Water Quality Assessment in the Hawkesbury Nepean River System, New South Wales

School of Computing, Engineering and Mathematics

WATER QUALITY ASSESSMENT IN THE HAWKESBURY NEPEAN RIVER SYSTEM, NEW SOUTH WALES

Kuruppu Arachchige Upeka Kanchnamalie Kuruppu

Supervisory panel:

Principal Supervisor : A/Prof. Ataur Rahman Co-supervisors : A/Prof. Arumugam Sathasivan Prof. Basant Maheshwari A/Prof. Gary Dennis

This thesis is presented for the degree of Master of Engineering (Honours) in the

Western Sydney University

02 May, 2016

Statement of Authentication

I hereby declare that this thesis is my own work and to the best of my knowledge it contains no materials previously published or written by another person, nor material which to a substantial extent has been accepted for the award of any other degree or diploma at Western Sydney University or any other educational institution, except where due acknowledgement is made in the thesis.

I also declare that the intellectual content of this thesis is the product of my own work, except to the extent that assistance from others in the project’s design and conception or in style, presentation and linguistic expression is acknowledged.

Signature……………………………………………………..

Date…………………………………………………………..

Abstract

Surface waters are the most vulnerable to pollution due to their easy accessibility for disposal of wastewaters. Both the natural processes as well as the anthropogenic influences together determine the quality of surface water. The

Hawkesbury Nepean River system (HNRS) is an icon of Australia’s largest city,

Sydney, with important ecological, social and economic values. Since European settlement, the reliance on this river system has steadily increased to meet the drinking water requirements of the population, and it now provides 97% of fresh drinking water to more than 4.8 million people living in and around Sydney.

HNRS has been placed under increasing pressure and the environmental health of the river system has suffered due to the increasing development and population growth over time. The river regulation has resulted in large volumes of water being extracted for drinking water, irrigation and industrial uses. There are a number of sewage treatment plants (STPs) located in the catchment, and stormwater runoff from agricultural and urban areas can also carry pollutants into the river system. Algal and introduced macrophyte blooms have commonly occurred in the past and are likely to continue to occur in the future unless serious intervention is made by the NSW Government.

Identifying the deteriorated section of a river and actual sources of pollution along different parts of the river helps to make suitable pollution prevention activities.

Therefore, this study attempts to investigate the state of the HNRS, using water quality data from the past 20 years. Therefore, the following objectives are

iii

primarily emphasized in this thesis:

 Assess the water quality in the HNRS.

 Assess the trend of water quality in the HNRS.

 Develop prediction equations to predict water quality from surrogate water

quality parameters.

 Assess the impact of land use on the water quality of the HNRS.

 Develop a water quality index for the river in order to conduct an overall

evaluation of the water quality of the river.

This thesis consists of a series of experimental and numerical studies. They include exploratory analysis, trend analysis, principal component analysis, factor analysis, regression analysis, and application of water quality index method to make an overall water quality assessment of the HNRS.

This study has found that the concentrations of total phosphorus, nitrogen oxides and chlorophyll along the HNRS are higher than those recommended by the

Australian and New Zealand Environment and Conservation Council (ANZECC) guidelines. An increasing trend for turbidity, chlorophyll-a, alkalinity, dissolved organic carbon, total iron, total aluminium, total manganese and reactive silicate has also been detected for majority of the monitoring stations. Application of the

Canadian Water Quality Index (WQI) method shows that the water quality at 9 stations fall under either the poor or marginal category. Stations N14 and N35

were found to be the most polluted stations in the HNRS among the 9 stations.

There are many sewage treatment plants discharging treated wastewater to upstream of N35. Also, the dominant land use in this part of the catchment includes rural, grazing, commercial gardening, intensive agriculture and urban and industrial activities. These land uses can be attributed to the low WQI at N35.

Water quality at station N14 should be improved due to dilution by high quality inflows from the Colo River and the undisturbed upstream catchment. The high pollutant levels at N14 need to be investigated to find the possible reasons and to devise controlling measures. Although an improvement in water quality can be seen at some stations downstream of the undisturbed parts of the catchment, there has been an overall water quality deterioration in the HNRS during the last decade.

The HNRS is a very important river system of Australia .The findings of this study would provide an important basis for better land use planning in the catchment of the HNRS, which would improve the overall state of the river water quality.

Dedications

To my loving husband Sameera, for his love, understanding, encouragement and great support

to my loving daughter Vinuki and son Imeth.

Acknowledgement

I would like to express my deep and sincere appreciation to my principle supervisor, Associate Professor Ataur Rahman, for his endless support, exceptional advice, guidance, supervision and encouragement throughout every stage of my Masters Research. I would like to thank my co-supervisors,

Associate Professor Arumugam Sathasivan, Professor Basant Maheshwari and Associate Professor Gary Dennis, for their valuable guidance and support. I would also like to acknowledge the assistance of Dr Md Mahmudul Haque in statistical analysis.

I would like to acknowledge Ms. Tracey Schultz and Mr. Ramen Charan at

Sydney Catchment Authority for their great support by providing the water quality data needed for this study.

I would like to thank my work supervisor Mr. Kiran KC and the University of

Western Sydney for providing the opportunity to undertake a Masters Research degree.

I am indebted to my parents, Mr. Sisira Kuruppu and Mrs. Manel Hyacinth, for their love, support and inspiration.

vii

PREFACE

This thesis is submitted in fulfilment of the requirements for the degree of Masters

Honours at The Western Sydney University, NSW, Australia. The work described herein was performed by the candidate from the School of Computing,

Engineering and Mathematics, Western Sydney University. The candidate was supervised by Associate Professor Ataur Rahman (as Principal Supervisor) during the period of March 2013 to October 2015. The thesis has been supported by papers and book chapters that have been submitted for consideration, accepted or published in internationally renowned journals and conferences. These papers and book chapters are listed below:

Book chapters

 Kuruppu, U., Haque, M.M., Rahman, A. (2016), Water quality in the

urban rivers: A case study for the Hawkesbury-Nepean River system in

Australia. In Water Resources: Problems and Solutions, Edited by

Jonathan Y.S. Leung, OMICS Group International – eBooks, USA.

(Accepted and in press).

Journal papers

 Kuruppu, U., Rahman, A. (2015). Trends in water quality data in the

Hawkesbury-Nepean River System, Australia, Journal of Water and

Climate Change, doi:10.2166/wcc.2015.120. (IF=1.044, 5-Year IF=1.00,

viii

relative ranking 52/81 in water resources category, ISSN: 2040-2244, Q2,

ERA 2010 ranking: B).

Conference papers

 Kuruppu, U., Rahman, A. (2013). An Exploratory Analysis of Water

Quality in the Nepean River, Australia, 35th IAHR World Congress.

September 8 to 13, 2013 Chengdu, China, 1-6.

 Kuruppu, U., Rahman. A., Haque, M.M., Sathasivan, A. (2013). Water

Quality Investigation in the Hawkesbury- Nepean River in Sydney Using

Principal Component Analysis, 20th International Congress on Modelling

and Simulation, 1 to 6 December, 2013, Adelaide, Australia, 2646-2652.

TABLE OF CONTENTS CHAPTER 1

INTRODUCTION ...... 1

1.1 Overview ...... 1

1.2 Background ...... 2

1.3 Expected Outcomes, Values and Benefits ...... 3

1.3.1 Why is this particular piece of research worth doing? ...... 3

1.3.2 What special groups stand to benefit? ...... 4

1.4 Research Questions ...... 4

1.5 Methodology ...... 5

1.6 Thesis Structure ...... 6

CHAPTER 2

LITERATURE REVIEW ...... 8

2.1 River Water Quality ...... 8

2.2 Hawkesbury-Nepean River System ...... 14

CHAPTER 3

DESCRIPTION OF METHODS ...... 18

3.1 Overview ...... 18

3.2 Preliminary data analysis – Boxplots (box-and-whisker plots) ...... 20

3.3 Principal Component Analysis and Factor Analysis ...... 21

3.4 Mann–Kendall statistical test and Sen’s slope analysis ...... 23

3.5 Regression analysis ...... 25

3.6 Water Quality index method ...... 26

3.7 Chapter Summary ...... 30

CHAPTER 4

THE STUDY AREA AND DATA ...... 32

4.1 Overview ...... 32

4.2 Description of land use in Hawkesbury Nepean River catchment and

information on treated waste water discharge to HNRS ...... 32

4.3 Data Requirements ...... 37

4.4 Water sampling and testing ...... 41

4.4.1 Location Selection and Characterisation ...... 41

4.5 Sampling locations ...... 41

CHAPTER 5

RESULTS ON ASSESSMENT OF RIVER WATER QUALITY ...... 48

5.1 Overview ...... 48

5.2 Preliminary water quality data analysis ...... 48

5.2.1 pH ...... 48

5.2.2 Temperature ...... 50

5.2.3 Dissolved Oxygen ...... 52

5.2.4 Conductivity ...... 53

5.2.5 Turbidity ...... 55

5.2.6 Phosphorus ...... 58

5.2.7 Nitrogen ...... 63

5.2.8 Alkalinity ...... 70

5.2.9 Suspended solids ...... 71

5.2.10 Algae and chlorophyll-a ...... 73

5.3 Results from principal component analysis (PCA) ...... 78

5.4 Long term trends in water quality data ...... 84

xii

5.5 Results from regression analysis for developing prediction equations for

water quality parameters...... 97

5.6 Results of water quality assessment by using Water Quality Index ..... 101

5.7 Comparison of measured water quality data with SCA data ...... 110

5.7.1 pH ...... 111

5.7.2 Dissolved Oxygen ...... 112

5.7.3 Electrical Conductivity...... 114

5.7.4 Turbidity ...... 115

5.7.5 Nitrogen Oxides ...... 117

5.7.6 Ammonical Nitrogen ...... 118

5.7.7 Temperature ...... 120

5.8 Chapter Summary ...... 121

CHAPTER 6

SUMMARY AND CONCLUSTIONS ...... 124

6.1 Summary ...... 124

6.1 Preliminary water quality data analysis ...... 124

xiii

6.2 Trend Analysis...... 125

6.3 Regression Analysis ...... 125

6.4 Application of Canadian Water Quality Index method ...... 126

6.5 Comparison of measured water quality data with SCA data ...... 127

6.6 Conclusion ...... 127

6.7 Limitations of the study ...... 127

6.7 Suggestions for Future Research ...... 128

REFERENCES ...... 129

xiv

TABLE OF FIGURES

Figure 1.1. Methodology...... 5

Figure 2.1: Land use in Hawksburn-Nepean catchment (BOM, 2013)...... 16

Figure 3.1. Components of a default boxplot...... 19

Figure 4.1. Schematic diagram of the HNRS with land use details...... 34

Figure 4.2. Locations of the 9 sampling stations adopted in the preliminary assessment...... 39

Figure 4.3. Locations of sampling stations...... 42

Figure 4.4. Sampling stations...... 43

Figure 5.1. Box plot of pH values at different measuring stations along the Hawkesbury Nepean River System...... 50

Figure 5.2. Box plot of measured temperature along the Hawkesbury Nepean River System...... 51

Figure 5.3. Box plot of DO along the Hawkesbury Nepean River System...... 52

Figure 5.4. Box plot of conductivity along the Hawkesbury Nepean River System for all sampling stations (with the scale up to 50 mS/cm)...... 54

Figure 5.5. Box plot of conductivity along the Hawkesbury Nepean River System for all sampling stations (with the scale up to 3 mS/cm)...... 55

Figure 5.6. Box plot of turbidity along the Hawkesbury Nepean River System (with the scale up to 500 NTU)...... 57

Figure 5.7. Box plot of turbidity along the Hawkesbury Nepean River System

(with the scale up to 100 NTU)...... 58

Figure 5.8. Box plot of total phosphorus along the HNRS (with the scale up to 0.4 mg/L)...... 60

Figure 5.9. Box plot of total phosphorus along the Hawkesbury Nepean River System (with the scale up to 0.2 mg/L)...... 61

Figure 5.10. Box plot of filterable phosphorus along the Hawkesbury Nepean River System (with the scale up to 0.25 mg/L)...... 62

Figure 5.11. Box plot of filterable phosphorus along the Hawkesbury Nepean River System (with the scale up to 0.05 mg/L)...... 63

Figure 5.12. Box plot of total nitrogen along the Hawkesbury Nepean River System...... 64

Figure 5.13. Box plot of total nitrogen along the Hawkesbury Nepean River System...... 65

Figure 5.14. Box plot of nitrogen oxidised along the Hawkesbury Nepean River System...... 66

Figure 5.15. Box plot of ammoniacal nitrogen along the Hawkesbury Nepean River System (with the scale up to 1.2 mg/L)...... 67

Figure 5.16. Box plot of ammoniacal nitrogen along the Hawkesbury Nepean River System (with the scale up to 0.5 mg/L)...... 68

Figure 5.17. Box plot of nitrogen TKN along the Hawkesbury Nepean River System...... 69

Figure 5.18. Box plot of alkalinity along the Hawkesbury Nepean River System...... 71

xvi

Figure 5.19. Box plot of suspended solids along the Hawkesbury Nepean River System (with scale up to 400 mg/L)...... 72

Figure 5.20. Box plot of suspended solids along the Hawkesbury Nepean River System (with scale up to 50 mg/L)...... 73

Figure 5.21. Box plot of algal total count along the Hawkesbury Nepean River System (with the scale up to 700,000 cells/mL)...... 74

Figure 5.22. Box plot of algal total count along the Hawkesbury Nepean River System (with the scale up to 200,000 cells/mL)...... 75

Figure 5.23. Box plot of chlorophyll-a along the Hawkesbury Nepean River System (with the scale up to 250 ug/L)...... 76

Figure 5.24. Box plot of chlorophyll-a along the Hawkesbury Nepean River System (with the scale up to 50 ug/L)...... 77

Figure 5.25. Median values of pH along the Hawkesbury Nepean River System. 87

Figure 5.26. Decreasing trend of DO at station N35...... 88

Figure 5.27. Decreasing trend of EC at station N14...... 89

Figure 5.28. Median values of chlorophyll-a along the Hawkesbury Nepean River System...... 93

Figure 5.29. Increasing trend of alkalinity at station N92...... 95

Figure 5.30. Increasing trends of reactive silicate at station N35...... 96

Figure 5.31. Plot of standardized residuals against estimate for Chlorophyll-a. 100

Figure 5.32. Plot of standardized residuals against estimate for total nitrogen. . 100

xvii

Figure 5.33. Plot of standardized residuals against estimate for total phosphorous...... 101

Figure 5.34. Change in WQI over time for 9 monitoring stations in HNRS...... 103

Figure 5.35. Average WQI along the HNRS...... 104

Figure 5.36. Scope, frequency and amplitude values at 9 monitoring stations in HNRS...... 105

Figure 5.37. pH at S1 and N67...... 111

Figure 5.38. pH at S2 and N57...... 111

Figure 5.39. pH at S3 and N44...... 112

Figure 5.40. pH at S1 and N67...... 112

Figure 5.41. pH at S2 and N57...... 113

Figure 5.42. pH at S3 and N44...... 113

Figure 5.43. Electrical conductivity at S1 and N67...... 114

Figure 5.44. Electrical conductivity at S2 and N57...... 114

Figure 5.45. Electrical conductivity at S3 and N44...... 115

Figure 5.46. Turbidity at S1 and N67...... 115

Figure 5.47. Turbidity at S2 and N57...... 116

Figure 5.48. Turbidity at S3 and N44...... 116

Figure 5.49. Nitrogen oxides at S1 and N67...... 117

Figure 5.50. Nitrogen oxides at S2 and N57...... 117

xviii

Figure 5.51. Nitrogen oxides at S3 and N44...... 118

Figure 5.52. Ammonical nitrogen at S1 and N67...... 118

Figure 5.53. Ammonical nitrogen at S2 and N57...... 119

Figure 5.54. Ammonical nitrogen at S3 and N44...... 119

Figure 5.55. Temperature at S1 and N67...... 120

Figure 5.56. Temperature at S2 and N57...... 120

Figure 5.57. Temperature at S3 and N44...... 121

xix

LIST OF TABLES

Table 4.1: Sewage treatment plants along the HWNRS ...... 35

Table 4.2: Water quality monitoring stations used in the preliminary assessment 38

Table 4.3: Water quality parameters considered in the preliminary assessment .. 40

Table 4.4: Water quality data at Blaxland Crossing ...... 45

Table 4.5: Water quality data at M4...... 46

Table 4.6: Water quality data at Weir Reserve ...... 47

Table 5.1: Principal components with eigenvalues > 1 ...... 78

Table 5.2: Component score coefficients for first three PCs (for monitoring stations) ...... 79

Table 5.3: Varimax rotated factor loadings (for first 5 factors) ...... 80

Table 5.4: Explained variance and eigenvalues (for water parameters) ...... 81

Table 5.5: Component loadings for first eight PCs (water quality parameters) .... 83

Table 5.6: Median values of water quality parameters and ANZECC (2000) guidelines ...... 85

Table 5.7: Mann-Kendal test results and yearly Sen’s slope ...... 86

Table 5.8: Correlations among water quality parameters at station N44 of the HNRS ...... 98

Table 5.9: Water quality parameters and ANZEC Guidelines for Fresh and Marine Water Quality ...... 102

Table 5.10: Amplitudes at 9 stations in different years ...... 106

Table 5.11: Water quality results at N14 ...... 107

xxi

SYMBOLS

ALK Alkalinity

B A constant

CHLA Chlorophyll-a

DO Dissolved oxygen

DOC Dissolved organic carbon

EC Conductivity field

ECOCC Enterococci

ECOL E. coli

F1 Scope

F2 Frequency

F3 Amplitude

AF Aluminium filtered

FI Iron filtered

FM Manganese filtered

xxii

FP Phosphorus filterable

F-A Factor analysis

LOR Lorenzen

MK Mann–Kendall n Length of the data set nes Normalised sum of excursions

NH-N Nitrogen ammonical

NO Nitrogen oxidised

PH pH r Pearson correlation coefficient

PHA Phaeophytin

Q Slop

R2 Coefficient of determination

RS Silicate reactive

SS Suspended solids

ti Number of ties of extent i

xxiii

TA Aluminium total

TCOL True colour

TEMP Temperature

TI Iron total

TKN Nitrogen TKN

TM Manganese total

TN Nitrogen total

TP Phosphorus total

TUR Turbidity

UV UV absorbing constituents

VFs Varifactors x Sequential data value

Z Standard test statistics

xxiv

ABBREVIATIONS

ANZECC New Zealand guidelines for fresh and marine water quality

BOM Burro of metrology

EPA Environmental protection authority

HDPE High density poly ethylene

HNRS Hawkesbury Nepean River system

IUCN International union for conservation of nature and natural

Resources

NSW New South Wales

PCA Principal Component analysis

SCA Sydney catchment authority

UK United Kingdom

US United States

WQA Water quality analyser

xxv

CHAPTER 01: Introduction

CHAPTER 1

INTRODUCTION 1

1.1 OVERVIEW

Society benefits immeasurably from rivers. They are the main water resource in many inland areas for drinking, irrigation and industrial purposes. Also, rivers provide a recreational value to the adjoining community by supporting boating, fishing and outdoor activities. Although rivers contain only about 0.0001% of the total amount of water on earth at any given time, they are vital carriers of water and nutrients to areas all around the earth. They are a critical component of the hydrological cycle, acting as drainage channels for surface water. The world's rivers drain nearly 75% of the earth's land surface. They act as habitats, and provide nourishment and means of transport to countless organisms; their powerful forces create majestic scenery; they provide travel routes for exploration, commerce and recreation; they leave valuable deposits of sediments, such as sand and gravel; they form vast floodplains where many of our cities are built; and their power provides much of the electrical energy (e.g. hydro-electricity) we use in our everyday lives. Water quality in the urban environment has become important in recent Australian urban development and water management (e.g. Van der

Sterren et al., 2009; Van der Sterren et al., 2015). This thesis focuses on the study of an important river in Australia, known as the Hawkesbury Nepean River system. This chapter presents the background of the research, expected outcomes, values and

CHAPTER 01: Introduction benefits, special groups stand to benefit, research questions, and methodology and the thesis structure.

1.2 BACKGROUND

This Masters thesis provides findings of a study examining the pollution level and their sources along different parts of the Hawkesbury-Nepean River system (HNRS), located in New South Wales,Australia.

The objectives of this study are to:

. Assess the water quality in the Hawkesbury-Nepean River system.

. Assess the trend of water quality in the Hawkesbury-Nepean River system.

. Develop prediction equations to predict water quality from surrogate water quality

variables.

. Develop a water quality index to describe the overall quality of the river.

