscous sublayer τ = τv = const.
u(δ) = 0.99u∞ Bottom shear stress, τb
Figure 3.1: Typical velocity and shear stress distribution within different flow regions (layer thickness is not to scale) of a turbulent bottom boundary layer.
The BBL can be subdivided into four regions (see Figure 3.1): (1) viscous sub-layer
(thickness δv) representing a thin laminar flow layer just above the bottom – in this layer there is almost no turbulence and the viscous shear stress is constant; (2) transition layer, where viscosity and turbulence are equally important and the flow is turbulent; (3) turbulent logarithmic layer, where the viscous shear stress can be neglected and the turbulent shear stress is constant and equal to the bottom shear stress; and (4) turbulent outer layer, where velocities are almost constant because of the presence of large eddies, which produce strong mixing of the flow and shear stress gradually reducing to zero at the free stream (outer edge of the boundary layer). In a well-mixed fully developed turbulent flow over a rough channel bed, the outer turbulent layer covers approximately 80 per cent of the BBL thickness (Granger, 1985).
A typical phenomenon of turbulent flow is the fluctuation of velocity. The instantaneous velocity consists of a mean and a fluctuating component, and can be written as follows:
U u u, V v v and W w w (3.1)
Ayub Ali 58 Chapter III A traversing system to measure hydraulic properties where U, V and W are instantaneous velocities; u, v and w are time-averaged velocities; and u′, v′ and w′ are instantaneous velocity fluctuations in longitudinal, transverse and vertical directions, respectively. Shear stress in laminar flow is defined as:
du (3.2) v dz
where τv is the viscous shear stress; ρ is the density of fluid; ν is the kinematic viscosity of fluid; and z is the elevation above the bed. On the other hand, shear stress in turbulent flow is defined as:
2 du t (3.3) dz
where τt is the turbulent shear stress, and η is a turbulent mixing coefficient (often called eddy viscosity). The eddy viscosity η is not a property of the fluid like ρ and ν, but is a function of the velocity. Turbulent velocity fluctuations generate momentum fluxes resulting in shear stresses (called Reynolds stresses) between adjacent parts of a flow (Tennekes and Lumley, 1972). The Reynolds stress (turbulent shear stress) is defined as:
t uw (3.4)
This can be measured with high precision velocity recording devices such as ADV and Laser Doppler Systems. Turbulence shear stress equals the bed shear stress when measured within the constant shear stress region (Figure 3.1).
Prandtl (1926) introduced the mixing length concept and derived the logarithmic velocity profile (also known as von-Kármán – Prandtl equation) for the turbulent logarithmic layer as
u* z uz ln (3.5) z0
Ayub Ali 59 Chapter III A traversing system to measure hydraulic properties
where u is the shear velocity defined as u b ; is the bed shear stress; z is the * * b 0 elevation where velocity is zero, usually known as roughness height; and is the von- Kármán constant = 0.4. Various expressions have been proposed for the velocity distribution within the transitional layer and the turbulent outer layer, none of which is widely accepted (Granger 1985; Crowe et al., 2005). However, by modifying the mixing length assumption, the logarithmic velocity profile also applies to the transitional layer and the turbulent outer layer. Under such conditions, measurement and computed velocities show reasonable agreement. Therefore, we have assumed a turbulent layer with the logarithmic velocity profile covers the transitional layer, the turbulent logarithmic layer and the turbulent outer layer (Figure 3.1). Once detailed velocity measurement over a water column is available, the time-averaged velocities of the BBL can be fitted to the logarithmic velocity profile (Equation (3.5)), and the unknown parameters (shear velocity and roughness height) can be estimated. Furthermore, bed shear stresses can be estimated by using several other methods utilising the velocity fluctuations (e.g. Kim et al., 2000; Pope et al., 2006).
3.3 Methods and Materials
3.3.1 Techniques for estimating bed shear stress
Commonly-employed techniques to estimate bed shear stress from velocity measurements include: (1) Log-Profile (LP); (2) Reynolds stress (RS); (3) Turbulent Kinetic Energy (TKE); and (4) Inertial Dissipation (ID) methods. The suitability, assumptions and limitations of these methods have been critically reviewed by Kim et al. (2000) and Pope et al. (2006). These authors concluded that the TKE approach was the most consistent and offered most promise for future development. However, they have suggested simultaneous use of several methods to estimate bed shear stress where possible, as all of these methods have both advantages and disadvantages; in this way, likely sources of errors can be identified.
The LP method fits velocity and height data into the von-Kármán – Prandtl equation (Equation (3.5)) and estimates shear velocity and roughness height. The shear velocity is used to calculate bed shear stress from
Ayub Ali 60 Chapter III A traversing system to measure hydraulic properties
2 b u* (3.6)
One of the main problems with this law of the wall approach (LP method) is that the theory is strictly valid only for steady flows (Cheng et al., 1999; Pope et al., 2006). Another fundamental feature of the LP method is that it is critically dependent upon precise knowledge of the elevations above the bed at which the sequence of current velocities are measured (Kabir and Torfs, 1992; Biron et al., 1998). While this may be straightforward for very smooth, fine-grained, abiotic sediments, this can be considerably problematic in the case of natural estuarine systems where grain size variation, bed forms and biota may conspire to increase bed roughness and make precise determination of elevation less certain (Kabir and Torfs, 1992; Wilcock, 1996).
The RS approach (Equation (3.4)) may appear to represent a suitable method of estimating bed shear stress for fully turbulent flow with a large Reynolds number (Dyer, 1986), and for cases where measurements close to the bed are available. However, it has been shown that this method may also be largely unsuitable in field or laboratory studies because of errors arising from any tilting of the velocity measuring device or to secondary flows (Kim et al., 2000). Moreover, the measurement must be within the turbulent logarithmic layer (constant stress region), and where density stratification is not important.
Turbulent Kinetic Energy (TKE) is the absolute intensity of velocity fluctuations from the mean velocity, i.e. the variances of the flow within an XYZ co-ordinate system, and is defined as:
1 TKE u'2 v'2 w'2 (3.7) 2
Simple relationships between TKE and shear stress have been formulated in turbulence models (Galperin et al., 1988), while further studies (Soulsby and Dyer, 1981; Stapleton and Huntley, 1995) have shown the ratio of TKE to shear stress is constant, i.e.:
t C1TKE (3.8)
Ayub Ali 61 Chapter III A traversing system to measure hydraulic properties
The proportionality constant C1 was found to be 0.20 (Soulsby and Dyer, 1981), while
C1=0.19 has been adopted by others (Soulsby, 1983; Stapleton and Huntley, 1995; Thompson et al., 2003). The main advantage of the TKE method over the LP method is that it does not require accurate knowledge of elevation above the bed, and is therefore less sensitive to conditions, where sediment erosion and deposition can alter sediment levels by several millimetres or more. Furthermore, in inter-tidal field studies some tilting of the acoustic sensor is almost inevitable, and this method is less sensitive to tilting. However, there are some potential disadvantages to the use of the TKE method. Firstly, the exact limits and dimensions of the sampling volume must be known so when measurements are made within the BBL (near the bed) the sampling volume is not mistakenly positioned partially within the bed (Finelli et al., 1999). Secondly, an inherent feature of all Doppler-based backscatter systems is Doppler noise, which is attributable to several sources, including positive and negative buoyancy of particles in the sampling volume; small-scale turbulence (at scales less than that of the sampling volume); and acoustic beam divergence, which in total may lead to high-biased estimates of turbulent energy from Acoustic Doppler devices (Nikora and Goring, 1998). Finally, accelerating and decelerating flows can cause errors in the TKE approach just as in the LP method. However, this may be corrected by detrending the velocity time-series. Similar to the second technique, the measurement must be taken within the turbulent logarithmic layer. Bed shear stress can also be estimated by using spectral analysis of turbulences and energy budgets.
For a log layer, a first-order balance between shear production P and energy dissipation ε is a fair assumption (e.g. Tennekes and Lumley, 1972; Nakagawa and Nezu, 1975)
u P uw 0 (3.9) z
u u Taking uw u 2 from the Reynolds stress method (Equation 3.4) and * from * z z the LP method (Equation (3.5)), we have:
1 3 u* z (3.10)
Ayub Ali 62 Chapter III A traversing system to measure hydraulic properties
The energy dissipation ε can be estimated from the inertial sub-range of spectral density distribution of the velocity (Grant and Madsen, 1986; Gross et al., 1994) measured at height z. Then the shear velocity can be estimated from Equation (3.10).
Most importantly, all of these methods require the measurement to be made within the constant stress turbulent logarithmic layer. The aforementioned four techniques were used in this study to estimate bed shear stress from velocity data.
3.3.2 Technique for estimating roughness height and drag coefficient
While fluid flows over a solid surface, it encounters friction termed as bottom friction
(or bed roughness). The roughness height z0 is most often estimated from recorded velocity profiles (Equation (3.5)) while bed shear stresses can be computed using velocities at different points in the water column and the heights of those points with reference to the bed. The velocities and corresponding elevations measured from a water column are plotted onto a logarithmic graph, and roughness height z0 and shear velocity are obtained from curve fitting (Wilkinson, 1986; Bergeron and Abrahams, 1992; Ke et al., 1994; Mathisen and Madsen, 1996).
The drag coefficient is also used to represent the bed roughness in numerical models.
The drag coefficient CD (at a referenced height zr) can be calculated using roughness height z0 (Gross et al., 1999; Bricker et al., 2005) from:
2 (3.11) C D lnz r z0 which depends upon bed sediment grain size and bed-form geometry. Therefore, the roughness height and drag coefficient can be estimated from the traverser-collected velocity profiles.
3.3.3 New traversing system
Instrument set-up
In order to easily and readily measure velocities within the BBL, a new traversing system comprising a flexible head ADV (Vector velocimeter; Nortek AS), an altimeter
Ayub Ali 63 Chapter III A traversing system to measure hydraulic properties
(Tritech Digital Precision Altimeter; model PA500/6-S; Tritech International Ltd), and a DC underwater motor (model P00625; Seaeye Marine Ltd.) was assembled on a tripod (see Figure 3.2). The tripod was made from hollow (to reduce weight) and thin (to minimise the flow blocking effect) aluminium pipe. Along one leg of the tripod, a track was fitted along which a small cart ran. The ADV probe and the altimeter were attached to the cart, which was moved along the track using the motor (fitted on top of the tripod). Expendable wooden plates were also fitted under the legs to prevent the tripod from sinking into the ground. The ADV measured the water velocity (mean and turbulent components), while the altimeter determined the height of the sampling point above the bed. The ADV was connected to a laptop computer for the purposes of controlling and data logging. To reduce any blocking effects, the ADV sensor head was kept 120 mm away from the leg. The altimeter provided a 0–5 VDC analogue signal, which was calibrated against the height and read directly into the ADV, thus ensuring simultaneous height and velocity measurement. The traversing motor was operated using an external 12 VDC power supply and control cable.
Power supply and Underwater motor control cable
Data and Track Cart communication cable
Altimeter ADV main ADV probe housing ADV battery case Wooden plate
Figure 3.2: New traversing system.
Ayub Ali 64 Chapter III A traversing system to measure hydraulic properties
The altimeter was attached vertically in a support frame on the cart and 120 mm away from a tripod leg (Figure 3.3). The ADV probe head was set 106 mm in front of the altimeter. Nortek (2004), the manufacturer of this ADV, reported the presence of weak spots close to the boundary where velocity data might be problematic. Initially the ADV was set up vertically looking downward; however, in this configuration the velocity data were found to be very noisy between 50 and 200 mm above the bed. To reduce the thickness of the problematic layer and to get closer to the bed, after testing various angles, a 45° inclination of the ADV head-unit with the vertical was selected. The sampling volume was 180 mm below the altimeter and 100 mm away from the ADV transducer. The main housing of the ADV in addition to its external battery housing was placed on a pipe screwed to the remaining two legs (see Figure 3.2). This helped to keep the ADV sensors pointing upstream when the instrument was lowered into the water column, thus minimising the frame blocking effect. Furthermore, data were only collected when the flow was approaching towards the frame.
Altimeter cable
Altimeter Support frame ADV cable 120 mm Tripod leg 106 mm
180 mm 45° ADV probe 100 mm
Sampling volume
Figure 3.3: Schematic set-up of Altimeter and ADV probe.
A special multi-cable was made to configure the instrumentation and view the data online. This consisted of four sub-cables, including: (1) an 8-pin cable connected to the ADV; (2) a 6-pin cable connected to the altimeter; (3) a 3-pin cable connected to the
Ayub Ali 65 Chapter III A traversing system to measure hydraulic properties underwater external battery (see Figure 3.2) for supplying power to the ADV and to the altimeter; and (4) an 8-pin data I/O cable connecting to the laptop on the boat.
Overall, it was found the system can be used to measure the hydrodynamic properties at different heights up to 1 m from the bed with the accuracy of elevation of ±2 mm and the accuracy of velocity of ±0.5% of the measured value.
Study site
The traversing system was tested and used within Coombabah Creek (Figure 3.4), which is a 17 km long, moderately impacted (Cox and Moss, 1999; Lee et al., 2006; Benfer et al., 2007; Dunn et al., 2007) sub-tropical tidal creek. The creek catchment (area 44 km2) is urbanised with residential, commercial and light industrial developments. It flows through Coombabah Lake and ultimately discharges into the Gold Coast Broadwater, a vitally important coastal system both economically and recreationally within southern Moreton Bay, Queensland, Australia. Coombabah Creek’s northern bank is lined with mangroves, whilst most of its southern bank is lined with concrete and rock walls belonging to residential developments. The lower section of the creek has an average width of approximately 100 m and an average depth of 4 m, with relatively steep banks on its southern side and only a few exposed sand banks at low tide. However, the upper section is approximately 200 m wide, with an average depth of 1 m and many exposed sand/mud banks at low tide. Episodically large inputs of freshwater occur during periods of heavy rainfall, predominantly during summer periods. Benfer et al. (2007) reported that Coombabah Creek developed inverse estuary characteristics during the summer months when rainfall events did not occur.
Altimeter calibration
As mentioned previously, a critical aspect of velocity profile measurements within the BBL is an accurate knowledge of the heights at which the velocity measurements are made. For this reason an altimeter was incorporated into the traversing system. The altimeter was calibrated in a laboratory tank where we could readily and accurately measure distances. Altimeter signals (read and logged as counts by the ADV) were
Ayub Ali 66 Chapter III A traversing system to measure hydraulic properties calibrated in the lab against the height within a water tank, and the following relationship was found:
a 0.16b 59.72 (3.12) where a is the height of altimeter above the bed (mm) and b is the measured altimeter signal (count), with a coefficient of determination (R2) of 0.99. The count (b) was the mean of 2 min altimeter signals at a constant height (a) with 1 Hz frequency, while the height was measured manually with a scale ruler. The mean standard deviation of the altimeter signals was 13 counts (equivalent to 2 mm of altimeter height). The minimum height the altimeter could measure was 150 mm, a high level of noise was evident when the height was <100 mm; this limitation was a consequence of the operational nature of the altimeter. To overcome this problem on the traverser, the altimeter was set >180 mm from the bed at the lowest traverser height.
Field measurement
After the set-up was fully tested, the traversing system was taken and deployed within Coombabah Creek, Gold Coast Broadwater (Australia) for field measurements (see Figure 3.4). Measurements covered a full range of ebb current during a spring tide. The mean water depth was 2.5 m. Velocities were measured from at least five elevations above the bed, with more measurements near the bottom. Data were also collected while moving the cart up and down, with an average speed of 2.0 cm/s throughout the full traverser range, together with the point measurements. Six profiles were measured with 30 min intervals taking 20 min to complete a single profiling cycle. A profiling cycle consists of the following steps:
Step 1: Lower the traverser into the water column;
Step 2: Align ADV probe along the streamline (pointing upstream);
Step 3: Move ADV to a desired elevation;
Step 4: Record data for two min;
Step 5: Move ADV to a new elevation;
Ayub Ali 67 Chapter III A traversing system to measure hydraulic properties
Step 6: Repeat steps 4-5 at least 5 times to complete a profile;
Step 7: Move ADV to the lowest/ highest point;
Step 8: Continue moving ADV up/down up to its limit;
Step 9: Repeat steps 7-8 in opposite direction.
This sampling routine permitted analyses of the different BBL property determining techniques.
Study site
Figure 3.4: Location map of the study site, Coombabah Creek and its adjacent estuaries (adapted from Benfer et al., 2007).
3.3.4 Data processing
Initially, raw ADV data were processed using ExploreV software supplied with Nortek ADV systems (Version 1.55 Pro, Nortek AS). This software was used to rotate the measured velocity from XYZ co-ordinate system to stream-wise, transverse and vertical coordinate systems. The preset 45° inclination angle and ADV recorded heading, pitch
Ayub Ali 68 Chapter III A traversing system to measure hydraulic properties and roll data were used to rotate the measured data. The direction of the main stream flow was measured at the site with a hand-held compass. Velocity data having a correlation score < 70, or Signal Noise Ratio (SNR) < 5, or velocity greater than three times their standard deviation were counted as a bad data. Less than 5% of all measurements were of sub-standard quality, and so were removed from further processing. Stationary ADV data were used to calculate mean velocities, variances, stresses and energy dissipation rates for the measurement points. These calculated parameters were then utilised in estimating the bed shear stresses by using four different methods.
1200 1200
1000 1000 Measured Measured 800 800 Fitted
) Fitted )
600 600 Height (mm Height (mm 400 400
200 200
0 0 0 0.1 0.2 0.3 0.4 0 0.1 0.2 0.3 0.4 Velocity (m/s) Velocity (m/s) (a) (b)
Figure 3.5: Sample of measured and fitted velocity profiles: (a) stationary ADV; and (b) moving ADV.
The four distinct methods described earlier in this paper were used to calculate the bed shear stresses with stationary ADV data. The mean velocities and their elevations were fitted into the logarithmic profile (Equation (3.5)); and shear velocity and roughness height were estimated for each profile (see Figure 3.5a). Some points measured within the weak spots or outside of the logarithmic layer were excluded from the Log-Profile; however, at least four points were used for a profile. Next, the shear velocity was used
Ayub Ali 69 Chapter III A traversing system to measure hydraulic properties
to calculate bed shear stress using Equation (3.6). The estimated roughness height z0 was used in Equation (3.11) to calculate drag coefficient and the standard height of 1 m was used as the reference height in this equation.
Turbulent shear stresses at various heights were estimated using Equations (3.4), (3.7) and (3.8). Energy dissipation rates (along with their heights) were used in Equation
(3.10) to estimate the shear velocity, u*. Estimated shear velocity was then used in Equation (3.6) to calculate the bed shear stresses.
