i
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
ii
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…………………………………………………………..
iii
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
iv
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
v
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.
vi
Dedications
To my loving husband Sameera,
for his love, understanding, encouragement and great support
&
to my loving daughter Vinuki and son Imeth.
vii
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.
viii
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,
ix
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.
x
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
xi
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
xii
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
xiii
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
xiv
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
xv
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
xvi
(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
xvii
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
xviii
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
xix
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
xx
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
xxi
Table 5.11: Water quality results at N14 ............................................................ 107
xxii
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
xxiii
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
xxiv
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
xxv
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
CHAPTER 01: Introduction
1
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
2
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
3
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
4
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
5
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
6
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
7
Chapter 6 presents a summary, conclusions and recommendations for further study.
CHAPTER 02: Literature Review
8
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
9
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
10
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
11
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
12
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
13
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
14
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
15
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
16
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
17
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
18
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
19
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
20
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
21
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
22
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
23
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
24
Where it 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 )t(f can be described as follows (equation 3.6):
BQt)t(f 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
25
kj
xxQ
kj
i
3.7
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
26
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
27
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:
𝐹1 =𝑇𝑜𝑡𝑎𝑙 𝑛𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑓𝑎𝑖𝑙𝑒𝑑 𝑣𝑎𝑟𝑖𝑎𝑏𝑙𝑒𝑠
𝑇𝑎𝑡𝑎𝑙 𝑛𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑣𝑎𝑟𝑖𝑎𝑏𝑙𝑒𝑠 × 100 3.8
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
28
the percentage of individual tests that do not meet the objectives (“failed tests”):
𝐹2 =𝑁𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑓𝑎𝑖𝑙𝑒𝑑 𝑡𝑒𝑠𝑡𝑠
𝑇𝑜𝑡𝑎𝑙 𝑛𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑣𝑎𝑟𝑖𝑎𝑏𝑙𝑒𝑠 × 100 3.9
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:
𝑒𝑥𝑐𝑢𝑟𝑠𝑖𝑜𝑛𝑖 = (𝐹𝑎𝑖𝑙𝑑 𝑡𝑒𝑠𝑡 𝑣𝑎𝑙𝑢𝑒𝑖𝑂𝑏𝑗𝑒𝑐𝑡𝑖𝑣𝑒𝑗
) − 1 3.10
For the cases in which the test value must not fall below the objective:
𝑒𝑥𝑐𝑢𝑟𝑠𝑖𝑜𝑛𝑖 = (𝑂𝑏𝑗𝑒𝑐𝑡𝑖𝑣𝑒𝑗
𝐹𝑎𝑖𝑙𝑑 𝑡𝑒𝑠𝑡 𝑣𝑎𝑙𝑢𝑒𝑖) − 1 3.11
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
29
𝑛𝑠𝑒 =∑ 𝑒𝑥𝑐𝑢𝑟𝑠𝑖𝑜𝑛𝑖𝑛𝑖=1
𝑁𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑡𝑒𝑠𝑡𝑠 3.12
3F 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.
𝐹3 = (𝑛𝑠𝑒
0.01𝑛𝑠𝑒 + 0.01) 3.13
The CWQI is finally calculated as:
𝐶𝑊𝑄𝐼 = 100 −
(
√𝐹1
2 + 𝐹22 + 𝐹3
2
1.732
)
3.14
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
30
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
31
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
32
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
33
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
34
Figure 4.1. Schematic diagram of the HNRS with land use details.
CHAPTER 04: Study Area and Data
35
Table 4.1: Sewage treatment plants (STP) along the HWNRS
STP Treatment level Completed date
Discharge
(ML/day) Discharge location
Picton Tertiary (includes additional Phosphorus
removal and disinfection)
30/06/2009 1.5 Re-used on-site for agricultural irrigation
Precautionary discharge to Stone Quarry
Creek
West
Camden
Tertiary (includes additional Phosphorus
removal and disinfection)
30/06/2006 10.7 Re-used at Agricultural Institute. Remainder
discharged via Matahill Creek to the Nepean
River
Wallacia Tertiary (includes additional phosphorus
and nitrogen removal and disinfection)
30/06/2009 0.8 Warragamba River
Penrith Tertiary (includes additional Phosphorus
and Nitrogen removal and disinfection)
30/06/2009 22.4 Re-used locally. Remainder transferred to St
Marys Advanced Water Treatment Plant.