The selection of the HNRS as the key focus for this study was based on a number of reasons. Firstly, it is the main source of fresh drinking water supply to more than 4.8 million people living in, and around, Sydney. Secondly, over 70% of the HNRS flows through extensive peri-urban areas in Western Sydney and, as a result, the river system clearly indicates a gradual degradation due to peri-urban pressures such as water extraction for agriculture, discharge of treated sewage and pollutants from humans.

CHAPTER 01: Introduction

Thirdly, the HNRS has a large number of interested stakeholders, making it easy to study the conflicting social issues towards its sustainable management. Finally, there is an existing historic and current water quality data set available for this river system, which can be used for analysis and comparison during the study.

1.3 EXPECTED OUTCOMES, VALUES AND BENEFITS

1.3.1 Why is this particular piece of research worth doing?

Being the main fresh drinking water supply for more than 4.8 million people; monitoring and assessing the water quality of the Hawkesbury-Nepean River system is of immense importance. Although many government organizations, researchers and environmental agencies have spent millions of dollars every year to monitor and collect water quality data along the river, the full capacity of the water quality data set has not been used to draw meaningful conclusions describing the state of the river. This is due to the complexity of analysing these data. Investigating the most important sampling stations and water quality parameters is needed for developing cost-effective monitoring programs. Also, this collected complex water quality data should be summarized in a way that can be easily understood by the public, water distributors, planners, managers and policy makers. Available data should be effectively used to understand the current state of the river and develop restoration plans, estimate the ecological risks associated with land use plans in a watershed, or select among pre- existing, alternative development options to minimise overall river degradation.

CHAPTER 01: Introduction

1.3.2 What special groups stand to benefit?

Findings of this research support the long term river health management strategies to achieve sustainable, river health goals, as well as short term advisory information for frequent river users. River management authorities will also benefit from the outcomes of this research. Further, it will be a good source of information for different research of this large river system while providing guidance for the selection of water quality indicators for efficient monitoring. The general public who have an interest in the

Hawkesbury-Nepean River system will receive important information that will help them to make informed decisions.

1.4 RESEARCH QUESTIONS

The following research questions were addressed in this study:

. Has the water quality in the Hawkesbury-Nepean River system (HNRS) improved

in recent years?

. Is it possible to link the water quality in the HNRS with land use changes?

. Is it possible to develop surrogate equations to predict water quality from easily

measureable water quality parameters?

. Does the water quality in this river meet the national standards (e.g. Sydney Water,

Sydney Catchment Authority and Australian and New Zealand water quality

guidelines)?

CHAPTER 01: Introduction

1.5 METHODOLOGY

Figure 1.1 presents the overall methodology adopted in this study. It consists of data collation and application of various statistical techniques to address the research questions. The main statistical techniques adopted in this thesis include box plot analysis, trend investigation, principal component analysis and regression analysis. A water quality index method has been used to make an overall assessment of water quality in the HNRS.

Figure 1.1. Methodology adopted in this study.

CHAPTER 01: Introduction

1.6 THESIS STRUCTURE

The research presented in this study has been organised into six chapters, as outlined below.

Chapter 1 presents a brief introduction to the proposed research, including the background and the importance of performing the proposed research. The aims, objectives and the research questions of the proposed research are also presented in this chapter.

Chapter 2 presents a literature review on previous and ongoing water quality, monitoring programs for the Hawkesbury-Nepean River system, and other similar international studies. The gaps in the research are identified.

Chapter 3 presents the description of the methods in detail and the underlying assumptions and limitations.

Chapter 4 presents the study area and data collation procedure. A summary of the measured data is also presented in this chapter.

Chapter 5 presents the preliminary data analysis. This applies Principal Component

Analysis (PCA) and Factor Analysis (FA) to reduce the dimensionality of the data set and multiple linear regression analysis to developing prediction equations using easily measurable parameters, and Mann-Kendall (MK) test and Sen’s slope estimator to assess the trends of water quality parameters.

CHAPTER 01: Introduction

Chapter 6 presents a summary, conclusions and recommendations for further study.

CHAPTER 02: Literature Review

CHAPTER 2

LITERATURE REVIEW 2

2.1 RIVER WATER QUALITY

Water quality in urban environments is important in terms of management of stormwater and receiving water quality of river systems (Van der Sterren et al., 2012). River water quality depends on various geologic, climatic, catchment and land use characteristics. Among these, climate and land use are the key drivers of water quality in a river system. But determining the relative influence of these factors on water quality remains a significant challenge for aquatic science and management (Interlandi and Crockett, 2003). Various pollutant sources related to industries, urbanization, agriculture and mining can have a strong impact on a river system (Kendall et al., 2007; Tian and Fernandez, 2000). In recent years, an increasing awareness has been noticed in different countries about the impacts of anthropogenic activities on river water quantity and quality (Dawson and Macklin, 1998; Ma et al., 2009;

Erturk et al., 2010; Tabari et al., 2011). Climate change and urbanisation are key factors affecting the future of water quality and quantity in urbanised catchments, and are associated with significant uncertainty (Astarair-Imani et al., 2012). Pollutant build up and wash off in connection with urban catchments have become a focus of current research in different countries (Rahman et al., 2002; Egodawatta et al., 2009; Van der Sterren et al., 2013; Van der

Sterren et al., 2014; Haddad et al., 2013). In this regard, the roles of rainwater harvesting systems and water sensitive urban designs have become relevant to control water quality in

CHAPTER 02: Literature Review urban rivers (Van Der Sterren et al., 2015; Eroksuz and Rahman, 2010; Rahman et al., 2012).

Rivers play a major role in assimilating or carrying off industrial and municipal wastewater, manure discharges and runoff from agricultural fields, roadways and streets, which are responsible for river pollution (Stroomberg et al., 1995; Vega et al., 1998). Applying the concept of health to rivers is a logical outgrowth of scientific principles, legal mandates, and changing societal values (Karr, 1999). Surface waters can be contaminated by human activities in two ways: (1) by point sources, such as sewage treatment discharge and industrial discharge; and (2) by non-point sources such as runoff from urban and agricultural areas. Non-point sources are especially difficult to detect since they generally encompass large areas in drainage basins and involve complex biotic and abiotic interactions (Solbe,

1986).

Over the past century, humans have changed many rivers dramatically, threatening river health. As a result, societal well-being is also threatened because goods and services critical to human society are being depleted. Having reliable information of water quality is essential for effective and efficient water management, as it provides information regarding the condition, or health, of rivers and their adjacent landscapes, and to diagnose causes of degradation. Based on this information, we can develop restoration plans, estimate the ecological risks associated with land use plans in a watershed, or select pre-existing, alternative development options to minimise river degradation.

Watershed management and catchment scale studies have become increasingly more important in determining the impact of human development on water quality both within the

CHAPTER 02: Literature Review watershed, as well as that of receiving waters. Although these studies have become more common in the past 20 years, they still leave many questions unanswered. For example, there is a dispute regarding whether land use of the entire catchment, or that of the riparian zone is more important in influencing the water quality, while all other factors remain constant

(Osborne and Wiley, 1988).

As water drains from the land surface, it carries residues from the land. Surface runoff, especially under the first flush phenomena, is an important source of non-point source pollution. Runoff from different types of land use may be enriched with different kinds of contaminants. For example, runoff from agricultural lands may be enriched with nutrients and sediments whereas runoff from highly developed urban areas may be enriched with rubber fragments, heavy metals, as well as sodium and sulphate from road de-icers at some locations. Moreover, through evapotranspiration, interception, infiltration, percolation and absorption, different types and coverage of vegetative surfaces can modify the land surface characteristics, water balance, hydrologic cycle, and the surface water temperature (LeBlanc et al., 1997).

As a result, the quantity of water available for runoff, streamflow and groundwater flow, as well as the physical, chemical and biological processes in the receiving water bodies can be affected. It is therefore, conceivable that there is a strong relationship between land-use types and the quantity and quality of water (Gburek and Folmar, 1999). Many peri-urban rivers draining from extensive urban and agricultural areas in Australia have become highly degraded over the past few decades and remain a sensitive issue in the agenda of river

CHAPTER 02: Literature Review management authorities (Pinto and Maheshwari, 2011). With the expansion of cities into peri-urban areas, there has been a rapid increase in the number of sewage treatment facilities that discharge treated effluent into peri-urban waterways. Similarly, land use patterns can alter the quality and quantity of nutrients and sediment-rich stormwater runoff during high rainfall events (Pinto et al., 2012). Algal blooms in Australian freshwaters cost the community between AUD180 and AUD240 million every year (Atech, 2000). Rivers that are severely impacted due to anthropogenic influence are said to be suffering from urban stream syndrome (Walsh et al., 2005). Hence, prediction of water quality for river health management and issuing short and long-term advisories on the suitability of water quality to a wide range of river users are important (Pinto et al., 2012).

The river water quantity is also controlled by the climate (e.g. precipitation and wind). Over the last decade, there has been a rising concern that global warming may be impacting, and may continue to significantly impact the temperature and precipitation patterns. For example, this was recognized by the Great Lakes Regional Assessment Team in their study of the potential impacts of climate change in the Great Lakes region (Sousounis and Bisanz 2000).

Eutrophication can be influenced by climate, including precipitation, temperature and solar radiation. Precipitation and temperature firstly act on water discharge, which is widely acknowledged to be a dominant factor influencing eutrophication in river systems (Lack,

1971). The largest algal blooms always occur during periods of low ﬂows and reduced velocity, when the residence times are longer (Bowes et al., 2012). Furthermore, seasonal rises in water discharge often coincide with a decline in eutrophication abundance (Lack,

1971). Air temperature also strongly inﬂuence water temperature (warmer air means warmer

CHAPTER 02: Literature Review water). Solar radiation is also a key factor for algal blooms (Whitehead and Hornberger,

1984) which is likely to vary in the future due to climate change and anthropogenic factors

(Stanhill and Cohen, 2001).

Long-term surveys and monitoring programs of water quality are an adequate approach to a better knowledge of river hydrochemistry and pollution, but they produce large sets of data which are often difficult to interpret (Dixon and Chiswell, 1996). Also, it is quite expensive to monitor a river for a large number of water quality parameters. Most discussions on trend detection focus on analysing a single variable, while routine monitoring programs ordinarily measure several variables. The problem of data reduction and interpretation of multi- constituent chemical and physical measurements can be approached through the application of multivariate statistical techniques and exploratory data analysis (Massart et al., 1988;

Wenning and Erickson, 1994). The usefulness of multivariate statistical tools in the treatment of analytical and environmental data is reflected by the increasing number of papers cited in

Analytical Chemistry Reviews bases on these techniques (Brown et al., 1994).

The identification of trends in water quality can also be used to either confirm the effectiveness of certain management actions, or to establish a need for new management interventions. Many water quality monitoring networks have been established in Australia with the primary objective of detecting temporal trends in water quality to meet ANZECC guidelines (ANZECC, 2000). Statistical tests for trend analysis provide evidence if a trend is detected, but not the reason and hence the reason for the change/trend should be investigated

(WQA, 2013). There are many previous research studies on spatial and temporal changes in

CHAPTER 02: Literature Review water quality in river systems, such as the Han River in South Korea (Chang, 2008), the

Struma River in Bulgaria (Astel et al., 2007), the Lake Tahoe basin in the USA (Stubblefield et al., 2007), the Amu Darya River in Central Asia (Crosa et al., 2006), the water bodies of

New Seine River in France (Meybeck, 2002) and the Frome River in the UK (Hanrahan et al.,

2003).

Traditionally the assessment of river water quality has been based solely on the measurement of physical, chemical and some biological characteristics. While these measurements may be efficient for regulating effluent discharges and protecting humans, they are not very useful for large-scale management of catchments, or for assessing whether river ecosystems are being protected. Measurements of aquatic biota, to identify structural or functional integrity of ecosystems, have recently gained acceptance for river assessment. Empirical evidence from studies of river ecosystems under stress suggests that a small group of biological ecosystem level indicators can assess the river condition. However, physical and chemical features of the environment affect these indicators, the structure and function of which may be changed by human activities. The term ‘river health’, applied to the assessment of river conditions, is often seen as being analogous with human health, giving many a sense of understanding.

Unfortunately, the meaning of ‘river health’ remains obscure. It is not clear what aspects of river health sets of ecosystem-level indicators actually identify, nor how physical, chemical and biological characteristics may be integrated into measures instead of simply being observations of causes and effects. Increased examination of relationships between environmental variables that affect aquatic biota, such as habitat structure, flow regime, energy sources, water quality and biotic interactions and biological conditions, are required in

CHAPTER 02: Literature Review the study of river health (Norris et al, 1999).

To assess the health of freshwater for biotic species and humans, various guidelines have been developed internationally e.g. IUCN Global Freshwater Initiative (International Union for Conservation of Nature and Natural Resources), Healthy Watershed Initiative in the US

(Young and Sanzone, 2002), Pressure, State, Response model in Australia (Commonwealth of Australia, 1996) and EU Water Framework Directive in Europe (Kaika, 2003), and

Australian and New Zealand Guidelines for Fresh and Marine Water Quality (ANZECC)

(ANZECC, 2000). One of the long-term goals of the monitoring program of stream water quality is to detect changes or trends in pollution levels over time and to identify, describe, and explain the major factors affecting such trends, and to devise a strategy to improve the overall water quality of a river system (Yu et al., 1993).

2.2 HAWKESBURY-NEPEAN RIVER SYSTEM

Sydney is the most populous city in Australia with a population of over 4.5 million. The surroundings of Sydney are highly urbanized as compared to the rest of Australia due to continued, high residential developments in the region over the past several decades.

Populations have moved away from city centres to the peri-urban surrounding areas at higher rates, resulting in exponential increases in commercial and residential developments. Even though the chemical composition of fresh surface water in the Hawkesbury-Nepean River

System (HNRS), located in New South Wales (NSW), Australia, has been extensively studied over the past 20 years. Majority of the monitored data is strongly biased towards pH,

CHAPTER 02: Literature Review conductivity, turbidity, dissolved oxygen, major ions and nutrients. These data are routinely monitored by state government authorities to provide information on the quality of Sydney’s potable water supply and sewage treatment plants. The full potential of data has not been used to examine the long-term spatial trends in the chemistry of the freshwater reaches of the

HNRS (Markich and Brown, 1998).

The Hawkesbury-Nepean River System (HNRS) is the main source of fresh drinking water supply to more than 4.8 million people living in, and around, Sydney. The HNRS system is a combination of two major rivers (Figure 2.1): the Nepean River (155 km) and the

Hawkesbury River (145 km) (Markich and Brown, 1998). The river system is complex in nature; the upper part contains poorly accessible gorges, the middle part is running through irrigated farm lands and the lower part has tidal slopes with deposited soil pockets (Diamond,

2004). The middle part of the river is being continuously influenced by increasing population growth, urbanization, industrialization and other human activities which cause contamination of the quality of the river water from different sources (e.g. sewage, stormwater, runoff from disused mines, toxic forms of blue-green algae, and waste from domestic and native animals).

Pinto and Maheshwari (2011) have shown that river health in peri-urban landscapes are prone to higher degrees of degradation. Within the HNR catchment, vegetation clearance has been continuously practised over the last 200 years causing increased subsurface and agricultural runoff and sediment loads into the river system (Thomas et al. 2000). Land use in the HNR catchment includes regions that are heavily peri-urbanised and industrialised, and which are important for recreational and agricultural activities and tourism (Baginska et al. 2003, Pinto and Maheshwari, 2010). Agricultural runoff contributes approximately 40% to 50% of

CHAPTER 02: Literature Review phosphorus loads and 25% of nitrate loads into the HNRS which are believed to have originated from agricultural and animal farms (Markich and Brown 1998).

Figure 2.1: Land use in Hawksbury-Nepean catchment (BOM, 2013).

This river system has been subjected to multiple disturbances since European settlement, including extensive clearing of over 37% of the catchment for agriculture, urban and industrial land use, nutrient enrichment associated with sewage, urban runoff and wastewater disposal, extractive industries, regulation and diversion of river flows, and mining (Gehrke

CHAPTER 02: Literature Review and Harris, 1996).

Unlike other natural rivers where flow is dominated by rainfall events, the flow of HNRS is highly regulated by impoundments and treated effluent discharges from sewage treatment plants. There are about 22 dams and 15 weirs situated along the HNRS. The major dam on this river is at Warragamba, which holds about 2.057 × 109 km3 of water, captured from a

9000 km2 catchment area (Turner and Erskine, 2005).

There are 18 sewage treatment plants along the HNRS discharging significant volumes of treated municipal wastewater into the river. The river system has been increasingly regulated since completion of the first diversion weirs in 1888, with the largest dam, Warragamba

Dam, completed in1960 to provide the main water supply for the Sydney metropolitan area.

Twenty-nine dams of 7m or more in height, and another 52 smaller structures, now regulate flows in the river system (Marsden and Gehrke, 1996).

Concern about the ability of the river to support the increasing water demands of a growing urban population has led to the enactment of legislation in New South Wales requiring the

Sydney Water Corporation to protect the aquatic environment by conducting its activities in an ecologically sustainable manner (WBC, 1994).

The above review highlights the complex nature of the land use and water quality interaction of the HNRS and it underlines the importance of assessing the water quality of this important river system, which is the focus of this study.

CHAPTER 03: Description of Methods

CHAPTER 3

DESCRIPTION OF METHODS 3

3.1 OVERVIEW

A regular water quality monitoring program generates reliable data which reflects the state of the water quality of a river. However, generating good data is not enough to meet the objectives of a water quality monitoring program; data must be processed and presented in a manner that provides the understanding of the spatial and temporal patterns in water quality parameters. The intent is to use a collected set of data to explain the current state of the water more widely and make necessary controls to overcome future water quality issues. Water quality data usually exhibit the following characteristics: non-normal distribution, presence of outliers, missing values, values below detection limits (censored), and serial dependence. It is essential to apply an appropriate, statistical methodology when analysing water quality data to draw valid conclusions, and hence it provides useful advices in water management. This chapter presents a detailed description of statistical methods used in this research to assess the water quality in the Hawkesbury-Nepean river system.

Different forms of graphs have been used to provide visual summaries of data quickly and clearly to describe important information contained in the data, and provide insight into the data. Graphs help to determine if more complicated modelling is necessary.

CHAPTER 03: Description of Methods

Three particularly useful graphical methods are boxplots, scatter plots, and Q-Q plots.

In this study, box plots have been used for exploratory data analysis as it provides summaries of a dataset.

Water quality monitoring programs generate complex multidimensional data.

Multivariate statistical techniques need to be used to extract useful information from this data. In this study, factor analysis and principal component analysis have been performed to identify the most significant water quality monitoring stations, and water quality parameters in the HNRS.

The rank based non-parametric Mann–Kendall (MK) statistical test has been used to assess the trend in the water quality time series data as these tests are more suitable for non-normally distributed data and censored data which are frequently encountered in hydro-meteorological time series (Yue et al., 2002). For the MK test, data is not needed to conform to any particular distribution and moreover, it has less sensitivity to data gaps (Tabari et al., 2011).

Pearson correlation coefficients have been used to examine the correlations among various pollutants. Multiple linear regression technique has been used to develop the prediction equations for water quality parameters such as Chlorophyll-a, total phosphorous and total nitrogen (which are difficult to measure) as a function of easily measurable water quality parameters. The plots of standardized residuals are examined and coefficient of determination (R2) and standard error of estimates are used to assess the adequacy of the developed prediction equations.

CHAPTER 03: Description of Methods

Water quality index method has been used to compare water quality parameters with respective regulatory standards, which gives a single indicator to describe the overall quality of a water body (Boyacioglu, 2010).

3.2 PRELIMINARY DATA ANALYSIS – BOXPLOTS

A boxplot is a very useful and convenient tool to provide summaries of a dataset and is often used in exploratory data analysis. A boxplot usually presents a dataset through five numbers: extreme values (minimum and maximum values), median (50th percentile), 25th percentile, and 75th percentile. It also indicates the degree of dispersion, the degree of skew and unusual values of the data (outliers). Furthermore, boxplots can display differences between different populations without making any assumptions of the underlying statistical distribution. Figure 3.1 illustrates the components of a default boxplot.

Figure 3.1. Components of a default boxplot.

CHAPTER 03: Description of Methods

3.3 PRINCIPAL COMPONENT ANALYSIS AND FACTOR ANALYSIS

The multidimensionality (i.e. different sampling stations and different parameters over time) of the data makes analysis more complicated. Principal component analysis

(PCA) and factor analysis (FA) are the two multivariate techniques with the central aim of reducing as much laity of a multivariate data set as much of possible, while still retaining their variation/useful information as much as possible. This objective is achieved by transforming the original variables to a new set of hypothetical variables, called principal components or factors (PC/F) that are uncorrelated. They are obtained as a linear combination of the original variables. Principal components or factors explain the original variance in a monotonically decreasing way (Kovács et al., 2012).