Therefore, the RS and the TKE methods provided shear stresses at different heights, and the shear stress in the constant stress layer was considered as the bed shear stress. On the other hand, the ID and the LP methods provided bed shear stresses directly.
Moving ADV data were utilised only in the logarithmic profile for calculating bed shear stress and roughness height; and subsequently drag coefficient. After removing the sub- standard data, velocities and elevations were fitted into Equation (3.5), similar to stationary ADV data (Figure 3.5b) and; shear velocity and roughness height were estimated. The estimated roughness height z0 was used in Equation (3.11) to calculate drag coefficient.
Ayub Ali 70 Chapter III A traversing system to measure hydraulic properties
0.5
0.0 Tide level (m)
-0.5 11:00 12:00 13:00 14:00 15:00 (a) Time
0.3
0.2
St Up Dow n
0.1 Depth-average velocity (m/s) 11:00 12:00 13:00 14:00 15:00 (b) Time
) 1.0 2
0.5
St Up Dow n Bed shear stress (N/m stress shear Bed 0.0 11:00 12:00 13:00 14:00 15:00 (c) Time
Ayub Ali 71 Chapter III A traversing system to measure hydraulic properties
6.0
4.0
2.0 St Up Dow n
Bed roughnessBed (mm) height 0.0 11:00 12:00 13:00 14:00 15:00
(d) Time
0.0058
0.0053 Drag coefficient St Up Dow n 0.0048 11:00 12:00 13:00 14:00 15:00
(e) Time
Figure 3.6: Time series of measured and estimated parameters: (a) tidal level; (b) mean velocity; (c) bed shear stress; (d) roughness height; and (e) drag coefficient.
3.4 Results and Discussion
Three sets of velocity and height data were measured for each profiling cycle: (1) keeping the ADV probe stationary at different heights; (2) moving the ADV probe upward; and (3) moving the ADV probe downward; to fit with the logarithmic profile. The flow properties were assumed to be steady during a profiling cycle, as it took a maximum of 20 min to complete the profiling cycle. Hence there are three sets of bed shear stress, roughness height and drag coefficient data available for each profiling cycle (Figure 3.6). Tide levels during the measurements are also shown in Figure 3.6 for the same time frame. Figure 3.6 shows that the bed shear stress follows the trend of the mean velocity; that is, high bed shear stresses during high flows and low bed shear stresses during low flows. Bed shear stress varied in the range of 0.43–0.56 N/m2 for velocities from 0.20–0.25 m/s. Results are fairly consistent with that reported by Cheng et al. (1999) for South San Francisco Bay; Kim et al. (2000) for York River Estuary;
Ayub Ali 72 Chapter III A traversing system to measure hydraulic properties and Sherwood et al. (2006) for Grays Harbor in Washington (silty bed estuaries). It can be seen that variations in bed roughness heights and drag coefficients are very small during the measurement period, which implies there was no significant change of bed material and bed forms during the ebb tide measurement period. Similar mean velocity, bed shear stress, roughness height and drag coefficient estimates were derived for stationary and moving ADV data.
1200 1200 )
) 800 800 Reynolds stress method ID method TKE method LP method Height (mm Height 400 400 Reference height (mm
0 0 0.0 0.5 1.0 0.0 0.5 1.0 (a) Turbulent shear stress (N/m2) (b) Bed shear stress (N/m2)
Figure 3.7: Sample profiles of shear stress: (a) turbulent shear stress estimated by Reynolds stress and TKE methods; and (b) bed shear stress estimated by ID method using dissipation rates at different heights and the same by LP method; (turbulence within the shaded layer could not be measured due to ADV limitations).
Turbulent shear stresses at different elevations (except within weak spots) were determined using stationary ADV data, and are presented in Figure 3.7(a). On the other hand, bed shear stresses estimated from dissipated energy recorded at various heights are shown in Figure 3.7(b) with referenced heights. A brief summary of bed shear stresses estimated by all four methods is given in Table 3.1. It can be seen from Figure 3.7(a) that the shear stresses from both the Reynolds stress and the TKE methods produced very similar shear stress variations. The highest shear stress was considered to be the bed shear stress; this was approximately 0.48 N/m2, and was observed at a height of about 160 mm above the bed. It then gradually reduced to about 0.20 N/m2 at a height of 1000 mm above the bed. Available shear stresses below 160 mm showed a
Ayub Ali 73 Chapter III A traversing system to measure hydraulic properties drastic reduction to one-fifth of the maximum value at about 20 mm above the bed. Bed shear stresses determined by the ID method (Figure 3.7(b)) provided quite similar values, with the average value being slightly lower than that from the LP method. The approximate height and thickness of flow layers during the study period were deduced from turbulent shear stress profiles (see Figure 3.7(a)). The turbulent outer layer was observed to start from approximately 160 mm above the bed, and extended beyond the measured layer. On the other hand, the thickness of the viscous sub-layer was less than 20 mm, since turbulence was still present at the lowest recorded height (20 mm).We measured velocity data at least at one point from the constant stress layer (turbulent shear stress was maximum, and quite similar to the bed shear stress derived from the LP method) and observed that the constant shear stress layer extended up to 160 mm from the bed. It is vital to precisely locate the turbulent logarithmic layer in estimating bed shear stress with the Reynolds stress and the TKE methods in the absence of the vertical profile.
Table 3.1: Bed shear stresses (N/m2) estimated by various methods.
Reynolds Profile LP TKE ID stress
1 0.44 0.55 0.39 0.25 2 0.47 0.28 0.47 0.36 3 0.56 0.60 0.56 0.45 4 0.56 0.37 0.48 0.51 5 0.50 0.45 0.36 0.33 6 0.43 0.48 0.61 0.44 Mean 0.49 0.46 0.48 0.39 Std 0.06 0.12 0.10 0.09
In summary, all four methods produced similar shear stress estimates. The mean bed shear stress estimated by the LP method was the highest (0.49 N/m2), followed by the TKE (0.48 N/m2) and the Reynolds stress (0.46 N/m2) methods. On the other hand, the estimated value derived by the ID method was the lowest (0.39 N/m2). However, the
Ayub Ali 74 Chapter III A traversing system to measure hydraulic properties variations are not large, and all shear stress estimates are within the error bands. The LP method produced quite similar bed shear stresses (Std. 0.06 N/m2) from all profiles, while the Reynolds stress method produced a more scattered value (Std. 0.12 N/m2). Therefore, the LP method was the most consistent method in relation to the ID, TKE and Reynolds stress methods.
The errors related to the shear velocity calculated from the logarithmic profile were estimated using Gross and Nowell (1983) formula:
1/ 2 1 1 R 2 err (t ) (3.13) / 2,n2 2 n 2 R where t is the Student’s t distribution for (1-α) confidence interval with n-2 degrees of freedom. Here n is the number of measurement points, and R is the regression correlation coefficient. An average error of ±30% with 95% confidence level was observed in shear velocity estimation. Moreover, Yu and Tan (2006) observed more than 3% difference of bed shear stress for 1 mm of error in height of near bed data.
The standard errors of the shear stresses estimated using the Reynolds stress method was estimated using the following Sherwood et al. (2006) formula:
2 1 Cuu C ww euw 1 2 (3.14) N Cuw
where Cuu and C ww are autocovariances of u′ and w′; Cuw is covariance of u′ and w′; N is the degrees of freedom, equal to the number of statistically independent realisations of the turbulence field (Soulsby, 1983; Bendat and Piersol, 1986), which was estimated as:
U T U n N (3.15) l zf s
where T is the sampling period and is equal to n/fs, where n is the number of samples
(=3840); and fs is the sampling frequency (32 Hz); l is the turbulence length scale, which scales with z, measurement elevation; and |U| is the mean speed. The mean
Ayub Ali 75 Chapter III A traversing system to measure hydraulic properties standard error was 0.05 (10% of the bed shear stress), with a 95% confidence interval of 0.09. Standard error of bed shear stress measured by ID method was estimated at various heights (see Figure 3.6(b)) using statistical formula and an average error was observed ±35%, with a 95% confidence limit. In the case of TKE, Garcia et al. (2006) predicted 26% of standard error from 32 sets of synthetic turbulent signal, which was validated with 80 sets of laboratory data. In addition to the statistical errors, there are several other sources of errors that were not determined in this study such as errors due to physical constraints of instrument (e.g. Doppler noise) and experimental set-up (e.g. tilting). In conclusion, all methods provided quite a similar value of bed shear stresses in view of associated error ranges.
A few limitations to this system were observed from this study: (1) the velocity data between elevations of 50 and 150 mm above the bed were noisy due to weak spots (Nortek, 2004), although this data can be used in the LP method as the mean values were unaffected; (2) velocity very near the bed was underestimated when the ADV sample volume partially penetrated into the bed, as reported by other studies (this data was not analysed here); (3) maximum traversing range of a metre may not be enough to cover the full boundary layer under all conditions; and (4) a relatively flat bed is essential for the best system stability. Future developments aim to fully automate the system to add another ADV so that, once deployed, the system can operate over a full tidal cycle.
3.5 Conclusions
This article described a new underwater traversing system that made estimation of bed shear stress and roughness height robust, and best use of all available techniques at the same time. The LP method was found to be the easiest and most useful, followed by the ID, TKE and RS methods for estimating bed shear stress within shallow estuaries and rivers. More importantly, the LP method estimated both bed shear stress and roughness height, both essential parameters for sediment (or pollutant) transport modelling at the same time, whereas the other three methods estimated only bed shear stress. Moreover, the other three methods require precise velocity measurement within the constant stress layer (within centimetres) near the bottom to determine the bed shear stress. Mean
Ayub Ali 76 Chapter III A traversing system to measure hydraulic properties velocity (after filtering noise) within the weak spot appeared reasonably accurate, and therefore was used in constructing the velocity profile. However, the same data could not be used for calculating turbulent shear stress due to the noise.
3.6 Acknowledgement
The authors would like to acknowledge the financial assistance of the Cooperative Research Centre for Coastal Zone, Estuary and Waterway Management. Acknowledgments are also made to Mr Johann Gustafson and the lab technicians for their assistance in making the traverser and collecting data from field with this system. Acknowledgements are also made to Dr M. Maraqa, Dr M. H. Azam, Dr M. F. Karim, and Mr Ryan Dunn for their suggestions on improving the manuscript.
3.7 References
Ackerman, J.D., Hoover, T.M., 2001. Measurement of local bed shear stress in streams using a Preston-static tube. Limnology and Oceanography 46 (8), 2080–2087. Bendat, J.S. and Piersol, A.G., 1986. Random data: analysis and measurement procedures : 2nd edition. John Wiley, New York, 566 pp. Benfer, N.P., King, B.A., Lemckert, C.J., 2007. Salinity observations in a subtropical estuarine system on the Gold Coast, Australia. Journal of Coastal Research SI 50, 646–651. Bergeron, N.E., Abrahams, A.D., 1992. Estimating shear velocity and roughness length from velocity profiles. Water Resources Research 28 (8), 2155–2158. Biron, P.M., Lane, S.N., Roy, A.G., Bradbrook, K.F., Richards, K.S., 1998. Sensitivity of bed shear stress estimated from vertical velocity profiles: the problem of sampling resolution. Earth Surface Processes and Landforms 23 (2), 133–139. Black, K.S., 1998. Suspended sediment dynamics and bed erosion in the high shore mudflat region of the Humber Estuary, UK. Marine Pollution Bulletin 37 (3–7), 122–133. Bricker, J.D., Inagaki, S., Manismith, S.G., 2005. Bed drag coefficient variability under wind waves in a Tidal Estuary. Journal of Hydraulic Engineering, ASCE 131 (6), 497–508. Cheng, R.T., Ling, C.H., Gartner, J.W., Wang, P.F., 1999. Estimates of bottom
Ayub Ali 77 Chapter III A traversing system to measure hydraulic properties
roughness length and bottom shear stress in South San Francisco Bay, California. Journal of Geophysical Research 104 (C4), 7715–7728. Cox, M., Moss, A., 1999. Nerang River, Tallebudgera, Currumbin and Coombabah Creeks: Water Quality Report 1999. Queensland Environmental Protection Agency, Brisbane, 26 pp. Crowe, C.T., Elger, D.F., Roberson, J.A., 2005. Engineering Fluid Mechanics, eighth ed. John Wiley & Sons Inc., USA, 656 pp. DHI, 2002. Mud Transport Module User Guide, MIKE21 MT, DHI Software. DHI Water and Environment, Copenhagen, Denmark, 110 pp. Douglas, J.F., Gasiorek, J.M., Swaffield, J.A., 1986. Fluid Mechanics, second ed. Longman Singapore Publishers Pty Ltd, Singapore, 746 pp. Dunn, R.J.K., Ali, A., Lemckert, C.J., Teasdale, P.R., Welsh, D.T., 2007. Short-term variability of physio-chemical parameters and the estimated transport of filterable nutrients and chlorophyll-a in the urbanised Coombabah Lake and Coombabah Creek system, southern Moreton Bay, Australia. Journal of Coastal Research SI 50, 1062–1068. Dyer, K.R., 1986. Coastal and Estuarine Sediment Dynamics, vol. xv. Wiley, Chichester, 342 pp. Feddersen, F., Trowbridge, J.H., Williams III, A.J., 2007. Vertical structure of dissipation in the nearshore. Journal of Physical Oceanography 37, 1764–1777. Finelli, C.M., Hart, D.D., Fonseca, D.M., 1999. Evaluating the spatial resolution of an acoustic Doppler velocimeter and the consequences for measuring near-bed flows. Limnology and Oceanography 44 (7), 1793–1801. Galperin, B., Kantha, L.H., Hassid, S., Rosati, A., 1988. A quasi-equilibrium turbulent energy model for geophysical flows. Journal of the Atmospheric Sciences 45 (1), 55–62. Garcia, C.M., Jackson, P.R., Garcia, M.H., 2006. Confidence intervals in the determination of turbulence parameters. Experiments in Fluids 40, 514–522. Granger, R.A., 1985. Fluid Mechanics. Holt, Rinehart and Winston, Tokyo, Japan. 884 pp. Grant, W.D., Madsen, O.S., 1986. The continental-shelf bottom boundary layer. Annual Review of Fluid Mechanics 18, 265–305.
Ayub Ali 78 Chapter III A traversing system to measure hydraulic properties
Gross, T.F., Nowell, A.R.M., 1983. Mean flow and turbulence scaling in a tidal boundary layer. Continental Shelf Research 2, 109–126. Gross, T.F., Williams, A.J., Terray, E.A., 1994. Bottom boundary layer spectrum dissipation estimates in the presence of wave motions. Continental Shelf Research 14 (10–11), 1239–1256. Gross, E.S., Koseff, J.R., Manismith, S.G., 1999. Three-dimensional salinity simulations of South San Francisco Bay. Journal of Hydraulic Engineering, ASCE 125 (11), 1199–1209. HydroQual, 2002. User Manual, a Primer for ECOMSED. HydroQual, Inc., Mahwah, N.J. 07430, USA, 188 pp. Jing, L., Ridd, P.V., 1996. Wave-current bottom shear stresses and sediment resuspension in Cleveland Bay, Australia. Coastal Engineering 29, 169–186. Kabir, M.R., Torfs, H., 1992. Comparison of different methods to calculate bed shear stress. Water Science and Technology 25 (8), 131–140. Ke, X.K., Collin, M.B., Poulos, S.E., 1994. Velocity structure and sea bed roughness associated with intertidal (sand and mud) flats and saltmarshes of the Wash, UK. Journal of Coastal Research 10, 702–715. Kim, S.C., Friedrichs, C.T., Maa, J.P.Y., Wright, L.D., 2000. Estimating bottom stress in tidal boundary layer from acoustic Doppler velocimeter data. Journal of Hydraulic Engineering, ASCE 126 (6), 399–406. Klen, T., 2006. Estuaries, an introduction to marine biology and oceanography.
Ayub Ali 79 Chapter III A traversing system to measure hydraulic properties
Nikora, V.I., Goring, D.G., 1998. ADV measurements of turbulence: can we improve their interpretation? Journal of Hydraulic Engineering, ASCE 124 (6), 630–634. Nortek, 2004. User Manual, Nortek Vector Current Meter. Nortek AS, Vangkroken 2, NO-1351 Rud, Norway, 84 pp. Osborne, P.D., Boak, E.H., 1999. Sediment suspension and morphological response under vessel generated wave groups: Torpedo Bay, Auckland, New Zealand. Journal of Coastal Research 15 (2), 388–398. Pope, N.D., Widdows, J., Brinsley, M.D., 2006. Estimation of bed shear stress using the turbulent kinetic energy approach – a comparison of annular flume and field data. Continental Shelf Research 26, 959–970. Prandtl, L., 1926. Uber die Ausgebildete Turbulenz. In: Proceedings of 2nd International Conference on Applied Mechanics, Zurich. pp. 62–75. Sherwood, C.R., Lacy, J.R., Voulgaris, G., 2006. Shear velocity estimates on the inner shelf off Grays Harbor, Washington, USA. Continental Shelf Research 26, 1995– 2018. Soulsby, R.L.,1983. The bottom boundary layer of shelf seas. In: Johns, B. (Ed.), Physical Oceanography of Coastal and Shelf Seas. Elsevier, Amsterdam, pp. 189– 266. Soulsby, R.L., Dyer, K.R., 1981. The form of the near-bed velocity profile in a tidally accelerating flow. Journal of Geophysical Research - Oceans and Atmospheres, 86 (NC9), 8067–8074. Stapleton, K.R., Huntley, D.A., 1995. Seabed stress determinations using the inertial dissipation method and the turbulent kinetic energy method. Earth Surface Processes and Landforms 20 (9), 807–815. Stips, A., Prandke, H., Neumann, T., 1998. The structure and dynamics of the bottom boundary layer in shallow sea areas without tidal influence: an experimental approach. Progress in Oceanography 41, 383–453. Tennekes, H., Lumley, J.L., 1972. A First Course in Turbulence. MIT Press, Cambridge, Massachusetts, USA. Thomsen, L., 1999. Processes in the benthic boundary layer at continental margins and their implication for the benthic carbon cycle. Journal of Sea Research 41, 73–86. Thompson, C.E.L., Amos, C.L., Jones, T.E.R., Chaplin, J., 2003. The manifestation of
Ayub Ali 80 Chapter III A traversing system to measure hydraulic properties
fluid-transmitted bed shear stress in a smooth annular flume - a comparison of methods. Journal of Coastal Research 19 (4), 1094–1103. Wilcock, P.R., 1996. Estimating local bed shear stress from velocity observations. Water Resources Research 32 (11), 3361–3366. Wilkinson, R.H., 1986. Variation of roughness length of a mobile sand bed in a tidal flow. Geo-Marine Letters 5, 231–239. You, Z.J., 2005. Estimation of bed roughness from mean velocities measured at two levels near the seabed. Continental Shelf Research 25, 1043–1051. Yu, G. and Tan, T.K., 2006. Errors in the bed shear stress as estimated from vertical velocity profile. Journal of Irrigation and Drainage Engineering 132 (5), 490–497.