Some excess discharged to Boundary Creek
St. Marys Tertiary (includes ultrafiltration, reverse
osmosis, de-carbonation, additional
Phosphorous and Nitrogen removal and
disinfection)
30/06/2009 33.5 Re-used locally and at Dunheved. Remainder
discharged to Nepean River. Some excess
discharged to South Creek
Winmalee Tertiary(includes additional phosphorus 30/06/2009 16.5 Unnamed creek to the Nepean River
CHAPTER 04: Study Area and Data
36
STP Treatment level Completed date
Discharge
(ML/day) Discharge location
and nitrogen removal and disinfection)
North
Richmond
Tertiary (includes additional phosphorus
removal and disinfection)
30/06/2009 0.9 Redbank Creek to the Hawkesbury River
Riverstone Tertiary(includes additional phosphorus
removal and disinfection)
30/06/2009 1.8 Eastern Creek to South Creek
Quakers Hill Tertiary (includes additional Phosphorus
and Nitrogen removal and disinfection)
30/06/1974 31.1 Re-used locally and at Ashlar Golf Course.
Remainder transferred to St Marys Advanced
Water Treatment Plant. Some excess
discharged to Breakfast Creek
Rouse Hill Tertiary (includes additional Phosphorus
and Nitrogen removal and disinfection)
also includes ultra-violet irradiation and
super-chlorination for reuse water
30/06/2009 15.3 Recycled back to households for non-drinking
use. Excess discharged to Second Ponds Creek
via wetlands to Cattai Creek
Castle Hill Tertiary (includes additional Phosphorus
removal and disinfection)
30/06/2009 6.5 Cattai Creek
CHAPTER 04: Study Area and Data
37
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
38
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
39
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
40
Table 4.3: Water quality parameters considered in the preliminary assessment
Water quality
Parameter
Abbreviation Units Min Max Median
pH PH 5.78 9.94 7.63
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
0mL
0.00 6100.00 13.00
Iron Filtered FI mg/L 0.01 3.43 0.09
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
O3/L
1.00 298.00 40.00
Aluminium Total TA mg/L 0.01 3.97 0.08
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
mL
0.00 8400.00 20.00
Phosphorus Total TP mg/L 0.01 0.18 0.01
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
mg/L
0.01 14.90 1.71
Dissolved Organic
Carbon
DOC mg/L 0.20 350.00 4.60
UV Absorbing
constituents
UV 0.01 0.93 0.12
CHAPTER 04: Study Area and Data
41
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
42
Figure 4.3. Locations of sampling stations in the HNRS (Reproduced from: Google maps).
S1
S3
S2
CHAPTER 04: Study Area and Data
43
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
44
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
45
Table 4.4: Water quality data at Blaxland Crossing
Date pH DO
(mg/L)
Tem
(deg C)
EC
(us/cm)
Turbidity /
(NTU)
NOx
(mg/L)
NH3 -N
(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
46
Table 4.5: Water quality data at M4
Date pH DO
(mg/L)
Tem
(deg C)
EC
(us/cm)
Turbidity
(NTU)
NOx
(mg/L)
NH3 -N
(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
47
Table 4.6: Water quality data at Weir Reserve
Date pH DO
(mg/L)
Tem
(deg C)
EC
(us/cm)
Turbidity
(NTU)
NOx
(mg/L)
NH3 -N
(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
48
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
49
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
50
lower pH values have not been specifically identified in this study.
N92
N881N8
6N8
5N7
5N6
7N6
41N64
N57
N44
N42
N35
N21
N14
E851
12
11
10
9
8
7
6
5
4
pH ANZECC upper limit
ANZECC lower limit
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
51
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.