Factor analysis (F-A) is similar to principal component analysis, but the two are not identical. In F-A, components extracted from PCA are rotated according to a mathematically established rule (i.e., varimax, equamax and quarimax) yielding easily interpretable new variables, called varifactors (VFs) (Pinto et al., 2013). F-A uses regression modelling techniques to test hypotheses producing error terms, while PCA is a descriptive statistical technique (Bartholomew et al., 2008). The difference between

PCs obtained in PCA and VFs obtained in F-A is that PCs are linear combinations of observable water quality parameters but VF are unobservable, hypothetical and latent variables (Alberto et al., 2001).

The differences between PCA and F-A are further illustrated by Suhr (2009) as follows:

CHAPTER 03: Description of Methods

. PCA results in principal components that account for a maximal amount of variance

for observed variables. F-A accounts for common variance in the data. PCA inserts

ones on the diagonals of the correlation matrix. F-A adjusts the diagonals of the

correlation matrix with the unique factors.

. PCA minimizes the sum of squared perpendicular distance to the component axis.

F-A estimates factors which influence responses on observed variables.

. The component scores in PCA represent a linear combination of the observed

variables weighted by eigenvectors. The observed variables in F-A are linear

combinations of the underlying and unique factors.

In this study, PCA was performed first to identify the most important water quality monitoring station(s) in the HNRS. For the purpose of this analysis, the median value of each parameter was used, as the median is better suited for a skewed distribution to describe the central tendency of the data. In this analysis stations with correlation coefficient greater than 0.9 were taken as principal water quality monitoring stations.

Equations for principal components were derived by considering the loadings of the variables (water quality monitoring stations).

An F-A was employed to further identify the monitoring stations that are important in revealing surface water quality variations. Varimax rotation was selected as the data rotation method, as it makes an orthogonal rotation of the factor axes to maximize the variance of the squared loadings of a factor on all the variables in a factor matrix which

CHAPTER 03: Description of Methods has the effect of differentiating the original variables by extracted factors. Each factor has either large or small loadings of any particular variable. A varimax solution was used to identify each variable with a single factor. This is the most common rotation option used in PCA and F-A. However, the orthogonality (i.e., independence) of factors is often an unrealistic assumption (Russell, 2002). In the second step, PCA was performed on water quality data to identify the principal components that explain most of the variance in the water quality data set.

3.4 MANN–KENDALL STATISTICAL TEST AND SEN’S SLOPE ANALYSIS

The MK test is based on the test statistics defined as follows (equation 3.1):

3.1

Where sgn(Ө) is taken as equation 3.2:

3.2

Where xi and xj are the sequential data values, n is the length of the data set, and E(S) and V(S) are as follows (equations 3.3 and 3.4):

3.3

3.4

CHAPTER 03: Description of Methods

Where ti is the number of ties of extent i. The standard test statistics Z is computed by equation 3.5.

3.5

Positive values of Z indicate increasing trends while negative values indicate decreasing trends. When testing either increasing or decreasing monotonic trends at an α significance level, the null hypothesis is rejected for absolute values of Z greater than Z

(1-α/2), obtained from the standard normal cumulative distribution table (Tabari and

Ahmadi, 2011).

Sen’s method uses a liner model to estimate the slope of the trend and variance of the residuals should remain constant over time (Drapela and Drapelpva, 2011). If a linear trend is present in a time series, the true slope (change per unit time) can be estimated by using a simple nonparametric procedure (Sen, 1968; Drapela and Drapelpva, 2011).

This liner model f ( t )can be described as follows (equation 3.6):

f ( t )  Qt  B 3.6

Where Q is the slope and B is a constant.

Slopes of all data pairs are calculated and the median value is taken as the Sen’s slope (equation 3.7):

CHAPTER 03: Description of Methods

x  x Q  j k 3.7 i j  k

3.5 REGRESSION ANALYSIS

Regression analysis generates an equation to describe the statistical relationship between one or more predictors and the response variable, and to predict new observations. Regression generally uses the ordinary least squares method, which derives the equation by minimizing the sum of the squared residuals.

Regression results indicate the direction, size, and statistical significance of the relationship between a predictor and response.

. Sign of each coefficient indicates the direction of the relationship.

. Coefficients represent the mean change in the response for one unit of change in the

predictor while holding other predictors in the model constant.

. P value for each coefficient tests the null hypothesis that the coefficient is equal to

zero (no effect). Therefore, low p-values suggest the predictor is a meaningful

addition to your model.

. The equation predicts new observations for given specified predictor values.

The aim of regression analysis is to construct mathematical models which describe or explain relationships that may exist between variables (Draper and Smith, 1981).

Pearson correlation coefficients are used in this study to examine the correlations

CHAPTER 03: Description of Methods among various pollutants. Multiple linear regression technique is used to develop the prediction equations for Chlorophyll-a, total phosphorous and total nitrogen as a function of easily measurable water quality parameters. The plots of standardized residuals are examined and the coefficient of determination (R2) and the standard error of estimates are used to assess the adequacy of the developed prediction equations.

3.6 WATER QUALITY INDEX (WQI) METHOD

First studies on WQI were done in 1848 in Germany (Sarkar and Abbasi, 2006; Lumb et al., 2011); Horton (1965) developed the first WQI based on 8 water quality parameters. Dede et al. (2013), used 5 WQI methods (Oregon WQI, Aquatic toxicity index, overall index of pollution, universal water quality index and CCME WQI) to evaluate surface water quality and concluded that CCME WQI is the only method that allows utilization of all the available parameters in the calculation of overall index value.

WQI can be used for tracking changes at one site over time, and for comparisons among sites in a river. It was simply developed to provide a broad overview of environmental performance (Khan et al., 2004).

Though the WQI provides a meaningful summarization of the quality of a water body, it is not a substitute for detailed analysis of water quality data and should not be used as a sole tool for management of water bodies (Al-Janabi et al., 2012).

CHAPTER 03: Description of Methods

The Canadian Council of Ministers of the Environment (CCME) Water Quality Index is based on a formula developed by the British Columbia Ministry of Environment,

Lands and Parks and modified by Alberta Environment. The Index incorporates three elements:

. Scope (F1)- the number of variables not meeting water quality objectives;

. Frequency (F2) - the number of times these objectives are not met;

. Amplitude (F3) - the amount by which the objectives are not met.

Scope (F1) - Scope assesses the extent of water quality guideline non-compliance over the time period of interest, which means the number of parameters whose objective limits are not met. It has been adopted directly from the British Columbia Water

Quality Index:

푇표푡푎푙 푛푢푚푏푒푟 표푓 푓푎푖푙푒푑 푣푎푟푖푎푏푙푒푠 퐹 = × 100 3.8 1 푇푎푡푎푙 푛푢푚푏푒푟 표푓 푣푎푟푖푎푏푙푒푠

Where, the variables indicate those water quality parameters whose objective values

(threshold limits) are specified and observed values at the sampling sites are available for the index calculation.

Frequency (F2) - The frequency (i.e. how many occasions the tested or observed values were off the acceptable limits) with which the objectives are not met, which represents

CHAPTER 03: Description of Methods the percentage of individual tests that do not meet the objectives (“failed tests”):

푁푢푚푏푒푟 표푓 푓푎푖푙푒푑 푡푒푠푡푠 퐹 = × 100 3.9 2 푇표푡푎푙 푛푢푚푏푒푟 표푓 푣푎푟푖푎푏푙푒푠

The formulation of this factor is drawn directly from the British Columbia Water

Quality Index.

Amplitude (F3) - The amount by which the objectives are not met (amplitude) represents the amount by which the failed test values do not meet their objectives, and is calculated in three steps. The number of times by which an individual concentration is greater than (or less than, when the objective is a minimum) the objective is termed as an “excursion” and is expressed as follows. When the test value must not exceed the objective:

퐹푎푖푙푑 푡푒푠푡 푣푎푙푢푒푖 푒푥푐푢푟푠푖표푛푖 = ( ) − 1 3.10 푂푏푗푒푐푡푖푣푒푗

For the cases in which the test value must not fall below the objective:

푂푏푗푒푐푡푖푣푒푗 푒푥푐푢푟푠푖표푛푖 = ( ) − 1 3.11 퐹푎푖푙푑 푡푒푠푡 푣푎푙푢푒푖

The collective amount, by which the individual tests are out of compliance, is calculated summing the excursions of individual tests from their objectives and then dividing the sum by the total number of tests. This variable, referred to as the normalized sum of excursions (nse) is calculated as:

CHAPTER 03: Description of Methods

∑푛 푒푥푐푢푟푠푖표푛 푛푠푒 = 푖=1 푖 3.12 푁푢푚푏푒푟 표푓 푡푒푠푡푠

F3 is then calculated by an asymptotic function that scales the normalized sum of the excursions from objectives ( nse ) to yield a value between 0 and 100.

푛푠푒 퐹 = ( ) 3.13 3 0.01푛푠푒 + 0.01 The CWQI is finally calculated as:

2 2 2 √퐹1 + 퐹2 + 퐹3 퐶푊푄퐼 = 100 − 3.14 1.732 ( ) The factor of 1.732 has been introduced to scale the index from 0 to 100. Since the individual index factors can range as high as 100, it means that the vector length can reach a maximum of 173.2 as shown below:

√1002 + 1002 + 1002 = √30000 = 173.2 3.15

The index produces a number between 0 (worst water quality) and 100 (best water quality). These numbers are divided into 5 descriptive categories to simplify the presentation:

. Excellent: (CCME WQI Value 95-100) – water quality is protected with a virtual

absence of threat or impairment; conditions very close to natural or pristine levels.

. Good: (CCME WQI Value 80-94) – water quality is protected with only a minor

degree of threat or impairment; conditions rarely depart from natural or desirable

CHAPTER 03: Description of Methods

levels.

. Fair: (CCME WQI Value 65-79) – water quality is usually protected but

occasionally threatened or impaired; conditions sometimes depart from natural or

desirable levels.

. Marginal: (CCME WQI Value 45-64) – water quality is frequently threatened or

impaired; conditions often depart from natural or desirable levels.

. Poor: (CCME WQI Value 0-44) – water quality is almost always threatened or

impaired; conditions usually depart from natural or desirable levels.

3.7 CHAPTER SUMMARY

The discretion of methods used to assess the water quality in the Hawkesbury Nepean

River system (HNRS) have been presented in this chapter. It describes the use of boxplots as the preliminary data analysis tool to identify the distribution of the water quality data. Thereafter, the use of principal component analysis (PCA) and factor analysis (FA) is presented to identify the most significant water quality monitoring stations in the Hawkesbury-Nepean River System. The mathematical formulation of rank-based non-parametric Mann–Kendall (MK) statistical test and Sen’s slope analysis have been presented, which were used to assess the trend in the water quality time series data. With the aim of developing prediction equations for complex water quality parameters, regression analysis was performed; the description of the method is

CHAPTER 03: Description of Methods presented in this chapter. Finally, it describes the water quality index method, which was used to identify and assess deteriorated sections in the Hawkesbury-Nepean River

System (HNRS). In addition, it identifies the water quality parameters contributing to poor water quality and tracks changes in water quality at different sites over time.

CHAPTER 04: Study Area and Data

CHAPTER 4

THE STUDY AREA AND DATA 4

4.1 OVERVIEW

The water quality data and land data for this research project have been collected from the Hawkesbury-Nepean River System (HNRS) and its catchment area. HNRS is the main source of fresh drinking water supply to more than 4.8 million people living in, and around Sydney. The HNRS system is a combination of two major rivers: the

Nepean River (155 km) and the Hawkesbury River (145 km). At present, the HNRS is under increasing pressures from peri-urbanisation and industrialization. This chapter provides a description of land use in the Hawkesbury Nepean River catchment, information on treated waste water discharge to HNRS, water quality parameters used for this analysis and the water sampling and testings.

4.2 DESCRIPTION OF LAND USE IN HAWKESBURY NEPEAN RIVER CATCHMENT AND INFORMATION ON TREATED WASTE WATER DISCHARGE TO HNRS

More than 1.3 billion litres of wastewater is collected daily and treated by Sydney

Water, following strict license conditions issued by the NSW Environment Protection

Authority (EPA), before it is re-used or discharged into rivers.

The wastewater transported to water recycling plants goes through many treatment

CHAPTER 04: Study Area and Data steps including filtration and disinfection to remove nearly all biodegradable organic material and nutrients. A schematic diagram of the HNRS with the land use details is

presented in Figure 4.1 and Table 4.1.

CHAPTER 04: Study Area and Data

Figure 4.1. Schematic diagram of the HNRS with land use details.

CHAPTER 04: Study Area and Data

Table 4.1: Sewage treatment plants (STP) along the HWNRS

Discharge Treatment level Completed date Discharge location STP (ML/day) Picton Tertiary (includes additional Phosphorus 30/06/2009 1.5 Re-used on-site for agricultural irrigation removal and disinfection) Precautionary discharge to Stone Quarry Creek West Tertiary (includes additional Phosphorus 30/06/2006 10.7 Re-used at Agricultural Institute. Remainder Camden removal and disinfection) discharged via Matahill Creek to the Nepean River Wallacia Tertiary (includes additional phosphorus 30/06/2009 0.8 Warragamba River and nitrogen removal and disinfection) Penrith Tertiary (includes additional Phosphorus 30/06/2009 22.4 Re-used locally. Remainder transferred to St and Nitrogen removal and disinfection) Marys Advanced Water Treatment Plant. Some excess discharged to Boundary Creek

St. Marys Tertiary (includes ultrafiltration, reverse 30/06/2009 33.5 Re-used locally and at Dunheved. Remainder osmosis, de-carbonation, additional discharged to Nepean River. Some excess Phosphorous and Nitrogen removal and discharged to South Creek disinfection) Winmalee Tertiary(includes additional phosphorus 30/06/2009 16.5 Unnamed creek to the Nepean River

CHAPTER 04: Study Area and Data

Discharge Treatment level Completed date Discharge location STP (ML/day) and nitrogen removal and disinfection) North Tertiary (includes additional phosphorus 30/06/2009 0.9 Redbank Creek to the Hawkesbury River Richmond removal and disinfection) Riverstone Tertiary(includes additional phosphorus 30/06/2009 1.8 Eastern Creek to South Creek removal and disinfection) Quakers Hill Tertiary (includes additional Phosphorus 30/06/1974 31.1 Re-used locally and at Ashlar Golf Course. and Nitrogen removal and disinfection) Remainder transferred to St Marys Advanced Water Treatment Plant. Some excess discharged to Breakfast Creek Rouse Hill Tertiary (includes additional Phosphorus 30/06/2009 15.3 Recycled back to households for non-drinking and Nitrogen removal and disinfection) use. Excess discharged to Second Ponds Creek also includes ultra-violet irradiation and via wetlands to Cattai Creek super-chlorination for reuse water

Castle Hill Tertiary (includes additional Phosphorus 30/06/2009 6.5 Cattai Creek removal and disinfection)

CHAPTER 04: Study Area and Data

4.3 DATA REQUIREMENTS

This research required long term water quality data, rainfall data and land use data from the catchment. Water quality data has been collected in-house laboratory testing and from Sydney Catchment Authority, as well as by field and laboratory testing for a period of one year. Rainfall data has been obtained from the Australian Bureau of

Meteorology. Land use data has been collected from NSW government departments and other available sources.

The locations of the selected water quality monitoring stations are presented in Table

4.2. Figure 4.2 and Table 4.3 illustrates various water quality parameters examined in this preliminary assessment.

CHAPTER 04: Study Area and Data

Table 4.2: Water quality monitoring stations used in the preliminary assessment

Site code Site Longitudes Latitudes

N92 Nepean River at Maldon Weir upstream of Stone quarry Creek and Picton Sewage Treatment Plant 150.62 -34.2

N75 Nepean River at Sharpes Weir downstream of Matahil Creek and Camden Sewage Treatment Plant 150.67 -34.03

N67 Nepean River at Wallacia Bridge upstream of Warragamba River 150.63 -33.86

N57 Nepean River at Penrith Weir upstream of Boundary Creek and Penrith Sewage Treatment Plant 150.68 -33.74

N44 Nepean River at Yarramundi Bridge upstream of Grose River 150.69 -33.61

N42 Hawkesbury River at North Richmond upstream of North Richmond Water Treatments Works 150.71 -33.59

N35 Hawkesbury River at Wilberforce upstream of Cattai Creek 150.83 -33.58

N21 Hawkesbury River at Lower Portland upstream of Colo River 150.88 -33.43

N14 Hawkesbury River at Wisemans Ferry downstream of Car Ferry 150.98 -33.38

CHAPTER 04: Study Area and Data

Figure 4.2. Locations of the 9 sampling stations adopted in the preliminary assessment (Reproduced from: http://www.lahistoriaconmapas.com/atlas/map-river/Cook-Islands- river-map.htm).

CHAPTER 04: Study Area and Data

Table 4.3: Water quality parameters considered in the preliminary assessment

Water quality Abbreviation Units Min Max Median pH PH 5.78 9.94 7.63 Parameter Lorenzen LOR ug/L 0.10 539.90 4.40 Iron Total TI mg/L 0.04 5.62 0.29 Phaeophytin PHA ug/L 0.10 25.20 0.80 Nitrogen TKN TKN mg/L 0.02 5.40 0.27 Temperature TEMP Deg C 8.10 30.60 19.50 Chlorophyll-a CHLA ug/L 0.20 253.10 5.10 E. coli ECOL orgs/10 0.00 6100.00 13.00 Iron Filtered FI mg/L 0.01 3.43 0.09 0mL True Colour TCOL 1.00 93.00 11.00 Nitrogen Total TN mg/L 0.08 5.90 0.45 Turbidity TUR NTU -0.60 380.00 3.85 Alkalinity ALK mgCaC 1.00 298.00 40.00 Aluminium Total TA mg/L 0.01 3.97 0.08 O3/L Manganese Total TM mg/L 0.00 0.48 0.03 Dissolved Oxygen DO mg/L 1.50 16.20 9.10 Enterococci ECOCC cfu/100 0.00 8400.00 20.00 Phosphorus Total TP mg/L 0.01 0.18 0.01 mL Suspended Solids SS mg/L 1.00 105.00 3.00 Nitrogen Oxidised NO mg/L 0.00 5.00 0.17 Aluminium Filtered FA mg/L 0.00 0.45 0.01 Manganese Filtered FM mg/L 0.00 0.35 0.01 Conductivity Field EC mS/cm 0.01 48.40 0.30 Nitrogen Ammonical NH-N mg/L 0.01 0.41 0.01 Phosphorus Filterable FP mg/L 0.00 0.11 0.01 Silicate Reactive RS SiO2 0.01 14.90 1.71 Dissolved Organic DOC mg/L 0.20 350.00 4.60 mg/L UV Absorbing UV 0.01 0.93 0.12 Carbon constituents

CHAPTER 04: Study Area and Data

4.4 WATER SAMPLING AND TESTING

In addition to the water quality data obtained from Sydney Water, as a part of this study, water samples were also collected from selected sampling stations fortnightly for a period of one year.

4.4.1 Location Selection and Characterisation

Samples were collected from three locations along the HNRS. The selection of the locations was governed by three factors. Firstly, these locations are largely exposed to impacts from extensive agricultural and urban activities. Secondly, pre-established river management authority monitoring stations were found in close vicinity of these locations, providing access to existing water quality data sets if required. Thirdly, the locations were easily accessible by a boat.

4.5 SAMPLING LOCATIONS

River water samples were collected fortnightly from three sampling stations, and testing for different water quality parameters was done following the standard methods.

Sampling stations are presented in Figure 4.3 and Figure 4.4.

CHAPTER 04: Study Area and Data

Figure 4.3. Locations of sampling stations in the HNRS (Reproduced from: Google maps).

CHAPTER 04: Study Area and Data

S1 - Blaxland Crossing S2 - M4 S3 - Weir Reserve

Figure 4.4. Sampling stations.

The digital water quality multi probes (HACH HQ 40D) were utilised to obtain the measurements of temperature (measured in degrees Celsius), pH, dissolved oxygen

(DO measured in milligrams per litre) and electrical conductivity (EC measured in micro Siemens per centimetre at 250C). Turbidity was measured using HACH 2000NT turbid meter. From each sub-site, 1 L of water sample was collected in an acid rinsed, high-density polyethylene bottle (HDPE) for laboratory analysis. Ammoniacal nitrogen

(NH3 -N) and Nitrogen Oxides (NOx) were measured in the laboratory. The Gallery

(Thermo Scientific), a high precision, chemistry automated analyser, was adopted for measuring NH3 -N, nitrite and NOx concentrations. It is a fully automated instrument that provides analyses on optical multi-cell cuvette which provides a discrete analysis.