Ayub Ali 81 Chapter III A traversing system to measure hydraulic properties
Ayub Ali 82 Chapter IV Hydrodynamics of a shallow estuarine system
CHAPTER IV
Numerical study of the hydrodynamics of a very shallow estuarine system - Coombabah Lake, Gold Coast, Australia
Abstract
Coastal wetlands and estuaries are important environments providing significant habitats for flora and fauna species – often supporting commercial and recreational fisheries. These systems also act as filters for contaminants and sediments, and the absorption of wave energy. As a consequence of the ecological significance and the potential for anthropogenic disturbances and inputs into Coombabah Lake estuary (Australia), the lake and surrounding wetlands have been the focus of recent scientific study efforts. This estuarine lake (~2 km2 in size) is a very shallow (mean depth < 1 m) estuarine system that experiences a tidal range of 1.2 m, thus resulting in the continual exposure of large mud flats at low tide. Variations in water column physio-chemical and biological parameters and nutrient concentrations of the benthic sediments have previously been attributed to the hydrodynamic regime, hydrologic events, and sediment sources. In this study, a three-dimensional (3D) hydrodynamic model with unstructured mesh is setup to simulate the hydrodynamic regime and Bottom Boundary Layer (BBL) properties. In particular, the sensitivity of calibration parameters for a very shallow estuarine model is investigated. Model results are verified by recent intensive measurements. The hydrodynamic regime of the lake was found to be favourable for settlement of suspended sediments. The results reveal the necessity to correctly measure
Ali, A., Zhang, H. and Lemckert, C.J., 2009. Numerical study of the hydrodynamics of a very shallow
estuarine system - Coombabah Lake, Gold Coast, Australia. Journal of Coastal Research, SI 56, 922-926
Ayub Ali 83 Chapter IV Hydrodynamics of a shallow estuarine system and use the appropriate bathymetry and bed roughness conditions in the numerical scheme for very shallow environments.
4.1 Introduction
Estuaries are of immense importance to many communities. It has been estimated that 60 to 80 % of the commercial marine fisheries resources depend on estuaries for part of or all of their life cycle (Klen, 2006). The characteristics of estuarine flow and sediment transport patterns are important as they play a critical role in the functionality and health of these systems. When bottom sediment is resuspended, trace metals, nutrients and organic contaminants are released into the water column, which in turn can limit the amount of light entering the water and reduce water quality (Morris and Howarth, 1998). Sediment settling can inhibit channel continuity by deposition in navigational areas. If any of these issues creates a significant problem, management strategies must be developed and implemented in order to rectify the situation and/or preserve the environment in a healthy state. These strategies usually involve the development of numerical models that must be based on sound scientific principles. However, many knowledge gaps still exist - resulting in most models relying on the use of approximations when determining boundary conditions and sediment transport dynamics.
The mechanisms that control the transport, resuspension and deposition of the fine suspended sediments or contaminants in tidal estuaries are extremely complex. They are directly influenced by highly variable hydrodynamic conditions near the bed and the characteristics of the transporting material itself. There have been many investigations of near-bed flows, resuspension and transport of sediments under natural field conditions, with the majority of these have involved non-cohesive sediments [e.g. (Soulsby et al., 1994) offshore sites (Williams et al., 1999; Nikora et al., 2002) or coastlines dominated by wind wave effects (Davies, 1985; Conley and Griffin, 2004)]. However, many estuaries have regions dominated by tidal mudflats. A quantitative understanding of sediment/contaminant transport is of fundamental importance for many engineering applications. However, the current theories describing its behaviour require further development (Black et al., 2002). This issue is also important for
Ayub Ali 84 Chapter IV Hydrodynamics of a shallow estuarine system ecological applications because suspended sediments may affect the health of aquatic ecosystems by degrading water clarity, transporting pollutants and smothering of the benthic communities. An accurate hydrodynamic model is a pre-requisite for simulating sediment/contaminant transport in aquatic systems.
As a consequence of the ecological significance and the potential for anthropogenic disturbances and inputs into Coombabah Lake, the lake and surrounding wetlands have been the focus of recent scientific effort (e.g. Hollingsworth and Connolly, 2006; Dunn et al., 2007a, 2007b and 2008; Knight et al., 2008; Ali et al., 2009). Dunn et al. (2007b) have investigated intratidal variability of physio-chemical and biological parameters in the Coombabah Lake and attributed the variations to the input sources, hydrodynamic regime, freshwater input, and tidal cycles.
The aim of this study was to set up a three-dimensional model and to simulate the hydrodynamic regime, in particular the BBL hydrodynamic properties in order to develop a better understanding of the sediment/contaminant transport processes within Coombabah Lake, Gold Coast, Australia. The sensitivity of calibration parameters of a very shallow estuarine model is investigated and discussed. The model is verified by intensive measurements made within the lake system itself.
Ayub Ali 85 Chapter IV Hydrodynamics of a shallow estuarine system
Figure 4.1: Location map of study site (left) and bathymetry of Coombabah Lake including measurement stations (right).
Ayub Ali 86 Chapter IV Hydrodynamics of a shallow estuarine system
4.2 Study site
Coombabah Lake is a sub-tropical estuarine system situated in southern Moreton Bay, south-east Queensland (Australia) (Figure 4.1), one of the fastest growing regions in the developed world (Skinner et al., 1998). The lake covers ~ 2 km2 (GHD, 2003) with an urbanised catchment area of 44 km2 characterized by residential, commercial and light industrial developments. The lake is a shallow body of water characterized by fine sediments (Dunn et al., 2007a; 2008) and located in the mid-tidal region of Coombabah Creek. The creek enters into the lake at south-west side (hereby referred to as the creek mouth) and leaves the same from the north-east side (hereby referred to as the lake entrance). Ultimately, Coombabah Creek discharges into the Gold Coast Broadwater, within southern Moreton Bay (Dunn et al., 2007b). With the exception of shallow channels, Coombabah Lake is characterised by a relatively flat bathymetry, with mean depths from 0 to 1 m. During periods of low water, large portions of the benthic lake sediments become exposed. Episodically large inputs of freshwater occur during periods of heavy rainfall, predominantly during summer periods. Despite its modest dimensions, Coombabah Lake is ecologically significant - being a valuable and recognized fish and migratory bird habitat. Dunn et al. (2008) carried out bed material analysis of the lake and concluded that the lake was dominated by mud (<63 μm) in the southern (landward) and sand (>63 μm) in the northern (seaward) regions.
4.3 Methods
4.3.1 Field Measurement
A field measurement campaign was conducted within Coombabah Lake from 1 to 10 November 2005. Tide levels were measured at eight stations (1-8) distributed over the entire lake (Figure 4.1) utilizing CTDs (NXIC-CTD; Falmouth Scientific, Inc.). Velocities were also measured at 30 cm above the bed at Stations 1 and 8 using Nortek ADVs (Vector velocimeter; Nortek AS). Station 1 had the maximum influence of freshwater flow; on the other hand, Station 8 had the maximum tidal influence. Station 8 also represents the deepest point within the lake followed by Station 4 and Station 1, which remain under water almost all the time. On the other hand, Station 5 represents the shallowest point followed by Station 7, which become dry during all low tides.
Ayub Ali 87 Chapter IV Hydrodynamics of a shallow estuarine system
Stations 2, 3 and 6 become dry only during spring tide lows. Bathymetry of the lake was measured by high accuracy (horizontal error < 0.50 m and vertical error < 0.20 m) hand- held GPS. However, the GPS was fitted on a pole and the bed level was measured from a boat. Therefore, the vertical error can be up to 0.50 m. Bathymetric points were very dense (spacing about 1.0 m) in steep areas and sparse (spacing about 100 m) in flat areas. A temporary weather station was also installed on a houseboat anchored within the lake during the study.
Another field study was conducted on 13 November 2007 to estimate bed shear stress and bed roughness within Coombabah Lake. A new traversing system was utilized to measure vertical velocity profiles at Stations 1, 4, 8, 9 and 10 during an ebb tide. The velocity and elevation data were fitted into the Prandtl’s Logarithmic velocity profile
(Prandtl, 1926). Shear velocity (u*) and bed roughness height (z0) were estimated from the best fit curve. Subsequently, the shear velocity and roughness height were utilized to 2 calculate bed shear stress (τb=ρu* , where ρ is the density of water) and roughness length
(ks=30z0) respectively. Thus calculated bed roughness lengths were utilized in calibration and bed shear stresses were utilized in verification of the model.
4.3.2 Numerical Model
A three-dimensional flexible mesh modelling system MIKE3 FM (DHI, 2008) was employed to simulate hydrodynamic properties with particular emphasis upon the investigation of the BBL. A horizontally unstructured grid was used with larger cells within flat areas and smaller cells within narrow channels. The cells were triangular in shape with approximately 1800 m2 for the largest and approximately 400 m2 for the smallest sizes amounting to 1548 cells in total. A vertically structured bottom-fitted sigma grid system was used with eight layers of variable thickness. The layers were thinner near the bottom for comparison with the observed data. Predicted tide levels with seasonal correction were applied as open boundary conditions. Measured time series of precipitation, evaporation and wind were applied as other forces in the model. A source point was also added at the creek mouth (Easting 534285 m, Northing 6911760 m) to represent the flow from the upstream of Coombabah Creek. Flow of the source was estimated from the measured velocity at Station 1 multiplied by average depth and width of the creek.
Ayub Ali 88 Chapter IV Hydrodynamics of a shallow estuarine system
4.4 Results and Discussion
4.4.1 Field Data
The lake experiences a mixed tidal regime, mainly of a semidiurnal nature. During the 2005 study, wind conditions were generally moderate but highly variable, ranging from 2 to 20 km/h with the average wind speed of 7 km/h and directed from south or south- east (blowing from the creek mouth to the lake entrance) (Ali et al., 2009). Light rainfall was recorded every day except on 7 November. Rainfall increased gradually from 0.5 mm/day on 1 of November to 18 mm/day (the maximum) on 6 of November. Evaporation varied between 5 and 10 mm/day, which is greater than the annual regional average (3.5 mm/day), and even greater than the monthly average for the same periods in other years (5 mm/day) recorded by the Australian Bureau of Meteorology (2007). Air pressure was dropped by approximately five millibar (mbar) during this period.
Bed shear stresses and bed roughness heights observed within Coombabah Lake-Creek system are presented in Table 4.1. Observed data shows bed shear stresses within the lake are significantly (by more than an order) less than that of the creek and lake entrance channel. Bed shear stresses within the lake increased almost geometrically along the main flow path landward to seaward from 0.02 N/m2 to 0.12 N/m2. The bed shear stresses within the creek also varied along the reach, high (0.78 N/m2) at the lower reach near model open boundary and low (0.55 N/m2) at the upper reach near lake entrance. The variations in bed shear stresses were expected since it depends on the flow velocity which is low in the upper reach and high in the lower reach within a tidal estuary if cross-sectional areas are the same. On the other hand, bed roughness was very similar all around the lake and the creek, though the bed materials are different (Dunn et al., 2008). The roughness length (ks) varied slightly, between 0.10 m and 0.13 m.
Table 4.1: Observed bed shear stress and bed roughness length. Bed shear stress, τ Roughness length, k Station b s (N/m2) (m) 1 0.02 0.11 4 0.06 0.13 8 0.12 0.12 9 0.55 0.11 10 0.78 0.10
Ayub Ali 89 Chapter IV Hydrodynamics of a shallow estuarine system
Bed shear stresses and bed roughness lengths were measured during ebb tide because of field constraint. The lake is very shallow and did not allow our boat to move freely for more than an hour. Therefore, the traverser was moved to Station 1 during flood tide and started measurement immediately after starting of the ebb tide to avoid the boat to be stranded. However, bed shear stresses and bed roughness lengths were assumed to be very much similar during flood and ebb tides when there is no rainfall.
4.4.2 Model Calibration
The model was calibrated against water levels and velocities measured from 1-10 November 2005. Bathymetric correction, bed roughness and eddy viscosity were tuned in the calibration processes. Correlation coefficients between simulated and measured water levels and velocities were used as model performance indicators. To examine the sensitivity to possible errors in the measured bathymetric levels the lake bathymetry was artificially lowered from 0 to 0.25 m in 0.05 m increments. The correlation coefficients between observed and simulated time series data were compared. The highest correlation was observed to occur when the bed was lowered by 0.15 m with respect to the measured elevation (Figure 4.2). However, a ±0.20 m error in bathymetric measurements is usual (Chia-chyang and Hsing-wei, 2003). Water levels at Stations 3 and 4, and velocity at Station 8 showed continuous improvement up to the maximum (0.25 m) lowering. In contrast, water level at Station 7 slightly deteriorated when bed level was lowered. The average correlation coefficient improved by 0.06 for a bed level lowering by 0.15 m. This shows how critical it is to get the Digital Elevation Model (DEM) correct for such shallow systems. Deeper systems are certainly not as sensitive to such elevation errors due to a greater volume/depth ratio.
Ayub Ali 90 Chapter IV Hydrodynamics of a shallow estuarine system
1.0
0.9 2
0.8
0.7
wl-1 wl-2 wl-3 wl-4 wl-5 Correlation coefficient, R coefficient, Correlation 0.6 wl-6 wl-7 wl-8 vel-1 vel-8
0.5 0.00 0.05 0.10 0.15 0.20 0.25 Bed level lowering (m)
Figure 4.2: Correlation coefficients between observed and simulated water levels and velocities for various bed level lowering.
The model was also tested against a series of bed roughness values. A range of roughness length from 0.02 m to 0.20 m (constant over entire model domain) was utilized in the calibration process. It was found to be the second most sensitive calibration parameter to the model. The observed roughness length 0.10 m (constant over the entire model domain) provided the best result. A bed roughness map generated based on the grain size distribution (0.06 m for muddy area and 0.12 m for the sandy area) was also tested and found the results quite similar to the constant roughness 0.06 m. However, almost all the measurement stations were located within muddy areas where the roughness did not vary on the roughness map. Similarly, the model was tested with various eddy viscosities and was found relatively less sensitive to it. The correlation coefficient changed by 0.02 when the eddy coefficient was changed by an order of 2. Finally, the k-ε eddy formula was selected with default values (DHI, 2008) of various parameters. Therefore, numerical models of shallow estuaries are highly sensitive to bathymetric accuracy followed by bed roughness and eddy viscosity.
Ayub Ali 91 Chapter IV Hydrodynamics of a shallow estuarine system
1.0 Observed Simulated
0.5
0.0 Tide level (m+MSL) level Tide
-0.5 1/11/05 0:00 3/11/05 0:00 5/11/05 0:00 7/11/05 0:00 9/11/05 0:00 11/11/05 0:00 Date and time
0.5 Observed Simulated
0.25
0 Velocity (m/s) -0.25
-0.5 1/11/05 0:00 3/11/05 0:00 5/11/05 0:00 7/11/05 0:00 9/11/05 0:00 11/11/05 0:00 Date and time
Figure 4.3: Comparison of tide level (top) and velocity (bottom) at Station 1.
Simulated water levels and velocities at Station 1 are compared with the measured data on Figure 4.3. Quite good agreement was observed between simulated and measured water levels. However, the simulated velocity was slightly low during both flooding and ebbing tides. Very low bed roughness (0.02 m) slightly improved velocity calibration but significantly deteriorated water level calibration. In general, the model reproduced the field water level and velocity quite well.
Ayub Ali 92 Chapter IV Hydrodynamics of a shallow estuarine system
4.4.3 Simulation Results
The numerical model described previously was utilized to analyse the hydrodynamic regime of the lake particularly the flow distribution within the lake. A snapshot of velocity vectors (Figure 4.4) shows a number of eddy circulations were generated at the onset of flood tide. Figure 4.4 shows the flood tide already travelled up to the middle of the lake through the left channel while the main channel experiencing ebb tides. More importantly, an eddy around the large island reveals the existence of flood and ebb dominated channels within the lake. The flow distribution between the main channel and the left channel near the lake entrance are plotted on Figure 4.5 to compare their conveyance capacities. Simulated results reveal the total flow is distributed by 60% and 40% between main channel and left channel, respectively. Figure 4.5 also demonstrates that the ebb tide takes longer time than the flood tide. On the other hand, the flood flow is stronger than the ebb flow. As a result, the flood tide can carry relatively more and heavier sediments than the ebb tide. The strong flood tide brings marine borne sediments (especially sand) into the lake entrance.
6914000
6913800
6913600
6913400
6913200
6913000
Northing (m) Northing 6912800
6912600
6912400
6912200
534000 534500 535000 535500 Easting (m)
Figure 4.4: Simulated flow field at the onsets of flood tide.
Ayub Ali 93 Chapter IV Hydrodynamics of a shallow estuarine system
The three-dimensional model facilitated estimation of bottom boundary layer properties such as bed shear stress and Turbulent Kinetic Energy (TKE). Simulated bed shear stresses (Figure 4.6) matched quite well with observed value (see Table 4.1). Figure 4.6 shows bed shear stresses around the mud dominated inner lake areas are an order of magnitude lower than that of the sand dominated lake entrance. Therefore, the hydrodynamic regime of the lake is conducive for the catchment generated suspended sediments to settle on the lake bed. The model replicated vertical TKE profile reasonably well except a region around mid-depth (Figure 4.7) where simulated value is approximately 50% higher than the observed value. However, the observed value may contain approximately 30% errors in it (Garcia et al., 2006). Moreover, the model itself inherits a number of uncertainties (Huang et al., 2001).
50
0 /s) 3
-50 Discharge (m -100 Main channel Left channel -150 1/11/05 0:00 2/11/05 0:00 3/11/05 0:00 4/11/05 0:00 Date and time
Figure 4.5: Simulated flow through bifurcation channels at Lake entrance (positive seaward).
Ayub Ali 94 Chapter IV Hydrodynamics of a shallow estuarine system
Figure 4.6: Simulated bed shear stress at ebb tide.
1.2
1.0 Observed 0.8 Simulated
0.6
0.4 Height above bed (m) bed above Height 0.2
0.0 0.000 0.001 0.002 0.003 0.004 Turbulent kinetic energy (m2/s2)
Figure 4.7: Observed and simulated turbulent kinetic energy.