N92N881N86N85N75N67N641N64N57N44N42N35N21N14E851
35
30
25
20
15
10
Te
mp
era
ture
(D
eg
C)
Figure 5.2. Box plot of measured temperature along the Hawkesbury Nepean River System.
CHAPTER 05: Results
52
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.
N92
N881N8
6N8
5N7
5N6
7N6
41N64
N57
N44
N42
N35
N21
N14
E851
20
15
10
5
0
Dis
so
lve
d O
xy
ge
n (
mg
/L)
ANZECC lower limit
Figure 5.3. Box plot of DO along the Hawkesbury Nepean River System.
CHAPTER 05: Results
53
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
54
N92N881N86N85N75N67N641N64N57N44N42N35N21N14E851
50
40
30
20
10
0
Co
nd
ucti
vit
y
(mS
/cm
)
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
55
N92
N881N8
6N8
5N7
5N6
7
N641
N64N57N44N42N35N21N14
E851
3.0
2.5
2.0
1.5
1.0
0.5
0.0
Co
nd
ucti
vit
y
(mS
/cm
)
ANZECC upper limit (0.35 mS/cm)
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
56
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
57
N85N75N67N641N64N57N44N42N35N21N14E851
500
400
300
200
100
0
Tu
rbid
ity
La
b/
Fie
ld (
NTU
)
Figure 5.6. Box plot of turbidity along the Hawkesbury Nepean River System (showing all
the observed data range).
CHAPTER 05: Results
58
N85
N75
N67
N641N6
4N5
7N4
4N4
2N3
5N2
1N1
4E8
51
100
80
60
40
20
0
Tu
rbid
ity
La
b/
Fie
ld (
NTU
)
ANZECC upper limit (20 NTU)
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
59
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
60
N92N881N86N85N75N67N641N64N57N44N42N35N21N14E851
0.4
0.3
0.2
0.1
0.0
Ph
osp
ho
rus T
ota
l (m
g/
L)
Figure 5.8. Box plot of total phosphorus along the HNRS (showing all the observed data
range).
CHAPTER 05: Results
61
N92
N881N8
6N8
5N7
5N6
7N6
41N64
N57
N44
N42
N35
N21
N14
E851
0.20
0.15
0.10
0.05
0.00
Ph
osp
ho
rus T
ota
l (m
g/
L)
ANZECC upper limit (0.05 mg/l)
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
62
N92N881N86N85N75N641N64N57N44N42N35N21N14E851
0.25
0.20
0.15
0.10
0.05
0.00
Ph
osp
ho
rus F
ilte
rab
le (
mg
/L)
Figure 5.10. Box plot of filterable phosphorus along the Hawkesbury Nepean River System
(showing all the observed data range).
CHAPTER 05: Results
63
N92N881N86N85N75N641N64N57N44N42N35N21N14E851
0.05
0.04
0.03
0.02
0.01
0.00
Ph
osp
ho
rus F
ilte
rab
le (
mg
/L)
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
64
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.
N92N881N86N85N75N67N641N64N57N44N42N35N21N14E851
7
6
5
4
3
2
1
0
Nit
rog
en
To
tal (m
g/
L)
Figure 5.12. Box plot of total nitrogen along the Hawkesbury Nepean River System.
CHAPTER 05: Results
65
N92
N881N8
6N8
5N7
5N6
7N6
41N64
N57
N44
N42
N35
N21
N14
E851
3.0
2.5
2.0
1.5
1.0
0.5
0.0
Nit
rog
en
To
tal (m
g/
L)
ANZECC upper limit (0.35mg/l)
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
66
N92
N881N8
6N8
5N7
5N6
7N6
41N64
N57
N44
N42
N35
N21
N14
E851
6
5
4
3
2
1
0
Nit
rog
en
Oxid
ise
d (
mg
/L)
ANZECC upper limit (0.25 mg/l)
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
67
N92N881N86N85N75N641N64N44N35
1.2
1.0
0.8
0.6
0.4
0.2
0.0
Nit
rog
en
Am
mo
nia
ca
l (m
g/
L)
Figure 5.15. Box plot of ammoniacal nitrogen along the Hawkesbury Nepean River System
(showing all the observed data range).