NH3 -N included free ammonia, ammonium and ammonia associated with chloramine determined by using colorimetric method. Available ammonia reacts with hypochlorite ions generated by the alkaline hydrolysis of sodium dichloroisocyanurate to form

CHAPTER 04: Study Area and Data mono-chloramine which reacts with salicylate ions in the presence of sodium nitroprusside, at around pH 12.6, to form a blue compound. The compound is measured spectrophotometrically at 660 nm. Nitrite is measured by reaction with sulphanilamide and N-(1-naphthyl)-ethylenediamine dihydorchloride to form a highly colored azo-dye, thus, the absorbance is measured spectrophotometrically at 540 nm or 520 nm. The determination of nitrate is done by catalytically reducing the nitrate ions into nitrite ions

(possibly by nitrate reductive enzyme in the presence of reduced nicotinaminde dinucleotide), the total nitrite ions are then measured by sulphanilamide method as the

NOx, and nitrate is obtained by deduction nitrite from the NOx. The analyser has the detection limit for NH3 -N, nitrite and NOx of 0.002 mg-N/L. Standard curves for NH3

-N, nitrite and NOx were calibrated for the range 0.0 to 1.0 mg-N/L using stock solutions of ammonium chloride, sodium nitrite and sodium nitrate, respectively. The experimental errors were 1.5% for NH3 -N, NOx measurement. The obtained water quality data are provided in Tables 4.4 to 4.6.

CHAPTER 04: Study Area and Data

Table 4.4: Water quality data at Blaxland Crossing

Date pH DO Tem EC Turbidity / NOx NH3 -N (mg/L) (deg C) (us/cm) (NTU) (mg/L) (mg/L) 22/02/13 7.36 8.3 25 253 10.5 0.352 0.005 08/03/13 7.32 7.9 23.6 190 9.2 0.281 0.011 22/03/13 7.27 7.2 22.5 175 9.5 0.256 0.016 05/04/13 7.32 8.2 22.6 202 6.1 0.307 0.013 19/04/13 7.23 8.6 21.4 215 7.0 0.289 0.004 03/05/13 7.08 7.9 19.2 226 6.2 0.264 0.026 17/05/13 7.01 8.8 16.6 233 8.0 0.257 0.016 31/05/13 7.21 9.5 16.5 228 7.0 0.214 0.019 14/06/13 7.34 10.4 15.2 194 12.6 0.236 0.004 28/06/13 7.48 10.9 14.4 150 38 0.256 0.018 12/07/13 7.52 9.9 14.6 162 8.6 0.254 0.012 26/07/13 7.67 10.7 13.6 175 7.2 0.266 0.004 09/08/13 7.85 11.3 15.5 206 5.7 0.307 0.005 23/08/13 7.89 10.4 16.4 216 6.4 0.298 0.012 06/09/13 7.91 9.2 20.5 263 5.2 0.275 0.014 20/09/13 7.94 8.7 21.8 290 6.3 0.268 0.004 04/10/13 7.94 10.5 21.1 338 6.1 0.277 0.01 18/10/13 7.63 8.7 22.1 328 4.2 0.192 0.034 01/11/13 7.44 7.8 23.9 309 6 0.178 0.033 15/11/13 7.46 7.5 24.6 287 6.2 0.167 0.042 24/01/14 7.51 6.8 26.7 278 5.1 0.035 0.056 07/02/14 7.32 6.1 25.4 271 5.4 0.128 0.052 21/02/14 7.21 5.8 24.8 263 5.8 0.218 0.053 07/03/14 7.53 7.9 25.5 248 3.6 0.246 0.056

CHAPTER 04: Study Area and Data

Table 4.5: Water quality data at M4

Date pH DO Tem EC Turbidity NOx NH3 -N (mg/L) (deg C) (us/cm) (NTU) (mg/L) (mg/L) 22/02/13 7.53 9.8 24.8 239 6.96 0.386 0.005 08/03/13 7.42 9.6 24.3 220 6.8 0.286 0.006 22/03/13 7.32 9.4 23.5 170 6.1 0.175 0.008 05/04/13 7.21 8.3 22.4 190 5.92 0.110 0.004 19/04/13 7.28 8.5 21.5 212 5.8 0.213 0.007 03/05/13 7.32 8.6 19.5 234 6.2 0.267 0.009 17/05/13 7.34 8.7 15.4 256 6.54 0.314 0.008 31/05/13 7.38 9.2 15.3 246 6.8 0.246 0.005 14/06/13 7.46 10.7 14.8 186 14.2 0.257 0.008 28/06/13 7.6 11.6 14.4 157 25 0.251 0.013 12/07/13 7.59 11.4 13.5 164 4.5 0.213 0.016 26/07/13 7.64 11.1 13.1 169 4.2 0.224 0.004 09/08/13 7.77 11.3 12.9 175 3.5 0.234 0.005 23/08/13 7.67 10.5 13 201 5 0.223 0.004 06/09/13 7.58 10.4 16.7 237 5.6 0.284 0.007 20/09/13 7.68 10.4 18 274 4.4 0.322 0.005 04/10/13 7.78 9.8 20.4 294 4.8 0.116 0.005 18/10/13 7.83 9.1 21.4 303 5.1 0.004 0.007 01/11/13 7.46 8.7 21.6 286 5.9 0.121 0.006 15/11/13 7.32 8.2 21.8 294 6 0.191 0.005 24/01/14 7.19 6.2 25.7 266 3.9 0.007 0.005 07/02/14 7.21 6.6 25.4 268 3.7 0.086 0.004 21/02/14 7.29 6.9 25.5 256 3.1 0.024 0.004 07/03/14 7.22 6.6 25.6 259 3.4 0.064 0.004

CHAPTER 04: Study Area and Data

Table 4.6: Water quality data at Weir Reserve

Date pH DO Tem EC Turbidity NOx NH3 -N (mg/L) (deg C) (us/cm) (NTU) (mg/L) (mg/L) 22/02/13 7.4 8.1 24.2 304 13.1 0.536 0.007 08/03/13 7.42 8.8 23.8 258 7.8 0.326 0.004 22/03/13 7.47 9 23.1 209 6.34 0.262 0.005 05/04/13 7.31 8.4 21 261 6.16 0.376 0.013 19/04/13 7.41 9.2 20.8 282 6.3 0.352 0.011 03/05/13 7.56 9.6 16.4 276 5.8 0.325 0.012 17/05/13 7.59 10.3 13.7 288 5.59 0.431 0.013 31/05/13 7.62 10.5 13.2 264 5.4 0.426 0.011 14/06/13 7.71 10.4 13.4 249 4.8 0.482 0.008 28/06/13 7.68 11.2 12.8 267 5.2 0.428 0.004 12/07/13 7.34 10.9 12.6 253 4.7 0.448 0.005 26/07/13 7.68 11.4 12.3 241 4.6 0.472 0.007 09/08/13 7.73 11.4 12 233 4.2 0.481 0.005 23/08/13 7.71 11.5 11.9 263 3.1 0.408 0.005 06/09/13 7.58 10.6 15.8 266 6.2 0.395 0.004 20/09/13 7.53 10.8 16.8 371 7.4 0.531 0.011 04/10/13 7.71 10.2 18.9 267 5.4 0.321 0.015 18/10/13 7.74 9.6 21.2 269 3.1 0.127 0.019 01/11/13 7.31 9.2 21.6 274 8.3 0.214 0.018 15/11/13 7.42 8.1 21.8 283 15 0.249 0.028 24/01/14 7.5 8 25.3 221 2.2 0.056 0.004 07/02/14 7.66 8.3 25.4 269 2.1 0.229 0.003 21/02/14 7.75 8.2 25.5 298 2.3 0.219 0.005 07/03/14 7.62 8.1 25.3 289 2.2 0.116 0.005

CHAPTER 05: Results

CHAPTER 5

RESULTS AND DISCUSSION ON ASSESSMENT OF RIVER WATER QUALITY 5

5.1 OVERVIEW

This chapter presents the results of the assessment of the water quality of the Hawkesbury

Nepean River System (HNRS) using the water quality data obtained from Sydney Catchment

Authority and the data sampled as a part of this study. At the beginning, preliminary data analyses were performed to explore the general characteristics of the water quality parameters along the HNRS. Principal components and factor analyses were then performed to identify the most significant water quality monitoring stations, water quality parameters and the correlations among the water quality parameters. Thereafter, long term water quality trends were identified by performing Mann–Kendall statistical test and Sen’s slope analysis.

Afterwards, prediction equations for various water quality parameters were developed using multiple linear regression analysis and finally, the water quality index method was used to make an overall assessment of the water quality in the HNRS.

5.2 PRELIMINARY WATER QUALITY DATA ANALYSIS

5.2.1 pH

Figure 5.1 presents the box plot of the observed pH values along the Hawkesbury Nepean

CHAPTER 05: Results

River System (HNRS). It can be seen that station N21 has the highest observed pH value

(11.40). Overall, N92 shows the highest levels of pH values, where median value has exceeded the ANZECC trigger value (upper limit). Higher pH refers to a higher alkaline condition, which is generally attributed to numbers of factors such as weathering of concrete, pavement and other building materials into smaller particles, that are then washed off from the landscape into streams. This could also be partially linked to higher algal growth in the river. Excess alkalinity can cause ammonia toxicity and algal blooms, altering water quality and harming aquatic life.

The lowest observed pH (4.32) can be seen at station N42. Furthermore, pH values are found to be below the ANZECC trigger value (lower limit) for stations E851, N21, N42, N44, N57,

N641, N67 and N92. Lower pH indicates an acidic condition, which can be caused by acid rain, leaching of surrounding acid rocks, mining activities within the catchment and certain wastewater discharges. Low pH can allow toxic elements and compounds to become more mobile and available for uptake by aquatic plants and animals.

The spread (i.e. standard deviation) of the measured pH values is the highest for station N92, followed by N57, while it is the lowest for N86. The skewness is the highest for station N86, followed by N881. Generally, skewness of pH data is very low for most of the stations.

Overall, the observed pH values mostly fall within the ANZECC guideline recommended upper and lower limits; however, there are more cases where pH values are higher than the

ANZECC recommended trigger value (upper limit) compared with the recommended lower limit. The worst case is seen for station N92 (Figure 5.1). The causes for observed higher and

CHAPTER 05: Results lower pH values have not been specifically identified in this study.

H 8 ANZECC upper limit p

6 ANZECC lower limit

1 4 1 5 2 4 7 4 1 7 5 5 6 1 2 5 1 2 3 4 4 5 6 4 6 7 8 8 8 9 8 N N N N N N N 6 N N N N 8 N E N N

Figure 5.1. Box plot of pH values at different measuring stations along the Hawkesbury Nepean River System.

5.2.2 Temperature

Figure 5.2 presents the box plot of observed temperature values along the HNRS.

Temperature does not show any outlier, and much skewness, which implies that the observed variability might be due to seasonal variations. However, the medians and range of the temperature values vary along the river, which can be attributed to changes in weather, shading stream bank vegetation, impoundments, discharge of cooling water, urban

CHAPTER 05: Results stormwater and groundwater inflows to the stream in different parts of the river system. The highest range of temperature (whiskers to 8°C and 33.7°C) can be seen at station N57 and the smallest range can be seen at station N881 (whiskers to 9.4°C and 25.4°C). Station N57 has the highest observed temperature (33.7°C) and station N75 has the lowest (7°C). Temperature values do not show much skewness. It is interesting to note that the river temperature does not follow the extremes of the surrounding land temperature which exceeds 40°C and falls below 2°C a few times in a year. Temperature mainly governs the biological activity in a river e.g. the higher the temperature the greater the biological activity.

)

g 25

(

e r

u 20

p m

e 15 T

E851 N14 N21 N35 N42 N44 N57 N64 N641 N67 N75 N85 N86 N881 N92

Figure 5.2. Box plot of measured temperature along the Hawkesbury Nepean River System.

CHAPTER 05: Results

5.2.3 Dissolved Oxygen

Figure 5.3 presents the box plot of dissolved oxygen (DO) along the HNRS. At all the stations, the 25th percentile DO values are higher than the minimum ANZECC recommended value (5mg/l) of DO, which implies a good water quality condition (in terms of organic pollution) along the river system. In a few cases, the DO values are found to be below the

ANZECC recommended value. The low DO conditions might have been caused by higher sewage discharge, agricultural runoff containing higher organic load and failing septic systems in the rural parts of the catchment.

20 )

L 15

(

e g

y 10

s s

i 5 ANZECC lower limit D

1 4 1 5 2 4 7 4 1 7 5 5 6 1 2 5 1 2 3 4 4 5 6 4 6 7 8 8 8 9 8 N N N N N N N 6 N N N N 8 N E N N

Figure 5.3. Box plot of DO along the Hawkesbury Nepean River System.

CHAPTER 05: Results

It can be seen that station N92, which is the most upstream station, has the highest observed

DO (18.8 mg/l). The spread of the boxes are very similar for all the stations and no remarkable skewness is noticed for any of the monitored stations.

5.2.4 Conductivity

Figure 5.4 and Figure 5.5 present the box plots of conductivity along the HNRS. Site N14 shows a much higher range and a median value of conductivity as compared with other stations. It shows a highest observed reading of conductivity of 48.400 mS/cm and whiskers of 0.009 mS/cm and 28.7 mS/cm. Also, station N21 shows a comparatively high conductivity values. The lowest observed conductivity (0.031 mS/cm) can be seen at N57. Minimum range of data distribution can be seen at stations N86 (whiskers to 0.08 and 0.11). Sampling stations, N42, N57, N64, N85, N86 and N881 show comparatively better conductivity values.

Sites N14, N21, N35, N67, N75 and N92 show a higher median value exceeding the

ANZECC recommended value. Discharges to streams can change the conductivity depending on the water chemistry. A failing sewage system would raise the conductivity because of the presence of chloride, phosphate and nitrate. In contrast, an oil spill would lower the conductivity. Site N14 needs to be further investigated to find the sources of pollutants which contribute to the observed higher conductivity values. However, this was not done in this thesis as it falls beyond its scope.

CHAPTER 05: Results

)

c /

S 30

(

v i

t 20

o C 10

E851 N14 N21 N35 N42 N44 N57 N64 N641 N67 N75 N85 N86 N881 N92

Figure 5.4. Box plot of conductivity along the Hawkesbury Nepean River System for all sampling stations (showing all the observed data range).

CHAPTER 05: Results

3.0

2.5

) m

c 2.0

(

1.5

t c

u 1.0

o C 0.5 ANZECC upper limit (0.35 mS/cm)

0.0

1 4 1 5 2 4 7 4 1 7 5 5 6 1 2 5 1 2 3 4 4 5 6 4 6 7 8 8 8 9 8 N N N N N N N 6 N N N N 8 N E N N

Figure 5.5. Box plot of conductivity along the Hawkesbury Nepean River System for all sampling stations (with the scale up to only 3 mS/cm).

5.2.5 Turbidity

Figure 5.6 and Figure 5.7 present the box plot of the observed turbidity data along the HNRS.

It can be seen that all the median turbidity values are below the ANZECC trigger value.

Sampling stations N14, N35 and N64 show comparatively high turbidity values during the considered period of study. Site N57 has the highest observed turbidity (437 NTU) value.

There are many outliers at all the sites above the ANZECC trigger value, which indicates that at many instances, the turbidity values in the HNRS are too high. These high turbidity values can occur in wet weather conditions. Turbidity often increases sharply during rainfall,

CHAPTER 05: Results especially in developed watersheds, which typically have relatively high proportions of impervious surfaces. The flow of stormwater runoff from impervious surfaces rapidly increases stream velocity, which increases the erosion rates of stream-banks and channels, which can increase turbidity. Turbidity can also rise sharply during dry weather if earth- disturbing activities are occurring in or near a stream without erosion control practices in place, where atmospheric deposition can increase the turbidity in the river. High turbidity values can also arise from large numbers of bottom feeders and excessive algal growth.

The lowest observed turbidity of 0 NTU can be seen at some stations. When the outlier data is ignored, sampling stations N14 and N35 show a comparatively higher range of data distribution (N14: whiskers to 0 and 34.3 and N35: whiskers to 5.2 and 39.1), minimum range of data distribution can be seen at N85 (whiskers to 0.77 and 5.67). At all the stations, the turbidity values are positively skewed.

CHAPTER 05: Results

500

400

)

(

l 300

a L

200

r u T 100

E851 N14 N21 N35 N42 N44 N57 N64 N641 N67 N75 N85

Figure 5.6. Box plot of turbidity along the Hawkesbury Nepean River System (showing all the observed data range).

CHAPTER 05: Results

100

)

(

d 60

a L

r u

T 20 ANZECC upper limit (20 NTU)

1 4 1 5 2 4 7 4 1 7 5 5 5 1 2 3 4 4 5 6 4 6 7 8 8 N N N N N N N 6 N N N E N

Figure 5.7. Box plot of turbidity along the Hawkesbury Nepean River System (with the scale up to only 100 NTU).

5.2.6 Phosphorus

Figure 5.8 and Figure 5.9 present the box plot of total phosphorus and filterable phosphorus along the HNRS, respectively. Stations N21, N35 and N44 show comparatively high total phosphorus and filterable phosphorus values. High phosphorus values can occur due to both natural and human factors. These include soil and rocks, wastewater treatment plants, runoff from fertilized lawns and cropland, failing septic systems, runoff from animal manure storage areas, disturbed land areas, drained wetlands, water treatment and commercial cleaning preparations. Since phosphorus is a key nutrient in most fresh water bodies, even a modest

CHAPTER 05: Results increase in phosphorus can, under the right conditions, set off a whole chain of undesirable effects in a stream including accelerated plant growth, algae blooms, low dissolved oxygen, and the death of certain fish, invertebrates, and other aquatic animals (Boman et. al., 2002).

It can be seen that station N35 has the highest, observed, total phosphorus value (0.380). The lowest, observed phosphorus (0.005) can be seen at many stations. When the outliers are overlooked, sampling station N35 shows a comparatively high range of data distribution

(whiskers to 0.005 and 0.123). The minimum range of data distribution can be seen at stations

N86 and N881 (whiskers to 0.005 and 0.014). All the total phosphorus data are positively skewed. Except stations N14, N21, N35 and N44, the total phosphorus values at all the other stations lie below the ANZECC trigger value.

CHAPTER 05: Results

0.4

) 0.3

(

t o

T 0.2

s o

h 0.1 P

0.0

E851 N14 N21 N35 N42 N44 N57 N64 N641 N67 N75 N85 N86 N881 N92

Figure 5.8. Box plot of total phosphorus along the HNRS (showing all the observed data range).

CHAPTER 05: Results

0.20

) 0.15

(

a t

o 0.10

p s

o 0.05 ANZECC upper limit (0.05 mg/l)

h P

0.00

1 4 1 5 2 4 7 4 1 7 5 5 6 1 2 5 1 2 3 4 4 5 6 4 6 7 8 8 8 9 8 N N N N N N N 6 N N N N 8 N E N N

Figure 5.9. Box plot of total phosphorus along the Hawkesbury Nepean River System (with the scale up to only 0.2 mg/L).

The sampling station N92 has the highest, observed, filterable phosphorus value (0.237). The lowest, observed, filterable phosphorus (0.001) can be seen at many stations. When the outliers are overlooked, sampling station N35 shows a comparatively higher range of data distribution (whiskers to 0.001 and 0.46). The minimum range of data distribution can be seen at stations N881 and N881 (whiskers to 0.001 and 0.006). All the filterable phosphorus data are positively skewed.

CHAPTER 05: Results

0.25

0.20

)

(

l 0.15

s 0.10

s o

h 0.05 P

0.00

E851 N14 N21 N35 N42 N44 N57 N64 N641 N75 N85 N86 N881 N92

Figure 5.10. Box plot of filterable phosphorus along the Hawkesbury Nepean River System (showing all the observed data range).

CHAPTER 05: Results

0.05

0.04

)

(

l 0.03

s 0.02

s o

h 0.01 P

0.00 E851 N14 N21 N35 N42 N44 N57 N64 N641 N75 N85 N86 N881 N92

Figure 5.11. Box plot of filterable phosphorus along the Hawkesbury Nepean River System (with the scale up to only 0.05 mg/L).

5.2.7 Nitrogen

Figures 5.12, 5.13, 5.14, and 5.15 present the box plots of total nitrogen (TN), nitrogen oxidised, ammoniacal nitrogen and total kjeldahl nitrogen (TKN) along the HNRS, respectively.