Ayub Ali 95 Chapter IV Hydrodynamics of a shallow estuarine system
4.5 Conclusions
Both observed and simulated velocity supports the classical tidal pumping of sediments landward within Coombabah Lake. The hydrodynamic regime of the lake is also conducive for the catchment generated suspended sediments to settle within the lake. Similar to Yang and Khangaonkar (2008), this study found an accurate bathymetry is the most important parameter for modelling hydrodynamics particularly within shallow estuaries. Bed roughness was found as the second most important parameter in the modelling exercise for this region.
4.6 Acknowledgement
The authors would like to acknowledge DHI Water and Environment, Denmark for their assistance providing MIKE modelling system available for this study. The authors would like to acknowledge the financial assistance of the Cooperative Research Centre for Coastal Zone, Estuary and Waterway Management. Acknowledgments are also made to Ryan J. K. Dunn, Johan Gustafson and the lab technicians (Griffith University) for their assistance with fieldwork.
4.7 References
Ali, A., Lemckert, C.J. and Dunn, R.J.K., 2009. Salt fluxes in a very shallow sub- tropical estuary. Journal of Coastal Research, in press. Australian Bureau of Meteorology, 2007. [online] Available from
Ayub Ali 96 Chapter IV Hydrodynamics of a shallow estuarine system
FM, Hydrodynamic Module. DHI Water and Environment, Horsholm, Denmark, 110p. Davies, A.G., 1985. Field observations of the threshold of sediment motion by wave action. Sedimentology (Oxford), 32 (5), 685–704. Dunn, R.J.K., Lemckert, C.J., Teasdale, P.R., and Welsh, D.T., 2007a. Distribution of nutrients in surface and sub-surface sediments of Coombabah Lake, southern Moreton Bay (Australia). Marine Pollution Bulletin, 54, 602-625. Dunn, R.J.K., Ali, A, Lemckert, C.J., Teasdale, P.R., and Welsh, D.T., 2007b. Short- term variability of physio-chemical parameters and the estimated transport of filterable nutrients and chlorophyll-a in the urbanised Coombabah Lake and Coombabah Creek system, southern Moreton Bay, Australia. Journal of Coastal Research, SI 50, 1062-1068. Dunn, R.J.K., Welsh, D.T., Teasdale, P.R., Lee, S.Y., Lemckert, C.J., and Meziane, T., 2008. Investigating the distribution and sources of organic matter in surface sediment of Coombabah Lake (Australia) using elemental, isotopic and fatty acid biomarkers. Continental Shelf Research, 28 (18), 2535-2549. Garcia, C.M., Jackson, P.R and Garcia, M.H., 2006. Confidence intervals in the determination of turbulence parameters. Experiments in Fluids, 40, 514-522. GHD (Gutteridge, Haskins and Davey Pty Ltd), 2003. Coombabah Creek Environmental Inventory. Gutteridge, Haskins and Davey Pty Ltd, Brisbane, Australia, 439p. Hollingsworth, A. and Connolly, R.M, 2006. Feeding by fish visiting inundated subtropical saltmarsh. Journal of Experimental Marine Biology and Ecology, 336, 88-98. Huang, J., Lai, Y.G. and Patel, V.C., 2001. Verification and validation of a 3-D numerical model for open-channel flows. Numerical Heat Transfer, Part B, 40, 431-449. Klen, T., 2006. Estuaries, An Introduction to Marine Biology and Oceanography,
Ayub Ali 97 Chapter IV Hydrodynamics of a shallow estuarine system
Queensland, Australia. Estuarine Coastal and Shelf Science, 76, 580-593. Morris, A.W. and Howarth, M.J., 1998. Bed stress induced sediment resuspension. Continental Shelf Research, 18, 1203-1213. Nikora, V., Goring, D. and Ross, A., 2002. The structure and dynamics of the thin near- bed layer in a complex marine environment: a case study in Beatrix Bay, New Zealand. Estuarine, Coastal and Shelf Science, 54(5), 915–926. Prandtl, L., 1926. Uber die Ausgebildete Turbulenz. Proceedings of 2nd International Conference on Applied Mechanics, Zurich, 62-75. Skinner, J.L., Gillam E. and Rohlin, C.J., 1998. The demographic future of the Moreton Region. In: (Tibbets, I.R., Hall, N.J. and Dennison, W.C. eds.) Moreton Bay and Catchment, School of Marine Science, University of Queensland, Brisbane, 67– 78. Soulsby, R.L., Atkins, R. and Salkield, A.P., 1994. Observations of the turbulent structure of a suspension of sand in a tidal current. Continental Shelf Research 14(4), 429–435. Yang, Z. and Khangaonkar, T., 2008. Modelling tidal circulation and stratification in Skagit River estuary using an unstructured grid ocean model. Ocean Modelling, doi:10.1016/j.ocemod.2008.07.004 Williams, J.J., Rose, C.P., Thorne, P.D., O’Connor, B.A., Humphrey, J.D., Hardcastle, P.J., Moores, S.P., Cooke, J.A. and Wilson, D.J., 1999. Field observations and predictions of bed shear stresses and vertical suspended sediment concentration profiles in wave–current conditions. Continental Shelf Research, 19(4), 507–536.
Ayub Ali 98 Chapter V Sediment dynamics of a shallow estuarine system
CHAPTER V
Sediment dynamics of a very shallow subtropical estuarine system – Coombabah Lake, Gold Coast, Australia
Abstract
This study investigated the sediment dynamics of a very shallow subtropical estuarine lake. Total suspended solids (TSS) concentrations, turbidity, salinity, and tide levels were measured at eight stations within the lake. In addition, velocity data were collected from two of the above eight stations. Meteorological data were also collected on-site during the study period. Data were analysed to determine the dominant sediment dynamic processes within the lake. Sediment transport was simulated using a three- dimensional numerical model to understand the influence of various physical processes. Sediment dynamics of the lake were found to be dominated by advection process driven by tides with wave and wind playing minor roles. Simulation results agreed well with field data and supported the aforementioned findings. Correlation between TSS and turbidity was also investigated and found very poor; therefore, the employed automatic data logging system (turbidity meters) was determined inappropriate for the estimation of TSS concentration in the very shallow subtropical estuarine system and that future studies, or those in similar systems, requires a different approach.
Ali, A., Lemckert, C.J., Zhang, H. and Dunn, R.J.K., 2009. Sediment dynamics of a very shallow sub-
tropical estuarine system – Coombabah Lake, Gold Coast, Australia. Estuarine, Coastal and Shelf
Science (under revision)
Ayub Ali 99 Chapter V Sediment dynamics of a shallow estuarine system
5.1 Introduction
Estuaries are of immense importance to many communities. It has been estimated that 60 to 80 % of commercial marine fishery resources depend on estuaries for part or all of their life cycle (Klen, 2006). The characteristics of estuarine flow and sediment conditions are important as they play a critical role in the functionality and health of these systems. When bottom sediment is resuspended trace metals, nutrients and organic contaminants can be released into the water column, which in turn can limit the amount of light entering the water and reduce water quality (Morris and Howarth, 1998). This issue is important for ecological applications because suspended sediments may affect the health of aquatic ecosystems by reducing water clarity, transporting contaminants and smothering benthic communities (Wood and Armitage, 1997; Morris and Howarth, 1998; Yamada and Nakamura, 2002; Norkko et al., 2006). If any of these issues create significant problems, management strategies must be developed and implemented in order to rectify the situation and/or preserve the environment in a suitable state. Such strategies usually involve a detailed understanding of the physical processes, in particular the sediment dynamics within the estuary determined from observed data and numerical models (e.g. van Maren and Hoekstra, 2004; Lopes et al., 2006; Ralston and Stacey, 2007).
Mechanisms which control the transport, resuspension and deposition of fine sediments (and sediment bound contaminants) in tide dominated estuaries are extremely complex. To date there have been many investigations of near-bed flows, resuspension and transportation of sediments under natural field conditions, with the majority of these involving non-cohesive sediments [e.g. (Soulsby et al., 1994) offshore sites (Williams et al., 1999; Nikora et al., 2002) or coastlines dominated by wave actions (Davies, 1985; Butt and Russell, 2000; Conley and Griffin, 2004)]. However, many estuaries have regions dominated by tidal mudflats. A quantitative understanding of sediment/contaminant transport is of fundamental importance for many engineering design and environmental decision making processes. However, current theories describing its behaviour require further development (Black et al., 2002).
Fundamental differences exist in estuarine environments (e.g. estuary type, mixing structure, water depth and tidal influence) that directly affect the way in which tropical
Ayub Ali 100 Chapter V Sediment dynamics of a shallow estuarine system and subtropical estuaries process material in comparison to temperate estuaries (Eyre, 1998; Day et al., 1989; Alongi, 1998). Therefore, knowledge of sediment dynamics in tropical and subtropical estuaries is important for understanding global sediment and biogeochemical budgets (Booth et al., 2000). In contrast to the tropical region, climatic conditions (e.g. temperature, wind and rainfall) in subtropical region vary seasonally that influence the sediment transport, biological and geochemical processes in estuarine systems.
Coombabah Lake, surrounded by two environmental reserves (Coombabah Lake Nature Reserve and the Ivan Gibbs Wetland Reserve), is an area of significant international importance under the Ramsar Convention on Wetlands and is classified as a Fish Habitat Area within the Moreton Bay Marine Park. The wetland serves as an important wildlife corridor between the Nerang State Forest and the coastline (Frank and Fielding, 2004). Over the past five decades, human activities surrounding Coombabah Lake have altered significantly. Urban development surrounding the lake included the clearing of land for golf courses and the construction of a sewage treatment plant on the eastern margin of the lake.
As a consequence of the ecological significance and the potential for anthropogenic disturbances and inputs into Coombabah Lake; the lake and surrounding wetlands have been the focus of recent scientific efforts (e.g. Frank and Fielding, 2004; Hollingsworth and Connolly, 2006; Dunn et al., 2007a and 2007b; Benfer et al., 2007; Burton et al., 2008; Dunn et al., 2008; Knight et al., 2008; Ali et al., 2009a, 2009b; Ali and Lemckert, 2009; and Dunn et al., 2009). Most efforts aimed at understanding the physical and biogeochemical processes within the lake. Dunn et al. (2007a) studied intra-tidal variability of selected water quality parameters and recommended intertidal and sub- tidal assessment. Previously Ali et al. (2009a) quantified salt fluxes of Coombabah Lake and determined the dominant mixing processes of this estuarine system and recommended further modelling studies. Subsequently, Ali et al. (2009b) carried out a hydrodynamic modelling study of the lake to understand the hydrodynamics of Coombabah Lake. However, knowledge regarding sediment dynamics within the system is still poor.
Ayub Ali 101 Chapter V Sediment dynamics of a shallow estuarine system
This study utilised 10 days of hydrodynamic and suspended sediment data, and a three- dimensional model with unstructured meshes utilising the MIKE3 FM (DHI, 2008a and 2008b) modelling system, to provide an understanding of the sediment dynamics (more specifically; the primary source, distribution, settling, resuspension and transport of cohesive sediments) within Coombabah Lake (and the surrounding areas). The model simulated the sediment dynamics within the shallow intertidal lake in an attempt to identify sediment/contaminant transport processes. Furthermore, the sensitivity of bed roughness on TSS concentration was also examined utilizing the numerical model.
5.2 Coombabah Lake
Coombabah Lake is a subtropical estuarine system situated in Gold Coast, Queensland, Australia. The lake covers ~ 2 km2 with an urbanised catchment area of 44 km2 characterised by residential, commercial and light industrial developments (GHD, 2003). The lake (Figure 5.1(a)) is a shallow body of water characterized by fine sediments with the southern (landward) and northern (seaward) surface sediments dominated by mud (< 63μm) and sands (> 63μm), respectively (Dunn et al., 2007b, 2008). The lake is located in the mid-tidal region of Coombabah Creek. During rainfall events creek waters enter the lake at the south-west side (hereby referred to as the creek mouth) and leave the lake from the north-east side (hereby referred to as the lake entrance). Ultimately, Coombabah Creek discharges into the Gold Coast Broadwater, within southern Moreton Bay.
Coombabah Lake is hydrologically open and affected by tidal flux with a tidal range of less than 1 m and predominantly a semidiurnal mixed tidal regime (Ali et al., 2009a). The lake is fed from the southwest by Coombabah Creek, which meanders ~ 15 km from its headwaters in Nerang State Forest, through residential areas, and into Coombabah Lake. The creek exits the lake at its northern end and flows into Saltwater Creek before joining the Coomera River, which discharges into Moreton Bay to the west of South Stradbroke Island.
Ayub Ali 102 Chapter V Sediment dynamics of a shallow estuarine system
(a) (b)
Figure 5.1: (a) Location map of the study site; and (b) Lake bathymetry with sampling (circle) and model boundary (triangle) locations (AHD means the Australian Height Datum).
Ayub Ali 103 Chapter V Sediment dynamics of a shallow estuarine system
Episodically, large inputs of freshwater occur during periods of heavy rainfall, predominantly during summer periods. During periods of minimal rainfall, peak velocities during flood/ebb tides at the creek mouth are ~ 0.25 m/s (Ali et al., 2009b). Alternatively, the peak velocity within the main channel near the lake entrance was ~ 0.5 m/s during the same period because of the difference in tidal prism and channel geometry. The freshwater added to the lake via the Coombabah Creek was ~ 2.5 m3/sec during the study period. With the exception of shallow channels, Coombabah Lake is characterised by a relatively flat bathymetry (Figure 5.1(b)), with a mean depth of ~ 0.5 m. During periods of low water, large portions of the surface sediments become exposed.
5.3 Methods
5.3.1 Field and Laboratory Measurements
In situ measurements were conducted within Coombabah Lake from November 1 to November 10, 2005. Water depth, salinity and turbidity were measured simultaneously at eight stations (1-8) (Figure 5.1(b)) utilizing moored CTD probes (NXIC-CTD; Falmouth Scientific, Inc., Cataumet, Massachusetts). CTD sensors were positioned at 15 cm above the bed and collected the data for 3.5 minutes (at a frequency of 10 Hz) in every 15 minutes. Of the lake sampling stations, Station 1 was most influenced by catchment flows (freshwater) in comparison to Station 8 which, was most influenced by the tidal (seawater) regime (Ali et al., 2009a, 2009b). Station 8 also represented the deepest point within the lake followed by Station 4 and Station 1, which all remained submerged for the majority of the time. Station 5 represented the shallowest sample location followed by Station 7, both of which become exposed during all low tides. Stations 2, 3 and 6 were exposed only during spring tide lows.
Current data were collected at Stations 1 and 8 employing Acoustic Doppler Velocimeters (Vector velocimeter; Nortek AS, Rud, Norway) positioned 30 cm above the bed. ADV sensors collected the data at burst intervals of 30 minutes, at a frequency of 32 Hz and 4096 samples per burst.
A water sample and associated turbidity recording was simultaneously collected at each station twice during a sampling event (once everyday). Water samples were collected
Ayub Ali 104 Chapter V Sediment dynamics of a shallow estuarine system manually using a sample pole and 500 mL low density polyethylene (Nalgene) sample bottle from a small flat bottom research vessel. Turbidity values were recorded using a handheld Nephelometer (NEP 160 Turbidity meter, McVan Instruments Pty. Ltd.). Water samples and turbidity values were collected and measured at a depth ~ 30 cm below the water surface. However, the sampling points at the majority of the stations represented the mid-depth because total water depths were ~ 0.5 m. All stations were sampled daily during high tide throughout the study period because of access limitation. It took ~ 45 minutes to complete sampling at all stations. Great care was taken during vessel movement and sample collection to avoid both wave action and the resuspension of bottom sediments. Since filter tubes (filled with oil) supplied with the CTD were used in front of the pressure sensors, wave data were lost because of filtration. However, wave heights and wave periods were visually estimated at all stations during water sampling. In addition, water samples for the determination of TSS concentrations were also collected hourly for thirteen hours on November 4, 2005 in order to determine the tidal variations.
To determine TSS, water samples were filtered through pre-weighed GF/F membranes (47 mm Ø, Millipore) and the membranes were air-dried in desiccators to constant weight (Balls, 1994). All TSS and the corresponding turbidity data were utilised to estimate correlation between the TSS and the turbidity.
The bathymetry of the lake was also measured using a high accuracy (horizontal error < 0.50 m and vertical error < 0.20 m) hand-held Trimble RTK GPS unit. Bathymetric points were very dense (spacing about 1.0 m) in steep areas and sparse (spacing about 100 m) in flat areas. A temporary weather station was also installed on a houseboat anchored within the lake during this study. Meteorological conditions namely air pressure, solar radiation, humidity, rainfall, and wind speed and direction within the lake environment were recorded every 15 minutes during the study period using a data logging weather station (WeatherMaster 2000; Environdata). Daily evaporation was estimated from the collected meteorological data using a modified Penman equation facilitated within Environdata EasiAccess software supplied with the weather station.
Ayub Ali 105 Chapter V Sediment dynamics of a shallow estuarine system
5.3.2 Numerical Model
A 3D flexible mesh modelling system MIKE3 FM (DHI, 2008a) was used to simulate sediment dynamics within Coombabah Lake. A horizontally unstructured grid was used with larger cells within flat areas and smaller cells within narrow channels. The cells were triangular in shape with ~ 1800 m2 for the largest and ~ 400 m2 for the smallest sizes, amounting to 1548 cells in total. A vertically structured bottom-fitted sigma grid system was used with eight layers of different thickness – thicker near the surface and thinner near the bed. The top three layers were equally spaced with each layer thickness of 0.2 times the total water depth, next three layers were equally spaced with each layer thickness of 0.1 times the total water depth and the bottom two layers were also equally spaced with each layer thickness of 0.05 times the total water depth. The hydrodynamic model is based on the solution of the 3D incompressible Reynolds averaged Navier- Stokes equations, subject to the assumptions of Boussinesq and hydrostatic pressure (DHI, 2008a). MIKE3 FM uses a standard advection-diffusion equation for simulating transport of scalar quantities (e.g. salinity, suspended sediment).
