A Case Study of the Hawkesbury-Nepean River System

DEVELOPMENT OF INDICATORS AND FRAMEWORK FOR ASSESSING RIVER HEALTH IN PERI-URBAN LANDSCAPES: A CASE STUDY OF THE HAWKESBURY-NEPEAN RIVER SYSTEM

Mihindukulasooriya Uthpala Ananda Pinto

A thesis submitted in fulfilment of the requirements for the degree of Doctor of Philosophy Degree

School of Science and Health

University of Western Sydney

Australia

FEBRUARY 2013

This thesis is dedicated my parents and beloved wife for their endless support and encouragement.

ACKNOWLEDGEMENT

It is with great pleasure that I express my sincere gratitude to my principal supervisor, Prof. Basant Maheshwari, for his continuous encouragement, advice, and guidance. He has been a source of generosity, insight and inspiration; guiding me in all my efforts throughout my candidature. I owe my research achievements to his enthusiastic supervision. I extend my profound gratitude to my co-supervisors Assoc. Prof. Charles Morris and Assoc. Prof. Surendra Shrestha who provided me with the unflinching encouragement, support, and feedback during the candidature. Successful completion of this thesis would not have been possible without your invaluable insights and comments on my work.

I gratefully acknowledge the University of Western Sydney and the Australian Government for granting me the Australian Postgraduate Award, which gave me the opportunity to be exposed to a new knowledge base. I appreciate the travel support given by the School of Science and Health for my attending national and international conferences. A large number of experts from government agencies involved in this project through provision of historical data and valuable feedback. Especially, I want to acknowledge Maree Abood and Shane Barter from Office of the Hawkesbury-Nepean, Bill Dixon, Diana Shanks and Paul Bennett from the Hawkesbury-Nepean Catchment Management Authority, David Makin from the New South Wales Department of Primary Industries and Tracey Schultz from Sydney Catchment Authority.

I would like to thank all current and previous technical staff, general staff and academics of University of Western Sydney including Prof. Robert Hodge for assistance and feedback on key informants’ interviews and community surveys, Adj. Assoc. Prof. Bruce Simmons for providing valuable resources about the river system and suggesting key people in the relevant areas, Prof. Richard Ollerton for providing feedback on time series analysis, Michael Franklin, Mark Emmanuel, Maree Gorham, Julie Langford and Liz Kabanoff for friendly assistance in laboratory analysis, Jocelyn Applebee for editing this thesis, and all others who directly or indirectly helped me during my candidature. I am also thankful to Derek Cannon for phytoplankton identification and feedback on Chapter 7. My gratitude also goes to Mary Howard for her ongoing feedback on various issues related to river health of the Hawkesbury Nepean River, Michael Miller and Gayle Miller for their generous support in data collection on the river.

As always, my heartfelt gratitude goes to my mother, father, and uncle for their love and constant support throughout my life and for inspiring me to pursue an academic career. I can never forget their warmth and inspiration. Finally, my most tender and sincere thanks go to my loving wife, Harshini Pinto, who has been a shadow behind all my success during the last three years. Her understanding throughout these years has meant more than I can ever express......

DECLARATION

Author: M. Uthpala A. Pinto

Degree: Ph. D

Date: Monday, 29 April 2013

I certify that the work presented in this thesis is, to the best of my knowledge and belief, original, except as acknowledged in the text, and that the material has not been submitted, either in full or in part, for a degree at this or any other institution. I certify that I have complied with the rules, requirements, procedures, and policy relating to my higher degree research award of the University of Western Sydney.

Author’s Signature

TABLE OF CONTENTS

CHAPTER 1 INTRODUCTION ...... 4

1.1 RIVER SYSTEMS AND COMMUNITIES ...... 4

1.2 THE EXPANSION OF PERI-URBAN LANDSCAPES ...... 4

1.3 ISSUES WITH RIVER HEALTH MEANING AND ASSESSMENT ...... 6

1.3.1 WHAT IS RIVER HEALTH ...... 6

1.3.2 RIVER HEALTH ASSESSMENT AND WATER QUALITY MONITORING ...... 7

1.4 OVERALL OBJECTIVES ...... 8

1.5 EXPECTED OUTCOMES, VALUES AND BENEFITS ...... 9

1.5.1 WHY IS THIS PARTICULAR PIECE OF RESEARCH WORTH DOING? ...... 9

1.5.2 WHAT SPECIAL GROUPS STAND TO BENEFIT? ...... 9

1.5.3 WHAT FURTHER AVENUES OF RESEARCH WILL THE INFORMATION OPEN UP? …………………………………………………………………………………10

1.6 RESEARCH QUESTIONS ...... 10

1.7 FORMAT OF THE THESIS ...... 11

CHAPTER 2 THE STUDY AREA ...... 14

SUMMARY ...... 14

2.1 THE HAWKESBURY-NEPEAN RIVER SYSTEM ...... 14

2.2 THE HAWKESBURY-NEPEAN CATCHMENT ...... 16

2.3 THE CATCHMENT IN TRANSITION- PRESSURES OF URBANISATION ...... 17

CHAPTER 3 MEANING OF RIVER HEALTH : COMMUNITY PERSPECTIVES ...... 19

SUMMARY ...... 19

3.1 GENERAL ...... 19

3.1.1 RIVER HEALTH VERSUS HUMAN HEALTH ...... 20

3.1.2 ISSUES WITH CURRENT RIVER HEALTH MEANINGS ...... 21

3.2 SURVEY METHODOLOGY ...... 22

3.2.1 FORMULATION OF QUESTIONNAIRE ...... 22

3.2.2 PILOT SURVEY AND HUMAN ETHICS CLEARANCE ...... 23

3.2.3 RECRUITMENT OF SURVEY PARTICIPANTS ...... 24

3.2.4 ANALYSIS OF DATA ...... 25

3.3 RESULTS AND DISCUSSION ...... 25

3.3.1 GENERAL ...... 25

3.3.2 SUSTAINING ECOLOGICAL INTEGRITY ...... 30

3.3.3 VISUAL APPEAL ...... 32

3.3.4 MAINTAINING HYDROLOGIC BALANCE ...... 34

3.3.5 RIVER WATER FIT FOR PURPOSE ...... 35

3.3.6 INFLUENCE OF AGE AND GENDER ...... 37

3.3.7 THE MEANING OF RIVER HEALTH: COMMUNITY VERSUS EXPERTS ...... 38

3.4 CONCLUDING REMARKS ...... 39

CHAPTER 4 MEANING OF RIVER HEALTH : KEY INFORMANTS’ PERSPECTIVES ...... 41

SUMMARY ...... 41

4.1 GENERAL ...... 41

4.2 DATA COLLECTION AND ANALYSIS ...... 42

4.2.1 DEVELOPING THE INTERVIEW QUESTIONS ...... 42

4.2.2 SELECTION AND BACKGROUND OF KEY INFORMANTS ...... 43

4.2.3 THE INTERVIEW PROCESS ...... 45

4.3 RESULTS AND DISCUSSION ...... 46

4.3.1 MEANING OF RIVER HEALTH ...... 46

4.4 IMPACTS ON PERI-URBAN RIVER HEALTH ...... 49

4.4.1 RESIDENTIAL DEVELOPMENTS ...... 49

4.4.2 LEISURE ACTIVITIES AND WATER QUALITY ...... 50

4.4.3 POLICY IMPACTS ...... 51

4.5 INDICATORS FOR RIVER HEALTH ASSESSMENT...... 52

4.6 RIVER MANAGEMENT - WHO IS MORE EFFECTIVE? ...... 55

4.7 COMMUNICATION OF RIVER HEALTH ...... 56

4.8 CONCLUDING REMARKS ...... 57

CHAPTER 5 MANAGING RIVER HEALTH: ISSUES AND CHALLENGES ...... 59

SUMMARY ...... 59

5.1 GENERAL ...... 59

5.2 DATA COLLECTION AND ANALYSIS ...... 60

5.2.1 ANALYSIS OF THE SURVEY DATA ...... 60

5.3 RESULTS ...... 61

5.3.1 PARTICIPANTS’ PROFILE ...... 61

5.3.2 WHAT DO PARTICIPANTS VALUE THE MOST OF A PERI-URBAN RIVER

SYSTEM? ...... 62

5.3.3 EFFECTS OF PARTICIPANT’S SOCIO-DEMOGRAPHIC CHARACTERISTICS ...... 64

5.3.4 RIVER HEALTH INDICATORS ...... 67

5.3.5 RIVER HEALTH IMPROVEMENT PROGRAMS AND COMMUNITY SATISFACTION …………………………………………………………………………………69

5.3.6 PRACTICAL ACTIONS AND RIVER HEALTH COMMUNICATION ...... 71

5.4 DISCUSSION ...... 73

5.4.1 COMMUNITY VALUES OF RIVER SYSTEMS ...... 73

5.4.2 FACTORS AFFECTING THE UNDERSTANDING OF RIVER HEALTH ...... 74

5.4.3 COMMUNITY ENGAGEMENT AND RIVER HEALTH MANAGEMENT ...... 76

5.4.4 COMMUNICATING ABOUT RIVER HEALTH ...... 77

5.5 CONCLUDING REMARKS ...... 78

CHAPTER 6 SPATIAL AND TEMPORAL TRENDS IN RIVER WATER QUALITY ...... 80

SUMMARY ...... 80

6.1 GENERAL ...... 81

6.2 DATA ANALYSES ...... 83

6.2.1 SCREENING THE DATA SET FOR ANALYSIS ...... 83

6.2.2 FACTOR ANALYSIS ...... 85

6.2.3 CLUSTER ANALYSIS ...... 87

6.2.4 TREND ANALYSIS ...... 88

6.3 RESULTS AND DISCUSSION ...... 89

6.3.1 DESCRIPTIVE STATISTICS OF THE DATA SET ...... 89

6.3.2 FACTOR ANALYSIS ...... 90

6.3.3 HIERARCHICAL AGGLOMERATIVE CLUSTER ANALYSIS ...... 93

6.3.4 TREND ANALYSIS ...... 97

6.4 CONCLUDING REMARKS ...... 103

CHAPTER 7 BIOTIC ASSEMBLAGES AS RIVER HEALTH INDICATORS ...... 104

SUMMARY ...... 104

7.1 GENERAL ...... 105

7.2 MATERIALS AND METHODS...... 106

7.2.1 LOCATION SELECTION AND CHARACTERISATION ...... 106

7.2.2 WATER QUALITY VARIABLES ...... 108

7.2.3 PHYTOPLANKTON AND BENTHIC-MACROINVERTEBRATE SAMPLING ...... 109

7.2.4 TAXA IDENTIFICATION ...... 110

7.2.5 DATA ANALYSIS ...... 110

7.3 RESULTS ...... 112

7.3.1 WATER QUALITY ...... 112

iii

7.3.2 THE COMPOSITION OF PHYTOPLANKTON COMMUNITIES ...... 118

7.3.3 THE COMPOSITION OF BENTHIC-MACROINVERTEBRATE COMMUNITIES ..... 123

7.4 DISCUSSION ...... 125

7.4.1 SPATIAL AND TEMPORAL TRENDS IN THE WATER QUALITY ...... 125

7.4.2 PHYTOPLANKTON COMMUNITIES ...... 128

7.4.3 BENTHIC-MACROINVERTEBRATES...... 130

7.5 CONCLUDING REMARKS ...... 132

CHAPTER 8 IMPACTS OF WATER QUALITY ON AQUATIC LIFE IN RIVER SYSTEM ...... 134

SUMMARY ...... 134

8.1 GENERAL ...... 134

8.2 MATERIALS AND METHODS...... 136

8.2.1 DATA COLLECTION ...... 136

8.2.2 STATISTICAL ANALYSIS...... 139

8.3 RESULTS ...... 139

8.4 DISCUSSION ...... 142

8.4.1 WATER TEMPERATURE AND PRAWN HARVEST ...... 142

8.4.2 DISSOLVED OXYGEN AND PRAWN HARVEST ...... 143

8.4.3 PRAWN HARVEST AND OTHER VARIABLES ...... 144

8.4.5 IMPLICATIONS FOR PRAWN INDUSTRY ...... 145

8.5 CONCLUDING REMARKS ...... 146

CHAPTER 9 KEY INDICATORS FOR RIVER HEALTH ...... 147

SUMMARY ...... 147

9.1 ISSUES IN RIVER HEALTH ASSESSMENT ...... 147

9.2 IDENTIFYING KEY VARIABLES ...... 149

9.3 MATERIALS AND METHODS...... 150

9.3.1 APPLICATION OF MULTIVARIATE TECHNIQUES ...... 150

9.3.2 DATA ANALYSIS ...... 150

9.4 RESULTS AND DISCUSSION ...... 152

9.4.1 VARIABILITY IN WATER QUALITY DATA ...... 152

9.4.2 IDENTIFICATION OF WATER QUALITY FACTORS ...... 157

9.4.3 KEY WATER QUALITY VARIABLES FOR MONITORING ...... 160

9.5 CONCLUDING REMARKS ...... 161

CHAPTER 10 DEVELOPING RIVER HEALTH ASSESSMENT TOOLS ...... 162

SUMMARY ...... 162

10.1 GENERAL ...... 162

10.1.1 COMPLEXITY OF ASSESSING EUTROPHICATION ...... 163

10.1.2 ENTEROCOCCI AS AN INDICATOR OF MICROBIAL WATER QUALITY ...... 164

10.2 MATERIAL AND METHODS ...... 166

10.2.1 DISCRIMINANT FUNCTION ANALYSIS APPROACH ...... 166

10.3 DATA COLLECTION AND ANALYSIS ...... 168

10.3.1 DATA COLLECTION ...... 168

10.3.2 MODEL DEVELOPMENT ...... 169

10.3.3 MODEL VALIDATION ...... 172

10.4 RESULTS AND DISCUSSION ...... 173

10.4.1 GENERAL ...... 173

10.4.2 ROLE OF KEY VARIABLES IN THE EUTROPHICATION RISK MODEL ...... 176 10.4.3 CHALLENGES OF MODELLING CHLOROPHYLL a ...... 178