Figures 5.12 and 5.13 show that station N75 has the highest observed TN (6.73) value. The lowest observed TN (0.01) can be seen at stations N851, N42, N85, N86 and N881. The sampling station N75 shows a comparatively higher range of data distribution (whiskers to

0.01 and 5.9). The minimum range of data distribution can be seen at N86 (whiskers to 0.1

CHAPTER 05: Results and 0.4). At all the stations, TN does not show any notable skewness. Most of the values lie above the ANZECC trigger value. Stations N85, N86 and N881 exhibit comparatively better values.

) 5

g m

( 4

o T

r t

i 2 N

E851 N14 N21 N35 N42 N44 N57 N64 N641 N67 N75 N85 N86 N881 N92

Figure 5.12. Box plot of total nitrogen along the Hawkesbury Nepean River System.

CHAPTER 05: Results

3.0

2.5

) L

/ 2.0

(

l a

t 1.5

n e

g 1.0

i N 0.5 ANZECC upper limit (0.35mg/l)

0.0

1 4 1 5 2 4 7 4 1 7 5 5 6 1 2 5 1 2 3 4 4 5 6 4 6 7 8 8 8 9 8 N N N N N N N 6 N N N N 8 N E N N

Figure 5.13. Box plot of total nitrogen along the Hawkesbury Nepean River System.

Figure 5.14 shows that station N75 has the highest observed Nitrogen oxides (NOx) (5,900 mg/L) value. The lowest observed NOx (0.002) can be seen at some stations. When the outliers are overlooked, sampling stations N35 and N75 show a comparatively higher range of data distribution (N35: whiskers to 0.01 and 3.00 and N75: whiskers to 0.002 and 5.00).

The minimum range of data distribution can be seen at N881 (whiskers to 0.002 and 0.233).

The median values of NOx at stations N35, N44 and N75 are above the ANZECC trigger value. Data does not show any notable skewness.

CHAPTER 05: Results

)

L /

g 4

(

s i

d 3

e 2

i N 1

ANZECC upper limit (0.25 mg/l) 0

1 4 1 5 2 4 7 4 1 7 5 5 6 1 2 5 1 2 3 4 4 5 6 4 6 7 8 8 8 9 8 N N N N N N N 6 N N N N 8 N E N N

Figure 5.14. Box plot of nitrogen oxidised along the Hawkesbury Nepean River System.

For nitrogen ammonical (NH4-N), (Figure 5.15 and 5.16) station N75 has the highest observed NH4-N value (1.070 mg/l). The lowest observed NH4-N (0) can be seen at N35.

When the outliers are ignored, sampling station N75 shows a comparatively higher range of data distribution (whiskers to 0.006 and 0.13). The minimum range of data distribution can be seen at N641 (whiskers to 0.005 and 0.03). NH4-N data is positively skewed and median values at all the stations are below the ANZECC recommended trigger value.

CHAPTER 05: Results

1.2

1.0

)

/ g

m 0.8

(

a i

n 0.6

n 0.4

t i

N 0.2

0.0

N35 N44 N64 N641 N75 N85 N86 N881 N92

Figure 5.15. Box plot of ammoniacal nitrogen along the Hawkesbury Nepean River System (showing all the observed data range).

CHAPTER 05: Results

0.5

0.4

)

(

l 0.3

o m

m 0.2

r t

i 0.1 ANZECC upper limit (0.1 mg/l) N

0.0

N35 N44 N64 N641 N75 N85 N86 N881 N92

Figure 5.16. Box plot of ammoniacal nitrogen along the Hawkesbury Nepean River System (with the scale up to only 0.5 mg/L).

For nitrogen TKN, Figure 5.17 shows that station N75 has the highest observed value (5.40 mg/L). The lowest observed TKN (0) can be seen at N35. Sampling station N75 shows a comparatively higher range of TKN values (whiskers to 0.01 and 1.00), while minimum range of data distribution can be seen at N881 (whiskers to 0.01 and 0.3). TKN data does not exhibit any notable skewness.

CHAPTER 05: Results

5 )

L 4

(

K 3

g o

r 2

i N

E851 N14 N21 N35 N42 N44 N57 N64 N641 N75 N85 N86 N881 N92

Figure 5.17. Box plot of nitrogen TKN along the Hawkesbury Nepean River System.

In all the forms of nitrogen, the sampling stations N35 and N75 show fairly high median and range. This can be due to the discharge from wastewater treatment plants, runoff from fertilized lawns and cropland, failing on-site septic systems, runoff from animal manure storage areas and industrial discharges that contain corrosion inhibitors. Though nitrates are essential plant nutrients, in excess amounts they can cause significant water quality problems.

Together with phosphorus, nitrates in excess amounts can accelerate eutrophication, causing dramatic increase in aquatic plant growth and changes in the types of plants and animals that live in streams. This, in turn, affects dissolved oxygen, temperature, and other indicators. In terms of water treatment, algal growth is highly undesirable as it makes the water toxic. The

CHAPTER 05: Results cost of treating water with algal content is too high and hence water authorities are highly vigilant to identify any early sign of nitrogen and phosphorous increase in raw water.

5.2.8 Alkalinity

Figure 5.18 presents the box plot of alkalinity along the HNRS. Alkalinity in streams is influenced by surrounding rocks and soils, salts, certain plant activities, and industrial wastewater discharges. Except stations N851, N86 and N881, alkalinity readings at all the other stations are higher than the ANZECC trigger value. These high alkalinity values can be due to many factors such as dissolved compounds in rain, soil, sediments, and bedrock and by-products from biological processes in the stream. It can be seen that station N92 has the highest observed alkalinity value (298 mg CaCO3/L). Also, it shows the highest range of data distribution (whiskers to 1.00 and 298.00). The lowest observed alkalinity (1 mg CaCO3/L) can be seen at many stations. The minimum range of alkalinity can be seen at N86 (whiskers to 5.00 and 16.00)

CHAPTER 05: Results

300

250

)

L /

3 200

a C

g 150

(

n i

l 100

l A 50

ANZECC upper value (20 mgCaCO3/l) 0

1 4 1 5 2 4 7 4 1 7 5 5 6 1 2 5 1 2 3 4 4 5 6 4 6 7 8 8 8 9 8 N N N N N N N 6 N N N N 8 N E N N

Figure 5.18. Box plot of alkalinity along the Hawkesbury Nepean River System.

5.2.9 Suspended solids

Figure 5.19 and Figure 5.20 present the box plot of suspended solids (SS) along the HNRS. It can be seen that many higher concentration of SS values have been plotted as outliers. As the turbidity and SS often increase sharply during rainfall, especially in developed watersheds, which typically have relatively high proportions of impervious surfaces. The flow of stormwater runoff from impervious surfaces rapidly increases stream velocities which increase the erosion rates of stream-banks and channels. This can also rise sharply during dry weather if earth-disturbing activities are occurring in, or near, a stream without erosion control practices in place. It can be seen that station N67 has the highest observed SS value

CHAPTER 05: Results

(360 mg/L). The lowest observed SS (.05 mg/l) can be seen at N75. When the outliers are ignored, sampling stations N14 and N35 show a comparatively higher range of data distribution (whiskers to 1.00 and 32.00), while minimum range of data distribution can be seen at N57 (whiskers to 1.00 and 4.00). Most of the SS values are found to be below the

ANZECC recommended trigger value.

400

) 300

(

i l

o 200

p s

u 100 S

E851 N14 N21 N35 N42 N44 N57 N64 N641 N67 N75 N85 N86 N881 N92

Figure 5.19. Box plot of suspended solids along the Hawkesbury Nepean River System (showing all the observed data range).

CHAPTER 05: Results

)

m (

e 20 ANZECC upper limit (20 mg/l)

s u

S 10

1 4 1 5 2 4 7 4 1 7 5 5 6 1 2 5 1 2 3 4 4 5 6 4 6 7 8 8 8 9 8 N N N N N N N 6 N N N N 8 N E N N

Figure 5.20. Box plot of suspended solids along the Hawkesbury Nepean River System (with scale up to only 50 mg/L).

5.2.10 Algae and chlorophyll-a

Figures 5.21 to 5.24, displaying box plots of total algal count and chlorophyll-a, clearly show the relationship between algal count and chlorophyll-a. It can be seen that the sampling stations N21 and N35 show a comparatively higher median and a range.

When considering Figures 5.21 and 5.22, it can be seen that station N92 has the highest, observed algal count (633,800 cells/mL). The lowest, observed algal count (314 cells/mL) can be seen at N14. When the outliers are overlooked, sampling stations N21 and N35 show a comparatively higher range of data distribution (N21: whiskers to 1900 and 180303 and N35:

CHAPTER 05: Results whiskers to 1011 and 163269). The minimum range of data distribution can be seen at

E851(whiskers to 1744 and 12097).

700000

600000

) L

m 500000

e c

( 400000

o C

300000

l 200000

l A 100000

E851 N14 N21 N35 N42 N44 N57 N64 N641 N67 N75 N85 N92

Figure 5.21. Box plot of algal total count along the Hawkesbury Nepean River System (showing all the observed data range).

CHAPTER 05: Results

200000 )

L 150000

(

n 100000

a 50000

l A

E851 N14 N21 N35 N42 N44 N57 N64 N641 N67 N75 N85 N92

Figure 5.22. Box plot of algal total count along the Hawkesbury Nepean River System (with the scale up to only 200,000 cells/mL).

Considering Figure 5.23 and Figure 5.24, it can be seen that station N21 has the highest observed chlorophyll-a (253.1 ug/L). The lowest observed chlorophyll-a (0 ug/L) can be seen at N57, N85 and N92. When the outliers are overlooked, sampling stations N21 and N35 show a comparatively higher range of data distribution (N21: whiskers to 0.1 and 46.3 and

N35: whiskers to 0.2 and 49.5). Except at stations N85 and N881, chlorophyll-values are positively skewed, and also, most of the observed values are above the ANZECC trigger value.

CHAPTER 05: Results

250

200

)

g u

( 150

h p

o 100

h C

E851 N14 N21 N35 N42 N44 N57 N64 N641 N67 N75 N85 N86 N881 N92

Figure 5.23. Box plot of chlorophyll-a along the Hawkesbury Nepean River System (showing all the observed data range).

CHAPTER 05: Results

)

/ g

u 30

(

y h

p 20

h C 10

ANZECC upper limit (5 ug/l)

1 4 1 5 2 4 7 4 1 7 5 5 6 1 2 5 1 2 3 4 4 5 6 4 6 7 8 8 8 9 8 N N N N N N N 6 N N N N 8 N E N N

Figure 5.24. Box plot of chlorophyll-a along the Hawkesbury Nepean River System (with the scale up to only 50 ug/L).

CHAPTER 05: Results

5.3 RESULTS FROM PRINCIPAL COMPONENT ANALYSIS (PCA)

When 15 monitoring stations were reduced to three principal components, it explained 95.2% of the total variance and the rest of the 12 components only accounted for 4.8%. Further, the first, second and third components (PC 1, PC 2 and PC3) accounted for about 79.6%, 8.8% and 6.6% of the total variance in the data set, respectively. Therefore, only the first three principal components are focused in this thesis as they contain the bulk of the data information.

Item PC 1 PC 2 PC 3 Eigenvalue 11.960 1.3328 0.993 Variance (%) 79.731 8.855 6.620 Cumulative variance (%) 79.731 88.586 95.206

Table 5.1: Principal components with eigenvalues > 1

Item PC 1 PC 2 PC 3 Eigenvalue 11.960 1.3328 0.993 Variance (%) 79.731 8.855 6.620 Cumulative variance (%) 79.731 88.586 95.206

The first component has almost equal loadings on all the stations (Table 5.2). Therefore, it is a measure of overall performance of the stations. It also shows an extremely high correlation with the stations. It accounts for 79.7% of the data variance (Table 5.1). Similarly, the second and third components have different loadings on different stations. Hence, PC 2 and PC 3 represent a difference among the stations. Loading reflects only the relative importance of a variable (station) within a component, and does not reflect the importance of the component

CHAPTER 05: Results itself (Davis, 1986).

The results of the first PCA identify three important components that account for 95.2% of the variance in the dataset.

Table 5.2: Component score coefficients for first three PCs (for monitoring stations)

Variable/station PC 1 PC 2 PC 3 E852 0.218 0.316 -0.386 N14 0.265 -0.221 -0.176 N21 0.243 -0.116 -0.338 N35 0.272 -0.081 0.098 N42 0.287 0.09 0.03 N44 0.249 0.094 0.489 N57 0.229 0.145 0.572 N64 0.279 -0.106 -0.197 N641 0.284 -0.031 -0.061 N67 0.278 -0.17 0.15 N75 0.274 -0.261 0.022 N85 0.284 -0.073 0.076 N86 0.234 0.489 0.006 N881 0.21 0.547 -0.207 N92 0.251 -0.374 -0.119

Table 5.3 demonstrates the rotated factor correlation coefficient (obtained from factor analysis) for 15 water quality monitoring stations. In this study, the factor correlation coefficient is considered to be significant if the value is greater than 0.7. This conservative

CHAPTER 05: Results criterion is selected because the study area is relatively large and the HNRS is deemed to be highly non-linear and dynamic in nature. As it can be clearly seen from Table 9.8 , water quality monitoring stations N14, N64, N641, N67, N75, N85, N86, N881 and N92 have coefficient values greater than 0.70, and hence these are considered to be the most important water quality monitoring stations.

Table 5.3: Varimax rotated factor loadings (for first 5 factors)

Variable/station Factor 1 Factor 2 Factor 3 Factor 4 Factor 5

E852 0.378 0.672 0.125 -0.19 0.594 N14 0.766 0.265 0.27 -0.452 0.081 N21 0.582 0.339 0.147 -0.717 0.098 N35 0.555 0.267 0.566 -0.524 0.118 N42 0.621 0.536 0.498 -0.266 0.07 N44 0.404 0.303 0.85 -0.128 0.058 N57 0.293 0.288 0.904 -0.099 0.033 N64 0.818 0.432 0.248 -0.27 0.094 N641 0.768 0.473 0.373 -0.185 0.07 N67 0.776 0.244 0.537 -0.174 0.082 N75 0.842 0.189 0.424 -0.244 0.116 N85 0.749 0.368 0.498 -0.175 0.11 N86 0.261 0.846 0.43 -0.12 0.037 N881 0.195 0.926 0.238 -0.184 0.074 N92 0.946 0.119 0.227 -0.16 0.111

CHAPTER 05: Results

The results of PCA on the water quality parameters dataset give eight principal components with eigenvalues > 1, explaining about 72.7% of the total variance in the data set. The first

PC (PC 1) accounts for 24.1% of the total variance of the data, which is highly correlated

(loading > 0.7) with total iron (TI), true color (TCOL), turbidity, aluminum total and UV absorbent. Whereas, the other seven PCs, although account for 12.7%, 8.3%, 7.3%, 6.6%,

5.2%, 4.4% and 3.8% variances, respectively, show little correlation (loading > 0.7) with none of the parameters (Table 5.4 and 5.5).

Principal components extracted for water quality parameters do not have a strong correlation when comparing with principal components extracted for the water quality monitoring stations. Monitoring stations are primarily controlled by hydrological conditions, while water quality parameters are controlled by a combination of hydrological, chemical, physical and biological conditions, so it is expected that the monitoring stations would have a higher correlation than the water quality parameters.

Table 5.4: Explained variance and eigenvalues (for water parameters)

Item PC 1 PC 2 PC 3 PC 4 PC 5 PC 6 PC 7 PC 8 Eigenvalue 6.75 3.55 2.34 2.06 1.85 1.46 1.24 1.07 Variance (%) 24.13 12.71 8.36 7.38 6.61 5.21 4.44 3.83 Cumulative variance (%) 24.13 36.84 45.20 52.59 59.20 64.41 68.85 72.69

CHAPTER 05: Results

The stations N14, N64, N641, N67, N75, N85, N86, N881 and N92 were found to be the most significant sampling stations explaining the most variation in the water quality data in the Hawkesbury-Nepean River System. This result might be used to reduce the number of sampling stations in the river system. Principal component analysis allowed deriving three principal components which explained more than 90% of the total variance in the data set.

The stations N14, N64, N641, N67, N75, N85, N86, N881, N92, N57 and N21were found to be the most significant sampling stations explaining the most variation in the water quality data in the Hawkesbury- Nepean River System. This result might be used to reduce the number of sampling stations in the river system. Principal component analysis allowed the derivation of three principal components which explained more than 90% of the total variance in the data set.

CHAPTER 05: Results

Table 5.5: Component loadings for first eight PCs (water quality parameters)

Variable PC 1 PC 2 PC 3 PC 4 PC 5 PC 6 PC 7 PC 8 PH -0.404 0.450 -0.092 0.065 -0.148 -0.036 0.351 -0.464 LOR 0.052 0.402 0.080 -0.599 0.517 0.092 0.207 -0.112 TI 0.907 -0.133 0.080 -0.078 -0.007 0.179 -0.016 -0.030 PHA 0.124 0.378 0.032 -0.295 0.045 0.081 -0.186 0.202 TKN 0.322 0.515 -0.239 -0.102 -0.038 -0.218 0.081 0.042 TEMP 0.059 0.168 0.487 -0.315 -0.260 -0.629 0.089 0.019 CHLA 0.085 0.507 0.080 -0.630 0.502 0.089 0.141 -0.059 ECOL 0.459 0.169 0.250 0.324 -0.167 0.170 0.426 0.183 FI 0.504 -0.554 -0.220 -0.236 -0.011 0.050 0.059 -0.161 TCOL 0.754 -0.342 -0.207 -0.016 0.195 -0.296 0.061 -0.134 TN 0.172 0.618 -0.665 -0.017 -0.110 -0.063 -0.073 0.250 TUR 0.748 0.294 0.288 0.203 0.013 0.205 0.101 0.055 ALK -0.236 0.482 -0.157 -0.003 -0.424 -0.122 0.143 -0.448 TA 0.802 0.220 0.170 0.272 0.138 0.052 -0.080 -0.064 TM 0.553 -0.207 0.038 -0.535 -0.383 0.264 0.017 0.017 DO -0.243 -0.042 -0.491 0.258 0.380 0.530 0.134 -0.213 ECOCC 0.450 0.201 0.225 0.328 -0.188 0.156 0.504 0.180 TP 0.700 0.428 0.063 0.107 0.066 -0.025 -0.166 -0.087 SS 0.605 0.409 0.361 0.033 0.076 0.220 -0.295 -0.069 NO 0.061 0.510 -0.694 0.026 -0.117 0.022 -0.123 0.283 FA 0.487 -0.288 -0.198 0.208 0.332 -0.232 0.008 -0.254 FM 0.486 -0.430 -0.171 -0.415 -0.405 0.287 0.096 -0.027 EC -0.046 0.116 0.305 0.046 -0.127 0.193 -0.554 -0.219 NH-N 0.498 -0.038 -0.345 -0.222 -0.477 0.044 -0.059 -0.144 FP 0.527 0.395 -0.082 0.275 -0.110 -0.132 -0.213 -0.259 RS 0.570 -0.352 -0.272 0.106 0.165 -0.202 -0.033 0.079 DOC 0.117 0.039 -0.041 -0.014 0.092 -0.170 0.004 0.266 UV 0.742 -0.155 -0.122 0.010 0.211 -0.303 0.096 -0.036

CHAPTER 05: Results

5.4 LONG TERM TRENDS IN WATER QUALITY DATA

Median values of various water quality parameters are compared against the ANZECC

(2000) guidelines for fresh water. The rank-based non-parametric Mann–Kendall (MK) statistical test is used to assess the trends in water quality parameters. The MK test is performed at a significance level of 0.05.

The median water quality parameters and the corresponding ANZECC (2000) trigger values are presented in Table 5.6, where the medians above the trigger values are marked in red. The trend test results for all the water quality parameters at each station are summarised in Table

5.7, in which the detected trends are represented by arrows, with an upward arrow to indicate an upward trend and a downward arrow for a downward trend. Dash (-) designates no detected significant trend.