CHAPTER 05: Results
68
N92N881N86N85N75N641N64N44N35
0.5
0.4
0.3
0.2
0.1
0.0
Nit
rog
en
Am
mo
nia
ca
l (m
g/
L)
ANZECC upper limit (0.1 mg/l)
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
69
N92N881N86N85N75N641N64N57N44N42N35N21N14E851
6
5
4
3
2
1
0
Nit
rog
en
TK
N (
mg
/L)
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
70
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
71
N92
N881N8
6N8
5N7
5N6
7
N641N6
4N5
7N4
4N4
2N3
5N2
1N1
4E8
51
300
250
200
150
100
50
0
Alk
alin
ity
(m
gC
aC
O3
/L)
ANZECC upper value (20 mgCaCO3/l)
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
72
(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.
N92N881N86N85N75N67N641N64N57N44N42N35N21N14E851
400
300
200
100
0
Su
sp
en
de
d S
olid
s (
mg
/L)
Figure 5.19. Box plot of suspended solids along the Hawkesbury Nepean River System
(showing all the observed data range).
CHAPTER 05: Results
73
N92
N881N8
6N8
5N7
5N6
7N6
41N64
N57
N44
N42
N35
N21
N14
E851
50
40
30
20
10
0
Su
sp
en
de
d S
olid
s (
mg
/L)
ANZECC upper limit (20 mg/l)
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
74
whiskers to 1011 and 163269). The minimum range of data distribution can be seen at
E851(whiskers to 1744 and 12097).
N92N85N75N67N641N64N57N44N42N35N21N14E851
700000
600000
500000
400000
300000
200000
100000
0
Alg
al To
tal C
ou
nt
(ce
lls/
mL)
Figure 5.21. Box plot of algal total count along the Hawkesbury Nepean River System
(showing all the observed data range).
CHAPTER 05: Results
75
N92N85N75N67N641N64N57N44N42N35N21N14E851
200000
150000
100000
50000
0
Alg
al To
tal C
ou
nt
(ce
lls/
mL)
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
76
N92N881N86N85N75N67N641N64N57N44N42N35N21N14E851
250
200
150
100
50
0
Ch
loro
ph
yll-
a (
ug
/L)
Figure 5.23. Box plot of chlorophyll-a along the Hawkesbury Nepean River System
(showing all the observed data range).
CHAPTER 05: Results
77
N92
N881N8
6N8
5N7
5N6
7N6
41N64
N57
N44
N42
N35
N21
N14
E851
50
40
30
20
10
0
Ch
loro
ph
yll-
a (
ug
/l)
ANZECC upper limit (5 ug/l)
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
78
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
79
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
80
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
81
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
82
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
83
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
84
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
85
Table 5.6: Median values of water quality parameters and ANZECC (2000) guidelines
Variable
N14 N21 N35 N42 N44 N57 N67 N75 N92
ANZECC
trigger
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
86
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
87
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
88
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.
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
0
2
4
6
8
10
12
14
N35 - Dissolved Oxygen (mg/L)
CHAPTER 05: Results
89
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.
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
0
10
20
30
40
50
60
N14 - Conductivity Field (mS/cm)
CHAPTER 05: Results
90
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
91
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
92
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
93
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
94
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
95
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
0
10
20
30
40
50
60
70
N92 - Alkanility (mgCaCO3)
CHAPTER 05: Results
96
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.
Figure 5.30. Increasing trends of reactive silicate at station N35.