10.4.4 ROLE OF KEY VARIABLES IN THE MICROBIAL RISK MODEL ...... 180

10.5 CONCLUDING REMARKS ...... 182

CHAPTER 11 BRINGING IT TOGETHER : A FRAMEWORK FOR ASSESSING RIVER HEALTH ...... 184

SUMMARY ...... 184

11.1 GENERAL ...... 184

11.1.1 UNDERSTAND THE RIVER SYSTEM ...... 186

11.1.2 IDENTIFY THE RIVER HEALTH INDICATORS ...... 1 88

11.1.3 DEVELOP AND APPLY PREDICTIVE TOOLS ...... 189

11.2 RIVER HEALTH ASSESSMENT FRAMEWORK IN PRACTICE – AN EXAMPLE OF

PRACTICAL CONTEXT ...... 190

11.2.1 UNDERSTAND...... 190

11.2.2 IDENTIFY ...... 192

11.2.3 DEVELOP AND APPLY...... 193

11.3 CONCLUDING REMARKS ...... 193

CHAPTER 12 CONCLUSIONS AND FUTURE RESEARCH ...... 194

SUMMARY ...... 194

12.1 MEANING OF RIVER HEALTH...... 194

12.2 VIEWS AND EXPECTATIONS OF STAKEHOLDERS ...... 195

12.3 THE KEY RIVER HEALTH INDICATORS ...... 195

12.4 RIVER HEALTH ASSESSMENT FRAMEWORK AND TOOLS ...... 196

12.5 SUGGESTIONS FOR FUTURE RESEARCH ...... 196

REFERENCES ...... 198

APPENDICES ...... 227

LIST OF FIGURES

FIGURE 1. HAWKESBURY-NEPEAN RIVER SYSTEM AND ITS CATCHMENT...... 17

FIGURE 2. MAP OF HAWKESBURY-NEPEAN RIVER SYSTEM INDICATING MAJOR

SUBURBS...... 23

FIGURE 3. DESCRIPTORS OF RIVER HEALTH IDENTIFIED BASED ON THE RESPONSES

OF SURVEY PARTICIPANTS...... 26

FIGURE 4. DIFFERENT DESCRIPTORS FOR DEFINING THE MEANING OF RIVER

HEALTH...... 26

FIGURE 5. EXAMPLES OF WORDS USED BY THE SURVEY PARTICIPANTS TO DESCRIBE

RIVER HEALTH IN THE SUSTAINING ECOLOGICAL INTEGRITY THEME...... 27

FIGURE 6. EXAMPLES OF WORDS USED BY THE SURVEY PARTICIPANTS TO DESCRIBE

RIVER HEALTH IN THE VISUAL APPEAL THEME...... 28

FIGURE 7. EXAMPLES OF WORDS USED BY THE SURVEY PARTICIPANTS TO DESCRIBE

RIVER HEALTH IN THE MAINTAINING HYDROLOGIC BALANCE THEME...... 29

FIGURE 8. KEY CONSTITUENTS OF RIVER HEALTH...... 48

FIGURE 9. MAJOR FACTORS AFFECTING RIVER HEALTH...... 50

FIGURE 10. INDICATORS OF RIVER HEALTH...... 54

FIGURE 11. IMPORTANCE OF THE DIFFERENT FUNCTIONS OF THE HAWKESBURY-

NEPEAN RIVER SYSTEM TO THE SURVEY PARTICIPANTS...... 63

FIGURE 12. VALUES OF THE HAWKESBURY-NEPEAN RIVER TO THE SURVEY

PARTICIPANTS...... 63

FIGURE 13. VIEWS OF THE SURVEY PARTICIPANTS FOR THE PRESENT CONDITION OF

RIVER HEALTH...... 65

FIGURE 14. SYMMETRIC PLOT OF CORRESPONDENCE ANALYSIS REPRESENTING THE

PRESENT STATUS OF THE HNR SYSTEM WITH PARTICIPANT’S (A) YEARS OF

RESIDENCY AND (B) AGE...... 66

FIGURE 15. INDICATORS OF GENERAL RIVER HEALTH PROPOSED BY THE SURVEY

PARTICIPANTS...... 69

FIGURE 16. COMMUNITY SATISFACTION OF RIVER HEALTH IMPROVEMENT

PROGRAMS...... 70

vii

FIGURE 17. SYMMETRIC PLOT OF CORRESPONDENCE ANALYSIS REPRESENTING THE

COMMUNITY SATISFACTION OF RIVER HEALTH IMPROVEMENT PROGRAMS WITH

PARTICIPANTS RESIDENCE TIME...... 70

FIGURE 18. PRACTICAL ACTIONS TO IMPROVE THE RIVER HEALTH PROPOSED BY

THE SURVEY PARTICIPANTS...... 72

FIGURE 19. PARTICIPANTS’ VIEW ON COMMUNICATING RIVER HEALTH TO WIDER

COMMUNITY...... 72

FIGURE 20. SCHEMATIC REPRESENTATION OF THE HAWKESBURY-NEPEAN RIVER

AND ITS TRIBUTARIES...... 84

FIGURE 21. PCA AND FA METHODOLOGY...... 87

FIGURE 22. DENDROGRAMS SHOWING CLUSTERING OF MONITORING STATIONS

BASED ON WATER QUALITY IN YEAR 1997 AND 2000...... 96

FIGURE 23. MULTIVARIATE MK TEST RESULTS FOR TEMPORAL TRENDS...... 101

FIGURE 24. LONG-TERM MEDIAN NOX VARIATION ALONG HNR...... 102

FIGURE 25. LONG-TERM MEDIAN CHLOROPHYLL A VARIATION ALONG HNR...... 102

FIGURE 26. MAP WITH KEY LOCATIONS ON THE HAWKESBURY-NEPEAN RIVER

SYSTEM...... 107

FIGURE 27. SPATIAL AND TEMPORAL TRENDS OF ABIOTIC WATER QUALITY

VARIABLES...... 114

FIGURE 28. SPATIAL AND TEMPORAL TRENDS OF BIOTIC WATER QUALITY

VARIABLES...... 115

FIGURE 29. SEASONAL RAINFALL PATTERNS ACROSS SAMPLING LOCATION...... 115

FIGURE 30. PCA EIGEN VECTOR PLOT OF WATER QUALITY VARIABLES...... 116

FIGURE 31. THE ABUNDANCE AND MEAN SHANNON DIVERSITY OF PHYTOPLANKTON

SPECIES ACROSS FOUR LOCATIONS IN DIFFERENT SEASONS. ERROR BARS

REPRESENT ±SE...... 120

FIGURE 32. DOMINANT PHYTOPLANKTON SPECIES DURING FEBRUARY, MAY AND

AUGUST...... 121

FIGURE 33. THE ABUNDANCE, MEAN SHANNON DIVERSITY AND EPT INDEX OF

BENTHIC MACROINVERTEBRATES. ERROR BARS REPRESENT ±SE...... 124

FIGURE 34. MAP OF THE HAWKESBURY-NEPEAN RIVER SYSTEM INDICATING THE

MAIN TRIBUTARIES AND DATA COLLECTION POINTS...... 138

FIGURE 35. SCHEMATIC DIAGRAM OF THE DISTRIBUTION OF DATA COLLECTION

POINTS ALONG THE HNR...... 138 viii

FIGURE 36. DENDROGRAM OF HIERARCHICAL AGGLOMERATIVE CLUSTER ANALYSIS

INDICATING THE SIMILARITIES AMONG WATER QUALITY VARIABLES, PRAWN

HARVEST AND RAINFALL USING WARD LINKAGE MEASURE AND SQUAD

EUCLIDEAN DISTANCES...... 140

FIGURE 37. THE PATTERN OF VARIATION OF (A) PRAWN HARVEST AND WATER

TEMPERATURE, (B) PRAWN HARVEST AND RAINFALL AND (C) TEMPERATURE AND

DISSOLVED OXYGEN...... 141

FIGURE 38. MAP WITH KEY LOCATIONS ON THE HAWKESBURY-NEPEAN RIVER

SYSTEM ...... 151

FIGURE 39. A SAMPLE OF PLOTS SHOWING THE EXTENT CORRELATION AMONG THE

SELECTED VARIABLES...... 153

FIGURE 40. MAP WITH KEY LOCATIONS ON THE HAWKESBURY-NEPEAN RIVER

SYSTEM...... 168

FIGURE 41. CALCULATION OF CUT-OFF SCORE FOR EUTROPHICATION RISK MODEL...... 171

FIGURE 42. CALCULATION OF CUT-OFF SCORE FOR MICROBIAL RISK MODEL. .... 171

FIGURE 43. RELATIONSHIP BETWEEN TN AND CHLOROPHYLL A...... 180

FIGURE 44. RELATIONSHIP BETWEEN TP AND CHLOROPHYLL A...... 180

FIGURE 45. THE FRAMEWORK FOR ASSESSING PERI-URBAN RIVER HEALTH

CONDITION...... 186

LIST OF TABLES

TABLE 1. A LAYOUT OF THE THESIS CHAPTERS...... 13

TABLE 2. THE EFFECT OF PARTICIPANTS’ DISTANCE OF RESIDENCE FROM THE RIVER

ON THE TOP 5 DESCRIPTORS OF RIVER HEALTH...... 30

TABLE 3. DESCRIPTORS AND MAJOR THEMES FOR DESCRIBING RIVER HEALTH. .... 32

TABLE 4. EFFECT OF PARTICIPANTS’ AGE ON THE TOP 5 DESCRIPTORS OF RIVER

HEALTH...... 38

TABLE 5. THE EFFECT OF PARTICIPANTS’ GENDER ON THE TOP 5 DESCRIPTORS OF

RIVER HEALTH...... 38

TABLE 6. LIST OF KEY INFORMANT INTERVIEW QUESTIONS...... 43

TABLE 7. AREAS OF EXPERTISE OF KEY INFORMANTS...... 45

TABLE 8. SOCIO-DEMOGRAPHIC FEATURES OF SURVEY PARTICIPANTS...... 61

TABLE 9. KEY QUESTIONS USED IN THE SURVEY QUESTIONNAIRE...... 62

TABLE 10. SIGNIFICANCE OF SOCIO-DEMOGRAPHIC FACTORS ON COMMUNITY VIEW

ON THE CONDITION OF RIVER...... 67

TABLE 11. SIGNIFICANCE OF SOCIO-DEMOGRAPHIC FACTORS INFLUENCING

COMMUNITY SATISFACTION OF RIVER HEALTH IMPROVEMENT PROGRAMS...... 71

TABLE 12. DISTANCES OF THE MONITORING STATIONS FROM OCEAN...... 85

TABLE 13. DESCRIPTIVE STATISTICS OF RIVER DATA...... 90

TABLE 14. THE VALUES OF THE CORRELATION MATRIX OF THE SIX PHYSICO-

CHEMICAL VARIABLES...... 93

TABLE 15. EIGENVALUES FOR THE FIRST THREE PCS FOR THE WATER QUALITY

DATA SET OF HNR...... 93

TABLE 16. ROTATED FACTOR LOADINGS FOR EXTRACTED PRINCIPAL COMPONENTS...... 93

TABLE 17. TWO CLUSTERS OBTAINED FROM DENDROGRAMS AT 25-RESCALED

DISTANCE CLUSTER COMBINE LEVEL (NON HIGHLIGHTED-CLUSTER 1,

HIGHLIGHTED- CLUSTER 2)...... 97

TABLE 18. PERMANOVA RESULTS...... 116

TABLE 19. DESCRIPTIVE STATISTICS OF WATER QUALITY VARIABLES...... 117

TABLE 20. PROPERTIES OF THE PCA PLOT FOR WATER QUALITY VARIABLES,

SHOWING THE VARIATION EXPLAINED BY THE INDIVIDUAL PCS AND CUMULATIVE

PERCENTAGE OF VARIATION...... 117

TABLE 21. PROPERTIES OF PCA PLOT FOR WATER QUALITY VARIABLES, SHOWING

VARIABLES AND EIGENVECTORS FOR FOUR PCS...... 118

TABLE 22. SIMPER RESULTS FOR PHYTOPLANKTON SPECIES...... 122

TABLE 23. SUMMARY OF BIOENV ROUTINE CONNECTING MULTIVARIATE

ENVIRONMENTAL PATTERNS WITH BIOTIC ASSEMBLAGES...... 123

TABLE 24. SIMPER RESULTS FOR BENTHIC SPECIES...... 125

TABLE 25. THE DESCRIPTIVE STATISTICS OF THE DATA SET...... 142

TABLE 26. PEARSON CORRELATIONS BETWEEN PRAWN HARVEST, FLOW, RAINFALL,

CHLOROPHYLL A (CHL), TEMPERATURE (TEMP), TOTAL NITROGEN (TN), TOTAL

PHOSPHORUS (TP), TURBIDITY (TURB), DISSOLVED OXYGEN (DO), SUSPENDED

SOLIDS (SS) AND REACTIVE SILICATES (SIL)...... 142

TABLE 27. HIGHLY CORRELATED VARIABLES IN THE STUDY...... 154

TABLE 28. STATISTICAL DESCRIPTIVE OF ORIGINAL WATER QUALITY VARIABLES (2008)...... 155

TABLE 29. STATISTICAL DESCRIPTIVE OF ORIGINAL WATER QUALITY VARIABLES (2009)...... 156

TABLE 30. RESULTS OF KMO AND BARTLETT'S TESTS...... 156

TABLE 31. TOTAL VARIANCE EXPLAINED BY THE VARIFACTORS...... 157

TABLE 32. ROTATED COMPONENT MATRIX...... 157

TABLE 33. PUBLISHED CHLOROPHYLL A GUIDELINES...... 171

TABLE 34. PUBLISHED ENTEROCOCCI GUIDELINES...... 172

TABLE 35. DESCRIPTIVE STATISTICS OF THE DATA SET USED TO DEVELOP THE

MODEL (2008-2009)...... 175

TABLE 36. THE STEPWISE STATISTICS OF THE DISCRIMINANT FUNCTION ANALYSIS...... 175

TABLE 37. MODEL DEVELOPMENT STATISTICS...... 175

TABLE 38. MODEL VALIDATION STATISTICS...... 176

LIST OF PLATES

PLATE 1. THE HAWKESBURY-NEPEAN RIVER AT SACKVILLE...... 15

PLATE 2. REDUCED RIVER CLARITY IS OBSERVED IN MANY PARTS OF THE RIVER

SYSTEM...... 33

PLATE 3. A TURF FARM NEAR WINDSOR, IRRIGATED EXTENSIVELY BY THE RIVER

WATER...... 36

PLATE 4. COLLAPSED RIVER BANK NEAR CATTAI...... 51

PLATE 5. BLUE-GREEN ALGAE WARNING SIGN AT SACKVILLE FERRY...... 57

PLATE 6. OIL SCUM ON THE SURFACE WATER AT PENRITH...... 68

PLATE 7. HNR IS USED BY RESIDENTS FOR RECREATIONAL ACTIVITIES...... 73

PLATE 8. BOUNDARY CREEK STP DISCHARGE POINT AT PENRITH...... 99

PLATE 9. QUALITY OF THE BED SEDIMENTS COLLECTED FROM THE HNR...... 108

PLATE 10. WATER QUALITY SAMPLING AT THE HNR...... 109

PLATE 11. COLLECTION AND IDENTIFICATION OF BENTHIC-MACROINVERTEBRATES...... 110

PLATE 12. DENSE GROWTH OF E. DENSA IN THE PENRITH WEIR POOL...... 127

PLATE 13. NEW HOUSING DEVELOPMENTS IN WESTERN SYDNEY REGION...... 136

xii

ABBREVIATIONS

ASU - Aerial Standard Units Agriculture and Resource Management Council of Australia and ARMCANZ - New Zealand ANZECC - Australian and New Zealand Environment and Conservation Council BDL - Below Detection Level 0C - Celsius CHL - Chlorophyll a cfu - Colony Forming Unit CSIRO - Commonwealth Scientific and Industrial Research Organisation DFA - Discriminant Function Analysis DO - Dissolved Oxygen EC - Electrical Conductivity EPA - Environmental Protection Authority Eqn. - Equation EU - European Union FA - Factor Analysis GPS - Global Positioning System HNCMA - Hawkesbury-Nepean Catchment Management Authority HACA - Hierarchical Agglomerative Cluster Analysis KMO - Keiser-Meyer-Olkin test L - Litre LGA - Local Government Area MK - Mann-Kendall statistical test ML - Mega Litre μS/cm - Microsiemens per centimetre mV - Milli volts mg - Milli gram ml - Milli litre MDS - Multi-Dimensional Scaling National Framework for the Assessment of River and Wetland FARWH - Health NTU - Nephelometric Turbidity Unit NSW - New South Wales ORP - Oxidation Reduction Potential (REDOX potential) NOx - Oxides of Nitrogen ppt - Parts per Thousand % - Percentage PERMANOVA - Permutational Multivariate Analysis of Variance PDFA - Predictive Discriminant Function Analysis PRIMER routine test for linking biota to multivariate environmental BIOENV - patterns PCA - Principal Component Analysis