A water column experiences a shear stress at the surface and another shear stress at the bottom. Surface shear stress is caused by the wind and the bottom shear stress is caused by the bed roughness. Since surface and bottom shear stresses are important for hydrodynamics and sediment transport of shallow water systems, the wind shear stress and the bottom shear stress related formulas are documented herein. The wind shear stress is defined as (DHI, 2008a):
s a cd u10 u10 (5.1)
where a is the density of air; cd is the drag coefficient of air; and u10 u10 ,v10 is the wind speed 10 m above the water surface. The friction velocity associated with the surface stress is given by:
2 a cd u10 Us (5.2) 0
Ayub Ali 106 Chapter V Sediment dynamics of a shallow estuarine system
The drag coefficient can either be a constant or depend on the wind speed. The empirical formula used for the parameterisation of the drag coefficient is:
ca u10 ua cb ca cd ca u10 ua ua u10 ub (5.3) ub ua cb u10 ub
where ca , cb , ua and ub are empirical factors. The default values for empirical factors
-3 -3 ca = 1.255x10 , cb = 2.425x10 , ua = 7 m/s and ub = 25 m/s (DHI, 2008a) have been used for this simulation. Transformation of shear stresses through the water column was modelled by using the k vertical eddy viscosity formula (Rodi, 1984). The bed shear stress b bx , by was determined by a quadratic friction law:
b c f u f u f (5.4) 0
where x and y are horizontal co-ordinates; 0 is the reference density of water; c f is the drag coefficient; and u f u f ,v f is the flow velocity at a distance zb above the bed and the drag coefficient is:
c f (5.5) zb ln z0
k where 0.4 is the von Kármán constant; z s is the bed roughness length scale 0 30 and k s is the roughness height.
Deposition occurs when the computed bed shear stress is less than the critical bed shear stress for deposition (a predefined value). On the other hand, erosion occurs when the computed bed shear stress is greater than the bed shear stress for erosion (another predefined value). For further details, readers are referred to the MIKE3 Users and Scientific Manuals (DHI, 2008a and 2008b).
Ayub Ali 107 Chapter V Sediment dynamics of a shallow estuarine system
A sediment transport model was setup for Coombabah Lake utilising a calibrated hydrodynamic model of Ali et al. (2009b). The hydrodynamic model was calibrated against water levels and velocities measured at Stations 1 and 8. Bathymetric correction, bed roughness and eddy viscosity were tuned in the calibration processes.
The bed sediments may potentially be resuspended by surface shear stress in shallow lake systems. On-site measured wind data was utilised for the simulation of surface shear stresses. Sediment transport modelling depends to a greater extent on the bed shear stress, and the bed roughness (Whitehouse et al., 2000). In most cases, numerical models (e.g. MIKE3, ECOMSED) simulate bed shear stress from near bed velocity and bed roughness using Equations (5.4) and (5.5). In the present study, bed shear stress and bed roughness data collected by Ali and Lemckert (2009) from the lake-creek system was used as the boundary conditions. Sediments start to move when simulated bed shear stress is greater than a pre-defined critical bed shear stress and start to settle down when simulated bed shear stress is less than another predefined critical bed shear stress. Critical bed shear stresses for erosion and deposition differ based on bed composition and depositional history. Since the model utilised the measured bed roughness, it was not changed during sediment transport model calibration. However, sensitivity of bed roughness on sediment concentration was tested.
Tide levels were predicted utilizing astronomical tidal constituents estimated by Ali et al. (2009a) near the open boundary of the model. Thus, predicted tide levels with a correction for low air pressure developed in this region during the study period, were applied as open boundary conditions. Measured time series of precipitation, evaporation and wind were applied as other forces in the model. A source point was also added at the creek mouth (See Figure 5.1(b); Easting 534285 m, Northing 6911760 m) to represent the flow from the upstream of Coombabah Creek. Flow of the source was estimated from the measured velocity at Station 1 multiplied by the mean depth and width of the creek.
The sediment transport model was utilized to simulate cohesive sediment only. Constant sediment concentrations were used at the open boundary and at the source point. A constant bed roughness height (ks) of 0.1 m obtained from the hydrodynamic model calibration of Ali et al. (2009b) was utilized in this sediment transport model.
Ayub Ali 108 Chapter V Sediment dynamics of a shallow estuarine system
5.3.3 Data Analysis
The relationship between TSS concentrations and turbidity was investigated using linear (Pearson) correlation analysis. No data was transformed. Statistical significance was set at α = 0.05. Statistical software used was SPSS for Windows (SPSS Inc.) version 11.5.
5.4 Results and Discussion
5.4.1 Field Data
A mixed tidal pattern was observed in water depth, velocity and salinity data, with higher velocity and salinity ranges occurring during spring tides at both Stations 1 and 8 (Figures 5.2(a-c)). The opposite trend was observed in the case of the tidal range, where lower ranges occurred during spring tide conditions, compared with the greater ranges observed during neap tide conditions. The spring semidiurnal tide also changed to a mixed type during neap tide conditions. Moreover, high water level during neap tide was higher than the high water level during spring tide period. Mean water depth and salinity were greater during neap tide conditions in comparison to the spring tide conditions (Figures 5.3(a-b)). Potential explanations of these peculiarities were twofold. Firstly, the presence of a low air pressure system that developed in the region during the neap tidal period resulted in an increase in water level by ~ 5 cm, and secondly, the occurrence of a recorded rainfall event within the lake catchment contributed increased freshwater flow into the shallow lake (see Figure 5.4). However, the actual rate of freshwater inflow is unknown. The increasing salinity trend was due to the added storage from the seawater, indicating the quantity of freshwater input was relatively small. Salinity of the lake followed the tide with high salinity during high tides and low salinity during low tides (Figure 5.2(c)). Alternatively, turbidity was lower during high tides and higher during low tides (Figure 5.2(d)). Observed turbidity at Station 1 was much higher (almost four times) than the turbidity observed at Station 8. On the other hand, salinity and flow velocity at Station 1 were lower than that of Station 8. Therefore, it can be concluded that the suspended sediments of the lake are catchment borne and enter the lake through the upper Coombabah Creek with catchment runoff. The turbid lake water mixes with relatively clear sea water during flood tides and propagates seaward during the ebb tide. In other words, relatively clear seawater pushes
Ayub Ali 109 Chapter V Sediment dynamics of a shallow estuarine system back turbid lake waters and reduces suspended sediment concentration during the flood tide. Observed sediment data reveals that advection is the dominant process of transporting sediment within Coombabah Lake. Ali et al. (2009a) previously identified advection as the dominant process of transporting salinity within Coombabah Lake. Mitchell at al. (2008) also identified advection via tide as the dominating factor of solute transport in a similar case study in United Kingdom. In contrast to the water depth and salinity, mean turbidity at Station 1 was high during spring tide and low during neap tide (see Figure 5.3). The opposite trend was observed at Station 8, where low turbidity during spring tide and high turbidity during neap tides were observed. High tidal range during neap tide increased mixing and flushing capacities of the lake and therefore reduced the turbidity at Station 1 compared to the spring tide periods. Station 8 experienced high turbidity during neap tide because of high mixing with sediment laden catchment runoff.
2.0 (a) 2.0
1.0 1.0
Water depth (m) depth Water 0.0 0.0
1.0 (b) 1.0 0.5 0.5 0.0 0.0 -0.5 -0.5 Velocity (m/s) -1.0 -1.0
40 (c) 40
30 30
20 20 Salinity (psu) Salinity 10 10
600 600 ) (d) Station 1 Station 8
400 400
200 200 Turbidity (FTU Turbidity 0 0 1/11/05 2/11/05 3/11/05 4/11/05 5/11/05 6/11/05 7/11/05 8/11/05 9/11/05 10/11/05 11/11/05 Spring tide Date Neap tide
Figure 5.2: Measured hydraulic and water quality data: (a) water depth; (b) flow velocity; (c) salinity; and (d) turbidity.
Ayub Ali 110 Chapter V Sediment dynamics of a shallow estuarine system
Wind conditions during the field campaign were generally moderate but highly variable, ranging from 2 to 20 km/h with a mean wind speed of 7 km/h directed from the south or south-east (blowing from the creek mouth to the lake entrance) (Figures 5.4(a-b)), which is the dominant wind direction for coastal south-east Queensland. Light rainfall was recorded every day except on November 7, 2005. Rainfall increased gradually from 0.5 mm/day on November 1, 2005 to 18 mm/day (the maximum) on November 6, 2005. Air pressure also dropped by approximately five milibar (mbar) during this period. In general, waves were small with approximate height of 0.2 m and period of 2 seconds.
1.5 1.5 (a)
1.0 1.0
0.5 0.5 Avg. depth (m)
0.0 0.0
35 35 (b) 30 30
25 25
20 20 Avg. salinity (psu) 15 15
300 300 ) 250 (c) Station 1 Station 8 250 200 200 150 150 100 100 50 50 Avg. Turbidity (FTU Turbidity Avg. 0 0 1/11/05 2/11/05 3/11/05 4/11/05 5/11/05 6/11/05 7/11/05 8/11/05 9/11/05 10/11/05 11/11/05 Spring tide Date Neap tide
Figure 5.3: Tidally averaged hydraulic and water quality data: (a) water depth; (c) salinity; and (d) turbidity.
Observed TSS concentrations within the lake varied temporally and spatially. TSS concentration was high during the spring tide period and low during the neap tide period. TSS gradually reduced by 50% (from 0.30 kg/m3 to 0.15 kg/m3) within a week during the study period (Figure 5.5). There were several reasons behind the reduction of TSS concentration. High tidal range during neap tide increased mixing and flushing capacity of the lake; thus reduced the TSS concentration within the lake. In addition,
Ayub Ali 111 Chapter V Sediment dynamics of a shallow estuarine system high water depth during neap tide reduced the wave induced resuspension of bed sediments that also favoured the sediment settling process.
Approximately 0.05 kg/m3 differences in TSS concentration was observed between two samples (Figure 5.5) collected from the same sites within a short time period (e.g. minutes apart) presumably as a result of the variability of TSS concentrations within the dynamic environment and to a lesser extent possible uncertainties in sampling and measurement processes.
30 30 (a) 20 20
10 10
Wind speedWind (km/h) 0 0
360 360 (b)
180 180 Wind directionWind (°) 0 0
20 20 (c)
10 10 Rainfall (mm/day)Rainfall 0 0
1025 (d) 1025 1020 1020
1015 1015
1010 1010
Air pressure (mbar) pressure Air 1005 1005 1/11/05 2/11/05 3/11/05 4/11/05 5/11/05 6/11/05 7/11/05 8/11/05 9/11/05 10/11/05 11/11/05 Spring tide Date and time Neap tide
Figure 5.4: On-site meteorological conditions: (a) wind seed; (b) wind direction; (c) rainfall; and (d) air pressure.
Mean TSS concentrations over the study period at Stations 1, 2, 3, 4, 5, 6, 7 and 8 were 0.236, 0.230, 0.281, 0.215, 0.228, 0.261, 0.266 and 0.166 kg/m3 respectively. Mean TSS concentrations reveal that landward and shallow stations experienced higher TSS
Ayub Ali 112 Chapter V Sediment dynamics of a shallow estuarine system concentration compared with seaward and relatively deep stations. The highest TSS concentration was observed at Station 3. The lowest TSS concentration was observed at Station 8. Being located near the sediment source (creek mouth), Station 1 experienced slightly higher TSS concentration compared with other deeper stations. Increased TSS concentrations at the shallower stations resulted from sediment resuspension by waves and winds.
0.60 1.0 Observed TSS Observed tide level
0.40 0.5 ) 3 TSS (kg/m
0.20 0.0 Total water depth (m)
0.00 -0.5 1/11/05 2/11/05 3/11/05 4/11/05 5/11/05 6/11/05 7/11/05 8/11/05 9/11/05 10/11/05 11/11/05 Spring tide Date Neap tide
Figure 5.5: Observed TSS concentrations (two data almost at the same time represent two samples) and total water depth at Station 4.
The turbidity against TSS concentration (Figure 5.6) shows an increasing linear relationship with several outlying points. The correlation between TSS and turbidity during this study was very poor (R2 = 0.48). However, the correlation was quite good for concentrations <0.25 kg/m3. The local and meteorological (e.g. wind and rainfall) influences on correlations were also examined. In an attempt to improve TSS and turbidity correlations, data were grouped for stations and dates; however correlations at individual stations remained poor. Coefficient of determination values varied between stations (R2 = 0.30; Station 6 to 0.68; Station 3). Day-wise data revealed a poorer correlation compared to station-wise data. Therefore, location was presumed more influential on TSS and turbidity correlations than meteorological parameters in this
Ayub Ali 113 Chapter V Sediment dynamics of a shallow estuarine system region, particularly in shallow estuaries. There are many other factors that affect the correlation. Variations from the trend line are due to the highly individualistic nature of the water with different loadings of suspended material and varying ratios of organic and inorganic matter (Davies-Colley and Smith, 2001). These two parameters are also taken from different apparatus, so are susceptible to variance arising from analysing different samples and volumes of water (Obrador and Pretus, 2008). However, the data collected in this study is best described by an exponential curve (R2 = 0.66; Figure 5.6). The correlation was still poor compared to previous studies (e.g. Lenhart et al., 2009; Lewis et al., 2002). Moreover, in general the correlations were not improved at individual stations and days even using exponential curve presumably due to the small data set.
80
70
60
50
40
Turbidity (NTU) Turbidity 30 y = 89.43x R2 = 0.48 20
y = 3.95e6.11x 10 R2 = 0.66 0 0.00 0.10 0.20 0.30 0.40 0.50
TSS (kg/m3)
Figure 5.6: Correlation between TSS and turbidity measured at Stations 1-8 during the study period.
The correlation between TSS and turbidity was poor because of different geophysical characteristics of the suspended and/or dissolved materials which in turn depend on the
Ayub Ali 114 Chapter V Sediment dynamics of a shallow estuarine system various factors such as watershed geology, soil, vegetation and land uses (Lewis et al., 2002). Moreover, the lake is a complex mixing zone where catchment runoff and seawater mixes with different ratios at different times and locations within the lake. Therefore, more works on TSS and turbidity relationships are required to determine the dynamic control of the system. A large number of TSS samples and turbidity recordings from a single station in a short period of time could reduce the number of variables and improve the TSS-turbidity relation. Hydrodynamic and sediment data along with weather conditions for a longer period of time were also required to quantify the influence of environmental variables on sediment dynamics.
In summary, the automatic data logging system (turbidity meters) was not suitable for estimating time series of TSS concentrations in this very shallow subtropical estuarine system because of a large number of variables.
5.4.2 Model Calibration
The sediment transport model was calibrated against TSS concentration measured from November 1 to November 10, 2005. Critical bed shear stresses, dispersion coefficients and sediment boundary conditions were changed during the calibration processes. Through a calibration process, this study selected the value of critical bed shear stresses for erosion and that for deposition as 0.4 N/m2 and 0.1 N/m2, respectively. The above bed shear stress values are within the range observed in mud dominated estuarine systems (Verney et al., 2006; Araújo et al., 2008, Tolhurst et al., 1999). Horizontal and vertical dispersion parameters were also refined through calibration process. The horizontal dispersion coefficient was selected as constant over time and space with a value of 1 m2/s. The vertical eddy viscosity selected from the hydrodynamic model calibration was used as the vertical dispersion coefficient with a scaling factor of 1. Due to the lack of observed data, constant TSS concentrations, which were determined through trial and error method, were used at source point and open boundary. Generally, estuarine systems carry catchment runoff during ebb tides and seawater during flood tides. Moreover, highest TSS concentration was observed during ebb tide and lowest TSS concentration was observed during flood tide (Figure 5.7(b)) in a tidal period. The highest and the lowest observed TSS concentration was 0.4 kg/m3 and 0.1 kg/m3 at Station 8 (near lake entrance), respectively. Therefore, a TSS concentration of 0.5 kg/m3
Ayub Ali 115 Chapter V Sediment dynamics of a shallow estuarine system at the upstream source point and that of 0.1 kg/m3 at the downstream open boundary were used in the model which provided better results. The TSS concentration used at the upstream source point was higher than the observed TSS concentration because of the concentration gradient from seaward to landward.
0.60 0.6 (a) Observed Simulated
0.40 0.4 ) 3 TSS (kg/m TSS
0.20 0.2
0.00 0 1/11/05 2/11/05 3/11/05 4/11/05 5/11/05 6/11/05 7/11/05 8/11/05 9/11/05 10/11/05 11/11/05 Spring tide Neap tide Date
0.6 1.0 (b) Observed TSS Simulated TSS Tide level ) 0.4 0.5 ) 3 TSS (kg/mTSS 0.2 0.0 Tide level (mAHD level Tide
0.0 -0.5 4/11/05 9:00 4/11/05 12:00 4/11/05 15:00 4/11/05 18:00 4/11/05 21:00 5/11/05 0:00 Date and time
Figure 5.7: Comparison of simulated and observed TSS: (a) entire study period; and (b) one tidal cycle.
Simulated and observed TSS concentrations of the entire study period matched reasonably well (Figure 5.7(a)) compared to the complexity of the sediment transport model. The model produced excellent results during the neap tide period. However,
Ayub Ali 116 Chapter V Sediment dynamics of a shallow estuarine system simulated TSS concentrations were lower than those observed during the spring tide period. The marine borne non-cohesive sediment potentially have been transported into the lake by strong spring tidal current which has not been considered in this modelling exercise. There is a slight phase difference between simulated and observed TSS near the spring tide period (Figure 5.7(b)). However, the magnitudes were quite similar. Tidal variation of TSS concentration can be driven by resuspension (Uncles and Stephens, 1998) or advection (van Maren and Hoekstra, 2004) or both (Pritchard, 2005). Therefore, this study undertook a modelling investigation to determine the dominant processes of sediment transport in this shallow estuarine lake.
5.4.3 Simulation Results
The numerical model described previously was utilized to characterise the dynamics of cohesive sediments within the lake in order to assess the contribution of the main forcings; tides, runoff, waves and winds. Additionally, the sensitivity of bed roughness on the sediment transport was also investigated. Simulation results demonstrated that the TSS concentrations during the ebb tide period (Figures 5.8(b, d and f)) were much higher than that during the flood tide (Figures 5.8(a, c and e)). The model replicated the observed TSS distribution reasonably well. TSS concentrations mapped during the flood tide period showed high concentrations (0.45 kg/m3) at shallow landward stations and low concentrations (0.15 kg/m3) at relatively deeper seaward stations, which are in agreement with the field observations. TSS concentrations near the eastern and western edges of the lake were higher during high tides. Observed salinity was also low at these shallow areas during high tides. Therefore, these areas experience relatively low mixing with seawater and low flushing by tidal flow. On the other hand, seawater which carries low sediments mixes with sediment laden catchment runoff within the channels and finally flushes them out of the system during ebb tides.
Higher TSS concentrations moved seaward during ebb tide. The highest concentration zone was formed around the creek mouth and extended up to the middle zone of the lake (Figure 5.8(b)). Therefore, the region of highest suspended sediments, the turbidity maxima, could be formed within the creek further upstream of the lake which moves back and forth under tidal influence. Model results support that advection is the main
Ayub Ali 117 Chapter V Sediment dynamics of a shallow estuarine system process of sediment transport within Coombabah Lake which is driven by the tide regime. Model produced almost similar results for both spring and neap tidal conditions.