CHAPTER 05: Results

Table 5.6: Median values of water quality parameters and ANZECC (2000) guidelines

ANZECC Variable trigger N14 N21 N35 N42 N44 N57 N67 N75 N92 values PH 7.47 7.60 7.50 7.69 7.70 7.77 7.79 7.88 8.21 6-8 TEMP 20.90 21.10 20.90 20.50 20.90 20.70 20.50 20.70 19.05 DO 7.85 8.70 8.00 9.00 8.60 9.15 8.40 9.00 9.47 mini 5 EC 6.28 0.34 0.40 0.25 0.34 0.28 0.57 0.54 0.36 0.35 SS 12.00 8.00 12.00 2.00 2.00 2.00 4.00 3.00 2.00 20 TUR 11.00 8.31 13.60 2.42 2.70 1.76 5.13 3.80 1.40 20 TCOL 8.00 10.50 12.00 10.00 11.00 9.00 10.00 11.00 10.00 15 TN 0.40 0.42 0.82 0.50 0.70 0.35 0.61 1.20 0.43 0.350 NO 0.12 0.05 0.44 0.26 0.40 0.07 0.26 0.77 0.20 0.250 NH-N 0.01 0.01 0.02 0.01 0.02 0.01 0.01 0.02 0.01 1.000 TKN 0.27 0.30 0.40 0.24 0.34 0.25 0.36 0.44 0.25 TP 0.02 0.02 0.04 0.01 0.02 0.01 0.02 0.02 0.01 0.050 FP 0.01 0.01 0.01 0.01 0.01 0.00 0.01 0.01 0.01 20 CHLA 8.55 18.90 18.20 5.10 6.20 3.80 5.80 8.50 3.00 5 ALK 44.75 32.75 46.00 32.00 49.00 40.00 81.50 83.50 122.00 20 DOC 4.00 4.30 5.00 4.00 4.90 4.40 5.10 5.20 4.00 TI 0.43 0.34 0.53 0.28 0.18 0.19 0.24 0.19 0.15 0.300 FI 0.05 0.05 0.05 0.12 0.07 0.08 0.05 0.05 0.08 TA 0.24 0.14 0.25 0.04 0.04 0.02 0.10 0.06 0.03 0.200 FA 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.020 TM 0.04 0.04 0.06 0.03 0.04 0.03 0.07 0.03 0.02 0.100 FM 0.01 0.00 0.01 0.01 0.02 0.01 0.02 0.00 0.00 RS 1.21 0.80 1.40 2.23 1.40 1.82 1.50 1.93 1.59 ECOL 12.00 5.00 23.00 11.00 46.00 55.00 22.00 23.00 9.00 ECOC 6.00 6.00 26.00 20.00 53.00 50.00 40.00 20.00 11.00

CHAPTER 05: Results

Table 5.7: Mann-Kendal test results and yearly Sen’s slope

pH TEMP DO EC SS TUR TCOL TN NO NH-N TKN TP FP CHLA ALK DOC TI FI TA FA TM FM RS ECOL ECOC N14 ↓ ↑ ↓ ↓ ↓ ↑ ↑ ↑ ↑ − − − − ↑ ↑ ↑ ↑ ↑ − − − − ↓ ↓ ↑ 0.065 0.307 0.109 3.502 0.317 1.040 1.391 0.003 0.003 1.625 6.742 0.442 0.039 0.005 0.029 1.131 0.723

N21 ↓ − ↓ ↓ − ↑ ↑ − − − ↓ − − ↑ ↑ ↑ ↑ ↑ ↑ − ↓ − ↑ ↑ ↑ 0.073 0.148 0.060 0.372 1.084 0.013 0.026 2.569 0.143 0.036 0.018 0.003 0.003 0.008 0.406 0.499

N35 ↓ ↑ ↓ ↓ ↑ ↑ ↑ ↓ ↓ − ↓ − − ↓ ↑ ↑ ↑ ↑ ↑ − ↑ ↑ ↑ ↑ ↑ 0.078 0.130 0.317 0.023 0.801 1.438 0.998 0.109 0.081 0.018 0.751 2.954 0.325 0.062 0.016 0.013 0.008 0.008 0.239 3.000 1.009

N42 ↓ ↑ ↓ ↓ − ↑ ↑ ↓ ↓ − ↓ − − ↑ ↑ ↑ ↑ ↑ ↑ − ↑ ↑ ↑ ↑ ↑ 0.047 0.143 0.060 0.023 0.484 1.084 0.049 0.047 0.008 0.497 2.785 0.122 0.042 0.018 0.013 0.003 0.003 0.122 1.856 3.206

N44 ↑ ↑ ↑ ↓ − ↑ ↑ ↓ ↓ ↓ ↓ − − ↑ ↑ ↑ ↑ ↑ ↑ − ↑ ↑ ↑ ↓ ↓ 0.013 0.281 0.185 0.031 0.380 1.040 0.096 0.081 0.003 0.010 1.022 4.228 0.096 0.036 0.016 0.008 0.003 0.003 0.096 1.999 2.600 N57 ↓ − ↑ ↓ − ↑ ↑ ↑ ↑ − − − − ↑ ↑ ↑ ↑ ↑ ↑ − ↑ − ↑ ↑ ↓ 0.073 0.083 0.010 0.364 0.634 0.013 0.003 0.736 2.642 0.224 0.031 0.013 0.005 0.003 0.101 12.667 0.702

N67 ↓ − − ↓ − ↑ ↑ ↓ ↑ ↓ ↓ − − ↑ ↑ − ↑ ↑ ↑ − ↑ ↑ ↑ ↑ ↑ 0.065 0.070 0.983 0.650 0.039 0.003 0.003 0.029 0.679 10.517 0.055 0.016 0.010 0.005 0.003 0.096 3.588 4.703

N75 ↓ ↑ ↓ ↓ − ↑ ↑ ↓ ↓ ↓ ↓ ↓ ↓ ↓ ↓ ↑ ↑ ↑ ↑ − ↑ − ↑ ↑ ↑ 0.125 0.354 0.231 0.081 0.676 0.541 0.530 0.421 0.005 0.104 0.003 0.003 0.198 12.467 0.075 0.055 0.021 0.013 0.005 0.026 6.323 3.182

N92 ↓ ↑ ↑ ↓ − ↓ ↑ ↓ ↓ ↓ ↓ − − ↓ ↑ ↑ ↑ ↑ ↑ − ↑ − ↑ ↑ ↑ 0.114 0.195 0.109 0.068 0.052 1.019 0.075 0.039 0.003 0.023 0.406 25.732 0.216 0.029 0.234 0.003 0.003 0.213 1.547 2.062

CHAPTER 05: Results

The median values of pH at all the stations except N92 are within the ANZECC recommended trigger values (i.e. between 6 and 8). At station N92, the pH is 2.6% above the upper limit of the trigger value. The change in the median value of pH along the river is presented in Figure 5.25, which shows that pH reduces from upstream to downstream of the

HNRS. This result shows an increasing acidification from upstream to downstream of the

HNRS. The pH shows a decreasing trend for all the stations except for station N44. The maximum decrease in trend is found for station N75 (0.125 per year). The overall decreasing trend of pH indicates an increasing acidification of water in the HNRS over the last decade.

Figure 5.25. Median values of pH along the Hawkesbury Nepean River System.

The median values of dissolved oxygen (DO) at all the 9 stations are above the ANZECC trigger value. The DO has a decreasing trend for stations N14, N21, N35, N42 and N75. Its maximum decreasing trend of 0.317 mg/L per year can be seen for station N35 (Figure 5.26).

CHAPTER 05: Results

The upstream of N35 is affected by quality and magnitude of flows coming from the South

Creek and discharge from North Richmond sewage treatment plant (STP). The dominant land use in this part of the catchment includes rural, grazing, commercial gardening, intensive agriculture and urban and industrial activities. These land uses can be attributed to the decreasing trend of DO at N35. The increasing trends of DO for N92 and N57 demonstrate the influence of natural undeveloped catchment conditions at upstream of these two stations.

N35 - Dissolved Oxygen (mg/L) 14

Figure 5.26. Decreasing trend of DO at station N35.

The median values for electrical conductivity are higher than the trigger values for stations

N14, N35, N67, N75 and N92. In particular, the median value of station N14 is 10 times higher than the ANZECC (2000) trigger value, which is significantly higher than any of the

CHAPTER 05: Results other stations. Figure 6 shows that the electrical conductivity value at station N14 has decreased significantly over time and the most current results in year 2012 are much smaller than those of 2002 to 2008. Electrical conductivity has a decreasing trend for all the stations, with a maximum decreasing trend of 3.5 mS/cm per year at station N14 (Figure 5.27). Its overall decreasing trend for all the 9 stations demonstrates an overall improvement of the

HNRS water quality with time, in terms of the total solids dissolved in water.

N14 - Conductivity Field (mS/cm) 60

Figure 5.27. Decreasing trend of EC at station N14.

The median suspended solids (SS) are within the ANZECC (2000) trigger value for all the 9 stations. For most of the stations, SS does not show any trend; however, it has a decreasing trend for station N14 (0.317mg/L per year) and an increasing trend for N35 (0.801 mg/L per

CHAPTER 05: Results year). The low SS levels in the river indicate that the river water is not notably polluted with particulate matter, which is a positive aspect of the water quality of the HNRS.

The median values for turbidity are well within the ANZECC (2000) trigger value for all the

9 stations. However, it has an increasing trend at all the stations except N92. It should be noted that station N92 is located at the most upstream part of the river among all the 9 stations. This part of the river has the lowest level of anthropogenic activity as it has the smallest degree of urbanisation and industrialisation. As a result, it has the lowest turbidity level (1.40 NTU) and an overall decreasing trend. The increasing trend of turbidity for the 8 out of 9 stations demonstrates the influence of increasing urbanisation and industrialisation within the downstream parts of the catchment that has intensified over recent times.

The median values for total nitrogen (TN) are above the ANZECC (2000) trigger value for 8 stations out of 9, which are N14 (14.2%), N21 (18.5%), N35 (134.2%), N42 (42.8%), N44

(100%), N67 (74.2%), N75 (242.8%) and N92 (2.8%). Also, the TN shows an increasing trend for stations N14 and N57. At stations N35, N42, N44, N67 and N74, the median values for oxidised nitrogen are above the ANZECC (2000) trigger value by 76.0%, 3.6%, 60.0%,

4.0% and 208.0% respectively. It has an increasing trend for N14, N57 and N67. Ammonical nitrogen shows a decreasing trend or no trend for all the stations. All the median values are within the ANZECC (2000) guidelines. Nitrogen TKN has decreasing trends for all the stations except for stations N14 and N57; the maximum slope of 0.104 mg/L can be seen for station N75. Considering the median values, it may be stated that NOx is the main contribution for the high value of TN. The median value of TN for N75 is 242% higher than

CHAPTER 05: Results the ANZECC (2000) trigger value, which appears to be associated with the intensive agricultural activities in the upstream catchment parts of station N75. The reduction of TN from N75 to N67 by 49%, and from N67 to N57 by 43%, can be attributed to the natural pristine undeveloped condition of the HNRS in between stations N75 to N67 and N57.

Furthermore, the agricultural activities at upstream of N44 has possibly increased the TN value at N44. The overall TN levels in the HNRS are notably higher than the ANZECC

(2000) trigger value, which is likely to make the river prone to eutrophication.

The medians total phosphorus and filterable phosphorus levels are within the ANZECC

(2000) trigger value for all the stations; however, for station N35, the total phosphorus level is very close to the trigger value (0.04 versus 0.05). No station shows a significant trend for total phosphorus except N75, which shows a decreasing trend.

The determination of photosynthetic chlorophyll pigments and their degradation products is one of the most frequently performed analyses in aquatic ecology (Gitelson, 1992). The median values for chlorophyll-a are above the ANZECC (2000) trigger value for 7 stations out of 9. Stations N92 and N57, which flow through natural undeveloped parts of the catchment, have median values within the ANZECC (2000) trigger value. The median value of chlorophyll-a for station N75 is 70% higher than the trigger value. It has been reduced by

68% while flowing through the natural undeveloped parts of the catchment between N75 and

N67. It is expected that water quality would be improved at N67 because of nutrient assimilation and loss processes while traveling this section of the catchment without further input of nutrients. The median value for chlorophyll-a has been further improved at station

CHAPTER 05: Results

N57 demonstrating further assimilation of nutrients while flowing through a pristine catchment part which is largely undeveloped. The Warragamba River joins the Nepean River in this section, carrying discharge from the Wallacia STP as well as environmental flow release from the Warragamba dam. Nutrients that enter via Matahil Creek from the West

Camden plant and via the Warragamba River from Wallacia plant experience long residence time and distance for assimilation, as well as dilution by low nutrient water from

Warragamba dam. When considering the median values for chlorophyll-a at stations N35 and

N21 (which are 374% and 378% higher than the ANZECC (2000) trigger value), it can be seen that, industrialization, urban developments and agricultural activities in the catchment have contributed in degrading the water quality. The land use at upstream parts of the catchment of N35 predominantly includes rural, grazing and market gardening, intensive agriculture, such as poultry farming, and both urban and industrial activities. Also, it receives water from the South Creek tertiary treated wastewater discharges originated from three

STPs. High nutrient levels, tidal influences, high residence times and low flows make the streams ideal for excessive algal growth and hence very high chlorophyll-a levels are noticed at N35 and N21. Figure 5.28 presents how the median values of chlorophyll-a have changed along the HNRS, which shows a remarkably high peak at stations N35 and N21. However, it is a good sign that station N35 shows a downward trend for clorophyll-a.

CHAPTER 05: Results

Figure 5.28. Median values of chlorophyll-a along the Hawkesbury Nepean River System.

Station N14 is located just before the confluence with the Macdonald River. The water quality of this station is influenced by flow from the Colo River and downstream of the

Hawkesbury River. The Colo River catchment is the best in terms of nutrient enrichments among all the other sub-catchments of the HNRS because it consists primarily of pristine and undisturbed catchment areas. About 80% of these catchments are national parks of the Blue

Mountains world heritage area. There are also limited upstream areas that support agricultural activities. Water quality at station N14 has been improved as expected because of dilution by high quality inflows from the Colo River and the undisturbed upstream catchment. Algae growth, and thus chlorophyll-a level, has directly been affected by the amount of nutrients in

CHAPTER 05: Results the river (e.g. Station 35 has very high chlorophyll-a level and it has the highest total phosphorous level and the second highest total nitrogen level among the 9 stations). Low levels of chlorophyll-a suggest a good river health; however, high levels are not necessarily bad; it is the long-term persistence of high levels that is a problem (NLWRA, 2008). It should also be noted that 6 out of 9 stations show an increasing trend for chlorophyll-a, indicating an overall deterioration of water quality in the HNRS over the last decade.

The median values of alkalinity are found to be above the ANZECC trigger value for all the 9 stations, N14 (123.7%), N21 (63.7%), N35 (130%), N42 (60%), N44 (145%), N57 (100%),

N67 (307.5%), N75 (317.5%) and N92 (510%). It has an increasing trend for 8 of the stations out of 9, with a high Sen’s slope. The maximum trend is found for station N92

(25.7mg/L per year) (Figure 5.29), which has a median value of 510% above the ANZECC trigger value. It should be noted that station N92 is located the most upstream among all the 9 stations, and the highest level of alkalinity at this station is somewhat unexpected, which needs further investigation (but not done in this study).

CHAPTER 05: Results

N92 - Alkanility (mgCaCO3) 70

Figure 5.29. Increasing trend of alkalinity at station N92.

Dissolved organic carbon shows an increasing trend for 8 out of the 9 stations. Organic carbon occurs as a result of decomposition of plant or animal materials. Total aluminium has an increasing trend for all the stations except for N14. The median values are within the

ANZECC (2000) trigger value for all the stations except N14 and N35, which are 20% and

25% above the ANZECC (2000) trigger value, respectively. Aluminium filtered does not show any trend for most of the stations. It was found in a study of “the water quality of

Roanoke River, Virginia”, that the sewage treatment plants were the most significant anthropogenic contributors of aluminium to the river (Butcher, 1988). Total manganese shows an increasing trend at all the stations except for N21 and N14. Its median values are

CHAPTER 05: Results within the ANZECC (2000) trigger value for all the stations. Manganese filtered shows an increasing trend for most of the stations. Reactive silicate shows an increasing trend for all the stations except for N14. It has the maximum increasing trend at N35 (Figure 5.30). The ratios between silicate and phosphorous, and silicate and nitrogen largely determine which algae would dominantly be present in the river water. Water moving over and through natural deposits is expected to dissolve a small amount of various silicate minerals. The overall increasing trends of aluminium, manganese and reactive silicate demonstrate the influence of intensified land use in recent years that has occurred along the HNRS.

N35 - Silicate Reactive (SiO2 mg/L) 7

Figure 5.30. Increasing trends of reactive silicate at station N35.

CHAPTER 05: Results

Trend analysis has been done for 9 water quality parameters and 25 sampling stations. Only dominant water quality parameters and stations of concern have been discussed in detail; however, the trends and their significance levels are presented in Table 5.7 for all the parameters and stations.

5.5 RESULTS FROM REGRESSION ANALYSIS FOR DEVELOPING PREDICTION EQUATIONS FOR WATER QUALITY PARAMETERS

Pearson correlation coefficients among various water quality parameters are provided in

Table 5.8. There are a number of high correlations which are of significance, as noted below.

Nitrogen total is highly correlated with nitrogen (oxidized) (Pearson correlation coefficient, r

= 0.976), which implies that most of the nitrogen in water remains in oxidized form. Nitrogen

(oxidized) has a strong negative correlation (r = -0.716) with temperature, which implies that nitrogen (oxidized) reduces as temperature increases. Total nitrogen (TN) is highly correlated with conductivity (r = 0.698), which implies that dissolved minerals in water are largely of nitrogen-based. Algal (total count) is highly correlated with suspended solids (SS) (r =

0.611). This implies that total SS in water contains a notable proportion of algae.

CHAPTER 05: Results

Table 5.8: Correlations among water quality parameters at station N44 of the HNRS

Nitrogen Algal Phosphor Nitrogen Temperat Chlorophy Nitrogen Dissolved Phosphor Suspende Nitrogen Conductiv pH Ammonia Total us TKN ure ll-a Total Oxygen us Total d Solids Oxidised ity cal Count Filterable

pH 1

Nitrogen TKN 0.166 1

Temperature 0.216 0.335 1

Chlorophyll-a -0.114 0.08 0.072 1

Nitrogen Total -0.138 0.182 -0.632 -0.162 1

Dissolved 0.443 -0.222 -0.29 0.387 0.049 1 Oxygen

Phosphorus -0.255 0.268 0.219 0.355 0.263 -0.047 1 Total

Suspended -0.048 -0.108 0.007 0.552 0.402 0.226 0.303 1 Solids

Nitrogen -0.177 -0.037 -0.716 -0.185 0.976 0.097 -0.321 -0.386 1 Oxidised

Conductivity 0.301 0.455 -0.223 -0.185 0.698 -0.005 -0.365 -0.382 0.608 1

Nitrogen 0.049 0.472 0.113 -0.199 0.124 -0.342 0.129 -0.188 0.021 0.343 1 Ammoniacal

Algal Total 0.103 0.023 0.213 0.819 -0.317 0.358 0.263 0.611 -0.329 -0.180 -0.135 1 Count

Phosphorus Filterable -0.026 0.306 0.303 -0.425 0.032 -0.361 0.219 -0.204 -0.034 0.080 0.422 -0.406 1

CHAPTER 05: Results

Prediction equations are developed using multiple linear regression analysis for chlorophyll- a, total nitrogen (TN) and total phosphorous (TP). The prediction equation for chlorophyll-a is presented by Equation 5.1. The multiple R of this equation is 0.827, coefficient of determination (R2) is 0.683 and the standard error of estimate is 4.312. The prediction equation for TN is expressed by Equation 5.2; the multiple R of this equation is 0.656, R2 is

0.430 and the standard error of estimate is 0.24. The prediction equation for TP is expressed by Equation 5.3, the multiple R in this case is 0.767, R2 is 0.589 and the standard error of estimate is 0.007. The plots of standardized residuals and predicted values of these three equations are shown in Figures 5.31, 5.32 and 5.33, respectively. Equations 5.4.1 to 5.4.3 show quite high R2 values. The plots of standardized residuals and predicted values do not show any trend, which indicate that the developed prediction equations satisfy the assumptions of the least squares regression quite well.

Chl-a = 3.995 + 80.051(TP) – 556.155(FP) – 3.274(EC) + 1.594(SS) + 10.128(TKN) (5.1)

TN =1.296 + 0.940(EC) + 0.003(SS) – 0.038(temp) (5.2)

TP = 0.031 – 0.004(pH) + 0.001(temp) + 0.003(SS) + 0.015(EC) (5.3)

CHAPTER 05: Results

Figure 5.31. Plot of standardized residuals against estimate for Chlorophyll-a.

Figure 5.32. Plot of standardized residuals against estimate for total nitrogen.

100

CHAPTER 05: Results

Figure 5.33. Plot of standardized residuals against estimate for total phosphorous.

5.6 RESULTS OF WATER QUALITY ASSESSMENT BY USING WATER QUALITY INDEX

For the calculation of CCME WQI, 12 water quality parameters were selected based on the importance and the availability of data. These selected water quality parameters and

Australian and New Zealand Guidelines for Fresh and Marine Water Quality (ANZECC) are presented in Table 5.9.