xiii

SIL - Reactive Silicate REDOX - Reduction Oxidation Potential STP - Sewage Treatment Plant SIMPER - Similarity Percentage SPSS - Statistical Package for Social Sciences SS - Suspended Solids SCA - Sydney Catchment Authority Temp - Temperature HNR - The Hawkesbury-Nepean River System TN - Total Nitrogen TP - Total Phosphorus TA - Trend Analysis TURB - Turbidity UWS - University of Western Sydney VF - Verifactors vUWS - Virtual University of Western Sydney

xiv

ABSTRACT

The main objectives of this thesis are to; (i) examine the concept of river health through views and expectations of stakeholders; (ii) identify and evaluate the key river health indicators related to the main functions and uses of the river system and, (iii) develop a flexible river health assessment framework which can be adapted to other river systems. The Hawkesbury-Nepean River system (HNR), a complex river in peri-urban landscape in south-eastern Australia, is used as a case study to achieve the objectives of this thesis. The first three chapters of this thesis are dedicated to understanding the social implications of river health including its meaning while the remaining chapters progressively explore the key indicators of river health assessment and attempt to develop a simple framework and tools for sustainable management of river health.

Throughout the study, the complexity in defining the concept of river health was highlighted. While the concept of river health was heavily influenced by the personal attitudes and participants’ level of attachment with the river system, both the community participants and key informants agreed on the importance of community satisfaction and ecological integrity as key components in the definition. The study indicated discharge of treated municipal effluent, surface run-off with nutrients from agricultural lands and the presence of exotic weeds (i.e., Egeria densa) as some of the major impacts on peri-urban river health. The community views on the condition of river health were mostly influenced by individual’s distance from the river and age of the participants. The study also indicated the need for sharing the management of river health between the community and the government agencies. However, the efficiency of river health improvement initiatives has been impaired by the presence of multiple organisations and their limited capacity to run river health monitoring and management programs uninterrupted to achieve effective river health outcomes. The community survey and the key interviews further highlighted the need of a suitable river health assessment framework due to a number of competing interests, views and uses associated with peri-urban rivers. Preferably, this framework needs to integrate multiple indicators (i.e., biological, ecological, hydrological, chemical) for various river uses and account for community expectations of the river system.

The environmental component of the study involved an understanding of the key water quality trends of the river system and how they influence the distribution and abundance of phytoplankton, benthic-macroinvertebrate communities, and the

harvest of economically important school prawns (Metapenaeus macleayi). Using historic water quality records collected between 1985 and 2008, and independent data collected in 2011, an attempt was made to characterise the river system and understand the discrete reaches influenced by point and non-point sources of pollution. The former identified two key reaches of the HNR, which showed high and low quality water, while the latter suggested how the water quality was influenced at multiple points due to geographical, natural, and anthropogenic factors. These changes were particularly prominent near confluences and impoundments and water quality tends to vary considerably over short distances in these areas. Out of 40 water quality variables collected from the polluted region, nine variables which statistically account for 50% of the variance in water quality were retained. Subsequent exploration of these variables revealed how they are related to the anaerobic fermentation (pH, turbidity, and dissolved oxygen), microbial pollution (Enterococci and Escherichia coli) and eutrophication (Chlorophyll a, algal biovolume, Manganese, Phaeophytin) aspects of the river system. Using dissolved oxygen, turbidity, and temperature variables, two predictive tools were developed to assess eutrophication and microbial risk in the river. These models intended to predict whether a particular location in the river system is heading to a high or low risk of algal blooms and whether the water quality is suitable for primary contact recreational activities such as swimming. When validated with an independent data set, both models predicted the observed risk category with a reasonable accuracy (above 50%).

The analysis of the primary and secondary data collected in this study indicated that water temperature has an important influence on phytoplankton community structure and downstream prawn harvesting (M. macleayi). The community patterns of benthic macroinvertebrates were influenced by water pH. The historical records revealed how the water temperature has significantly increased since the 1980s and it is expected that future rising temperatures due to climate change may have a significant influence on the phytoplanktons and commercially viable fisheries industry in the river system. Aphanocapsa holsatica and Chironomid larvae appeared as the important indicators for upstream and downstream site differences in water quality. A seasonal succession of phytoplanktons indicated that summer, autumn and winter samples were dominated by Cyanophyta, Chlorophyta and a mix of species respectively. As the final stage of this study, a four-step framework (understand, identify, develop and apply) was proposed to assess river systems based on the key findings of this study. The framework starts by understanding the

social and environment aspects related to river health. This is further evaluated in conjunction with multivariate statistical methods to identify key indicators to assess river health. Particular attention is given to retain variables that are cost effective, easy to measure and less labour intensive for routine monitoring purposes while providing valuable information on the condition of the river system. Finally, this information is utilised to develop reach specific river health assessment tools addressing the key services of the river system (i.e., irrigation, recreation). The framework developed in this thesis indicates a higher degree of flexibility, as it does not advocate a single method of assessment for rivers in different landscapes, considers local knowledge in great detail and attempts to develop tools for key river uses. Overall, the river water quality data analysed during this study helped in identifying a number of key indicators for routine and rapid monitoring purposes. In particular, selected key indicators were incorporated into a framework and predictive tools for assessing eutrophication and microbial risk for recreational activities

* * * * *

CHAPTER 1

INTRODUCTION

1.1 River Systems and Communities

The human recognition of the qualitative aspects of water resources plays a key role in the way we define, measure and manage environmental assets in the landscape. Rivers are one of the most intensively human intervened ecosystems on earth since the settlement of early civilisations. Rivers such as the Nile in Egypt, the Euphrates, and the Tigris in Mesopotamia, the Indus and the Ganges in India and the Yellow River in China have served communities by providing water for agriculture, transport and survival. River systems also play a key role in the environment and society by providing a range of ecosystem functions such as shelter and a food source for an array of biological species, aid in flood management and ecological refuge development (Hoelzl, 2007, Meyer, 1997). Socially, river systems accommodate communities by providing a medium for transport, recreation, tourism, worship, ecosystem services and a place to experience the serenity of nature.

Due to the close association of rivers with society, many river ecosystems are vulnerable to human activity stress, no longer able to efficiently and sustainably provide goods and services (i.e., purification of water, food production and recreational experiences) expected of them by society, particularly in urban and rural regions of the world (Naiman et al., 1995). River systems will continue to be in a constant battle to maintain their natural equilibrium and integrity due to climate change scenarios and rapid expansion of human population and extreme weather conditions. In particular, the expansion of the peri-urban zone over the last decade posed a series of new types of anthropogenic threats on the water quality and quantity of river systems.

1.2 The expansion of Peri-urban Landscapes

A peri-urban region is a diffused territory existing between the urban and rural townships, and river systems in such regions are often used to build dams to store freshwater, extract water for potable and agricultural purposes and discharge treated and untreated municipal effluent originating from urban townships (Adell,

1999, Ford, 1999, Buxton et al., 2006). American geologists coined the term ‘urban- fringe’ in 1940s to describe the landscape between towns and cities. It has since been widely articulated by many researchers as ‘pseudo-suburbs’, ‘rurban- periphery’ and more recently ‘rural-urban-fringe’ to describe the region between countryside and town (Wissink, 1962, Firey, 1946, Pryor, 1968, Thomas, 1974, Sinha, 1993). The environmental, economic and socio-demographic characteristics of the peri-urban regions indicate a considerable degree of distinctness from its neighbouring urban and rural counterparts. This view is well described by Herbers (1989) and Nelson and Dueker (1990) as ‘there is emerging across the continental United States a new form of urban development. It extends far into the rural countryside but within the limits of commuting range to urban and suburban employment opportunities. It is settled by households willing to spend large amounts of their income and commuting budget in pursuit of splendid isolation’.

At present, our knowledge about the peri-urban expansion in Asian countries is scarce, however evidence emerging from Europe, United Kingdom (UK), Latin America, Australia and Africa indicates how new developments and human settlement are gradually expanding into peri-urban fringes (Errington, 1994, Budds and Minaya, 1999, da Gama Torres, 2011, Low Choy et al., 2008, Briggs and Mwamfupe, 2000). The peri-urban regions also expand in size to accommodate the communities who migrated into these diffused territories in search of a prosperous lifestyle and who mostly work in nearby townships thus creating a range of competing and conflicting land-use issues (Nelson and Dueker, 1990, Barr, 2003, Buxton et al., 2006). In Australia, the peri-urban fringes are particularly chosen by ‘forced re-locators’ (low income residents seeking less expensive housing) and ‘free agents’ (retirees, professionals and alternative life-stylers who wish to experience a new way of life) (Buxton and Choy, 2008). The peri-urban settlements are a distinct form of inhabitancy and the state and local authorities are struggling to manage problems relating to new developments on existing infrastructure, spatial fragmentation of land and the demand for limited open space (Buxton et al., 2006, Errington, 1994).

The importance of peri-urban fringes as strategic locations for population growth and their significance in relation to rural, urban and recreational land-use types has now been identified. However, less attention has been given to better understand how such growth is influencing the wellbeing of natural watercourses especially the health of river systems. In particular, greater amounts of uncertainty exist in the

multiple meanings associated with the concept of river health and the variety of qualitative and quantitative methods proposed to assess river health.

1.3 Issues With River Health Meaning and Assessment

1.3.1 What is River Health

The use of the term ‘health’ to describe the condition of rivers is appealing to many politicians and water resource managers around the world (Hart et al., 1999). Nevertheless, the meaning itself is complex and associated with an inherent difficulty in defining, due to competing interests of multiple stakeholders. Should we describe it in the context of social, economic, or environmental values, or should we describe it merely by biological, chemical, or physical indicators? The concept of river health is often influenced by social, political and scientific objectives and therefore in reality it can be quite complex to define and measure. Historically, the use of this term has been restricted to researchers and river health management authorities. Many researchers, also refrained from providing an explicit description of river health; instead they described suitable criteria that assist with river health assessments (Bunn et al., 1999, Boulton, 1999, Norris and Thomas, 1999) while others provided integrated descriptions (Vugteveen et al., 2006).

The river health concept was originally derived from ecosystem health to indicate the condition of a stream to the general public and river managers (Hearnshaw et al., 2005, Karr, 1999). The ecosystem health has been defined as a function of primary productivity, nutrient turnover and changes to species diversity (Rapport et al., 1985). Costanza et al., (1992) defined a healthy system as one that ‘maintains its organisation and autonomy over time and resilient to stress’. Although the term river health was articulated by adapting the behaviour of ecosystem health, some researchers argue that a river is healthy if a single species fisheries unit is sustained but not healthy if other recreational fisheries are lost (Norris and Thomas, 1999). Others, argue whether diversity itself is a sign of good health because the diverse ecosystems are often vulnerable to environmental stress (Chapman, 1992). Alternatively there is evidence that resilience of biological systems is extremely difficult to quantify and interpret (Karr and Thomas, 1996). While, these definitions target river managers, only a few researchers acknowledge the importance of social value acceptance in this concept (Vugteveen et al., 2006). The adaption of the ecosystem health metaphor to define river health poses numerous problems

because in reality, river health is dissimilar to ecosystem health as it is a rapidly changing system, possibly at a faster rate than ecosystem health, and strongly influenced by evolving contemporary social awareness of the environment. Considering the unique geographical nature of the peri-urban zone and the ongoing ambiguity associated with developing a proper meaning for river health, there have been ongoing conflicts and debates among stakeholders, government agencies and the community at large for the adoption of effective river health improvement initiatives for the sustainability of peri-urban river systems. This uncertainty in the meaning of river health is also reflected in many present river health assessment methodologies.

1.3.2 River Health Assessment and Water Quality Monitoring

The assessment of river health mainly involves an in-depth understanding of the condition of the river-ecosystem due to anthropogenic and natural stressors particularly through a well-designed and cost-effective monitoring program. Many government agencies collect a standard group of variables regularly for routine assessments of river conditions. These indicators are mostly used to identify a system departing from the normal values. However, such deviations need to be of an extreme nature to be clearly detected by the indicators of interest (Boulton, 1999). In contrast to routine monitoring, data on biotic assemblages (i.e., phytoplanktons, macroinvertebrates) are also collected intermittently or seasonally to be used with specific river health assessment approaches. However, routine monitoring of multiple variables is an expensive task and measuring variables for different river health assessment methodologies becomes complex when available methods have a plethora of philosophies associated with underlying approaches.