(a) (b)
(c) (d)
(e) (f)
Figure 5.8: Simulated TSS concentrations: (a, b) with no waves and winds; (c, d) with waves; and (e, f) with winds (left panel during flood tides and right panel during ebb tides).
Ayub Ali 118 Chapter V Sediment dynamics of a shallow estuarine system
Modelling results also revealed that the catchment runoff was the main source of TSS during this study period. A constant wave force of 0.2 m height and 2 s period directed from the south-east (observed within the lake) was used in the sediment transport model and it was found that only 5% of the total TSS was borne from the bed resulting from resuspension by waves during low tides (Figures 5.8(d) and 5.9). Though, wave action is important in the mixing and sediment transport processes of shallow water environments (Zhang et al., 2004), no influence of waves on TSS concentrations was observed during high tides within Coombabah Lake for the conditions encountered. The waves were small because of a limited fetch and their influence diminished before reaching the bed sediments during high tides. Resuspension of sediment by waves occurred during low tides only.
0.5
0.4 ) 3
0.3
0.2 TSS concentration (kg/m concentration TSS 0.1 Observed With wave No wave and wind With wind 0 1/11/05 2/11/05 3/11/05 4/11/05 5/11/05 6/11/05 7/11/05 8/11/05 9/11/05 10/11/05 11/11/05 Spring tide Neap tide Date
Figure 5.9: Simulated TSS concentrations at Station 4 without and with wave and wind activity included.
Similar to the waves, wind induced currents also increased TSS concentrations during low tides (Figures 5.8(f) and 5.9). The effect was higher during stronger wind and lower during weaker wind as previously reported by Dunn et al. (2007a). Spatial distribution of TSS concentrations during low tide changed quite significantly because of wind forces (Figure 5.8(f)). The highest TSS concentration zone was enlarged and extended
Ayub Ali 119 Chapter V Sediment dynamics of a shallow estuarine system almost up to the lake entrance by the south-easterly wind. On average, wind generated currents increased the TSS concentration by 7% within the lake system. However, TSS concentrations at Station 4 increased by approximately 25% during neap tide (Figure 5.9) because of the extension of the high turbidity zone pushed by the south-easterly wind. Although wind and wave have minor effect on TSS concentrations during high tide, inclusion of these forces provided improved and more realistic results. However, very strong wind influences the TSS concentration during both high and low tides. Doubling the wind force in the model increased the TSS concentration within the lake by approximately three times the observed typical concentrations.
0.5
0.4 ) 3
0.3
0.2 TSS concentrationTSS (kg/m
0.1
Observed Normal Increased by 50% Decreased by 50%
0 1/11/05 2/11/05 3/11/05 4/11/05 5/11/05 6/11/05 7/11/05 8/11/05 9/11/05 10/11/05 11/11/05 Spring tide Neap tide Date
Figure 5.10: Simulated TSS concentration for various bed roughness values.
Sensitivity of bed roughness to sediment concentration (Figure 5.10) revealed that the TSS concentration changed by 7% for a 50% change of bed roughness. Moreover, bed roughness did not influence TSS concentration during high tides (< 1%).
Although, total water flow was distributed by 60% and 40% between the main channel and left channel (see Figure 5.1) (Ali et al., 2009b), respectively, interestingly sediment discharge was distributed the other way. TSS was distributed by 43% and 57% between the main channel and left channel, respectively. TSS concentrations were always higher
Ayub Ali 120 Chapter V Sediment dynamics of a shallow estuarine system in the left channel than that in the main channel, which caused higher sediment discharges through the left channel than that through the main channel similar to another subtropical estuary in Bangladesh (Ali et al., 2007). Although flood tides have a higher sediment carrying capacity, in this instance it carries less sediment because the main sediment source is located upstream. However, the flood tide may potentially transport higher quantities of non-cohesive sediment which has not been considered in this modelling exercise.
5.5 Conclusions
Intra-tidal variations of sediment concentrations is relatively higher than sub-tidal variation within Coombabah Lake. Shallow landward parts of the lake experience low flushing and therefore retain high suspended sediment concentrations compared with the more well-defined central channels. In contrast to a general tendency, the lake experienced higher mixing as well as more flushing during neap tide than the spring tide period because of higher tidal range (i.e. higher energy) during the neap tide period. Catchment runoff was determined as the main source of cohesive sediment input for the lake system. During the study period, wind generated currents contributed 7% of the sediment resuspension within the lake. On the other hand, wave action contributed 5% of the sediment resuspension but was limited to periods of low tide. Waves were not important in this shallow lake system because of a limited fetch and young nature of the waves. Sediment dynamics of this subtropical estuarine lake system was dominated by advection processes driven by tides, rather than by local resuspension and weak mixing process.
A large number of TSS samples and turbidity recordings from a single station in a short period of time are recommended to reduce the number of variables in determining TSS- turbidity relationship. In contrast, hydrodynamic and sediment data along with weather conditions for a longer period of time are also recommended to quantify the influence of environmental variables on sediment dynamics.
Ayub Ali 121 Chapter V Sediment dynamics of a shallow estuarine system
5.6 Acknowledgments
The authors would like to acknowledge DHI Water and Environment, Denmark for their assistance making the MIKE modelling system available for this study. The authors would like to acknowledge the financial assistance of the Cooperative Research Centre for Coastal Zone, Estuary and Waterway Management. Acknowledgments are also made to Peta Williams, Nathan Benfer and the Griffith School of Engineering laboratory technicians (Griffith University, Gold Coast Campus) for their assistance with fieldwork.
5.7 References
Ali, A. and Lemckert, C.J., 2009. A traversing system to measure bottom boundary layer hydraulic properties, Estuarine Coastal and Shelf Science, 83 (4), 425-433. Ali, A., Mynett, A.E. and Azam, M.H., 2007. Sediment Dynamics in the Meghna Estuary, Bangladesh: A Model Study. Journal of Waterway, Port, Coastal and Ocean Engineering, 133 (4), 255-263. Ali, A., Lemckert, C.J. and Dunn, R.J.K., 2009a. Salt fluxes in a very shallow sub- tropical estuary. Journal of Coastal Research. doi: 10.2112/08-1118.1. Ali, A., Zhang, H. and Lemckert, C.J., 2009b. Numerical Study of the Hydrodynamics of a Very Shallow Estuarine System - Coombabah Lake, Gold Coast, Australia. Journal of Coastal Research, SI 56, 922-926. Alongi, D.M., 1998. Coastal Ecosystem Processes. CRC Press, Boca Raton, Florida, USA, 420pp. Araújo, M.A.V.C., Teixeira, J.C.F. and Teixeira, S.F.C.F., 2008. Application of laser anemometry for measuring critical bed shear stress of sediment core samples. Continental Shelf Research, 28, 2718–2724. Balls, P.W., 1994. Nutrient inputs to estuaries from nine Scottish east coast rivers; influence of estuarine processes on inputs to the North Sea. Estuarine, Coastal and Shelf Science, 39, 329-352. Benfer, N.P., King, B.A. and Lemckert, C.J., 2007. Salinity observations in a subtropical estuarine system on the Gold Coast, Australia. Journal of Coastal Research, SI 50, 646-651.
Ayub Ali 122 Chapter V Sediment dynamics of a shallow estuarine system
Black, K.S., Tolhurst, T.J., Paterson, D.M. and Hagerthey, S.E., 2002. Working with natural cohesive sediments. Journal of Hydraulic Engineering, 128 (1), 2-8. Booth, J.G., Miller, R. L., McKee, B.A., and Leathers, R.A., 2000. Wind-induced bottom sediment resuspension in a microtidal coastal environment. Continental Shelf Research, 20, 785-806. Burton, E.D., Sullivan, L.A., Bush, R.T., Powell, B., 2008. Iron-sulfide and trace element behaviour in sediments of Coombabah Lake, southern Moreton Bay (Australia). Marine Pollution Bulletin, 56, 1353-1376. Butt, T. and Russell, P., 2000. Hydrodynamics and cross-shore sediment transport in the swash-zone of natural beaches: a review. Journal of Coastal Research, 16 (2), 255- 268. Conley, D.C. and Griffin, J.G., 2004. Direct measurements of bed stress under swash in the field. Journal of Geophysical Research-Oceans, 109 (C3) (art. no.-C03050). Davies, A.G., 1985. Field observations of the threshold of sediment motion by wave action. Sedimentology (Oxford), 32 (5), 685-704. Davies-Colley, R.J. and Smith, D.G., 2001. Turbidity, Suspended Solids and Water Clarity: A Review. Journal of the American Water Resources Association, 37 (5), 1085-1101. Day, J. W., Hall, C., Kemp, W. M. and Yanez-Arancibia, A., 1989. Estuarine Ecology. John Wiley and Sons, New York, USA, 558pp. DHI (Danish Hydraulic Institute), 2008a. Scientific Documentation, MIKE 21 & MIKE 3 Flow Model FM, Hydrodynamic and Transport Module. DHI Water and Environment, Horsholm, Denmark, 52pp. DHI (Danish Hydraulic Institute), 2008b. User Guide, MIKE 3 Flow Model, Mud Transport Module. DHI Water and Environment, Horsholm, Denmark, 88pp. Dunn, R.J.K., Ali, A., Lemckert, C.J., Teasdale, P.R., and Welsh, D.T., 2007a. Short- term variability of physio-chemical parameters and the estimated transport of filterable nutrients and chlorophyll-a in the urbanised Coombabah Lake and Coombabah Creek system, southern Moreton Bay, Australia. Journal of Coastal Research, SI 50, 1062-1068. Dunn, R.J.K., Lemckert, C.J., Teasdale, P.R., and Welsh, D.T., 2007b. Distribution of nutrients in surface and sub-surface sediments of Coombabah Lake, southern
Ayub Ali 123 Chapter V Sediment dynamics of a shallow estuarine system
Moreton Bay (Australia). Marine Pollution Bulletin, 54, 606-614. Dunn, R.J.K., Welsh, D.T., Jordan, M.A., Teasdale, P.R., Lemckert, C.J., 2009. Influence of natural amphipod (Victoriopisa australiensis) (Chilton, 1923) population densities on benthic metabolism, nutrient fluxes, denitrification and DNRA in sub-tropical estuarine sediment. Hydrobiologia, 628, 95-109. Dunn, R.J.K., Welsh, D.T., Teasdale, P.R., Lee, S.Y., Lemckert, C.J., and Meziane, T., 2008. Investigating the distribution and sources of organic matter in surface sediment of Coombabah Lake (Australia) using elemental, isotopic and fatty acid biomarkers. Continental Shelf Research, 28 (18), 2535-2549. Eyre, B., 1998. Transport, Retention and Transformation of Material in Australian Estuaries. Estuaries, 21 (4A), 540-551. Frank, T.D. and Fielding, C.R., 2004. Sedimentology and geochemistry of an urban coastal lake system: Coombabah Lake Nature Reserve, Gold Coast, Queensland. Australian Journal of Earth Sciences, 51, 261-271. GHD (Gutteridge, Haskins and Davey Pty Ltd), 2003. Coombabah Creek Environmental Inventory. Gutteridge, Haskins and Davey Pty Ltd, Brisbane, Australia, 439p. Hollingsworth, A. and Connolly, R.M., 2006. Feeding by fish visiting inundated subtropical saltmarsh. Journal of Experimental Marine Biology and Ecology, 336, 88-98. Klen, T., 2006. Estuaries, An Introduction to Marine Biology and Oceanography,
Ayub Ali 124 Chapter V Sediment dynamics of a shallow estuarine system
GTR-184, 107-118. Lopes, J.F., Dias, J.M. and Dekeyser, I., 2006. Numerical modeling of cohesive sediments transport in the Ria de Aveiro lagoon, Portugal. Journal of Hydrology, 319, 176-198. Mitchell, S.B., Burgess, H.M., Pope, D.J. and Theodoridou, A. 2008. Field studies of velocity, salinity and suspended solids concentration in a shallow tidal channel near tidal flap gates. Estuarine, Coastal and Shelf Science, 78, 385-395. Morris, A.W. and Howarth, M.J., 1998. Bed stress induced sediment resuspension. Continental Shelf Research, 18, 1203-1213. Nikora, V., Goring, D. and Ross, A., 2002. The structure and dynamics of the thin near- bed layer in a complex marine environment: a case study in Beatrix Bay, New Zealand. Estuarine, Coastal and Shelf Science, 54 (5), 915-926. Norkko, J., Hewitt, J.E. and Thrush, S.F., 2006. Effects of increased sedimentation on the physiology of two estuarine soft-sediment bivalves, Austrovenus stutchburyi and Paphies australis. Journal of Experimental Marine Biology and Ecology, 333, 12-26. Obrador, B. and Pretus, J.L., 2008. Light regime and components of turbidity in a Mediterranean coastal lagoon. Estuarine, Coastal and Shelf Science, 77, 123-133. Pritchard, D., 2005. Suspended sediment transport along an idealised tidal embayment: settling lag, residual transport and the interpretation of tidal signals. Ocean Dynamics, 55, 124-136. Ralston, D.K. and Stacey, M.T., 2007. Tidal and meteorological forcing of sediment transport in tributary mudflat channels. Continental Shelf Research, 27, 1510- 1527. Rodi, W., 1984. Examples of Calculation Methods for Flow and Mixing in Stratified Fluids. Journal of Geophysical Research, 92 (C5), 5305-5328. Soulsby, R.L., Atkins, R. and Salkield, A.P., 1994. Observations of the turbulent structure of a suspension of sand in a tidal current. Continental Shelf Research 14 (4), 429-435. Tolhurst, T.J., Black, K.S., Shayler, S.A., Mather, S., Black, I., Baker, K. and Paterson, D.M., 1999. Measuring the in situ Erosion Shear Stress of Intertidal sediments with the Cohesive Strength Meter (CSM). Estuarine, Coastal and Shelf Science,
Ayub Ali 125 Chapter V Sediment dynamics of a shallow estuarine system
49, 281-294. Uncles, R.J. and Stephens, J.A., 1998. Sediment transport in the Humber-Ouse Estuary, UK, during May 1994. In Proceedings of the 8th International Biennial Conference on Physics of Estuaries and Coastal Seas (Physics of Estuaries and Coastal Seas edited by Dronkers & Scheffers), Balkema, Rotterdam, The Netherlands, pp. 73-82. van Maren, D.S. and Hoekstra, P., 2004. Seasonal variation of hydrodynamics and sediment dynamics in a shallow subtropical estuary: the Ba Lat River, Vietnam. Estuarine, Coastal and Shelf Science, 60, 529-540. Verney, R., Brun-Cottan, J.-C., Lafite, R., Deloffre, J. and Taylor, J. A., 2006. Tidally- induced Shear Stress Variability above Intertidal Mudflats in the Macrotidal Seine Estuary. Estuaries and Coasts, 29(4), 653–664. Whitehouse, R., Soulsby, R.L., Roberts, W. and Mitchener, H., 2000. Dynamics of estuarine muds, Thomas Telford, London, 210pp. Williams, J.J., Rose, C.P., Thorne, P.D., O’Connor, B.A., Humphery, J.D., Hardcastle, P.J., Moores, S.P., Cooke, J.A. and Wilson, D.J., 1999. Field observations and predictions of bed shear stresses and vertical suspended sediment concentration profiles in wave–current conditions. Continental Shelf Research, 19 (4), 507-536. Wood, P.J. and Armitage, P.D., 1997. Biological effects of fine sediment in the lotic environment. Environmental Management, 21, 203-217. Yamada, H. and Nakamura, F., 2002. Effect of fine sediment deposition and channel works on periphyton biomass in the Makomanai River, Northern Japan. River Research and Applications, 18, 481-493. Zhang, H., Madsen, O.S., Sannasiraj, S.A. and Chan, E.S., 2004. Hydrodynamic model with wave-current interaction in coastal regions. Estuarine, Coastal and Shelf Science, 61, 317-324.
Ayub Ali 126 Chapter VI Conclusion and recommendations
CHAPTER VI
Conclusions and recommendations
This study quantified water circulation and mixing processes within a very shallow subtropical estuarine system Coombabah Lake in Gold Coast, Queensland, Australia from measured flow velocity and salinity data. Residual water transport within the shallow waters of Coombabah Lake was identified as the dominant factor influencing residual salt transport. Furthermore, this study indicates the net salt flux alternates frequently in contrast to the steady state condition. Additionally, this study has identified that advective flux was of primary importance in the movement of salts within Coombabah Lake (contributing 65% of the total salt flux in this shallow subtropical estuary) and that the lake was characterised by greater average water depth during neap tide phases, which aids in the diffusion of salts, with a 25% dilution occurring within the lake system. The lake was not an inverse type of estuary during this study because of freshwater input.
This study developed and tested a new underwater traversing system that made estimation of bed shear stress and roughness height robust. The newly developed traverser also made best use of all available techniques of bed shear stress estimation at the same time. However, the system was best suitable for unstratified, steady and shallow environments. The Log-Profile method was found to be the easiest and most useful method for estimating bed shear stress within shallow estuaries and rivers, followed by the Inertial Dissipation, Turbulent Kinetic Energy and Reynolds Stress methods. More importantly, the Log-Profile method estimates both bed shear stress and roughness height, both essential parameters for sediment (or pollutant) transport modelling at the same time, whereas the other three methods estimate only bed shear stress. Moreover, the other three methods require precise velocity measurement within the constant stress layer (within centimetres) near the bottom to determine the bed shear
Ayub Ali 127 Chapter VI Conclusions and recommendations stress. Mean velocity (after filtering noise) within the weak spot appeared reasonably accurate, and therefore was used in constructing the velocity profile. However, the same data could not be used for calculating turbulent shear stress due to the noise.
Both observed and simulated velocity supports the classical tidal pumping of sediments landward within Coombabah Lake. The hydrodynamic regime of the lake is also conducive for the catchment generated suspended sediments to settle within the lake. This study found an accurate bathymetry is the most important parameter for modelling hydrodynamics particularly within shallow estuarine systems. Bed roughness was found as the second most important parameter in the modelling exercise for this region.
Intra-tidal variation of sediment concentrations were found relatively higher than sub- tidal variation within Coombabah Lake. Shallow landward parts of the lake experience low mixing and tidal flushing, and therefore retain high sediment concentrations compared with the channels. Catchment runoff was determined as the main source of cohesive sediment for the lake system. Wind generated currents contributed 7% of the sediment resuspension within the lake. On the other hand, wave action contributed 5% of the sediment resuspension but was limited to periods of low tide. Waves were not important in this shallow lake system because of a limited fetch and young nature. Sediment transport of this subtropical estuarine lake system is dominated by advection processes driven by tides, rather than by local resuspension and weak mixing process.