101

CHAPTER 05: Results

Table 5.9: Water quality parameters and ANZEC Guidelines for Fresh and Marine Water Quality

Non-compliance Water Quality Parameter Value1 Value 2 Unit if: pH <> 6 8 Nitrogen Total > 0.35 mg/L Phosphorous Total > 0.05 mg/L Chlorophyll > 5 µg/L Dissolved Oxygen < 5 mg/L Turbidity > 20 NTU Iron Total > 0.3 mg/L Aluminium Total > 0.2 mg/L True Colour > 15 Alkalinity > 20 Suspended Solids > 20 Conductivity > 0.35 mS/cm

WQIs were primarily developed for each year for the 9 sampling locations to investigate the water quality changes along the HNRS over time. Figure 5.34 shows an improvement of water quality over time for most of the stations. Also, it shows a marginal water quality with

WQI in between 45 and 64 for all the stations except N14 and N35, which have WQIs less than 40 at many years.

Medians of CCME WQI values for the 21 years range from 33 to 57. All the monitoring stations indicate marginal or poor water quality. Water quality at N21, N42, N44, N57 and

N92 is frequently threatened or impaired. WQIs at N14, N35, N67 and N75 are below 40 and thus indicate that water quality is almost always threatened or impaired at these stations

(Figure 5.35).

102

CHAPTER 05: Results

Figure 5.34. Change in WQI over time for 9 monitoring stations in HNRS (Reproduced from: http://www.lahistoriaconmapas.com/atlas/map-river/Cook-Islands-river-map.htm).

103

CHAPTER 05: Results

60 Marginal Poor

55 WQI 50 45 40 35 30 25 20 15 10 5 0 N14 N21 N35 N42 N44 N57 N67 N75 N92

Water quality monitoring stations

Figure 5.35. Average WQI along the HNRS.

Scope, frequency and amplitude values at the 9 monitoring stations are presented in Figure

5.36. At N35, nearly 90% of water quality values are beyond the guidelines. N35 shows the highest frequency among the 9 monitoring stations; it also shows high amplitude (46.3). The upstream of N35 is affected by quality and magnitude of flows coming from the South Creek

(which carries discharges from St. Marys Sewage treatment plant (STP), Riverstone STP,

Quakers Hill STP, McGraths Hill STP, and South Windsor STP) and discharge from North

Richmond STP. The dominant land use in this part of the catchment includes rural, grazing, commercial gardening, intensive agriculture and urban and industrial activities. The low

WQI at N35 can be attributed to these land uses. .

104

CHAPTER 05: Results

100.0 90.0 80.0 70.0 60.0 F1 50.0 F2 40.0 F3 30.0 20.0 10.0 0.0 N14 N21 N35 N42 N44 N57 N67 N75 N92

Figure 5.36. Scope, frequency and amplitude values at 9 monitoring stations in HNRS.

At N14, 81% of water quality data are outside the guidelines. Also, it has an amplitude of

70%. From 1993 to 2008, amplitudes are greater than 60%. Table 5.10 presents the amplitudes at 9 stations in different years. The years with higher amplitudes (greater than

60%) are indicated in red.

Further data exploration was done at N14 as it shows the worst WQI among the 9 stations.

Table 5.11 presents details of percentage failed tests for different water quality parameter (the total number of tests, number of failed tests, and percentage failed for each parameter for different year). Total nitrogen, chlorophyll-a, total iron, total aluminium, alkalinity and conductivity are the water quality parameters which exceeded the guidelines on many occasions.

105

CHAPTER 05: Results

Table 5.10: Amplitudes at 9 stations in different years

Index Period N14 N21 N35 N42 N44 N57 N67 N75 N92 2013 39.0 31.2 39.9 29.5 34.8 30.1 35.5 24.7 20.2 2012 44.4 32.1 43.4 31.4 35.4 23.0 26.4 20.6 16.8 2011 49.9 24.8 34.7 7.2 14.8 18.8 21.9 15.5 7.5 2010 50.8 40.4 35.7 17.1 18.7 22.9 25.1 24.4 22.3 2009 52.3 33.1 37.4 10.8 22.3 35.4 31.1 36.8 31.3 2008 65.9 32.0 42.6 24.9 30.6 49.7 41.0 53.4 43.5 2007 71.0 38.5 43.6 22.3 33.4 50.5 46.4 57.2 55.8 2006 81.6 43.2 45.3 15.9 22.2 41.5 47.0 60.9 50.7 2005 76.3 43.4 44.7 13.9 23.8 32.3 41.1 53.9 39.3 2004 82.9 45.2 46.6 15.2 26.7 35.3 37.0 55.4 38.9 2003 80.2 39.3 43.7 17 26.7 41.3 37.7 54.8 43.3 2002 78.7 37.7 41.4 19.2 26.3 29.7 30.6 52.8 40.1 2001 78.2 35.3 38.9 16.7 23.4 3.6 29.4 40.0 26.0 2000 87.1 35.2 41.0 16.3 22.0 6.9 25.8 49.3 30.2 1999 65.1 42.7 50.4 22.5 38.2 18.8 35.5 55.6 19.9 1998 75.0 39.9 50.7 19.5 22.8 10.6 30.6 49.2 20.7 1997 81.2 46.4 58.3 21.5 27.0 7.4 34.5 59.1 19.7 1996 69.2 39.0 54.7 20.6 30.4 9.4 33.3 60.1 13.5 1995 80.4 37.7 58.0 18.2 24.3 12.2 27.8 50.3 7.9 1994 85.3 60.1 60.9 20.1 20.7 2.0 18.0 51.4 17.8 1993 74.8 61.5 31.4 29.2 4.1 23.4 24.5 18.9

106

CHAPTER 05: Results

Table 5.11: Water quality results at N14

Iron Total Aluminium Total True Colour Alkalinity Suspended Solids Conductivity

Number Percent Number Percent Number Percent Number Percent Number Percent Number Percent Index Number Number Number Number Number Number of Failed Failed of Failed Failed of Failed Failed of Failed Failed of Failed Failed of Failed Failed Period of Tests of Tests of Tests of Tests of Tests of Tests Tests (%) Tests (%) Tests (%) Tests (%) Tests (%) Tests (%) 2013 4 3 75.0 4 2 50.0 4 3 75.0 4 4 100.0 4 0 4 3 75.0 2012 14 10 71.4 14 8 57.1 14 5 35.7 14 11 78.6 14 2 14.3 13 10 76.9 2011 13 9 69.2 13 7 53.8 13 5 38.5 13 9 69.2 13 2 15.4 12 8 66.7 2010 12 9 75.0 12 7 58.3 12 5 41.7 12 11 91.7 12 1 8.3 9 7 77.8 2009 13 8 61.5 13 6 46.2 13 3 23.1 13 13 100.0 13 1 7.7 11 8 72.7 2008 12 9 75.0 12 9 75.0 12 5 41.7 12 12 100.0 12 1 8.3 11 10 90.9 2007 12 9 75.0 12 8 66.7 12 3 25.0 12 11 91.7 12 4 33.3 13 12 92.3 2006 13 5 38.5 13 6 46.2 13 0 13 13 100.0 13 2 15.4 12 11 91.7 2005 12 7 58.3 12 6 50.0 12 2 16.7 12 12 100.0 12 4 33.3 12 12 100.0 2004 13 6 46.2 13 5 38.5 13 2 15.4 13 13 100.0 13 1 7.7 13 13 100.0 2003 13 6 46.2 13 6 46.2 13 0 13 13 100.0 13 4 30.8 12 12 100.0 2002 14 10 71.4 14 11 78.6 14 4 28.6 14 12 85.7 14 8 57.1 14 13 92.9 2001 6 5 83.3 6 5 83.3 6 1 16.7 6 6 100.0 17 8 47.1 17 17 100.0 2000 0 0 0 0 0 0 0 0 27 1 3.7 21 18 85.7 1999 0 0 0 0 0 0 0 0 26 1 3.8 12 11 91.7 1998 0 0 0 0 0 0 0 0 24 7 29.2 24 20 83.3 1997 0 0 0 0 0 0 0 0 23 3 13.0 23 23 100.0 1996 0 0 0 0 0 0 0 0 26 6 23.1 19 19 100.0 1995 0 0 0 0 0 0 0 0 26 4 15.4 26 25 96.2 1994 0 0 0 0 0 0 0 0 23 8 34.8 24 24 100.0 1993 0 0 0 0 0 0 0 0 16 1 6.3 8 8 100.0

25 < percentage Failed < 50 = , , PercentageFailed >= 50 =

107

CHAPTER 05: Results

From Table 5.11, it can be seen that water quality at N14 is poor with respect to Nitrogen,

Chlorophyll a, Iron, Aluminium and conductivity. Nitrogen is a nutrient used by plants within natural ecosystems, with minimal leakage into surface or groundwater (Vitousek et al., 2002).

Nitrogen concentrations in streams generally increase due to discharge of sewage water, pollutant wash off from urban and agricultural land, and atmospheric deposition. Increased nitrogen may result in overgrowth of algae, which can decrease the dissolved oxygen content of water, thereby harming or killing fish and other aquatic species. Controlling of nitrogen load in the urban river systems is viewed as a priority by many river management authorities as this affects algal growth. The HNRS has seen a number of episodes of algal blooms in the past, causing public concerns. For examples, the shallow mid Nepean River section was affected heavily by aquatic weed Egeria densa (Roberts et al., 1999). The Berowra Creek estuarine section of the river was infested by toxic dinoflagellate algal blooms (SMEC,

1997).

The long-term persistence of elevated levels of Chlorophyll-a is a concern to water authorities. An excessive growth often leads to poor water quality, noxious odours, oxygen depletion, human health problems and fish kills. It may also be linked to harmful (toxic) algal blooms. Poor water quality associated with high chlorophyll concentrations needs to be distinguished from the natural variation observed with the seasons, with latitude, and those associated with hydrodynamic features (e.g. upwelling). However, there is very little information to make this distinction (Ward, 1998). Observed increases in the concentrations of chlorophyll may be related to increased nutrient concentrations, decreased flow/changed hydrodynamics (increased residence times) and/or decreased turbidity (increased light

108

CHAPTER 05: Results penetration) (i.e. the increasing eutrophication status). The high Chlorophyll-a level at N14 needs to be investigated to find the possible reasons and to devise controlling measures.

If the alkalinity level is too high, the water can be cloudy, which inhibits the growth of underwater plants i.e. it may restrict algal growths. A higher alkalinity may raise the pH level, which in turn, can harm or kill fish and other aquatic organisms which are too sensitive to higher pH levels. High alkalinity may result from the presence of the bicarbonate ion, which is derived from the dissolution of carbonates by carbonic acids due to factors such as weathering of limestone and dolomite rocks mainly composed of calcite. The high alkalinity level at N14 in HNRS needs further investigation.

There are a number of factors that can lead to high conductivity levels. For examples, streams that run through clay catchments may have a higher conductivity level due to the presence of clay particles that ionize when enter into the river system. Groundwater inflows can have the same effects if it contains clay particles. An underperforming STP could raise the conductivity level in the effluent because of the presence of chloride, phosphate and nitrate.

109

CHAPTER 05: Results

5.7 COMPARISON OF MEASURED WATER QUALITY DATA WITH SCA DATA

Measured water quality data were compared with SCA data (Figure 5.37 – 5.57) considering the closest site for different sampling location as follows.

S1 – Blaxland Crossing with N67

S2 – at M4 Bridge with N57

S3 - Weir Reserve with N44

When compared, measured pH data with SCA data, N57 (SCA station) and S2 (self- monitoring) show similar values for the monitored year. The other 2 stations do not show similar values. Measured and SCA data for DO show a considerable similarity for all the 3 stations. Only S2 and N57 show similar values for electrical conductivity. Turbidity values for S1 and S2 sampling stations are similar to the SCA data. At S3, measured data does not show comparatively higher values in June at S3 as compared to SCA data. For nitrogen oxides and nitrogen as ammonia, only S2 and N57 show similar results. Measured and SCA temperature values are almost similar for all the 3 sampling stations. Sampling stations for self-monitoring are not exactly the SCA sampling locations, though they are the closest points. This may be the reason for the variations of water quality data. Differences in sampling procedures (ex, water depth) and collection time also can cause variations in the water quality data. However, when considering overall data sets, they show a considerable similarity.

110

CHAPTER 05: Results

5.7.1 pH

S1 Vs N67 8 7.8 7.6

7.4

pH 7.2 7 6.8 6.6 6.4

Date

Figure 5.37. pH values at S1 and N67.

S2 Vs N57 8 7.8 7.6 7.4

7.2

7 pH 6.8 6.6 6.4 6.2 6

Date

Figure 5.38. pH values at S2 and N57.

111

CHAPTER 05: Results

S3 Vs N44 8 7.9 7.8 7.7 7.6

7.5

7.4 pH 7.3 7.2 7.1 7

Date

Figure 5.39. pH values at S3 and N44.

5.7.2 Dissolved Oxygen

S1 Vs N67 12 10

DO 6 4 2 0

Date

Figure 5.40. pH values at S1 and N67.

112

CHAPTER 05: Results

S2 Vs N57 14 12 10

8 DO 6 4 2 0

Date

Figure 5.41. pH values at S2 and N57.

S3 Vs N44 14 12 10

6 DO 4 2 0

Date

Figure 5.42. pH values at S3 and N44.

113

CHAPTER 05: Results

5.7.3 Electrical Conductivity

S1 Vs N67 500

400

300 EC 200

100

Date

Figure 5.43. Electrical conductivity at S1 and N67.

S2 Vs N57 400

300

200 EC 100

Date

Figure 5.44. Electrical conductivity at S2 and N57.

114

CHAPTER 05: Results

S3 Vs N44 400

300

200 EC 100

Date

Figure 5.45. Electrical conductivity at S3 and N44.

5.7.4 Turbidity

S1 Vs N67 80 70 60

50 40 30

Turbidity 20 10 0

Date

Figure 5.46. Turbidity at S1 and N67.

115

CHAPTER 05: Results

S2 Vs N57 30 25

10 Turbidity 5 0

Date

Figure 5.47. Turbidity at S2 and N57.

S3 Vs N44 40 35 30

25 20 15 Turbidity 10 5 0

Date

Figure 5.48. Turbidity at S3 and N44.

116

CHAPTER 05: Results

5.7.5 Nitrogen Oxides

S1 Vs N67 0.7 0.6

0.5 0.4

0.3 NOx mg/L NOx 0.2 0.1 0

Date

Figure 5.49. Nitrogen oxides at S1 and N67.

S2 Vs N57 0.45

0.4

0.35 0.3 0.25

NOx mg/L NOx 0.2 0.15 0.1 0.05 0

Date

Figure 5.50. Nitrogen oxides at S2 and N57.

117

CHAPTER 05: Results

S3 Vs N44 0.6 0.5

0.4

0.3

0.2 NOx mg/L NOx 0.1 0

Date

Figure 5.51. Nitrogen oxides at S3 and N44.

5.7.6 Ammonical Nitrogen

S1 Vs N67 0.06

0.05

0.04 N mg/L N

- 0.03

0.02 NhH 0.01 0

Date

Figure 5.52. Ammonical nitrogen at S1 and N67.

118

CHAPTER 05: Results

S2 Vs N57 0.02

0.016

0.012

N mg/L N -

0.008 NH3 0.004

Date

Figure 5.53. Ammonical nitrogen at S2 and N57.

S3 Vs N44 0.06

0.05

0.04 N mg/L N

- 0.03

NH3 0.02 0.01 0

Date

Figure 5.54. Ammonical nitrogen at S3 and N44.

119

CHAPTER 05: Results

5.7.7 Temperature

S1 Vs N67 30

20 15

10 Temperature 5 0

Date

Figure 5.55. Temperature at S1 and N67.

S2 Vs N57 30

25 20 15

Temperature 10 5 0

Date

Figure 5.56. Temperature at S2 and N57.

120

CHAPTER 05: Results

S3 Vs N44 30

20 15

10 Temperature 5 0

Date

Figure 5.57. Temperature at S3 and N44.

5.8 CHAPTER SUMMARY

The results of the assessment of the water quality in the Hawkesbury Nepean River system

(HNRS) have been presented in this chapter. It has been found that the water quality parameters vary along the length of the HNRS. Preliminary analyses from the box plots and principal component analysis of the water quality parameters have shown that many of the water quality parameters are highly correlated and some of the monitoring stations do not provide any independent information.

121

CHAPTER 05: Results demonstrate increasing trends at most of the stations, and total phosphorus, suspended solids, filterable aluminium, ammonia nitrogen and filterable phosphorus do not show any trend at most of the stations. The median values for chlorophyll-a, total nitrogen and alkalinity are above the ANZECC (2000) trigger values for most of the stations. The increasing trend of turbidity, chlorophyll-a, alkalinity, dissolved organic carbon, total iron, total aluminium, total manganese and reactive silicate and the exceedance of the ANZECC (2000) trigger values for chlorophyll-a, total nitrogen and alkalinity indicate an overall water quality deterioration in the HNRS during the last decade. The parameters such as phosphorus, suspended solids and ammonia nitrogen do not show any marked change over the period of this study. Although an improvement in water quality can be seen at some stations downstream of the undisturbed parts of the catchment, there has been an overall water quality deterioration in the HNRS during the last decade.

Three prediction equations have been developed for three important water quality parameters

(chlorophyll-a, total nitrogen and total phosphorous) for the HNRS. These equations generally present a high co-efficient of determination values and satisfy the assumptions of least squares regression analysis. These equations can be used to estimate chlorophyll-a, total nitrogen and total phosphorous from easily measurable water quality parameters.

Application of Canadian Water Quality Index method has shown that water quality at the 9 stations fall under either poor or marginal category, based on the Canadian Water Quality

Index (CWQI) categorisation where the CWQI values are found to be in the range of 33 to

57. Marginal water quality is found for 5 stations and poor water quality is found for the

122

CHAPTER 05: Results remaining 4 stations. None of the stations were found to have good quality water. Stations

N14 and N35 were found to be the most polluted stations in the HNRS among the 9 stations.

With detailed investigation at station N14, it was found that the higher values of water quality parameters: Nitrogen, Chlorophyll a, Iron, Aluminium, Alkalinity and Conductivity have contributed to the poor water quality condition at N14.

Comparison of self-monitored water quality data SCA data obtained from nearby sampling stations show a considerable similarity.

123

CHAPTER 06: Conclusion

CHAPTER 6

SUMMARY AND CONCLUSIONS 6

6.0 SUMMARY

The Hawkesbury Nepean River System (HNRS) is one of the most important rivers in

Australia as it supplies water to over 4 million people in Sydney. HNRS has multiple and complex land uses. Hence, the water quality of this river is of great significance. In this study, a total of 9 water quality parameters have been used from 15 water quality monitoring stations plus one-year self-monitoring to assess the quality of water in the HNRS.

6.1 PRELIMINARY WATER QUALITY DATA ANALYSIS

From the preliminary data analysis, it has been found that the average concentrations of some water quality parameters, such as total nitrogen and alkalinity, are higher than those recommended by the Australian and New Zealand Environment and Conservation Council

(ANZECC) guidelines at most of the monitoring stations along the HNRS. The higher levels of total nitrogen might be attributed to runoff from agricultural lands, urban areas and sewage treatment plants. A higher value of total phosphorus at station N35 is observed. Station N14 shows notable higher conductivity values and N35 shows higher values of nitrogen oxides and chlorophyll-a levels. Also, station N21 has high levels of chlorophyll-a which is found to be above the recommended guidelines. High nitrogen, phosphorus and chlorophyll-a levels at many stations appear to be a sign of deteriorated water quality in the HNRS.

124

CHAPTER 06: Conclusion

6.2 TREND ANALYSIS

From the trend analysis, a general pattern of downward trends of pH, nitrogen TKN, alkalinity, dissolved oxygen and electrical conductivity has been detected. Total iron, filterable iron, true colour, total aluminium, reactive silicate and dissolved organic carbon demonstrate an increasing trend at most of the stations. In addition, total phosphorus, suspended solids, filterable aluminium, ammonia nitrogen and filterable phosphorus do not show any trend at most of the stations. The median values for chlorophyll-a, total nitrogen and alkalinity are above the ANZECC (2000) trigger values for most of the stations. The increasing trend of turbidity, chlorophyll-a, alkalinity, dissolved organic carbon, total iron, total aluminium, total manganese and reactive silicate and the exceedance of the ANZECC

(2000) trigger values for chlorophyll-a, total nitrogen and alkalinity indicate an overall water quality deterioration occurred in the HNRS during the last decade. The parameters such as phosphorus, suspended solids and ammonia nitrogen do not show any marked changes over the period of this study. Although an improvement in water quality can be seen at some stations at downstream of the undisturbed parts of the catchment, trend analysis shows an overall water quality deterioration in the HNRS during the last decade.