Many river health methodologies are based on single perspective studies (i.e., biological, floral and faunal species), ecological function based studies (i.e., plant respiration and photosynthesis rates) and composite studies (i.e., based on water quality and macroinvertebrate indices). The single perspective studies are based on the philosophy that each species is capable of reflecting the environmental stress, as a response indication, which can be monitored over different time scales. The single perspective, biological assessments also involve the study of distribution and abundance of a wide range of faunal and floral groups such as macroinvertebrates, fish, diatoms and algae species (Chessman, 1995, Harris, 1995, Turner and Rabalais, 2003, Whitton and Kelly, 1995). Ecological function based methods on the

other hand are indirect methods of assessing river health and examining how biological communities function and interact with the environment (Boulton, 1999).

Composite indices incorporate either the quality of multiple aspects of a chosen ecosystem (e.g., Index of Stream Condition) or sub-levels of a single environmental component (e.g., Water Quality Index) (ISC, 2006, Vugteveen et al., 2006). Many predictive models compare the quality of an impacted site with a reference site on the basis that biological species, which share a similar habitat at different sites, are more alike than species which belong to different habitats but are found at the same site (Pearson and Norris, 1996). For example, the River Invertebrate Prediction And Classification System (RIVPACS), Index of Biotic Integrity (IBI) and the Australian River Assessment System (AUSRIVAS) are three widely used predictive models that utilise macroinvertebrates and fish assemblages data for making site-specific predictions (Wright, 1995, Karr, 1991, Barmuta et al., 2002). On the other hand, ranking matrices such as visually based habitat assessments (HABSCORE and Rapid Appraisal of Riparian Condition), are the simplest and quickest to perform and results can be obtained within 30-60 minutes (Barbour et al., 1999, Jansen and Robertson, 2001). The current application of ranking scores is mostly restricted for riparian habitat assessments.

A number of similar techniques have been tested by other researchers to obtain a glimpse of the patterns of degraded river systems across the globe, but none have indicated a high level of reliability and consistency dealing with the increased environmental complexity (Holdway et al., 1995, Hunsaker, 1990, Arakel, 1995, Prati et al., 1971, Cude, 2001, Turak et al., 2001). This uncertainty remains at its highest level when assessing peri-urban rivers due to our limited understanding of the mixed nature of pressures originating from both urban and rural landscapes.

1.4 Overall Objectives

This PhD research project provides findings of a study examining the meaning of river health and steps for development of a river health assessment framework using social and environmental data. The study is based on the Hawkesbury- Nepean River (HNR) that flows through peri-urban regions of Western Sydney, New South Wales (NSW), Australia.

The three specific objectives of this study are to,

i. Examine the complexity in dealing with the concept of river health and explore views and expectations of competing stakeholders; ii. Identify and evaluate the key river health indicators for the main functions and uses of the river system; and iii. Develop a suitable river health assessment framework and assessment tools to assist in sustainable management of river health.

The selection of the HNR as the key focus for this study was based on a number of reasons. Firstly, over 70% of the HNR 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. Secondly, the HNR 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.

1.5 Expected Outcomes, Values and Benefits

1.5.1 Why is this particular piece of research worth doing?

A wide variety of river health assessment methodologies available today ignore the unique qualities (e.g. physical, hydrological, and anthropogenic) of peri-urban regions. Many other techniques use ‘reference’ sites (sites with increased water quality) to benchmark the assessment sites. However, the increasing anthropogenic pressures on river systems will deteriorate the sites referred to as ‘reference’ sites and in the future we will need alternative approaches to assess river systems. A robust framework that holistically looks at both the social and environmental aspects of river health and is independent from the reference site will provide a clear basis for cost effective and less labour intensive method to assess river systems based on local conditions.

1.5.2 What special groups stand to benefit?

The river health assessment framework and tools developed in this thesis support the longer term river health management strategies to achieve sustainable river health goals as well as short term advisory information for frequent river users. In

particular, both active (e.g., farmers, fishers) and passive (e.g., tourists, ferry commuters, foreshore residents, catchment residents) river users will immensely benefit from the tools and framework developed. River management authorities will also benefit from the outcomes of the framework. Further, it will help characterise different reaches of a large river system while providing guidance for the selection of water quality indicators for efficient monitoring.

1.5.3 What further avenues of research will the information open up?

The tools and framework developed in this thesis will encourage future researchers to evaluate the different water uses such as irrigation, water extraction for drinking and water extraction for stock on river health in the longer term. This thesis also highlights the importance of community engagement when developing tools to assess river health and these findings can be strengthened by identifying community-based indicators for river health. Further, the framework can be expanded to develop a range of practical tools to assess rivers under changing climatic and socio-economic scenarios.

1.6 Research Questions

The research questions for this thesis were formulated through consultation with a number of stakeholders from government agencies and frequent river users. A desktop review for published river health assessment methods was conducted simultaneously to find knowledge gaps in the literature. The project gained the interest of a number of government agencies (e.g., The Office of Hawkesbury- Nepean, Hawkesbury-Nepean Catchment Management Authority (HNCMA), Sydney Catchment Authority (SCA), and Hawkesbury Trawl Association) and stakeholders who agreed to help in this project through data and feedback.

The following research questions were addressed during this study.

I. What is the meaning of river health from the point of view of river users and

regulators and what functions does a river perform in a peri-urban

landscape?

II. What are the competing views and expectations of stakeholders on peri-

urban river health?

III. What are key river health indicators that are robust, practical, and cost

effective to monitor a peri-urban river system?

IV. Can we develop a river health assessment framework and tools that are

reliable, robust and easily implemented by river managers or daily river users

for routine monitoring and river health management?

1.7 Format of the Thesis

This thesis presents results and recommendations of a multi-stage study conducted in Western Sydney, Australia towards the development of a river health assessment framework. There are 12 chapters in this thesis and Table 1 illustrates the flow of the chapters. Each chapter is dedicated to a specific stage of the study that produced significant outcomes towards the development river health assessment framework. Most chapters (Chapter 3-10) comprise of summary, methods, results, discussion, and concluding remarks. The materials contained in Chapters 3, 6, 8, 9 and 10 have been published in the Journal of Water Policy, Environmental Monitoring and Assessment, Shellfish Research and Water Research respectively (see Appendix - A - E ). Similarly, the material contained in Chapters 4 and 5 are under peer review in the Journal of Environmental Management and Geographical Research. Some content of Chapter 3-5 has been accepted for a poster presentation by the Integrated Water Resource Management Conference held in Dresden, Germany in 2011 (see Appendix - F).

Chapters 1 and 2 provide rationale and objectives of this study with a short review of the HNR and its catchment. Chapters 3 and 4 investigate the meaning of river health from the point of views of the general community and key informants. Chapter 5 investigates the contemporary river values and how community demographic factors affect the social and environmental benefits appreciated by the community.

Chapters 6-10 are focused on examining the environmental aspects of river health. An analysis of spatial and temporal trends in river water quality based on historic water quality data records is provided in Chapter 6. It characterises major water quality trends and classifies different reaches of the river system based on long-term water quality data. Chapter 7 identifies water quality variables important for the community structure of phytoplankton and benthic macroinvertebrate species of the river system. Similarly, Chapter 8 investigates using the harvest of school prawn as

an example, the impacts of water quality variables on aquatic life in the river system. In Chapter 9, a review of various principles behind the selection of different river health assessment techniques is presented together with a proposed multivariate analysis to identify key indicators for river health monitoring. In Chapter 10, an attempt is made to develop and validate predictive tools to assess eutrophication and microbial risk of river water by combining the indicator variables identified in Chapter 9. Chapter 11 explains a simple four-step river health assessment framework and how the findings from previous chapters can be used for managing river health. In the last chapter (Chapter 12), the key conclusions of the study are summarised and suggestions for future research are provided.

Table 1. A layout of the thesis Chapters.

Chapter Chapter title Content number Presents, -a brief introduction to the study, Chapter : 1 Introduction -research aims and questions, -expected outcomes, values, and benefits of the study.

The Hawkesbury- Provides a description of the study site and current Chapter : 2 Nepean River System stressors affecting river health.

Provides an analysis of the concept of river health Meaning of river health : Chapter : 3 and problems associated with the current meaning Community perspectives based on community views.

Provides an in depth analysis of the concept of river The key informants Chapter : 4 health and various issues related to the management meaning of river health of river health from key informants views.

Investigates the contemporary river values and how Managing river health : community demographic factors are affecting the Chapter : 5 Issues and challenges social and environmental benefits appreciated by the community.

Spatial and temporal Provides a detailed analysis of historic water quality Chapter : 6 trends in river water data records collected from the SCA for effective quality river health monitoring.

Investigates the influence of water quality indicators Biotic assemblages as Chapter : 7 on the community composition of phytoplankton and river health indicators benthic-macroinvertebrate species.

Impacts of water quality Investigates how the river health is affecting the Chapter : 8 on the harvest of school commercially viable Metapenaeus macleayi. prawns

Key indicators for river Identifies a suite of key water quality indicators for Chapter : 9 health efficient monitoring and assessment of river health.

Presents two predictive tools and validation results Developing river health Chapter : 10 assessing the eutrophication and microbial risks of assessment tools river waters.

Framework for Describes a simple four-step framework for the Chapter : 11 assessing river health assessment of river health. status

Conclusions and further Describes the conclusions of the study and Chapter : 12 research suggestions for future research.

* * * * *

CHAPTER 2

THE STUDY AREA

Summary

The water quality data and social data for this research project have been collected from the HNR and its suburbs in the catchment. The HNR is a large river system in NSW, Australia and a major proportion of this river system flows through peri-urban fringes of the main suburbs in Western Sydney region. The river system is rich in fisheries resources, provides drinking water for residents and meet irrigation water demands to a high level. At present, the HNR is under rising pressures from peri- urbanisation and used to discharge significant amounts of treated effluent originating from households and industries in the region.....

2.1 The Hawkesbury-Nepean River System

The influence of urban factors, particularly the discharge of effluent and stormwater, set peri-urban rivers such as HNR apart from rivers in dominantly rural catchments. This study was based on the HNR system, which is the main source of water supply for the Sydney Metropolitan area (330 34’ 14.72” S, 1510 20’ 16.36” E to 340 11’ 31.59” S, 1500 43’ 11.57” E) (Plate 1). The HNR originates at the Woronora Plateau south-west of Sydney and enters the Pacific Ocean at Broken Bay 30 km north of Sydney after passing through extensive peri-urban landscapes over its 300 km length (Howell and Benson, 2000). The river system itself is complex in nature, the upper one-third being edged with many poorly accessible gorges, the middle one- third running through irrigated farm lands and the lower one-third having tidal slopes with alluvial soil pockets (Diamond, 2004).

Plate 1. The Hawkesbury-Nepean River at Sackville.

The HNR system is a combination of two major rivers, the Nepean River (155 km) and the Hawkesbury River (145 km) (Markich and Brown, 1998). The Nepean becomes the Hawkesbury River at the Grose River confluence near a rural town of Yarramundi, NSW at an elevation of 39 m above sea level (D.E.C.C.W., 2010). Almost all major reaches (e.g., South Creek, Cattai Creek, Grose River, Colo River) of the river system flows through peri-urban landscape in the catchment. The estimated average dry weather flow in the HNR is about 1970 ML/day and river exchanges about 14062 ML/day of tidal flows at Brooklyn. Due to this large tidal exchange, the water quality in the lower parts of the river remains relatively high (Ball and Keane, 2006).

There are 22 large dams and 15 weirs established along the HNR system. The major dam on this river is at Warragamba which holds about 2.031x109 m3 of water captured from a 9051 km2 catchment area (Turner and Erskine, 2005). There are four other dams, viz., Cataract, Cordeaux, Avon and Nepean which determine the flow pattern of the Nepean River. The Warragamba dam and its intermittent environmental flow releases have considerably contributed to the natural flow pattern of the HNR since its construction between 1948 and 1960. Historically, the HNR is subject to alternating flood and drought dominated regimes since the early days with an interval of 40 years. During flood-dominated regimes, river widened

and depth decreased while during drought-dominated regimes river width decreased and depth increased (Warner, 1994, Simmons and Scott, 2006).

The entire length of the HNR system supports a variety of recreational activities for both residents and tourists. Between Windsor and Wisemans Ferry, a number of caravan parks and picnic grounds have been established (SPCC, 1983). The river system also supports economically significant industries related to fish and prawn, animal products (i.e., meat, eggs, wool) and different agricultural and horticultural crops (Gavin et al., 1998).

2.2 The Hawkesbury-Nepean Catchment

The catchment of the HNR disperses over 21,710 km2 of the outer western peri- urban suburbs, which are under constant pressure from rising population (Figure 1). Due to a large number of peri-urban activities, the HNR catchment presents some particular challenges in terms of water quality and health of the HNR system. The catchment of the HNR receives an annual rainfall of about 800-1400 mm with an average maximum summer temperature of 300C and minimum winter temperature of 90C (C.S.I.R.O., 2007). Geological records suggest that majority of the catchment soil consists of, the Narrabeen Group (conglomerate sedimentary rocks and shale), Hawkesbury sandstone (quartz-rich sandstone) and the Wianamatta Group (green, grey and black shale) (Markich and Brown, 1998). At present, land-use in the catchment includes regions that are urbanised, industrial, recreational, agricultural and scenically attractive and are dispersed along the lush banks of the HNR (Baginska et al., 2003). The influence of urbanisation, particularly the discharge of effluent and stormwater, set peri-urban rivers such as HNR apart from rivers in dominantly rural catchments. The catchment is currently managed by the HNCMA and the SCA.

Figure 1. Hawkesbury-Nepean River system and its catchment.

Area in white around the catchment indicates the river catchment.

2.3 The Catchment in Transition- Pressures of Urbanisation

The European settlement in Australia and Western Sydney has a relatively short time span of 200 years since 1788. The surroundings of HNR system is one of the first areas to be chosen by the settlers outside of Parramatta River basin. Small colonisation began in late 1700s near Windsor, which then gradually spread into the entire catchment (Thoms et al., 2000). The present population in Western Sydney is around 1.09 million and there will be an additional 1.5 million people living in the HNR catchment by the year 2020 (Diamond, 2004). Due to pressures from land acquisition for residential development and increased agricultural activities to accommodate the rising population, the HNR catchment presents some key challenges in terms of management of water quality and health of the HNR system.