This study recommends some improvements to the newly developed traversing system. This study also recommends further investigations on Coombabah Lake-Creek system for simulating the impact of land use changes within the catchment on the aquatic ecosystem health. The recommendations are summarised below:
Automation of the traversing system incorporating Procedural Language Code (PLC) programming.
Reduction of the size of the traversing system to improve it’s portability.
An intensive TSS sampling and turbidity recordings from a single station to reduce the number of variables in determining TSS-turbidity relationship.
Ayub Ali 128 Chapter VI Conclusions and recommendations
Continuous investigation of salt fluxes for a long period to include seasonal variations at any particular location within the lake.
Investigation of the variations of characteristic sediment transport parameters within shallow estuarine systems.
Preparation of maps of a number of systems as a database for future modelling.
Modelling pollutant/nutrient export from Coombabah Lake-Creek catchment.
Modelling water quality of Coombabah Lake-Creek system.
Ayub Ali 129 Chapter VI Conclusions and recommendations
Ayub Ali 130 Appendix A Short-term variability of physio-chemical parameters
APPENDIX A
Short-term Variability of Physio-chemical Parameters and the Estimated Transport of Filterable Nutrients and Chlorophyll-a in the Urbanised Coombabah Lake and Coombabah Creek System, Southern Moreton Bay, Australia
Abstract
Extensive urban development has occurred within the Coombabah Creek catchment and surrounds over the last two decades, resulting in concerns about degrading water quality. Water quality was investigated three times during summer and three times during autumn/winter at two sites: within Coombabah Lake and at several km downstream within Coombabah Creek. Physio-chemical parameters, suspended solids, 3- - - - chlorophyll-a and filterable nutrients (reactive PO4 , NO2 , NO3 and NH3 ) were measured hourly for 13 hours in order to compare the water quality under different tidal conditions at each site. Bathymetric and hydrological data were also collected, which allowed short-term nutrient loads to be estimated. From these measurements, the net transport of filterable nutrients and chlorophyll-a into Coombabah Lake and the intratidal variability of Coombabah Creek and Lake waters were determined. Physio- chemical parameters, suspended solids and chlorophyll-a concentrations demonstrated
Dunn, R.J.K., Ali, A., Lemckert, C.J., Teasdale, P.R., and Welsh, D.T., 2007. Short-term variability of
physio-chemical parameters and the estimated transport of filterable nutrients and chlorophyll-a in the
urbanised Coombabah Lake and Coombabah Creek system, southern Moreton Bay, Australia. Journal of
Coastal Research, SI 50, 1099-1105
Ayub Ali A-1 Appendix A Short-term variability of physio-chemical parameters characteristic cyclic variations with the influence of tidal and diurnal cycles apparent. Despite elevated nutrient concentrations, chlorophyll-a values indicated an oligotrophic- mesotrophic environment, with concentrations ranging between 0.11-3.53 μg L-1. Maximum concentrations occurred during ebb tides, coinciding with periods of increased solar radiation. Elevated total suspended solids were observed during increased current velocities during low tide periods and greater wind speeds within the lake. Filterable nutrient concentrations and transport estimations also demonstrated tidal influences, with increased concentrations observed during sampled high tide phases, indicating increased inputs of nutrients originating from sources other than the creek 3- - and lake. Observed reactive PO4 , NOx and NH3 concentrations exceeded Broadwater sub-region values within the Queensland Water Quality Guidelines (2006)
A.1 Introduction
Coastal wetlands and estuaries are spatially diverse systems providing important habitats for flora and fauna including migratory and indigenous fish and bird species in addition to often supporting commercial and recreational fisheries (Stumpf and Haines, 1998). They also act as filters for contaminants and sediments helping to moderate water quality (Faulkner, 2004), absorb wave energy and provide cultural and recreational benefits (Lee et al., 2006). However, coastal wetlands and estuaries are under pressure from rapidly-increasing urban populations in coastal areas (Pauchard et al., 2006). Continued population growth within coastal regions ensures there will be ongoing impacts on coastal wetland ecosystems (Lee et al., 2006). Within coastal waters, the primary point of entry for nutrients is from terrestrial sources (Pereira-Filho et al., 2001) and for this reason, high primary productivity and biological abundance are often observed in these regions (Alongi, 1998). In densely populated urban regions nutrient supply is greater due to the entry of both domestic and industrial waste and urban drainage (Lee et al., 2006). The increase of nutrient concentrations within coastal waters can elicit either positive or negative responses in the ecological health of systems, including the alteration of species richness and abundance (Faulkner, 2004); productivity (Nixon, 1992); and fishing yields (Cederwall and Elmgren, 1980). Anthropogenic inputs of nutrients may lead to excessive eutrophication, especially where the circulation is restricted, such as in bays and coastal lagoons (Lin et al., 2006).
Ayub Ali A-2 Appendix A Short-term variability of physio-chemical parameters
Alterations in chemical characteristics and water quality within coastal systems occurs due to changes in biogeochemical flows (Pereira-Filho et al., 2001), as a consequence of modified land uses resulting in various ecological consequences (Lee et al., 2006). Investigations have previously assessed environmental changes caused by anthropogenic inputs of nutrients and organic material (Carmouze and Vasconcelos, 1992) and time- and tide-series observations of physio-chemical, nutrient and sediment parameters (e.g. Dittmar and Lara, 2001) in coastal systems. However, fewer studies (e.g. Pereira-Filho et al.; (2006)) have quantified these entries. Australia’s eastern seaboard is constantly changing from demands by populace for growth, coupled to the economics of increased tourism and development, where ~84% of the population lives within the coastal region (Lee et al., 2006). One such region is the Coombabah Lake- Creek system in southern Moreton Bay, south-east Queensland, Australia, one of the fastest-growing regions in the developed world (Skinner et al., 1998). The catchment and local surroundings have undergone rapid urban expansion, including waterfront and golf course developments. Coombabah Lake is the largest estuarine lake in southern Moreton Bay covering ~2 km2 (Gutteridge et al., 2003). Despite its modest dimensions, Coombabah Lake is ecologically important as it is a valuable fish (Queensland Fisheries Act 1994) and migratory bird habitat (Chinese-Australia Migratory Bird Agreement (1974) and Japan-Australia Migratory Bird Agreement (1986)). In addition to being ecologically significant, the lake system is unique within southern Moreton Bay as it behaves as an inverse estuarine lake during summer periods (Benfer, pers. comm.). The objectives of this study were to observe the intratidal variability of water quality parameters (WQP); total (TSS), organic (OSS) and mineral (MSS) suspended solids; 3- - - - filterable nutrients (reactive PO4 , NO2 , NO3 and NH3 ); and chlorophyll-a (chl-a) concentrations in the shallow urbanised Coombabah Creek-Lake waters and estimate the transport of filterable nutrients and chl-a at the Coombabah Lake entrance. This study provides the first account of physical, biological and chemical intratidal variability and estimates of filterable nutrient and chl-a transportation within the Coombabah Creek-Lake system. Such information permits an initial understanding of the system’s behaviour that will assist future management decisions in this ecologically and economically important region.
Ayub Ali A-3 Appendix A Short-term variability of physio-chemical parameters
A.2 Methodology
A.2.1 Site description
Coombabah Creek (6914320 m N, 536570 m E; Figure A.1) is an urbanised moderately impacted (Cox and Moss, 1999) sub-tropical tidal creek. The creek is ~17 km in length, and originates from the Nerang State Forest. The catchment is urbanised with residential, commercial and industrial development and has an area of 44 km2. The creek flows through Coombabah Lake (6912750 m N, 534400 m E; Figure A.1), a shallow body of water characterised by fine sediments located in the mid-tidal region of the creek, with urban development positioned to the east and along the southern and western shorelines. Coombabah Creek ultimately discharges into the Gold Coast Broadwater, a vitally important coastal system both economically and recreationally within southern Moreton Bay. The depth of the creek ranges from ~3 to <1 m, and the width varies between ~90 to ~200 m. With the exception of shallow channels, the lake is characterised by a relatively flat bathymetry (Lee et al., 2006). Depth within the lake ranges from typically <0 to ~1 m relative to mean water level at low tide, with large portions of the lake becoming exposed during this period. Episodically large inputs of freshwater occur during periods of heavy rainfall, predominantly occurring during summer periods.
Ayub Ali A-4 Appendix A Short-term variability of physio-chemical parameters
Figure A.1: Study area of Coombabah Lake and Coombabah Creek in southern Moreton Bay, Australia (inset).
A.2.2 Experimental design
Physio-chemical parameters were monitored in situ and water samples were collected at Coombabah Lake and Coombabah Creek during summer (November, 2005) and autumn-winter sample periods (May-June, 2006). Three sampling events occurred at each site during both seasonal sample periods. Sample collection and in situ measurements were conducted at 1 hour intervals for typically 13 hours at both sample locations, capturing ebb and flood tides. Sampling commenced during the morning high tide and incorporated both neap and spring tide periods. Hydrological parameters,
Ayub Ali A-5 Appendix A Short-term variability of physio-chemical parameters conductivity, temperature and turbidity were measured continuously during selected summer lake sampling periods. Hydrological data and water samples were collected at the creek and lake simultaneously during 4/11/05 and 11/11/05. Lake samples were collected in the main channel at the north-eastern entrance (6913140 m N, 535070 m E; Figure A.1). This site is ~160 m in width and is the only connection between the catchment and creek waters through the lake during low tide. Lake sediments have been characterised in a previous study (Lee et al., 2006). A time lag of three hours (Lee et al., 2006) was used to synchronise tidal observations between high tides occurring at the Gold Coast tidal station and Coombabah Lake-Creek system. An assessment of sample depth was also conducted at the lake entrance during summer sample periods. Water quality parameters, chl-a, suspended solids and nutrient concentrations were compared with the Broadwater sub-region values within the Queensland Water Quality Guidelines (QWQG) (2006). Statistical significance was measured at α = 0.05.
A.2.3 Water sample collection and analysis
Surface water samples for the determination of TSS, OSS, MSS, chl-a and filterable nutrient concentrations were collected ~0.30 m below the water surface using a pole sampler and an acid-washed (10% v/v HCl), sample-rinsed 500 ml low density polyethylene (LDPE) sample bottle (Nalgene). Additional samples from varying depths were also collected at the lake using an adjustable custom-built multi-port sampler. Samples were collected at heights of ~0.06, 0.30 and 1.2 m above the surface sediments through acid-washed tubing and collected by vacuum into acid-washed, sample-rinsed 500 ml LDPE bottles. Furthermore, nutrient and chl-a samples were also collected from cross-channel surveys at the lake entrance sample site. These cross-channel sites were nearly equally spaced across, and were used to detect spatial variability of analyte concentrations passing through the sites at any given time. If the concentrations from each of the cross-channel sites were found to be equivalent, then a single mid-channel sample site could be used as a representative sample of all waters passing through the sampled lake site. This cross-channel survey was conducted twice (4/11/05 and 11/11/05). Nutrient samples were immediately filtered through pre-washed, pre-ashed GF/F membranes (25 mm Ø, Millipore) and transferred into 10 ml sterilised polystyrene sample-rinsed tubes. Samples were stored frozen (-20 oC), typically one hour after
Ayub Ali A-6 Appendix A Short-term variability of physio-chemical parameters collection awaiting analysis. Filterable nutrient concentrations were determined by an automated nutrient analyser (Easychem Plus Random Access analyser; Systea Analytical Technologies). Natural filtered seawater references produced by the National Low Level Nutrient Collaborative Trials were used as quality assurance. Recoveries were good – averaging 93.2% for all filterable nutrients from the filtered seawater certified references. Suspended sediment concentrations were determined gravimetrically. TSS samples were filtered through pre-washed, pre-ashed GF/F membranes (47 mm Ø, Millipore), before membranes were air-dried in a desiccator to constant weight (Balls, 1994). Membranes were afterwards ashed (550 oC, 2 h) for the determination of OSS as loss of weight. MSS were determined as the difference between TSS and OSS. Chl-a samples were filtered through pre-washed, pre-ashed GF/C membranes (25 mm Ø, Millipore) immediately after collection. Membranes were then stored frozen in foil-wrapped glass vials. Chl-a concentrations were determined spectrophotomtrically (665, 750 nm) after acetone extraction and calculations according to Lorenzen (1967).
A.2.4 In situ water quality parameters
Physio-chemical data were collected at the surface water collection depth. Water temperature (oC), pH, dissolved oxygen (mg L-1), salinity and reduction-oxidation (redox) potential (mV) were recorded in situ with a multi-probe analyser (TPS 90- FLMV; TPS Pty. Ltd.), calibrated daily. Turbidity (NTU) was also recorded in situ using a nephelometric turbidity meter (Analite 160; McVan Instruments).
A.2.5 Field instrumentation
During an extended summer sampling period (31/10/05 to 12/11/05) a submersible sensor base was positioned at the lake entrance sample site consisting of: 1) conductivity, temperature, depth (CTD) and turbidity gauge (NIXIC-CTD-ADC; Falmouth Scientific, Inc.); 2) 3-dimensional open water current meter, Acoustic Doppler Current Profiler (ADCP) (Aquadopp® side-seeing profiler 2 MHz; Nortek AS); 3) high resolution 3-dimensional acoustic Doppler velocimeter (ADV) (Vector fixed stem velocimeter; Nortek AS); and 4) a submersible tide gauge (XR-420-TG; Richard Brancker Research Ltd.). CTD and turbidity high frequency data were collected
Ayub Ali A-7 Appendix A Short-term variability of physio-chemical parameters using time-averaged data (3.5 minute bursts at a frequency of 10 Hz) obtained every 15 minutes. ADV data collection were obtained using time-averaged data, with burst intervals of 30 minutes at a frequency of 32 Hz and 4096 samples per burst. Submersible tide gauge data were acquired continuously at 1 Hz. Velocity profiles were also undertaken along two Coombabah Creek transects (located approximately 1) 6915290 m N, 537450 m E and 2) 6914360 m N, 536800 m E), using a vessel-mounted ADCP (Workhorse Monitor; RD Instruments) during 4/11/05 and 11/11/05. Water sample collections and in situ WQP measurements were made along these transects. Transects were measured over consecutive ebb-flood tide phases.
A.2.6 Transport estimations
Transport estimations were determined during summer sample periods at the lake entrance. Cross-channel surveys indicated that concentrations were very similar, permitting the use of a single sample point. Calculations of water volume transport were obtained from current velocity and depth measurements. The area of the entrance cross- section was determined during a bathymetric survey described in Lee et al., (2006). Mass transport values were derived from measurements of water flow and the depth- averaged analyte concentration using the formula:
fi qi ci (A.1) where f, q and c are instantaneous flux, flow and concentration values, respectively. For instantaneous flux calculations, concentrations were converted from µg l-1 to g m-3, then multiplied by the instantaneous water volume transport (m3 s-1) at that time to find flux in g s-1. Nutrient concentrations measured along the depth profile were used to provide depth-average concentrations. During this study positive (+) and negative (-) values represent transport out of and into the lake, respectively.
A.2.7 Meteorological measurements
Meteorological conditions for sampling periods were obtained from a weather station (Australian Bureau of Meteorology station 040764, Gold Coast Seaway) positioned ~7 km south-east of Coombabah Lake. Conditions during the 4/11/05 and 11/11/05 were
Ayub Ali A-8 Appendix A Short-term variability of physio-chemical parameters recorded using a data logging weather station (WeatherMaster 2000; Environdata) positioned within the lake.
Figure A.2: Water depth (left) and sample velocity profiles (right) of the sampled Coombabah Lake entrance channel.
A.3 Results and discussion
A.3.1 Hydrological data
Observed tidal levels at the lake entrance displayed mixed tides with a maximum range of 1.00 m during the spring tide and 0.68 m during neap tide periods (Figure A.2). Tidal asymmetry was observed with ~5.5 h rising period and ~7.5 h falling period, indicating potential pumping of sediments from downstream Coombabah Creek and other external southern Moreton Bay sources. The lake entrance maximum depth-average velocity observed during flood and ebb tide periods was 0.80 m s-1 and 0.65 m s-1, respectively (Figure A.2). Observed lake entrance peak flows were 58.3 m3 s-1 and -57.2 m3 s-1. Additionally, ADCP transect data revealed peak flows of 164 m3 s-1 and -142 m3 s-1 during ebb and flood tide phases at the Coombabah Creek transect sites during the summer sample periods. Higher maximum and lower minimum peak flows within the creek were observed at the downstream transect site compared with the upstream transect location. Seasonal variability of flow volumes with respect to the magnitude of mean flow is to be expected within the lake-creek system. This current study does not
Ayub Ali A-9 Appendix A Short-term variability of physio-chemical parameters take into account such seasonal events, including periods of heavy rainfall, in which case flow values would be expected to increase significantly.
A.3.2 Variability within Coombabah Lake and Creek
Physio-chemical water quality parameters
Surface WQPs measured within the lake-creek system showed characteristic cyclic variations with the influence of tidal and diurnal cycles apparent. Variability ranged between 0.38-151% relative standard deviation (% rsd) (Table A.1). Due to the range of most riverine (~7-7.6) and ocean (~8.2) waters, pH values were characterised by the lowest variability. Observed pH and salinity values were similar within the lake-creek system. Mean pH values complied with the Broadwater values within the QWQGs of 8.0-8.4. Water temperature varied due to time of day, water column depth and season, with values ranging from 25.6-29.6 oC during the summer and 15.3-19.6 oC during the autumn-winter sample periods. Maximum temperatures within the lake-creek waters occurred between 1500-1600 h. Dissolved oxygen (DO) concentrations were significantly greater within the creek compared with the lake waters (p <0.001, df = 162) (Table A.1), possibly being due to a higher oxygen demand within the shallower lake waters. Low DO levels may be explained by high concentrations of organic matter within the lake sediments, as oxygen is consumed during the consumption of organic matter. DO concentrations exhibited seasonality within the creek with significantly greater concentrations observed during the summer sampling events (p = 0.006, df = 78). DO concentrations were significantly greater (p = 0.003, df = 162) during daylight hours compared with night hours, presumably due to the net production of primary producers during hours of solar radiation. DO saturation values within the lake ranged from 53-79%, which is below the lower Broadwater QWQG values. In comparison, the creek DO water saturation values ranged from 62-105%. Mean creek values during the summer (103%) and autumn-winter (80%) periods were comparable with the Broadwater QWQG values of 90-105%. These results support previous findings of Cox and Moss (1999).