0
1
2
3
4
5
6
7
N35 - Silicate Reactive (SiO2 mg/L)
CHAPTER 05: Results
97
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
98
Table 5.8: Correlations among water quality parameters at station N44 of the HNRS
pH
Nitrogen TKN
Temperature
Chlorophyll-a
Nitrogen Total
Dissolved Oxygen
Phosphorus Total
Suspended Solids
Nitrogen Oxidised
Conductivity
Nitrogen Ammonia
cal
Algal Total Count
Phosphorus
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 Oxygen 0.443 -0.222 -0.29 0.387 0.049 1
Phosphorus Total -0.255 0.268 0.219 0.355 0.263 -0.047 1
Suspended Solids -0.048 -0.108 0.007 0.552 0.402 0.226 0.303 1
Nitrogen Oxidised -0.177 -0.037 -0.716 -0.185 0.976 0.097 -0.321 -0.386 1
Conductivity 0.301 0.455 -0.223 -0.185 0.698 -0.005 -0.365 -0.382 0.608 1
Nitrogen Ammoniacal 0.049 0.472 0.113 -0.199 0.124 -0.342 0.129 -0.188 0.021 0.343 1
Algal Total Count 0.103 0.023 0.213 0.819 -0.317 0.358 0.263 0.611 -0.329 -0.180 -0.135 1
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
99
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
100
Figure 5.31. Plot of standardized residuals against estimate for Chlorophyll-a.
Figure 5.32. Plot of standardized residuals against estimate for total nitrogen.
CHAPTER 05: Results
101
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.
CHAPTER 05: Results
102
Table 5.9: Water quality parameters and ANZEC Guidelines for Fresh and Marine Water
Quality
Water Quality Parameter Non-compliance
if: Value1 Value 2 Unit
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).
CHAPTER 05: Results
103
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).
CHAPTER 05: Results
104
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. .
05
1015202530354045505560
N14 N21 N35 N42 N44 N57 N67 N75 N92
WQ
I
Water quality monitoring stations
Marginal Poor
CHAPTER 05: Results
105
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.
0.0
10.0
20.0
30.0
40.0
50.0
60.0
70.0
80.0
90.0
100.0
N14 N21 N35 N42 N44 N57 N67 N75 N92
F1
F2
F3
CHAPTER 05: Results
106
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
CHAPTER 05: Results
107
Table 5.11: Water quality results at N14
25 < percentage Failed < 50 = , , PercentageFailed >= 50 =
Index
Period
Number
of Tests
Number
of Failed
Tests
Percent
Failed
(%)
Number
of Tests
Number
of Failed
Tests
Percent
Failed
(%)
Number
of Tests
Number
of Failed
Tests
Percent
Failed
(%)
Number
of Tests
Number
of Failed
Tests
Percent
Failed
(%)
Number
of Tests
Number
of Failed
Tests
Percent
Failed
(%)
Number
of Tests
Number
of Failed
Tests
Percent
Failed
(%)
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
Aluminium Total True Colour Alkalinity Suspended Solids ConductivityIron Total
CHAPTER 05: Results
108
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
CHAPTER 05: Results
109
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.
CHAPTER 05: Results
110
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.
CHAPTER 05: Results
111
5.7.1 pH
Figure 5.37. pH values at S1 and N67.
Figure 5.38. pH values at S2 and N57.
6.4
6.6
6.8
7
7.2
7.4
7.6
7.8
8
pH
Date
S1 Vs N67
6
6.2
6.4
6.6
6.8
7
7.2
7.4
7.6
7.8
8
pH
Date
S2 Vs N57
CHAPTER 05: Results
112
Figure 5.39. pH values at S3 and N44.
5.7.2 Dissolved Oxygen
Figure 5.40. pH values at S1 and N67.
7
7.1
7.2
7.3
7.4
7.5
7.6
7.7
7.8
7.9
8
pH
Date
S3 Vs N44
0
2
4
6
8
10
12
DO
Date
S1 Vs N67
CHAPTER 05: Results
113
Figure 5.41. pH values at S2 and N57.
Figure 5.42. pH values at S3 and N44.