Within the HNR catchment, vegetation clearance has been a common practise over the last 200 years causing increased subsurface and agricultural runoff and sediment loads into the river system (Thoms et al., 2000). Unlike other natural rivers

where flow is dominated by rainfall events, the flow regime of HNR is highly regulated by impoundments and treated effluent discharge from a number of sewage treatment plants (STP). There are 18 STPs along the HNR discharging significant volumes of treated municipal wastewater into the river (Howard, 2009). The main point source pollutions are attributed to STPs, mining activities and discharged industrial effluent while diffuse sources of pollution are related to urban runoff and agricultural activities associated with farms and market gardens.

The HNR system was one of the first catchments exposed to extensive human impact by sand gravel extraction for construction in NSW (Thoms et al., 2000). Much of the sand gravel extracted from the river was used for construction activities in Sydney. Due to extensive dredging and soil extraction from the banks, the width of the Nepean River has increased at many locations (Turner and Erskine, 2005). Further, construction of water storage sites and weirs have considerably reduced flood events of the river, altered natural flow regimes and, most importantly, reduced natural mixing of water in weir pools (Turner and Erskine, 2005).

Due to ongoing human influence aggravated by the peri-urban activities, the condition of HNR system has changed considerably from its original state. These influences are well documented in the context of macroinvertebrates, diatoms, fish assemblages, riparian vegetation, and fluvial sediment quality suggesting the need for strategic river health management action to improve the river’s health (Gavin et al., 1998, Simonovski et al., 2003, Growns and Growns, 2001, Growns et al., 2003).

* * * * *

CHAPTER 3

MEANING OF RIVER HEALTH : COMMUNITY

PERSPECTIVES

Pinto, U., Maheshwari, B., Shrestha, S. & Morris, C. (2012). Understanding the meaning of river health for a community: perspectives from the peri-urban region of the Hawkesbury-Nepean Catchment, Australia. Water Policy, 1-18.

Summary

This chapter explores what river health means to ordinary citizens in the community through a survey of residents (n = 302) living in the peri-urban region of the HNR catchment. Community responses concerning the meaning of river health included explanations that were simple and used everyday words but integrated a number of perspectives of the river as a natural and community resource, which have often been lacking in descriptions pursued by experts and government agencies. A considerable proportion of participants surveyed related river health to its ecological integrity, visual appeal, hydrologic balance and ability to serve the community. The description of river health was not really affected by participants’ age, gender or the distance they live from the river. The findings of this chapter provide a number of insights that can assist in the engagement of communities in future river monitoring and management programs.

3.1 General

Community plays a significant role in the environment, and for this reason there is a need to include community dimensions for managing complex ecological systems such as peri-urban river system (Delaney, 2010). Increasingly, researchers are encouraged to include community models to better engage in policy debates surrounding the implications of their work (Queenborough and Cooke, 2010). If we can establish a common understanding of the term ‘river health’ from the community point of view, it will greatly facilitate both government agencies and river users to engage in discussions by providing a common language for reviewing and implementing management decisions; the result will be a win-win situation for all involved (in agriculture, industry and tourism). Therefore, a better understanding on

community consensus on river health will greatly assist in creating this receptive platform. The above discussion further suggests that researchers and river managers tend to project their expert view about river health and this has often limited any meaningful engagement between communities and other groups involved in river management.

Vugteveen et al. (2006) provide an in-depth analysis of the various ways that river health has been described previously in conjunction with ecosystem health and suggest the importance of understanding its proper meaning from an ecological, social and economic context. Not surprisingly, the literature that comprehensively encompasses the social and cultural aspects of river health and a sustainable river future is quite limited. Needless to say, there is a pressing need to develop explicit methodologies and strategic frameworks to assess how communities value the products and services of rivers.

3.1.1 River Health versus Human Health

The word ‘health’ is prominently used in connection with human health but there are a number of parallels between river health and human health. This analogy is particularly helpful to people who are not expert in a particular discipline or not directly associated with the management of a river system, and the analogy could help in effective engagement of a community in river matters. The meaning of human health, according to the Encyclopaedia Britannica, is ‘the extent of an individual’s continuing physical, emotional, mental, and social ability to cope with his environment’ (Britannica, 2010). Following this definition, river health refers to a dynamic entity, which shares social, environmental, and hydrological relationships with its surrounding environment. On the other hand, the term ‘health’, according to the Merriam-Webster Dictionary, is (a) the condition of being sound in body, mind or spirit, especially with freedom from physical disease or pain, and (b) the general condition of the body (Merriam-Webster, 2003). In general, the dictionary meaning of the term denotes a state of wellbeing, vitality, and prosperity, and this term makes more sense if considered in conjunction with human health.

The concept of river or ecosystem health is similar to a patient who may indicate symptoms of various abnormalities in the human body (Fairweather, 1999b, Schofield and Davies, 1996). This concept makes more sense to a layperson because rivers are dynamic entities like humans; they can become unhealthy and indicate symptoms of abnormality analogous to those of a human patient. While this

concept has been narrowly accepted (Rapport, 1989) some researchers oppose it, arguing that bringing the human health analogy to a river’s health is contentious because human health is rather an evolved state compared to that of a river (Wicklum and Davies, 1995). Suter (1993) argued that the use of the term health in relation to a river or ecosystem is not appropriate because health is a property of an organism. In this arena, Fairweather (1999a) describes river health in terms of a veterinary approach because river condition comes in various forms and because a river, like an animal, cannot complain of ill health on its own behalf; in Fairweather’s view, the veterinary model is more suitable than the human health metaphor because ecosystems appear in many distinct forms (similar to animals brought to a veterinarian) and we must intervene rather than wait for the patient to come to the doctor.

3.1.2 Issues with Current River Health Meanings

A critical examination of current river health meanings has revealed that almost all published descriptions are by researchers and river managers who have some formal training and experience in hydrology, aquatic ecology, limnology or catchment management. No descriptions are available from a community point of view, although the need to include social dimensions into a description has been strongly advocated (Vugteveen et al., 2006, Meyer, 1997). For example, what a healthy river means to a farmer, a person involved in fishing or someone who just passes by the river regularly can be quite different. The language used to define river health needs to be simple and easy to understand by all in the community, each of whom may have different educational, ethnic and socio-economic backgrounds. The descriptions proposed in the literature are somewhat related to ecological interactions. Queenborough & Cooke (2010) encourage ecologists to learn the terminology and concepts of social science if they want to influence human behaviour in implementing ecologically sound strategies and concepts for river health management.

Further, there seems to be an inherent difficulty associated with the integration of different aspects of river health related to the social and spiritual wellbeing of communities, into one composite description. For example, the river Ganges is a living goddess for Hindus where they can obtain liberation from sins by performing rituals and by dying by the river (Eck, 1982). Similarly, the river Amazon serves as the life blood for many Peruvian and Brazilian tribal communities (Smith, 1999). In general, rivers need to be shared by communities and biota, and the allocation of

water can be properly attuned only if we understand what communities need to derive from their rivers and what they value most about the rivers and environment. Much emphasis has been given in the past to ecological and biotic functions and their ability to cope with a changing river environment. A similar emphasis is also required to explore a community’s expectations and values for a healthy river system.

The main objective of this chapter, using the peri-urban region of the Hawkesbury– Nepean catchment as a case study, is to understand the meaning of river health at grass-roots level from a range of perspectives, including peoples’ age, gender, and location in the catchment.

3.2 Survey Methodology

3.2.1 Formulation of Questionnaire

A review of previous survey reports related to the HNR was conducted to understand what had already been done in this area and to identify critical knowledge gaps. A draft questionnaire was developed and was then reviewed by two key local water management authorities (the HNCMA and the Office of the Hawkesbury Nepean) to avoid duplication of effort and to ensure the relevance of the study to future river management strategies.

The survey was targeted at residents living in different suburbs in the catchment of the HNR and participants were selected from suburbs in 21 Local Government Areas (LGAs) (see Figure 2 for major LGAs). The target group for the survey included people over 18 years of age with an interest in sharing their views on river issues in general and river health in particular. A survey questionnaire was made available on-line to the survey participants. Hard copies of the questionnaires were also made available, thus enabling those participants without internet access to respond to the questionnaire. There were 25 open-ended, four-point Likert scale and short answer type questions (See Appendix - G for the questionnaire and cover letter). The first six questions collected the basic demographics of the participants. The next nine questions requested the participants to provide their views on the meaning of river health, significance of this river system to their life, and activities that are impacted or promoted by the conditions of river systems. The remaining questions gathered community views on river values, health indicators and

management issues. The main aim of these questions was to better understand and verify the answers provided by the survey participants for previous questions.

Figure 2. Map of Hawkesbury-Nepean River System indicating major suburbs.

3.2.2 Pilot Survey and Human Ethics Clearance

A draft of the survey questionnaire was developed after several iterations. A pilot survey using the questionnaire was conducted with ten people selected locally. The pilot survey was used to assess how different people responded to the draft questionnaire and to find out the ease of on-line access to the questionnaire through the survey web page. As a result of the pilot survey, some questions were revised to improve their clarity and new ones were added to ensure that all relevant data would be obtained. The University of Western Sydney (UWS) Ethics Committee was also involved in improving the questionnaire and ensuring that the UWS Human Ethics guidelines were met.

3.2.3 Recruitment of Survey Participants

The survey participants were recruited using four different methods: (i) advertising in local newspapers; (ii) publishing the questionnaire link on the vUWS (virtual UWS) Discussion Board; (iii) a letter box drop-off of the questionnaire; and (iv) inviting participants through internal mailing lists of organizations.

With the help of the Media Unit at UWS, all local newspapers in 21 suburbs in the catchment were sent a press release about the survey. As a result of the press release, a number of local newspapers (Camden Advertiser, Hawkesbury Gazette, Cumberland Courier, Hornsby and Upper North Shore Advocate, Hornsby Advocate, Hills News, Penrith Press, Penrith City Star and Macarthur Chronicle) published a news story about the project and provided the web link for accessing the survey questionnaire (See Appendix - H).

A pop-up announcement and an on-line survey link were published on the vUWS Discussion Board across UWS’s six campuses for a period of two weeks. The announcement on vUWS appeared every time a student or staff member logged onto their e-Learning account. In this case, the target was to seek views of students and staff with different educational backgrounds.

An electronic copy of the survey questionnaire was sent to the Western Sydney Regional Organization of Councils, Office of the Hawkesbury Nepean, HNCMA, Department of Environment, Climate Change and Water, SCA, Department of Water and Energy, Blue Mountains Conservation Society and municipal councils in 21 LGAs to advertise the survey on their intra-network, and requesting people to participate in the survey. This was mainly aimed at expert professionals in different fields. The survey was also announced in the daily UWS staff newsletter (e-Update), delivered to staff (>4000) across all six campuses of the University.

A total of 550 hard copies of the survey questionnaire, with self-addressed stamped envelopes, were delivered to homes in 21 LGAs across Western Sydney. The selection of the suburbs was based on the population ratio of the 2006 census data. Individual homes were selected quasi randomly ensuring a representative spread across the region. Two LGAs (Oberon and Upper Lachlan) were ignored because only one and two survey forms, respectively, were required to be delivered to these areas due to their considerably lower populations when compared to other LGAs.

3.2.4 Analysis of Data

All responses related to the meaning of river health were reviewed individually and were grouped into 19 descriptors. The descriptors were identified such that a response from an individual participant would correspond to one of the descriptors. It should be noted that responses from some participants covered more than one aspect of the meaning of river health and therefore responses from those participants were counted in more than one descriptor. The responses were indexed alongside participants’ details, and analysis was made of the effects of the participants’ age, gender and location on their understanding of river health. After grouping the responses into descriptors, four emergent themes for the meanings of river health were identified.

3.3 Results and Discussion

3.3.1 General

In this survey, a total of 302 completed questionnaires from both online and paper versions were received. Just over one-third (37%) of the survey participants were male, and most (80%) of the participants were born and raised in Australia. Further, nearly half (47%) of the participants included in the survey have been living in their current residence for more than 10 years. An analysis of words used by the participants to express river health indicates that the responses can be related to 19 descriptors, as shown in Figure 3. A number of themes were then developed by examining different descriptors (Figure 4). The main themes which emerged from the analysis were: (i) visual appeal; (ii) sustaining ecological integrity; (ii) maintaining hydrologic balance; and (iv) river water fit for purpose.

140 126 117 121 120 95 100 87

80 70 56 60

Counts 42 34 37 40 12 14 15 16 17 20 5 5 7 9 0

Figure 3. Descriptors of river health identified based on the responses of survey participants.

Visual appeal

Sustaining Meaning Maintaining ecological of River hydrologic integrity Health balance

River water fit-for- purpose

Figure 4. Different descriptors for defining the meaning of river health.

While the total number of descriptors used for the meaning of river health was 19, the three prominent descriptors used by the participants were: ‘visually unpleasant’, ‘human activity influences’ and ‘faunal species’. About one-third (30%) of the

participants invariably used one of the above three descriptors to explain their meaning of river health.

Figure 5 shows a summary in relation to the ‘sustaining ecological integrity’ theme, while Figure 6 shows a summary of the views held by survey participants when describing river health in relation to the ‘visual appeal’ theme. Interestingly, participants placed high importance on both flora and fauna species living in and out of the river system when describing river health. In particular, they mentioned seasonal aquatic birds that roam around healthy river systems.

Ability to sustain life

Presence Diversity of aquatic of flora plants Sustaining and fauna ecological intergrity Presence of fish and Presence macro- of birds invertebrates

Figure 5. Examples of words used by the survey participants to describe river health in the Sustaining Ecological Integrity theme.

No dead animals or plants

Free from Low debris turbidity Visual appeal No oil on Free from surface herbicides

Free from sewage discharge

Figure 6. Examples of words used by the survey participants to describe river health in the Visual Appeal theme.

The survey results revealed that community perceptions of river health are greatly dependent on what people observe in and around the river system on a regular basis (Figure 3 & 6). A healthy river for the community, therefore, is a waterway that is aesthetically appealing and free from toxic and foreign substances in the water body. The toxic and foreign substances according to the survey participants include a range of pollutants, which have mostly originated due to human activities occurring in the landscape. Only five participants included terms related to irrigation when defining the concept of river health. This may be associated with the unique nature of peri-urban landscapes, where the community is more attached to a river system by aspects such as its aesthetic appeal, recreational use, and drinking water supplies, while the use of a river for agricultural irrigation water is a lower priority. This also indicates the composition of the peri-urban communities, where more people are involved in non-agricultural activities when compared to rural communities in which a relatively large number of people are associated with agricultural irrigation.

Figure 7 presents a summary of different descriptors associated with the theme ‘maintaining hydrologic balance’, particularly the quantity and quality of flow to maintain different functions of river system. The three important descriptors in this

theme are water flow, water quality and water depth. In most instances, survey participants related these aspects to aquatic biota. They said that water quality and quantity should be adequately maintained to sustain the life of aquatic species. In general, participants emphasised that for a river to be healthy, it has to have visual appeal, be able to sustain biodiversity, have the capacity to maintain its hydrologic integrity and be fit for specific uses.