6.3 REGRESSION ANALYSIS

Using the regression analysis, three prediction equations have been developed for three important water quality parameters (chlorophyll-a, total nitrogen and total phosphorous) for the HNRS. These equations generally present a high co-efficient of determination values and

125

CHAPTER 06: Conclusion satisfy the assumptions of least squares regression analysis. These equations can be used to estimate chlorophyll-a, total nitrogen and total phosphorous from easily measurable water quality parameters.

6.4 APPLICATION OF CANADIAN WATER QUALITY INDEX METHOD

Application of Canadian Water Quality Index (CWQI) method shows the water quality at the

9 stations fall under either poor or marginal category based on the CWQI categorisation, where the CWQI values are found to be in the range of 33 to 57. Marginal water quality is found for 5 stations and poor water quality for the remaining 4 stations. None of the stations are found to have good quality water. Stations N14 and N35 are found to be the most polluted stations in the HNRS among the 9 stations. With detailed investigation at the station N14, it is found that the higher values of water quality parameters: Nitrogen, Chlorophyll a, Iron,

Aluminium, Alkalinity and Conductivity have contributed to the poor water quality condition at N14. There are many sewage treatment plants that discharge reed waste water to upstream of station N35. Also, the dominant land use in this part of the catchment includes rural, grazing, commercial gardening, intensive agriculture and urban and industrial activities. The low WQI at N35 can be attributed to these land uses. Water quality at station N14 should be improved because of dilution by high quality inflows from the Colo River and the undisturbed upstream catchment. The high pollutant levels at N14 need to be investigated to find the possible reasons and to devise controlling measures. `

126

CHAPTER 06: Conclusion

6.5 COMPARISON OF MEASURED WATER QUALITY DATA WITH SCA DATA

In general, the self-monitored water quality data is similar to SCA data obtained from nearby sampling stations.

6.6 CONCLUSION

The following conclusions can be drawn from this study.

 The concentrations of total phosphorus, nitrogen oxides and chlorophyll-a are higher than those recommended by the Australian and New Zealand Environment and Conservation Council (ANZECC) guidelines.  An increasing trend has been detected for turbidity, chlorophyll-a, alkalinity, dissolved organic carbon, total iron, total aluminium, total manganese and reactive silicate for majority of the monitoring stations.  Application of Canadian Water Quality Index method shows the water quality at 9 stations fall under either poor or marginal category.  Stations N14 and N35 are found to be the most polluted stations in the HNRS among the 9 stations.  Although an improvement in water quality can be seen at some stations at downstream of the undisturbed parts of the catchment, there has been an overall water quality deterioration in the HNRS during the last decade.  The developed prediction equations for three important water quality parameters (chlorophyll-a, total nitrogen and total phosphorous) can be used to predict these water quality parameters for the HNRS.

127

CHAPTER 06: Conclusion

6.7 LIMITATIONS OF THE STUDY

The study has a number of limitations as noted below:

 Self-monitoring was not conducted for all the selected water quality parameters due to limited laboratory facilities.  Water quality modelling could not be conducted due to the complex land use and the large catchment size.  The seasonality effects on water quality were not investigated.

6.8 RECOMMENDATIONS FOR FUTURE RESEARCH

 An artificial intelligence based model can be developed to increase the overall prediction accuracy of various water quality parameters along the HNRS where easily measurable water quality parameters can be used to predict other water quality parameters.  Specific land use data can be obtained and a close monitoring program can be developed to link water quality and land use characteristics.  Automatic water sampling probes and the telemetry system should be developed to provide real-time assessment of water quality for this very important river system.  A random sampling technique should be developed by a joint group of water authorities and universities to uncover any major water quality deterioration in the HNRS in future.

128

CHAPTER 06: Conclusion

References

Alberto, W. D., Marı́a del Pilar, D., Marı́a Valeria, A., Fabiana, P. S., Cecilia, H. A. & Marı́a de los Ángeles, B. (2001). Pattern recognition techniques for the evaluation of spatial and Temporal variations in water quality. A case study: Suquı́a River basin (Córdoba–Argentina). Water Research, 35(12), 2881-2894.

Al-Janabi, Z. Z., Al-Kubaisi, A., & Alobaidy, A. H. M. J. (2012). Assessment of water quality of Tigris River by using water quality index (CCME-WQI). Journal of Al-Nahrain University, 15(1), 119-126.

ANZECC (2000). Australian and New Zealand guidelines for fresh and marine water quality, Australian and New Zealand Environment and Conservation Council. http://www.environment.gov.au/water/publications/quality/nwqms-guidelines-4-vol1.html

Astaraie-Imani, M., Kapelan, Z., Fu, G. & Butler, D. (2012). Assessing the combined effects of urbanisation and climate change on the river water quality in an integrated urban wastewater system in the UK. Journal of Environmental Management, 112, 1-9.

Astel, A., Tsakovski, S., Barbieri, P. & Simeonov, V. (2007). Comparison of self-organizing maps classification approach with cluster and principal components analysis for large environmental data sets. Water Research, 41(19), 4566-4578.

Atech (2000). Cost of algal blooms. LWRRDC Occasional paper 26/99. Land and Water Resources Research and Development Corporation, Canberra, ACT, Australia, 42 pp.

Baginska, B., Pritchard, T. & Krogh, M. (2003). Roles of land use resolution and unit-area load rates in assessment of diffuse nutrient emissions. Journal of Environmental Management, 69(1), 39-46.

Bartholomew, D. J., Steele, F., Galbraith, J., & Moustaki, I. (2008). Analysis of multivariate

129

CHAPTER 06: Conclusion social science data. CRC press.

Boman, B. J., Wilson, P. C., & Ontermaa, E. A. (2002). Understanding water quality parameters for citrus irrigation and drainage systems. University of Florida IFAS Extension, Circ, 1406.

BOM (2013). Australian government – bureau of meteorology. http://www.bom.gov.au/water/nwa/2011/sydney/

Bowes, M. J., Gozzard, E., Johnson, A. C., Scarlett, P. M., Roberts, C., Read, D. S., ... & Wickham, H. D. (2012). Spatial and temporal changes in chlorophyll-a concentrations in the River Thames basin, UK: are phosphorus concentrations beginning to limit phytoplankton biomass?. Science of the Total Environment, 426, 45-55.

Boyacioglu, H. (2010). Utilization of the water quality index method as a classification tool. Environmental Monitoring and Assessment, 167(1-4), 115-124.

Brown, P. J., Le, N. D. & Zidek, J. V. (1994). Multivariate spatial interpolation and exposure to air pollutants. Canadian Journal of Statistics, 22(4), 489-509.

Chang, H. (2008). Spatial analysis of water quality trends in the Han River basin, South Korea. Water Research, 42(13), 3285-3304.

Commonwealth of Australia (1996). Australia State of the Environment.

Crosa, G., Froebrich, J., Nikolayenko, V., Stefani, F., Galli, P. & Calamari, D. (2006). Spatial and seasonal variations in the water quality of the Amu Darya River (Central Asia). Water Research, 40(11), 2237-2245.

Davis, J.C. (1986). Statistics and data analysis in geology. New York: Wiley, 2, 646.

Dawson, E. J. & Macklin, M. G. (1998). Speciation of heavy metals on suspended sediment

130

CHAPTER 06: Conclusion under high flow conditions in the River Aire, West Yorkshire, UK. Hydrological Processes, 12(9), 1483-1494.

Dede, O. T., Telci, I. T., & Aral, M. M. (2013). The use of water quality index models for the evaluation of surface water quality: a case study for Kirmir Basin, Ankara, Turkey. Water Quality, Exposure and Health, 5(1), 41-56.

Draper, N. R., Smith, H., & Pownell, E. (1966). Applied regression analysis. New York: Wiley, 3, 85-96.

Drápela, K. & Drápelová, I. (2011). Application of Mann-Kendall test and the Sen's slope estimates for trend detection in deposition data from Bílý Kříž (Beskydy Mts., the Czech Republic) 1997-2010. Beskydy, 4(2), 133-146.

Diamond, R. (2004). Water and Sydney’s future-balancing the value of our rivers and economy: Final report of the Hawkesbury–Nepean River Management Forum. Department of Infrastructure, Planning and Natural Resources, Sydney, NSW.

Dixon, W. & Chiswell, B. (1996). Review of aquatic monitoring program design. Water Research, 30(9), 1935-1948.

Egodawatta, P., Thomas, E. & Goonetilleke, A. (2009). Understanding the physical processes of pollutant build-up and wash-off on roof surfaces. Science of the Total Environment, 407(6) 1834-41.

Eroksuz, E. & Rahman, A. (2010). Rainwater tanks in multi-unit buildings: A case study for three Australian cities. Resources, Conservation and Recycling, 54(12), 1449-1452.

Erturk, A., Gurel, M., Ekdal, A., Avsan, C., Ugurluoglu, A., Seker, D. Z., Tanik, A. & Ozturk, I. (2010). Water quality assessment and meta modal development in Melen watershed – Turkey. Journal of Environmental Management, 91, 1526-1545.

131

CHAPTER 06: Conclusion

Gburek, W. J. & Folmar, G. J. (1999). Flow and chemical contributions to streamflow in an upland watershed: A base flow survey. Journal of Hydrology, 217(1), 1-18.

Gehrke, P. C. & Harris, J. H. (1996). Fish and fisheries of the Hawkesbury-Nepean River System. Final Rapport to the Sydney Water Corporation.

Gitelson, A. (1992). The peak near 700 nm on radiance spectra of algae and water: Relationships of its magnitude and position with chlorophyll concentration. International Journal of Remote Sensing, 13(17), 3367-3373.

Haddad, K., Egodawatta, P., Rahman, A. & Goonetilleke, A. (2013). Uncertainty analysis of pollutant build-up modelling based on a bayesian weighted least squares approach. Science of the Total Environment, 449, 410-417.

Hanrahan, G., Gledhill, M., House, W. A. & Worsfold, P. J. (2003). Evaluation of phosphorus concentrations in relation to annual and seasonal physico-chemical water quality parameters in a UK chalk stream. Water Research, 37(15), 3579-3589.

Horton, R. K., (1965). An index-number system for rating water quality. Journal of Water Pollution Control Federation. 37(3), 300–306.

Interlandi, S. J. & Crockett, C. S. (2003). Recent water quality trends in the Schuylkill River, Pennsylvania, USA: A preliminary assessment of the relative influences of climate, river discharge and suburban development. Water Research, 37(8), 1737-1748.

Kaika, M. (2003). The water framework directive: A new directive for a changing social, political and economic European framework. European Planning Studies 11, 299-316.

Karr, J. R. (1999). Defining and measuring river health. Freshwater Biology, 41(2), 221-234.

Kendall, C. A. R. O. L., Elliott, E. M. & Wankel, S. D. (2007). Tracing anthropogenic inputs of nitrogen to ecosystems. Stable Isotopes in Ecology and Environmental Science, 2(1), 375-

132

CHAPTER 06: Conclusion

449.

Khan, A. A., Paterson, R., & Khan, H. (2004). Modification and application of the Canadian council of ministers of the environment water quality index (CCME WQI) for the communication of drinking water quality data in Newfoundland and Labrador. Water Quality Research Journal of Canada, 39(3), 285-293.

Kovács, J., Tanos, P., Korponai, J., Székely, I. K., Gondár, K., Gondár-Sőregi, K. & Hatvani, I. G. (2012). Analysis of water quality data for scientists. Water Quality Monitoring and Assessment. 65-94.

Lack, T. J. (1971). Quantitative studies on the phytoplankton of the Rivers Thames and Kennet at Reading. Freshwater Biology, 1(2), 213-224.

LeBlanc, R. T., Brown, R. D. & FitzGibbon, J. E. (1997). Modeling the effects of land use change on the water temperature in unregulated urban streams. Journal of Environmental Management, 49(4), 445-469.

Lumb, A., Sharma, T. C., Bibeault, J. F., & Klawunn, P. (2011). A comparative study of USA and Canadian water quality index models. Water Quality, Exposure and Health, 3(3-4), 203- 216.

Ma, J., Ding, Z., Wei, G., Zhao, H. & Huang, T. (2009). Sources of water pollution and evolution of water quality in the Wuwei basin of Shiyang River, North Chaina. Journal of Environmental Management, 90, 1168-1177.

Markich, S. J. & Brown, P. L. (1998). Relative importance of natural and anthropogenic influences on the fresh surface water chemistry of the Hawkesbury–Nepean River, southeastern Australia. Science of the Total Environment, 217(3), 201-230.

Marsden, T. J., & Gehrke, P. C. (1996). Fish passage in the Hawkesbury-Nepean River system. Fish and Fisheries of the Hawkesbury-Nepean River System. NSW Fisheries

133

CHAPTER 06: Conclusion

Research Institute, Sydney, 200-218.

Massart, D. L., Vandeginste, B. G. M., Deming, S. N., Michotte, Y. & Kaufman, L. (1988). Chemometrics: A textbook, 20-21.

Meybeck, M. (2002). Riverine quality at the Anthropocene: Propositions for global space and time analysis, illustrated by the Seine River. Aquatic Sciences, 64(4), 376-393.

NLWRA (2008). The national land and water resources audit final report: 2002-2008. Australian government. http://lwa.gov.au/files/products/national-land-and-water-resources- audit/pn21537/pn21537.pdf.

Norris, R. H., & Thoms, M. C. (1999). What is river health?. Freshwater Biology, 41(2), 197- 209.

Osborne, L. L. & Wiley, M. J. (1988). Empirical relationships between land use/cover and stream water quality in an agricultural watershed. Journal of Environmental Management, 26(1), 9-27.

Pinto, U., Maheshwari, B. L. & Grewal, H. S. (2010). Effects of greywater irrigation on plant growth, water use and soil properties. Resources, Conservation and Recycling, 54(7), 429- 435.

Pinto, U. & Maheshwari, B. L. (2011). River health assessment in peri-urban landscapes: An application of multivariate analysis to identify the key variables. Water Research, 45(13), 3915-3924.

Pinto, U., Maheshwari, B. L., Shrestha S. & Morris C. (2012). Modelling eutrophication and microbial risks in peri-urban River systems using discriminant function analysis .Water Research, 46, 6476-6488.

Pinto, U., Maheshwari, B. L. & Ollerton, R. L. (2013). Analysis of long-term water quality

134

CHAPTER 06: Conclusion for effective river health monitoring in peri-urban landscapes: A case study of the Hawkesbury–Nepean River system in NSW, Australia. Environmental Monitoring and Assessment, 185(6), 4551-4569.

Rahman, A., Thomas, E. C., Bhuiyan, S., & Goonetilleke, A. (2002). Modelling pollutant washoff from south east Queensland catchments Australia.

Rahman, A., Keane, J., & Imteaz, M. A. (2012). Rainwater harvesting in Greater Sydney: Water savings, reliability and economic benefits. Resources, Conservation and Recycling, 61, 16-21.

Roberts, D. E., Church, A. G. & Cummins, S. P. (1999). Invasion of Egeria into the Hawkesbury-Nepean River, Australia. Aquatic Plant Management, 37, 31-34.

Russell, D. W. (2002). In search of underlying dimensions: The use (and abuse) of factor analysis in personality and social psychology bulletin. Personality and Social Psychology Bulletin, 28(12), 1629-1646.

Sarkar, C. & Abbasi, S. A. (2006). A new software for generating water quality indices. Environmental Monitoring and Assessment, 119, 201-231.

Sen, P. K. (1968). Estimates of the regression coefficient based on Kendall's tau. Journal of the American Statistical Association, 63(324), 1379-1389.

SMEC (1997). Options for sewage treatment and effluent disposal for West Hornsby and Hornsby Heights sewage treatment plants. SMEC Australia.

Solbe, J. F. (1986). Effects of land use on fresh waters: Agriculture, forestry, mineral exploitation, urbanisation.

Sousounis, P. J., & Bisanz, J. M. (2000). Preparing for a changing climate: The potential consequences of climate variability and change. Great Lakes Overview.

135

CHAPTER 06: Conclusion

Stanhill, G., & Cohen, S. (2001). Global dimming: a review of the evidence for a widespread and significant reduction in global radiation with discussion of its probable causes and possible agricultural consequences. Agricultural and forest meteorology, 107(4), 255-278.

Stroomberg, G. J., Freiriks, I. L., Smedes, F., & Cofino, W. P. (1995). Quality assurance and quality control of surface water sampling. Quality Assurance in Environmental Monitoring. Sampling and Sample Pretreatment, 51.

Stubblefield, A. P., Reuter, J. E., Dahlgren, R. A. & Goldman, C. R. (2007). Use of turbidometry to characterize suspended sediment and phosphorus fluxes in the Lake Tahoe basin, California, USA. Hydrological Processes, 21(3), 281-291.

Suhr, D. D. (2005). Principal component analysis vs. exploratory factor analysis. SUGI 30 proceedings, 203, 230.

Tabari, H., Marofi, S. & Ahmadi, M. (2011). Long-term variations of water quality parameters in the Maroon River, Iran. Environmental Monitoring and Assessment, 177(1-4), 273-287.

Thoms, M. C., Parker, C. R., & Simons, M. (2000). The dispersal and storage of trace metals in the Hawkesbury River valley. River Management: The Australasian Experience, 197-219.

Tian, J., & Fernandez, G. (1999). Seasonal trend analysis of monthly water quality data. Proceedings of the 7th Annual Western users of SAS software regional users group, 229-234.

Turner, L. & Erskine, W. D. (2005). Variability in the development, persistence and breakdown of thermal, oxygen and salt stratification on regulated rivers of southeastern Australia. River Research and Applications, 21(2‐3), 151-168.

Vega, M., Pardo, R., Barrado, E. & Debán, L. (1998). Assessment of seasonal and polluting effects on the quality of river water by exploratory data analysis. Water Research, 32(12), 3581-3592.

136

CHAPTER 06: Conclusion

Van der Sterren, M., Rahman, A., Shrestha, S., Barker, G. & Ryan, G. (2009). An overview of on-site retention and detention policies for urban stormwater management in the greater Western Sydney region in Australia, Water International, 34(3), 362-372.

Van der Sterren, M., Rahman, A. & Dennis, G.R. (2012). Implications to stormwater management as a result of lot scale rainwater tank systems: a case study in Western Sydney, Australia, Water Science and Technology, 65.8, 1475-1482.

Van der Sterren, M., Rahman, A. & Dennis, G. (2013). Quality and quantity monitoring of five rainwater tanks in Western Sydney, Australia. Journal of Environmental Engineering, 139, 332-340.

Van der Sterren, M., Rahman, A. & Ryan, G. (2014). Modeling of a lot scale rainwater tank system in XP-SWMM: A case study in Western Sydney, Australia, Journal of Environmental Management, 141, 177-189.

Van der Sterren, M., & Rahman, A. (2015). Single lot on site detention requirements in New South Wales Australia and its relation to holistic storm water management. Sustainability of Water Quality and Ecology, 6, 48-56.

Vitousek, P. M., Cassman, K., Cleveland, C., Crews, T., Field, C. B., Grimm, N. B. & Sprent, J. I. (2002). Towards an ecological understanding of biological nitrogen fixation. Biogeochemistry, 57(1), 1-45.

WBC (1994). Water Board (Corporatisation) Act. http://www.austlii.edu.au/au/legis/nsw/num_act/wba1994n88323.pdf.

Walsh, C. J., Roy, A. H., Feminella, J. W., Cottingham, P. D., Groffman, P. M., & Morgan, R. P. (2005). The urban stream syndrome: Current knowledge and the search for a cure. Journal of the North American Benthological Society, 24(3), 706-723.

Ward, S. N. (1998). On the consistency of earthquake moment release and space geodetic strain rates: Europe, Geophysical Journal International, 135(3), 1011-1018.

137

CHAPTER 06: Conclusion

Wenning, R. J. & Erickson, G. A. (1994). Interpretation and analysis of complex environmental data using chemometric methods. TrAC Trends in Analytical Chemistry, 13(10), 446-457.

Whitehead, P. G., & Hornberger, G. M. (1984). Modelling algal behaviour in the River Thames. Water Research, 18(8), 945-953.

WQA (2013). eWater Ltd. Water Board Corporatisation Act, (1994). http://www.austlii.edu.au/au/legis/nsw/num_act/wba1994n88323.pdf. ung, T.F. & Sanzone, S. (2002). A Framework for assessing and reporting on ecological condition. US Environmental Protection Agency, EPA Science Advisory Board.

Yue, S., Pilon, P. & Cavadias, G. (2002). Power of the Mann–Kendall and Spearman's rho tests for detecting monotonic trends in hydrological series. Journal of Hydrology, 259(1), 254-271.

Yu, Y. S., Zou, S. & Whittemore, D. (1993). Non-parametric trend analysis of water quality data of rivers in Kansas. Journal of Hydrology, 150(1), 61-80.

138