0
2
4
6
8
10
12
14
DO
Date
S2 Vs N57
0
2
4
6
8
10
12
14
DO
Date
S3 Vs N44
CHAPTER 05: Results
114
5.7.3 Electrical Conductivity
Figure 5.43. Electrical conductivity at S1 and N67.
Figure 5.44. Electrical conductivity at S2 and N57.
0
100
200
300
400
500
EC
Date
S1 Vs N67
0
100
200
300
400
EC
Date
S2 Vs N57
CHAPTER 05: Results
115
Figure 5.45. Electrical conductivity at S3 and N44.
5.7.4 Turbidity
Figure 5.46. Turbidity at S1 and N67.
0
100
200
300
400
EC
Date
S3 Vs N44
0
10
20
30
40
50
60
70
80
Tu
rbid
ity
Date
S1 Vs N67
CHAPTER 05: Results
116
Figure 5.47. Turbidity at S2 and N57.
Figure 5.48. Turbidity at S3 and N44.
0
5
10
15
20
25
30
Tu
rbid
ity
Date
S2 Vs N57
0
5
10
15
20
25
30
35
40
Tu
rbid
ity
Date
S3 Vs N44
CHAPTER 05: Results
117
5.7.5 Nitrogen Oxides
Figure 5.49. Nitrogen oxides at S1 and N67.
Figure 5.50. Nitrogen oxides at S2 and N57.
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
NO
x m
g/L
Date
S1 Vs N67
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
0.45
NO
x m
g/L
Date
S2 Vs N57
CHAPTER 05: Results
118
Figure 5.51. Nitrogen oxides at S3 and N44.
5.7.6 Ammonical Nitrogen
Figure 5.52. Ammonical nitrogen at S1 and N67.
0
0.1
0.2
0.3
0.4
0.5
0.6
NO
x m
g/L
Date
S3 Vs N44
0
0.01
0.02
0.03
0.04
0.05
0.06
Nh
H-N
mg/L
Date
S1 Vs N67
CHAPTER 05: Results
119
Figure 5.53. Ammonical nitrogen at S2 and N57.
Figure 5.54. Ammonical nitrogen at S3 and N44.
0
0.004
0.008
0.012
0.016
0.02
NH
3-N
mg/L
Date
S2 Vs N57
0
0.01
0.02
0.03
0.04
0.05
0.06
NH
3-N
mg/L
Date
S3 Vs N44
CHAPTER 05: Results
120
5.7.7 Temperature
Figure 5.55. Temperature at S1 and N67.
Figure 5.56. Temperature at S2 and N57.
0
5
10
15
20
25
30
Tem
per
atu
re
Date
S1 Vs N67
0
5
10
15
20
25
30
Tem
per
atu
re
Date
S2 Vs N57
CHAPTER 05: Results
121
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.
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
0
5
10
15
20
25
30
Tem
per
atu
re
Date
S3 Vs N44
CHAPTER 05: Results
122
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
CHAPTER 05: Results
123
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.
CHAPTER 06: Conclusion
124
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.
CHAPTER 06: Conclusion
125
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
CHAPTER 06: Conclusion
126
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. `
CHAPTER 06: Conclusion
127
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.
CHAPTER 06: Conclusion
128
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.
CHAPTER 06: Conclusion
129
References
Alberto, W. D., Marıa del Pilar, D., Marıa Valeria, A., Fabiana, P. S., Cecilia, H. A. & Marıa
de los Ángeles, B. (2001). Pattern recognition techniques for the evaluation of spatial and
Temporal variations in water quality. A case study: Suquıa 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
CHAPTER 06: Conclusion
130
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
CHAPTER 06: Conclusion
131
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 Kríz (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.
CHAPTER 06: Conclusion
132
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-
CHAPTER 06: Conclusion
133
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, south-
eastern 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
CHAPTER 06: Conclusion
134
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
CHAPTER 06: Conclusion
135
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.
CHAPTER 06: Conclusion
136
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.
CHAPTER 06: Conclusion
137
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.
CHAPTER 06: Conclusion
138
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.