High DO, low EC and neutral pH

Never runs Good out of water water quality Maintaining hydrologic balance Sufficient Free water flowing depth

Steady flow

Figure 7. Examples of words used by the survey participants to describe river health in the Maintaining Hydrologic Balance theme.

(DO = Dissolved oxygen, EC = Electrical conductivity)

The analysis of the responses received from the participants indicated that the top five descriptors were: ‘unpleasant objects’, ‘human activity influences’, ‘faunal species’, ‘biodiversity’ and ‘flow rate’ (Figure 3). Further, catchment health as a surrogate measure of river health was not a highly important descriptor for the participants to express their definition of river health. Distance had an influence on the order of the top five descriptors (Table 2). Participants living within 10 km of the river tended to include ‘flow’ to describe river health when compared with the participants living further than 10 km away from the river. It should be noted that the 10 km distance was used here as a guide only.

Table 2. The effect of participants’ distance of residence from the river on the top 5 descriptors of river health.

Order Distance from the HNR System <10 km >10 km 1 Unpleasant objects Faunal species 2 Human activity influences Human activity influences 3 Flow rate Unpleasant objects 4 Faunal species Biodiversity 5 Biodiversity Floral species

3.3.2 Sustaining Ecological Integrity

The theme ‘sustaining ecological integrity’ included seven descriptors: ‘faunal species’ (counts = 117), ‘biodiversity’ (95), ‘floral species’ (70), ‘river bank stability’ (37), ‘aquatic birds’ (16), ‘catchment health’ (12) and ‘pristine conditions’ (5) (see Table 3). Floral and faunal species, and biodiversity, encompassed phrases referring to diversity of aquatic species, their wellbeing, and attachment to the aquatic ecosystem. The riverbank stability descriptor consisted of various health attributes of the riparian zone: riverbank stability, health of riparian vegetation and the condition of the floodplain. Participants perceived the formation and quality of levees to have a major influence on river health, as they support the well-being of the ecological interactions occurring in the river system.

Participants used an array of terms, ranging from technical jargon (e.g., ecologically functioning ecosystem) to colloquial terms (e.g., lots of fish), to express the meaning of river health under this theme. The responses covered most of the key aspects of an aquatic environment such as ‘diversity of biota’, ‘biologically balanced ecosystem’ and ‘healthy marine life’. Some participants also described a healthy river in terms of biological interactions. For example, they used terms such as ‘stable ecosystem’, ‘ecologically functioning ecosystem’ or ‘sustainable ecosystem’ for this purpose. The use of terms related to biota by the participants to describe a healthy river indicates that the community is aware of the importance of aquatic species for a healthy river. Further, the biodiversity aspects and complex ecological interactions are not as directly visible to the community when compared to aspects described in the ‘visual appeal’ theme. However, members of the community seem to understand the importance of flora and fauna in a healthy river system and their role in maintaining a delicate ecological balance. Survey participants indicated their willingness to appreciate the value of the flora and fauna in and around the river system. The following quote from a survey participant is an example of a typical response on this aspect:

‘A healthy river for me should be ecologically functioning to support an array of aquatic life’.

The survey participants also mentioned that they wish to see an abundance of fish and macroinvertebrates as well as aquatic birds around a healthy river. The survey results confirm the instinctive awareness of the community about surrogate indicators of river health.

Some participants related a healthy river to aspects that are broad or somewhat abstract in nature. For example, one of the frequent descriptions participants used in relation to this theme was that healthy rivers are resilient to stress and are rich in self-purification capacities. Further, they stated that healthy rivers have room to adapt, absorb, and buffer the landscape during periods of dynamic variation while changing with the environment that surrounds them. The following two quotes capture this view of river health:

‘A healthy river is a living organism that is resilient, robust and extremely adaptable to broad and dynamic changes’ and

‘A…resilient riparian landscape that sustains natural assets.’

Overall, the theme ‘sustaining ecological integrity’ broadly encapsulated the community view of ecological functions, diversity, and sustainability of biotic species. Their main point here was that the community wants to see a variety of flora and fauna flourishing within the healthy river system. An earlier study by Tucker et al., (2006) also reported that the community in the Hawkesbury–Nepean catchment values abundant native wildlife and plants when describing the health of the river system.

Table 3. Descriptors and major themes for describing river health.

Theme Descriptor Response count Sustaining ecological Faunal species 117 integrity Biodiversity 95 Floral species 70 River bank stability 37 Aquatic birds 16 Catchment health 12 Pristine conditions 5 Visual appeal Unpleasant objects 126 Human activity influences 121 Aquatic weeds 56 Maintaining hydrologic Flow rate 87 balance Water quality 34 Water depth 9 River water fit-for-purpose Recreation 42 Drinkable 17 Resilience 15 Odours 14 Microbial contamination 7 Irrigation 5

3.3.3 Visual Appeal

The visual appeal theme consisted of three descriptors: ‘unpleasant objects’ (counts = 126), ‘human activity influences’ (121) and ‘aquatic weeds’ (56) (Table 3). The majority of survey participants in this theme described a healthy river as a waterway free from visually unpleasant objects. This was followed by descriptions that contained phrases describing at least one type of direct or indirect human activity influences.

Plate 2. Reduced river clarity is observed in many parts of the river system.

When describing the term river health, using different aspects of visual appearance, survey participants used two key words: ‘clean’ and ‘clear’. The term ‘clean’ often refers to a flowing stream which is free from any type of floating materials whereas ‘clear’ is used to describe the clarity of the river water. In the participants’ view the river water should be clear enough (low turbidity) and one should be able to see some depth in the river water (Plate 2).

There were 56 counts of terms related to weeds, algae, and presence of alien species in and around the river. In general, the presence of alien floral and faunal species is considered aesthetically unpleasant and seen as strong signs of an unhealthy river system. Three concerned survey participants described a healthy river by highlighting the visually appealing nature as follows:

‘A river that is free of non-native (introduced) flora and fauna. A river that is unaffected by unburnt fuel and oil from motorised water craft and the rubbish left behind by the people who drive them. A river where the native wildlife can be safe from the above people. A river where urban runoff and rubbish is captured and treated before entering the river system.’

‘A healthy river looks clean and smells clean! It sustains native fauna and flora. It is not contaminated by foreign weeds or chemicals.’

‘A clear or semi clear water, where you can see the bottom of it, or at least under it partway, clean and free of pollution, with a healthy river-side growth.’

In 1998 the Healthy Rivers Commission of NSW called for submissions from the general public for an independent inquiry into the health of the HNR system. According to the submissions received, it was evident that water quality, river flow and its ability to support regional commercial activities and social amenities are important aspects of river health (Healthy Rivers Commission, 1998). Community comments further stated that people were not particularly concerned about the accurate measurements of the various parameters but they were rather interested in both longer- and shorter-term progress of river health targets. The community views in the survey further confirmed the finding of the Commission but also indicated that the visual appeal of the river is more important than the specific water quality aspects, commercial activities, and flow rate. In summary, community perception of river health revolves around its visual appearance, which depends on dissolved and floating foreign substances and the influence of human activities, and the extent of exotic floral and faunal species.

3.3.4 Maintaining Hydrologic Balance

The theme ‘maintaining hydrologic balance’ included three key descriptors to express the meaning of river health: ‘flow rate’ (counts = 87), ‘water quality’ (34) and ‘water depth’ (9) (Table 3). Participants considered that the flow needs to be sufficient to sustain aquatic life and flush out the pollutants, which would otherwise accumulate in the river system due to slow moving water and low flow conditions. Water quality aspects mentioned by the participants included pH, dissolved oxygen, turbidity and salinity. Participants often described water depth as water level. They stated that for a river to be healthy it should maintain a sufficient water depth at anytime of the year to sustain the health of the river. Overall, flow rate was the most important variable in this theme.

In the hydrologic integrity theme, many survey participants stated that a healthy river needs to have sufficient flow on a regular basis followed by improved water quality and increased water depths (Figure 3 & 7). They used adjectives such as ‘steady’, ‘free’, ‘uniform’, ‘sufficient’, ‘good’ and ‘sustainable’ to describe the river flow. Environmental flows have a profound effect on downstream biota because the water released from dams and which subsequently flows over weirs results in modified levels of water temperature, nutrients and dissolved oxygen.

The other two parameters in this theme were ‘water quality’ and ‘water depth’. In almost all instances, survey participants related the water quality aspect of river

health to recreational activities (i.e., swimmable, boating and fishing), aquatic life (i.e., to sustain flora and fauna in the river) and human consumption (i.e., drinking). Water level on the other hand, in the view of survey participants’, should be maintained at an optimal level to support and sustain aquatic life. Nevertheless, the participants were of the view that river characteristics such as water flow, water quality and water level are integral parts of a healthy river system and should be maintained at appropriate levels suitable to sustain aquatic life, and routinely provide the services expected from a healthy river by the community. This can be summed up by the following participant’s views:

‘A healthy river to me is any river with continuous and uniform flow of good quality water and maintaining ecological diversity of flora and fauna with a minimum or zero degradation of local environment while meeting the agricultural, aesthetic and socio economic requirements of communities.’

The environmental flow regimes in the HNR, similar to many other rivers in Australia, are impacted by weirs and dams. However, the peri-urban nature of the HNR catchment, particularly effluent discharges from sewage treatment plants and extraction of river water by industries, has significantly affected the hydrologic behaviour of the river system. The amount and timing of environmental flows are often influenced by government policies, drought and the politics of water cycle management. For example, since the establishment of the dam, the minimum amount of water required to be released from the Penrith Weir in the HNR was set at 50 ML/day to meet riparian needs and the basic landholder rights of downstream river users (Diamond, 2004).

3.3.5 River Water Fit for Purpose

The main descriptors for ‘river water fit for purpose’ were ‘recreation’ (counts = 42), ‘drinkable’ (17), ‘resilience’ (15) ‘odours’ (14), ‘microbial contamination’ (7) and ‘irrigation’ (5). This theme particularly emphasised the community use of the river for recreational activities. The repeating word in this theme is ‘swimmable’. Community members assume a healthy river contains reasonably good quality water, which will allow them to swim without catching an illness. Further, in their view, a healthy river must provide an ideal platform for the community to enjoy leisure activities. The following two quotes from participants capture this theme:

‘When I think of a healthy river I picture clear waters, no litter on the grounds, people about, picnic tables, barbeque areas, children’s playground, walking pathways to see the sight of a beautiful river: Nature at its best!’

‘A river where you can safely drink and swim in the water’.

The HNR provides drinking water for over four million people living in the greater Sydney, Illawarra and Blue Mountains regions, supports 43,000 recreational anglers and attracts over 10 million tourists per year (HNCMA, 2007). However, only 17 survey participants included the aspect of drinkability when describing river health. It remains unclear why many participants did not use words related to the human consumption of water when exploring the concept of river health. However, the inclusion of swimming, taken together with the drinking aspect, indicates that a healthy river needs to meets a certain water quality standard for direct human consumption and contact.

Only five participants out of 302 stated that a healthy river should meet the irrigation needs of farmers (Figure 3).Supply of water for irrigation is one of the prime services of the HNR, particularly for growing vegetables, cut flowers, orchards, turf and dairy pastures, with an annual market value of AUS$ 10.6 million (Clarke, 2006) (Plate 3). If the river is unhealthy, the agricultural water requirement would be amongst the first to be severely affected. However, participants did not seem to be aware of this fact.

Plate 3. A turf farm near Windsor, irrigated extensively by the river water.

Major recreational areas in the HNR lie between Windsor and Sackville. The leisure activities on the riverbanks have not been greatly impacted over time although the river water has become unsuitable for swimming due to microbial contamination and potential toxins leaching from algal blooms (Kimmerikong, 2005). At present, no place in the HNR system is recommended for primary recreation activities due to degraded water quality.

3.3.6 Influence of Age and Gender

Four descriptors of river health (‘unpleasant objects’, ‘human activity influences’, ‘faunal species’ and ‘biodiversity’) appeared in all age groups (Table 4).

The flow rate descriptor was only mentioned by participants aged 41 and above. The top five descriptors used by survey participants aged 18–25 and 26–40 to describe river health were similar, with some minor difference in the ranking order. It is noted that in Table 4, for male participants who were aged between 41 and 60, there was more than one descriptor in some ranks (e.g., there were two descriptors in rank 1: Unpleasant objects, Flow rate). This was because those descriptors received equal number of responses from the participants, and therefore the descriptors were included in the same rank.

Six other descriptors of river health were ranked with low importance across all age groups: ‘irrigation’, ‘pristine conditions’, ‘microbial contamination’, ‘water depth’, ‘catchment health’ and ‘aquatic birds’. The descriptor ‘drinkable’ only featured in responses from three age groups: 18–25, 26–40 and over 61. Further, the survey participants aged 61 and over did not use the descriptors ‘odour’, ‘catchment health’, ‘irrigation’, ‘microbial contamination’, ‘water level’, ‘pristine conditions’ and ‘water quality’.

The overall pattern of river health descriptors for both males and females were similar with a minor change in the ranking. Interestingly, the descriptor ‘floral species’ was among the top five descriptors expressing river health for males, while ‘faunal species’ was among the top for females (Table 5). The overall pattern of the least important descriptors of river health remained the same, regardless of gender.

Table 4. Effect of participants’ age on the top 5 descriptors of river health.

*Where two descriptors received an equal number of participant responses, both were included in the same rank.

Top five river health descriptors Order 18–25 26–40 41–60 Over 61 Human activity Faunal Unpleasant objects, Flow rate 1 influences species Flow rate* Unpleasant Biodiversity Faunal species, Faunal species 2 objects Human activity influences* Faunal species Unpleasant Biodiversity, Aquatic Unpleasant objects 3 objects weeds Biodiversity Human activity Recreational Human activity 4 influences influences

5 Floral species Floral species River bank stability No answer

Table 5. The effect of participants’ gender on the top 5 descriptors of river health.

*Where two descriptors received an equal number of participant responses, both were included in the same rank.