Ayub Ali A-10 Appendix A Short-term variability of physio-chemical parameters
Table A.1: Minimum (min), maximum (max), mean, standard deviation (sd), and percent relative standard deviation of physio-chemical parameters, chl-a and suspended solids within the Coombabah Lake-Creek system (n = 42). Lake Creek Summer min max mean sd % rsd min max mean sd % rsd pH 7.97 8.09 8.04 0.03 0.38 7.84 8.57 8.10 0.14 2.60 salinity 29.5 36.8 33.2 2.11 6.37 30.5 36.2 34.2 1.68 4.93 redox 42.0 285 182 53.4 29.3 -29.0 158 27.4 42.6 151 temperature 25.6 29.1 27.4 1.04 3.53 26.2 29.6 27.9 0.83 2.72 DO 4.52 6.06 5.46 0.34 6.14 5.94 8.18 6.84 0.49 7.17 chl-a 0.11 3.53 1.48 0.94 63.2 0.11 3.32 1.60 0.92 56.3 turbidity 6.25 36.0 16.53 7.90 47.8 - - - - - TSS 93.2 520 227 112 50.4 158 437 286 63.9 22.4 Autumn-Winter min max mean sd % rsd min max mean sd % rsd pH 8.00 8.37 8.12 0.09 1.09 8.01 8.24 8.11 0.07 0.83 salinity 29.6 35.5 33.5 1.86 5.55 30.3 36.2 34.1 2.01 5.88 redox 34.0 280 170 49.6 29.8 -26.0 162 51.3 37.4 74.6 temperature 15.4 19.6 17.8 1.33 17.1 15.3 19.4 17.7 1.24 6.82 DO 5.01 6.12 5.54 0.34 6.12 5.90 7.03 6.58 0.33 4.97 chl-a 0.11 2.14 1.07 0.59 55.0 0.11 2.99 1.05 0.74 70.4 turbidity 6.27 34.38 14.2 6.97 49.0 - - - - - TSS 93.8 513 212 104 50.4 27.2 240 127 51.5 40.1
Ayub Ali A-11 Appendix A Short-term variability of physio-chemical parameters
Chlorophyll-a
Chl-a concentrations within the lake-creek system were comparable with concentrations previously reported in southern Moreton Bay (Abal and Dennison, 1996; Moss and Cox, 1999). Mean concentrations were below the Broadwater QWQG value (2.5 µg L- 1), and suggested an oligotrophic-mesotrophic system based on chl-a concentrations (Quinn, 1991). Mean chl-a concentrations within the lake and creek waters were not significantly different, nor was there any significant difference in concentrations within the lake and creek waters during neap and spring-tide sampling events. Increased and maximum chl-a concentrations were observed during morning ebb tides, occurring during periods of increased solar radiation. Significantly greater concentrations were observed during daylight hours (p< 0.001, df = 156) compared with samples collected at night-time. Benthic algae associated with mangrove vegetation on the margins of the lake and upper Coombabah Creek flushed during the ebb tide may have contributed to the increased concentrations. Furthermore, the summer period – characterised by longer daylight hours and greater water temperatures – demonstrated significantly greater chl-a concentrations within the lake and creek waters compared with the autumn-winter concentrations (p = 0.039, df = 156), suggesting greater phytoplankton biomass during the summer period.
Suspended solids
Knowledge of suspended sediment dynamics is essential for quantifying fluxes of substances and determining the fate of pollutants (Regnier and Wollast, 1993). Bottom water lake entrance turbidity values ranged from 13.67-403.69 NTU during the summer period and surface waters ranged from 6.25-36.0 NTU, with an average of 15.4±7.50 NTU. Mean turbidity values exceeded the Broadwater QWQG value (6 NTU). Shallow estuarine systems tend to have periodically high turbidity values due to tide- and wind- induced resuspension of bottom sediments and, as such, turbidity is not always a useful health indicator of such systems. Lake turbidity and TSS values significantly correlated (r = 0.906, p <0.001, n = 83). TSS concentrations ranged between 93-520 mg L-1 and 27-440 mg L-1 with mean values of 226.8 mg L-1and 202.6 mg L-1 for the lake and creek, respectively. TSS concentrations displayed changes in relation to tidal activity. Increased current velocities and shallower water depths during low tide periods resulted
Ayub Ali A-12 Appendix A Short-term variability of physio-chemical parameters in an increase of TSS concentrations within the lake-creek system, with maximum concentrations occurring during low tide. This was more pronounced within the shallower lake sample site. Turbidity values and TSS concentrations within the system remained relatively constant (~50% rsd) with greater variability within Coombabah Creek. TSS concentrations were not significantly different between seasons within the lake-creek system. TSS concentrations were influenced by both tidal action and locally- generated wind waves. Wave heights of 200 mm are commonly observed in Coombabah Lake during strong wind speeds. Maximum TSS concentrations observed at the lake entrance did not coincide with maximum flow values, but rather with increased wind speeds observed during the sample events. The maximum fetch at the lake entrance is ~1700m, with winds from a south-westerly direction. Northerly winds also create wave actions of interest. During the summer sampling period, increased lake TSS concentrations occurred during periods of elevated wind speeds (11/11/05). Maximum hourly wind speeds ranged from 8-30.5 km h-1 from the north-east during this sample period, with an overall mean of 16.8 km h-1. Wind speeds greater than 15 km h-1 were sustained for ~9.5 h and coincided with elevated TSS concentrations. Winds from the north with reduced speeds (max. 15.2 km h-1, mean 7.48 km h-1) (4/11/05) corresponded with lower TSS concentrations compared with periods of increased wind speeds during the same tide phases. Results suggest that prolonged winds with speeds exceeding 15 km h-1 from a northerly direction have the ability to directly influence TSS concentrations at the lake entrance and shallower regions. Resuspension events within the lake may impact nutrient cycling (Tengberg et al. 2003). The influence of bioturbation should not be underestimated within the system. OSS and TSS significantly correlated within the system (r = 0.990, p <0.001, n = 153), with OSS contributing 9.98±0.05% to the TSS concentrations under all sample conditions. Organic suspended sediments originate largely from phytoplankton in marine environments; however, no relationship was observed between OSS and chl-a concentrations. This is not uncommon in estuarine and coastal environments as part of the OSS is organic dentritus originating from terrestrial runoff (Pereira-Filho et al., 2001).
Ayub Ali A-13 Appendix A Short-term variability of physio-chemical parameters
Figure A.3: Coombabah Lake-Creek system filterable nutrient concentration box-plot representation.
Filterable nutrients
Surface water filterable lake and creek nutrient concentrations were not significantly 3- different through the duration of this study. Reactive PO4 concentrations within the lake-creek system ranged from 1.30-18.2 µg L-1, with a mean concentration of -1 3- 6.98±4.21 µg L (Figure A.3). Elevated reactive PO4 concentrations exceeded the -1 - - Broadwater QWQG concentration (6 µg L ). Mean NO2 and NO3 concentrations within the lake-creek system were 2.13±1.80 µg L-1 and 17.1±8.76 µg L-1, respectively. - - Mean NOx (NO2 + NO3 ) concentration greatly exceeded the Broadwater QWQG (4 µg -1 L ). Maximum NOx concentrations were observed at the lake entrance during high tide -1 - phases, with concentrations ranging between 9.75-46.8 µg L . NH3 surface - - concentrations dominated the lake-creek dissolved inorganic nitrogen (DIN: NO2 +NO3 - +NH3 ) concentrations with a mean contribution of 37.9±16.1% and a mean concentration of 20.6±9.42 µg L-1, which exceeded the Broadwater QWQG value (8 µg L-1). Variability of the surface water concentrations included: 166% rsd for reactive 3- - - - PO4 ; 188% rsd for NO2 ; 195% rsd for NO3 ; and 218% rsd for NH3 . This was despite the physio-chemical parameters, which have an influence on nutrient concentrations, being much less varied. Several observations indicate that the high variations of
Ayub Ali A-14 Appendix A Short-term variability of physio-chemical parameters filterable nutrient concentrations in the lake-creek system were caused by dynamic processes within the estuarine system, rather than analytical aspects. Firstly, the nutrient concentrations measured were typical of concentrations reported in Australian and local coastal waters (Abal and Dennison 1996; Cox and Moss, 1999; Moss and Cox, 1999; Water Ecoscience, 2003). Secondly, nutrient concentrations plotted against time (Figure A.4) demonstrated a pattern with respect to the high and low tide phases. Maximas occurred during high water periods and minimas occurred during periods of low water (Figure A.4). Therefore, concentrations seem to be responding to natural cycles, and indicate genuine changes. The maximas and minimas observed are likely due to the mixing of two end-member waters with different concentrations, suggesting that the dominant sources of filterable nutrients are external to the lake-creek system. This observation may be explained by the hydrology of the Broadwater. During flood tides, waters entering the lake-creek system originate from the heavily urbanised Paradise Point area and Coomera River. The lower reaches of Coomera River are highly developed, including a number of canal and golf course developments, and upstream land uses include crop growing, dairy farming and cattle grazing. The ebb tide brings waters from Paradise Point and Coomera River past the creek entrance and flood waters, then enter the urbanised creek before entering the lake. Additionally, the initial ebbing waters during high water may also aid in the downstream transportation of potentially elevated concentrations from upstream golf course developments and other catchment nutrient sources. This is admittedly one of several possible interpretations of the results. No significant differences were observed between periods of no rainfall and low recorded rainfall within the lake and creek waters. During heavy rainfall events industrial and commercial development, waterfront housing and impervious surfaces within the catchment and lake-creek foreshores act as point and non-point nutrient sources (Gutteridge et al., 2003). No significant differences in filterable nutrients were - identified between seasons within the lake-creek system. Lake NO3 (p = 0.035, F =
3.122) and NOx (p = 0.022, F = 3.510) concentrations demonstrated significant differences between the four sample depths. Mean concentrations showed a general trend of greater concentrations in the surface waters (~0.3 m below the water surface) compared with the deepest sampled waters (~0.06 m above surface sediments). This
Ayub Ali A-15 Appendix A Short-term variability of physio-chemical parameters may be explained by the consumption of filterable nutrients in the overlying waters by benthic planktonic and bacterial communities (Alongi, 1994).
3- Figure A.4: Coombabah Lake filterable reactive PO4 mean concentrations during summer ebb and flood tides.
A.3.3 Filterable nutrient and chl-a transport estimations
Tides are major agents of transport in most coastal environments. The transport of filterable nutrients and chl-a at the lake entrance sample site during ebb-flood tide phases were determined using Equation A.1. Mean instantaneous nutrient transport loads during the summer sample periods were all negative values, indicating transportation of filterable nutrients into the lake. Lake entrance transport estimations -1 3- -1 - included: -0.010±0.171 g s for reactive PO4 ; -0.014±0.068 g s for NO2 ; - -1 - -1 - 0.126±0.569 g s for NO3 ; and -0.152±0.757 g s for NH3 . Observed mean flux values for surface waters chl-a were determined as 0.016±0.066 g s-1, with a maximum transportation load determined as 0.167 g s-1. Estimates were derived solely from directly-measured values of flow and concentration collected across the range of measured flow values encountered during the summer sample periods. These estimations do not detail seasonal variations. The results do, however, provide estimations of filterable nutrients and chl-a transportation and their short-term summer
Ayub Ali A-16 Appendix A Short-term variability of physio-chemical parameters variability at the entrance sample site. Flow-averaged mean chl-a transport rates were greater during the ebb tide phases at the lake entrance. Additionally, flow-averaged transport rates during sampling events demonstrated greater fluxes during flood tide - - - periods for NO2 , NO3 , NH3 than ebb tide periods, and were very similar for reactive 3- PO4 during ebb and flood periods. Such results indicate that external sources transported nutrients into the lake during the sampling period. Even if the results obtained in this study are not representative of the whole year, they indicate a potential for eutrophication of lake waters as a result of nutrient input from external sources within southern Moreton Bay. Additionally, waters that enter the lake through a secondary shallower entrance channel during high tide periods would also increase nutrient loads entering the lake system.
A.4 Conclusion
Physio-chemical parameters, suspended solids and chl-a concentrations within the lake- creek system demonstrated cyclic variations, with parameter variability ranging between 0.38-151% rsd. pH and chl-a values complied with the Broadwater QWQG sub-region values, and both DO and TSS values failed to abide by the specified guideline values. Tides and winds in excess of 15 km h-1 were identified as influential in elevating lake entrance suspended sediment concentrations. Filterable nutrient concentrations demonstrated values typically encountered in Australian and local coastal waters. Nutrient concentrations were tidally influenced, with increased concentrations occurring 3- - during sampled high tide phases. Reactive PO4 , NOx and NH3 concentrations exceeded Broadwater QWQG sub-region values. Mean negative transport values indicated nutrient inputs from external sources into the lake system during the sample period. Continued assessment is required to provide a better understanding of the hydrodynamics and biogeochemical processes in the ecologically-significant region of Coombabah Lake and Coombabah Creek, which is an important link to material transfer to the Gold Coast Broadwater, southern Moreton Bay.
A.5 Acknowledgements
The authors would like to acknowledge the financial assistance of the Cooperative Research Centre for Coastal Zone, Estuary and Waterway Management.
Ayub Ali A-17 Appendix A Short-term variability of physio-chemical parameters
Acknowledgments are also made to N. Benfer, D. Dunn, K. Dunn, T. Dunn, C. Scraggs, and P. Williams for their efforts in fieldwork and to M. Jordan for laboratory assistance.
A.6 References
Abal, E.G. and Dennison, W.C., 1996. Seagrass depth range and water quality in southern Moreton Bay, Queensland, Australia. Marine and Freshwater Research, 47, 763-771. Alongi, D.M., 1998. Coastal Ecosystem Processes. Boca Raton: CRC Press, 419p. Alongi, D.M., 1994. The role of bacteria in nutrient recycling in tropic mangrove and other coastal benthic ecosystems. Hydrobiologia, 285(1-3), 19-32. Balls, P.W., 1994. Nutrient inputs to estuaries from nine Scottish east coast rivers; influence of estuarine processes on inputs to the North Sea. Estuarine, Coastal and Shelf Science, 39, 329-352. Carmouze, J.-P. and Vasconcelos, P., 1992. The Eutrophication of the Lagoon of Saquarema, Brazil. Science of the Total Environment, Supplement, 851-859. Cederwall, H. and Elmgren, R., 1980. Biomass increase of benthic macrofauna demonstrates eutrophication of the Baltic Sea. Ophelia, Supplement, 1, 287-304. Cox, M. and Moss, A., 1999. Nerang River, Tallebudgera, Currumbin and Coombabah Creeks: Water Quality Report 1999. Brisbane, Queensland Environmental Protection Agency, 26p. Dittmar, T. and Lara, R.J., 2001. Driving Forces Behind Nutrient and Organic Matter Dynamics in a Mangrove Tidal Creek in North Brazil. Estuarine, Coastal and Shelf Science, 52, 249-259. Faulkner, S., 2004. Urbanisation impacts on the structure and function of forested wetlands. Urban Ecosystems, 7, 89-106. Gutteridge, Haskins and Davey Pty. Ltd., 2003. Coombabah Creek Environmental Inventory. Brisbane, GHD Pty Ltd, 439 p. Lee, J., Connolly, R.M., Dale, P.E.R., Dunn, R.J.K., Knight, J.M., Lemckert, C.J., McKinnon, S., Powell, B., Teasdale, P.R., Welsh, D.T. and Young, R., 2006. Impact on urbanisation on coastal wetlands: a case study of Coombabah lake, South-east Queensland. Brisbane, Technical Report No. 54, CRC for Coastal Zone, Estuary and Waterway Management, 219p.
Ayub Ali A-18 Appendix A Short-term variability of physio-chemical parameters
Lin, H.-J., Dai, X.-X., Shao, K.-T., Su, H.-M., Lo, W.-T., Hsieh, H.-L., Fang, L.-S., and Hung, J.-J., 2006. Trophic structure and functioning in a eutrophic and poorly flushed lagoon in south-western Taiwan. Marine Environmental Research, 62, 61- 82. Lorenzen, C.J., 1967. Determination of chlorophyll and pheo-pigments: Spectrophotometric equations. Limnology and Oceanography, 12, 343-346. Moss, A. and Cox, M., 1999. Southport Broadwater and adjacent Pacific Ocean: Water quality study 1979-1998. Brisbane, Queensland Environmental Protection Agency, 28p. Nixon, S.W., 1982. Nutrient Dynamics, Primary Production and Fisheries Yields of Lagoons. Oceanologica Acta, Supplement, 357-371. Pauchard, A., Aguayo, M., Peña, E., and Urrutia, R., 2006. Multiple effects of urbanization on the biodiversity of developing countries: The case of a fast- growing metropolitan area (Concepión, Chile). Biological Conservation, 127, 272-281. Pereira-Filho, J., Schettini, C. A. F., Rörig, L., and Siegle, E., 2001. Intratidal Variation and Net Transport of Dissolved Inorganic Nutrients, POC and Chlorophyll a in the Camboriú River Estuary, Brazil. Estuarine, Coastal and Shelf Science, 53, 249-257. Quinn, J.M., 1991. Guidelines for the control of undesirable biological growths in water. Consultancy Report, 6213/2, DSIR Water Quality Centre, Hamilton, New Zealand. Queensland Environmental Protection Agency, 2006. Queensland Water Quality Guidelines. Brisbane, QLD Environmental Protection Agency Publication, 121p. Regnier, P. and Wollast, R., 1993. Distribution of trace metals in suspended matter of the Scheldt estuary. Marine Chemistry, 43, 3-19. Skinner, J.L., Gillam, E., and Rohlin, C.-J., 1998. The demographic future of the Moreton region. In: Tibbetts, I.R., Hall, N.J. and Dennison, W.C. (eds), Moreton Bay and Catchment. Brisbane, Australia: School of Marine Science, The University of Queensland, pp. 67-80. Stumpf, R. P. and Haines, J. W., 1998. Variations in tidal level in the Gulf of Mexico and implications for tidal wetlands. Estuarine, Coastal and Shelf Science, 46, 165-
Ayub Ali A-19 Appendix A Short-term variability of physio-chemical parameters
73. Tengberg, A., Almroth, E. and Hall, P., 2003. Resuspension and its effects on organic carbon recycling and nutrient exchange in coastal sediments: in situ measurements using new experimental technology. Journal of Experimental Marine Biology and Ecology, 285-286, 119-142. Water Ecoscience, 2003. Victorian Water Quality Monitoring Annual report: 2001. Victoria, Australia: Water Ecoscience Pty Ltd., Report Number: 720/03, 191p.
Ayub Ali A-20