Top five river health descriptors Bottom five river health descriptors Order Male Female Order Male Female Faunal species, Unpleasant Pristine Water 1 Human activity objects 1 condition quality influences* Unpleasant objects Human activity Water quality Microbial 2 influences 2 contaminati on Biodiversity Faunal Microbial Irrigation 3 species 3 contamination 4 Flow rate Biodiversity 4 Irrigation Not Sure Floral species Flow rate Water depth Pristine 5 5 condition

3.3.7 The Meaning of River Health: Community versus Experts

The responses from the survey participants about the meaning of river health have clearly indicated the inherent complexity of this concept. Participants’ responses covered multiple dimensions of ecosystem complexity, including biotic and abiotic interactions, hydrological and riparian characteristics. The survey has clearly indicated that a layperson’s meaning of river health differs considerably from that of an expert. Thus, the meaning of river health needs to incorporate community expectations and the concerns of a range of stakeholders. When investigating the

experience of humans, the present survey has emphasised that it is worthwhile examining how characteristics of a social nature are attributed to physical and biological environments. Obviously, an angler’s, water researcher’s or tourist’s description of river health are different and can be attributed to their experiences.

Experts consider rivers to be natural systems able to maintain integrity at an optimal level while adapting to various environmental conditions (Karr, 1996, Meyer, 1997). However, the community survey has highlighted that the community at large sees a river as a natural system which provides a range of services to both society and biota. Thus, the community believes that a river is healthy if it sustains ecological integrity, has visual appeal, maintains a hydrologic balance in the system and provides a range of services to the community. A clear difference between expert and community member views on river health is that the former are more compact and concise, while the latter are descriptive on several aspects of the river system. From the point of view of community members, a healthy river is a service provider to both human communities and biotic communities.

The community survey provides an opportunity to strengthen the existing knowledge on the concept of river health pursued by experts, researcher, and water managers. In particular, water managers can use the key findings of this survey to make their river improvement strategies more effective in a number of ways. For example, they can use it as evidence for the need to concentrate on river health assessment methodologies that particularly examine the function of the aquatic biota rather than their mere presence. Maintaining a river’s visual appeal, hydrological balance and sustaining river water fit-for-purpose are the other key components that need to be included in future river management strategies to meet community expectations. Furthermore, this survey has highlighted the need to explicitly include community satisfaction-based indicators in river health monitoring and analysis. Routine data collection including community-based indicators for river health assessment will not only allow river managers to evaluate river improvement programs more effectively but will also provide a communication pathway or a basis for dialogue between river managers and river users.

3.4 Concluding Remarks

In contrast to the meaning of river health pursued by experts and government agencies, the practical meaning of river health from the point of view of community members is multi-faceted and related to river ecology, aesthetics, hydrology and

environmental services. The meaning of river health for the community has emerged as a simple definition and covered four major themes: ecological integrity, visual appeal, hydrologic balance and the river’s ability to serve the community. Further, community members’ descriptions of river health were straightforward and used everyday words but integrated a number of perspectives of the river as a natural and community resource, which is often lacking in descriptions pursued by experts and government agencies. The description of river health was not much affected by the participant’s age, gender or by the distance they lived from the river. Another notable finding is that the words used to describe the meaning of river health are related to five key descriptors: ‘unpleasant floating objects’, ‘broader human activity influences’, ‘flora’, ‘fauna’ and ‘biodiversity’. Overall, the findings of this chapter provided a number of insights that can assist in the engagement of communities in future river monitoring and management programs.

* * * * *

CHAPTER 4

MEANING OF RIVER HEALTH : KEY INFORMANTS’

PERSPECTIVES

Summary

In this chapter, key informant technique was used to examine the issues, problems, and factors affecting the management of peri-urban river systems, involving active users, researchers, managers, and enthusiasts associated with the river. The analysis of interview data resulted in a number of insights that may be important in developing strategies for sustainable management for the HNR and other peri-urban river systems. In particular, the key informants’ suggested that it is difficult to define the term ‘river health’ in the context of peri-urban landscapes due to a multitude of interests and stakeholders. A number of organisational objectives, personal interests and the level of engagement of people with the river system influence the meaning of river health. The interviews also indicated that there is a perception that the health of the HNR system has been impacted considerably by on-going urbanisation in the catchment, increased recreational activities, and changes in government policies. Further, there was a view that indicators for river health assessment are often ambiguous and some of them tended to be holistic while others were narrowly based and thus limited their usefulness. The key informants highlighted the need for the assessment of river health based on ecological and social functions and the development of suitable indicators for this purpose. In addition, there was a general view that the river health management needs to be a shared responsibility between government agencies and river users, and a single entity is required to effectively facilitate the needs of different stakeholders and protect river health....

4.1 General

To date, there have been several studies conducted on the HNR system through community consultation investigating scenic quality, community acceptance of integrated catchment management and intentions for environmental behaviour (Scenic Quality, 1996, John et al., 1996, Tucker et al., 2006). However, the meaning of river health and explicit descriptions of issues that impact the management of

health of peri-urban rivers from multiple stakeholder perspectives do not exist. Therefore, the two main objectives of this chapter are (i) to explore the concept of river health with key informants having different backgrounds and perspectives, and (ii) to examine major issues related to the assessment and sustainable management of peri-urban river health.

4.2 Data Collection and Analysis

To gain an in-depth understanding of the meaning of river health, its major problems and explore key challenges in river health management, a qualitative data collection technique was employed over a period of six months in 2010. The key informant method is an anthropological data collection technique that examines a particular matter concerning a community at large through one-to-one consultation with individuals (viz., community leaders, professionals and local residents) who are knowledgeable about the study topic (Tremblay, 1957, Marshall, 1996).

4.2.1 Developing the Interview Questions

A pool of draft questions was developed from a desktop study in consultation with two key river management agencies, namely, the HNCMA and the Office of the Hawkesbury-Nepean. The questions focussed on understanding the social, environmental, and economic aspects of river health in a peri-urban context. From an initial list of 24 questions, 14 were chosen to be included in the final version (see Table 6). Questions in the interviews included direct questions, non-direct questions and hypothetical-scenario based questions. The questions were further reviewed for consistency, clarity and flow by an expert sociologist who had over 20 years related experience. After several iterations, the questionnaire was submitted to the Human Ethics Committee of the UWS for approval. The committee requested further clarification for some questions and suggested changes to the consent form and the key informants’ information sheet prior to granting permission to conduct the interviews.

Table 6. List of key informant interview questions.

1. Why is this river important to you? 2. From your previous knowledge and expertise how would you define the true health of a river ? 3. What do you think about the current state of the health of the Hawkesbury- Nepean River ? 4. What do you see as important signs of an unhealthy river? 5. How can we measure the health of river and which methods are good do you think? 6. In your view, what types of indicators we must use to understand the condition of this river? 7. What do you see as problems in the current methodologies to measure river health by various government agencies? (e.g., Macroinvertebrates, water quality, indices)? 8. In your opinion, what will be an effective way to communicate the condition of river health to river uses, community, and other interested groups? (e.g., report cards, display boards, annual reports)? 9. Hawkesbury-Nepean River is an important river in this region and it supports the social, economic and environmental well-being of people in Western Sydney. What socio-economic groups and what businesses benefit from a healthy river? 10. In your view, what activities would be most affected, if the river flow reduced considerably in the next 20 years? 11. If the health of this river does not improve in the future, what consequences will it bring to the region? 12. Are you aware of any major activity in the catchment or in the river that may have an impact on the river health? Overall, what are the main impacts on the health of the HNR ? 13. In your view, who will be more effective in improving the health of the HNR? Can you nominate a few government or community initiatives that worked well and shown some improvements in river health? 14. Do you think the actions taken by the authorities to improve river health of the HNR are satisfactory? In your view, what practical actions should be taken to improve the health of the HNR for future generations?

4.2.2 Selection and Background of Key Informants

Key informant interview is about seeking useful insights and provocative ideas, not just some facts about the topic under study. Therefore, the participants need to be selected based on their experience and knowledge to provide the needed information. Further, it is necessary to interview people who have competence in broader environmental, social and economic aspect of river health and those who

can communicate their experiences effectively by providing reliable information. The number of key informants should be sufficient to ensure adequate representation of different types of river management experiences and point of views. The selection depends on the time and resources availability of suitable key informants.

It should be noted that these interviews were per se not a community survey on river health aspects, but its aim was to analyse and elicit competing views from a selected key informants who have a significant experience and in-depth knowledge of some of the river health issues. In addition, it is emphasised here that there is no particular statistical requirement as to the number of key informants required for the interviews, but rather the total number of the informants included in the study need to together represent various river users and river managers. In a key informant based study the number of informants to be included need to represent most aspects of the problem under study to help build a rich picture of the situation (Kumar et al., 1993). Further, depending upon the problem under study and other factors, the number of key informants in past studies generally varied from 12 to 339 and there are no particular traditional statistical tests required for the analysis of interview data (Barker et al., 2005, Alem et al., 1999, de Toledo Piza Peluso and Blay, 2004).

An initial list of potential key informants for the interviews was prepared by consulting a range of government agencies, river users, researchers who have been working in the region for the past 10 years or more in Australia. The key informants were then short-listed based on their expertise and involvement in matters relating to the management of river health. Selected key informants were first contacted through telephone, details of the study were explained, and a written consent was requested. Out of 25 initial contacts, 14 key informants were available to take part in the interviews. Their areas of expertise are summarised in Table 7.

Most key informants selected were residents of the Hawkesbury-Nepean catchment. Their backgrounds were quite diverse and all had significant experience related to river and environmental management. Depending on key informant’s expertise, a few had influence or direct involvement in the policy matters related to the management of the HNR system and its catchment. Overall, there were four groups of key informants, viz., active river users, river researchers, river managers, and river enthusiasts. The active river users and river researchers had more physical contacts and frequent river visits than river managers and river enthusiasts.

The HNR system was important for the informants for one or more of these reasons: river being their workplace, they are engaged in river-research or management activities, they live close to the river and they are living in the river catchment for some time. Key informants’ whose workplace is the river system or had a business that depended on river flow were mostly associated with fisheries, boating industry, and farming. Their direct involvement with the river was considerably greater than other key informants. The informants who are involved in research had significant experience in river hydrology and catchment health for a number of years. Overall, all the informants who were living in the catchment were very interested in the river as a local resource and icon. River managers were from major government agencies responsible for initiating and planning policy instruments and catchment management programs.

Table 7. Areas of expertise of key informants.

Participant Area of expertise/interest number 1 River health and commercial fishing 2 commercial fishing 3 River educator 4 Indigenous perspective and values 5 River health and community engagement 6 Boating industry 7 River tourism 8 Environmental management and policy 9 River researcher (environmental, social and economic) 10 River researcher (water quality) 11 River researcher (catchment management and water quality) 12 River management (including water policy) River management (including catchment health and indigenous 13 aspects) 14 Riparian management (Non-governmental organisation perspectives)

4.2.3 The Interview Process

The interview began with general questions and proceeded to questions, which are more detailed where deep thinking was required by the key informants. All interviews were guided through a written list of questions to ensure that key aspects of the research topic have been covered. To ensure consistency, all the interviews were conducted by the same person and prior to the first interview the interviewer received training from a researcher with a significant experience in key informant

technique within the university. The key to the interviews was to establish good rapport with the informant early during the process, be neutral and objective, keep them on track, and encourage them to elaborate on a certain response while they are speaking. Each interview was initiated by reading out the purpose and process of the interview in accordance with the UWS human ethic guidelines and introducing the interviewer to the key informant.

The broad topic area covered during the interview included, (i) the concept of river health to key informants; (ii) the main causes and impacts of river health; (iii) the indicators of river health; (iv) problems associated with assessment and evaluation of river health (v) management of healthy rivers for future generations; and (vi) communication of river health matters. Each conversation lasted between one and two hours, was audio recorded and conducted systematically by ticking off the question list. The audio recordings of interviews were then transcribed and verbatim were subjected to content and thematic analysis.

4.3 Results and Discussion

4.3.1 Meaning of River Health

During the interview, key informants provided a variety of perspectives on the meaning of river health but mostly related to their life or work experience with the river system. For example, the informants related the meaning to in-stream health (i.e., health of macrophyte, fish, bivalves, algae and water quality) and community satisfaction of the river (i.e., provision of environmental and social services) (Figure 8). The former includes observable differences while the latter reflects an unobservable yet real sense of contentment indicating how the concept of river health is manifested in ecological functions as well as in community satisfaction. By no means, the concept of river health is restricted to a single definition. In particular, participants highlighted the lack of clarity and consistency in river health meaning among different government agencies. The following comment encapsulates this view from one of the informants:

‘...the definition of river health is quite different from department to department and even indeed within departments the definitions are different, so if you are in the business of extracting water, you know river health relates to how much water is coming down the river. If you are in a government department that is looking at water quality, and then the health is about nutrient loads...’

It was also noted how the definition varied based on informants’ physical attachment with the river system. For example, the informants who occasionally visited the river system had their meaning mostly based on ‘whole-entity’ (i.e., normal function and healthy catchment) while the key informants who frequently visited or worked with the river on a daily basis articulated the meaning of river health based on aspects related to visual amenity (i.e., water quality, presence of weed and health of fish). This shows how the meaning of river health varies with people depending upon their levels of attachment to the river system through employment, management and research activities (Table 7). The following comments by a key informant highlight how the meaning varies with their profession:

‘...Well from my perspective as a fisherman, healthy fish in the river mean healthy river.’

‘...The true health of the river really has to take a holistic view of the river, it has to have the whole catchment wide view of the river and it has to include not only the river channel and riparian zones but it has to take into account all the sub- catchments of creeks and rivers that are entering that main channel and the land uses that are contained within them.’

‘...River health is a term used by many people, has different meanings for everybody and I guess a healthy river probably is one to me, that obviously has a visual component.’

Overall the key informants’ definition of a river health resembles the view by Vugteveen et al., (2006). By reviewing a number of previous definitions related to health of ecosystem and rivers, they suggested that river health in particular must ‘consider societal functioning next to ecological functioning’. On the other hand the definitions provided by key informants who had a distinct attachment to the river system are similar to previously proposed river health definitions based on river- ecosystem characteristics such as ecological integrity, resilient to stress and sustainable turnovers of energy, nutrient and organic matter (Karr, 1999, Costanza et al., 1992, Haevey, 2001).

In the past, an analogy between human health for river health has been articulated, but there are strong arguments against this idea (Fairweather, 1999a, Schofield and Davies, 1996). Firstly, health is not an observable ecological property and secondly bringing the human health analogy to ecosystem health is ‘contentious’ because the human health is rather an evolved state (Suter II, 1993, Wicklum and Davies, 1995). Thirdly, river ecosystems can maintain their own life without intervention of humans and this makes the human health analogy difficult to adopt for river health (Rapport, 1989). Although the human health analogy was discouraged by past researchers,

the analysis of key informant interviews suggests that to a layperson this concept makes more sense because rivers are living entities and like humans they become unhealthy and can indicate symptoms of abnormality similar to a patient.

In summary, the comments provided by the key informants suggested some inherent difficulties associated with a single definition for the concept of river health. The definition need to be flexible depending upon the organisational objectives, employment status, physical attachment and personal enthusiasm of the person who defines it and the purpose for which it is used. A healthy river does not necessarily need be a pristine river. Rather, health is a dynamic state which fluctuates within a certain upper and lower threshold limits until it shows some symptoms of ‘sickness’. In this sense, the inclusion of a suite of key measurable parameters (viz., clarity, odour and primary production) and major social functions (viz., swimming and drinking) into the existing definitions seems appropriate. This will not only make the definition of river health more meaningful from the point of view of scientific community and river managers but also assist in engaging interested community into river health improvement initiatives.

Community satisfaction

Sustained Consistance visual in river flow amenity regimes

Key constituents of river health definition

Provision of Healthy envrionmenta catchment l and social and services tributaries Persistence of native floral and faunal communities

Figure 8. Key constituents of river health.

4.4 Impacts on peri-urban river health

4.4.1 Residential Developments

The key informants identified a range of catchment and river system related activities that have major impacts on the present river condition (Figure 9). At catchment level, land acquisition for housing developments, increased urban runoff from impervious surfaces and discharge of treated effluent were deemed critical factors. The effects of such activities are pronounced in the reaches of the HNR system, which passes through the peri-urban fringes of Western Sydney. Under the Metropolitan Strategy, the government of NSW is planning to accommodate 220,000 new dwellings in the North-West and South-West Growth regions of Western Sydney in the next 10-20 years (NCOSS, 2005). The gradual expansion of residential zones into peri-urban landscape will have severe consequences on the quality and quantity of run-off entering the river system. Most key informants claimed that the discharged effluent from municipal treatment facilities along the river is likely to be one of the main causes of the high levels of nutrients in the river water. At present, 18 sewage treatment plants release significant volumes of treated municipal wastewater into the river system (Howard, 2009). However, one key informant strongly argued that the diffuse point source pollution issue is greater in peri-urban catchment than the pollution associated with sewage treatment plants alone. In this key informant’s view, river quality is sometime considerably affected when there is a heavy rain event and large amount of nutrients are often discharged into the river system. Consequently, heavy loads of nitrogen and phosphorus can impact on the nutrient balance in the river system causing frequent eutrophication events.

• Rising population • Urban run-off • Effluent discharge • Long-wall mining • Urbanisation CATCHMENT • Lack of community awareness about river and catchment interactions

• Changes in flow • Dams and weirs • Diversion and extraction of water • Inter-catchment water transfer RIVER • Wake-board boating • Discharge of waste from house boats on river • Alteration of foreshore morphology

Figure 9. Major factors affecting river health.

4.4.2 Leisure Activities and Water Quality

The informants also pointed out that increasingly there are impacts on the river from leisure activities such as wake-board boats and powerboat usage. During winter months, the river is ‘stable’, but leisure activities become frequent during summer months and the riverbed is constantly stirred increasing the dissolved oxygen levels and the release of nutrients that are attached to sediment. It is now becoming evident from increased dispersal of invasive aquatic weed species in the river system that power boating activities could influence river health (E. densa and Elodea). Wake-board boating further damages the riverbank morphology through extensively large waves causing permanent damage to the foreshores eroding the riparian banks (Plate 4). Key informants mentioned considerable growth of aquatic weeds, lack of river clarity and presence of odours and discoloration of water as a result of boating activity during certain time of the year.

Plate 4. Collapsed river bank near Cattai. 4.4.3 Policy Impacts

The amount and timing of environmental flows are often influenced by government policies, the politics of water cycle management and drought. Key informants asserted that the HNR is currently stressed due to the numerous flow regimes implemented over the last decade and the number of dams and weirs on the river system which extracted water for irrigation of turf and market gardens. Further, the storage in the Warragamba dam and its intermittent environmental flow releases has significantly influenced the flows of the lower HNR since its construction during 1950s. Since building the dam, the 1.4 km reach immediately after the dam wall lost its environmental flows (Green and Erskine, 2000). A minimum flow release of 50 ML/day was set during late nineteens to achieve riparian needs and landholder rights of downstream river users. However during 2006, the flow release was halved to cope up with the ongoing drought condition (Ball and Keane, 2006). When the natural flow in river system is modified due to flow releases from dam, the river geomorphology is altered and consequently the breeding pattern of fish and survival mechanisms of flora species are affected (Varley, 2002), and thus resulting in some indirect effects on river health.

In summary, there were three main types of human pressures on peri-urban river system such as the HNR, viz., new developments, extreme leisure activities, and environmental impacts due to policy and legislative changes. Given the countless

pressures on river health, about half the number of key informants had a view that the current state of the HNR is improving although the rest suggested the condition is either poor or fluctuating. Nevertheless all agreed on the opportunities existing through the right level and kind of efforts to convert the deteriorated river into a relatively stable state. Perhaps, a legislative framework that acknowledges the unique qualities of peri-urban landscape may be more effective in managing such pressures.

4.5 Indicators for River Health Assessment

During the interviews, key informants were asked to comment on different types of indicators of an unhealthy river system. This question was asked twice, in slightly two different forms, to confirm the consistency of indicators mentioned by the informants. There was a consensus on surrogate indicators (e.g., macroinvertebrates and water quality) among key informants for the assessment of river health. However, the views of key informants were polarised between holistic and reductionist paradigms.

Key informants who favoured a holistic approach commented on two levels of indicators, viz., broad-scale and fine-scale indicators. The broad-scale indicators are related to functional value of the water for human consumption, for example, for swimming or drinking. Fine-scale indicators are the parameters (e.g., pH, Electrical conductivity and Escherichia coli) that are directly related to properties of water and often influence the broader-scale indicators. However, the two types of indicators are inter-related; for instance, if the water is suitable for swimming it is evident that a suite of fine-scale indicators such as pH, Electrical conductivity, algal bio-volume or E. coli are likely to be at a safer level. The usefulness of broad scale indicator is summed by the following thought from one of the informants:

‘You can talk about biological oxygen demands, the number of algae per litre, Escherichia coli per mL of water and so on but these sorts of measurements go straight over the top of people’s heads. But indicators such as whether you can swim in river and things that used to be in the river water aren’t there anymore mean something to the people and they start taking notice.’

The fine-scale indicators, suggested by the key informants, can be related to three groups, viz., fauna, aesthetics, and water quality (Figure 10). Interestingly, the usefulness of many such indicators mentioned by informants during interviews have been acknowledged as the main indicators for ecosystem health in Sustainable

River Audit (SRA) developed for the Murray-Darling river system in Australia (Peter et al., 2008).

Further, it is reasonably evident from key informant interviews the river health assessment requires multiple indicators, and therefore a single indicator cannot stand as a universal indicator of river health. The need for multiple indicators of river health is succinctly described in the following comment of a key informant:

‘... the first indication of an unhealthy river is the disappearance or decline in the population of fish, frogs and other aquatic lives. This is then followed by birds and so on and it is rather a chain reaction. But people just think of river health in terms of water quality. If you haven't got any fish in the river and it is polluted then you will not have the kingfishers, the sea eagles, the kookaburras who rely on some of the aquatic life for their food.’

About half of the key informants emphasised aquatic faunal species with particular attention on Australian freshwater bass (Macquaria novemaculeata) as indicator of river health. Simply, the presence of a variety of active and healthy fish and the return of certain species will indirectly indicate an improvement in river environment that is suitable for survival of fish. Healthy fish, according to one key informant, are fish that are free from sores created by red spot virus (Epizootic ulcerative syndrome). One key informant in the fishing industry recalled an experience in the lower Hawkesbury River:

‘When I was 20 years of age you could go down around the Colo River and below it and you could look down and see the Bass swimming and you could almost pick out which Bass you wanted to throw the hook into and catch it. Now, you will be lucky to see anything less than 20 cm below the water surface...’

Another key informant referred to the possibility of using fish as an indicator of river health and stated:

‘… if the fish are not in the river then there is something wrong in the river! It has always been a very good river for fish and eels and so forth’.

Comments provided by the key informants to follow up questions regarding the problems in river health assessment methodologies are interesting to investigate. Firstly, they suggested that the ‘snapshot approach’ for river health assessment currently being practised is not suitable. This may provide enough data for annual reports but the true health of the river system may not be reflected in such assessment. Secondly, they insisted on integrating different methods of assessments with particular attention to ecological function based methods, because all information needs to be treated holistically to understand the health of

rivers. Finally, they suggested benchmarking all indicators and being consistent with monitoring and assessments.

Over the past 5-10 years, a number of different river health assessment methods have been attempted in Australia (Schofield, 2009). Each technique is based on a different spatio-temporal data collection methodology; depended on health of selected ecosystem components (i.e., water quality, macroinvertebrates, and sediment heath) and provided the end result as an index value. The issue of consistency in assessment methods is often a problem when correlating old assessment records with current results.

Overall, the key informants provided mixed views on the indicators to assess peri- urban river health. There is no clear consensus as to what criteria must be followed when selecting indicators for this purpose. The ‘community satisfaction’ appeared as a suitable indicator for river health from a broader perspective although practically it is difficult to quantify and relate to the actual river health. Nevertheless, an indicator framework can be developed using broad-scale indicators and the guidelines provided by the Australian New Zealand Environmental Conservation council as a starting point (ANZECC, 2000).

Indicators of water Indicators of fauna Indicators of nutrients quality

•Red spot virus on fish •Taste and odour •Aquatic weeds •Decline in fish •Oil scum •Changes in dissolved population •Discolouration oxygen •Absence of benthic •Sedimentation •Changes in CO2 fauna •River clarity •Changes to nitrogen •Presence of toxicants and phosphorus cycle (i.e., pesticides, endocrine disruptors) •Altered flow regime

Figure 10. Indicators of river health.

4.6 River Management - Who is More Effective?

When the key informants were asked who would be more effective in improving the health of the HNR, many without a second thought agreed upon the notion of a shared responsibility. The Healthy Rivers Commissions report also agrees that people have an ‘ownership’ in their regional rivers (Healthy Rivers Commission of NSW, 2000). However, for the community to be more involved in river health issues there needs to be a clear consensus on contribution and level of obligation from those to be involved.

Key informants further insisted the need for an independent government body with sufficient enforcement powers to act promptly upon river issues. At present, there are a few government agencies looking at different aspects of the river system and this creates problems as to who has the full authority to act on river problems quickly and efficiently. For a day-to-day river user, the availability of multiple river management agencies has created confusion as to who does what. This has implications when the river users want to discuss or report issues to authorities regarding the river system. Further, they claimed that ‘buddy-buddy’ relationships, which exist among the government agencies, restrict each organisation to act independently and effectively on river issues.

Often, different government agencies are involved in administering regulations and laws towards river management and there are occasional conflicting motives or goals of the agencies involved. For example, the agency that manages urban effluent and discharges into a river system often has different goals to the one that monitors and regulates pollution, and therefore this has the potential to create conflict between the agencies when acting on specific river management issues. Several key informants also highlighted that there is a lack of opportunities for community members, who use the river on daily basis, to voice their concerns and suggestions on river matters.

An examination of the river rehabilitation programs that were in place over the past decade indicate that such programs were either not long enough or were not conducted vigorously to obtain significant improvements in the river system. For example the Department of Urban Affairs and Planning conducted an extensive scenic quality study in the late 1990s to understand reaches that have ‘high scenic quality’ and ‘regional scenic significance’ but no evidence has been found on its continuation (Scenic Quality, 1996). Overall, the key informants insisted that a

majority of the government programs did not continue in a timely fashion over a sufficient duration to achieve its desirable targets. Some key informants criticised the inability of the current river management authorities to maintain initiatives that lead to long-term improvements. One informant stated,

‘...there is nothing that really stands out that was done in a timely fashion. Most of what we see is just incremental; what you need is some kind of transformational change.’

In light of the above discussion, it is becoming evident that managing the health of a peri-urban river is a difficult and complex task due to the shared nature of the responsibility among the community, government, and industry. This complexity can be relaxed with the input of high quality data to support management decisions and through implementation of an independent government body with legislative powers to act on the river issues at spatial and temporal scales. A majority of informants even suggested an idea of a ‘river-keeper’ ‘commissioner’ or ‘ombudsman’ to keep a close eye on the river system and prosecute polluters as required.

4.7 Communication of River Health

According to key informants the condition of the river as well as the effectiveness of river health improvement programs need to be communicated to active and passive and distant and proximate river users. This is because the health of the river is not only important for those who actively use it on a regular basis and live along the river system but also for the majority of the distant community who contribute to poor river health by being inconsiderate in what they put down the sinks in their homes. The full extent of this understanding and how well this message can be delivered is still not fully understood and contemporary river management strategies have often overlooked this communication aspect. The following comments show how key informants felt about the importance of electronic media for communicating the condition of river health,

‘... we do have news reports from most TV stations on all sorts of aspects and programs to entertain us, but it would be not a great imposition to include a two minute segment on the river report, when that is ultimately vital to our long-term welfare.’

For the HNR, a variety of communication tools have been attempted in the past viz., display boards and radio announcements. In recent years, algal alert display boards are erected at certain sections along the river to notify the river users about toxic algal blooms. Key informants raised three main issues with the signboards erected

on the riverbanks of the HNR, which indicate the algal hazard of that particular reach (Plate 5). Firstly, this information can only be seen by active river users and does not go to the majority of community members in the catchment. Secondly, the sign boards lack up-to-date information on the river in broader prospects (i.e., water quality, suitability for swimming). Finally, key informants suggested signboards tend to be static unless they updated regularly to reflect the dynamic state of river condition.

Plate 5. Blue-green algae warning sign at Sackville Ferry. 4.8 Concluding Remarks

In key informants’ view, the health of peri-urban rivers is influenced by pressures from new housing developments, policy amendments, and intense leisure activities promoted by seasonal changes. The meaning of peri-urban river health remains a complex concept, and it is greatly influenced by personal attitude and one’s level of attachment to the river system. Where a proper definition is required, one must include both ecological and social descriptors into the definition. A clear consensus on the selection of indicators for river health assessment is lacking.

Key informants stressed the need to develop broad-scale indicators and frameworks for the assessment of river health in the context of peri-urban landscapes. At present many government initiatives on river health improvement do not achieve the

desired river health targets effectively because the initiatives are often operated on a smaller scale and at slower speed than the land-use and other changes that are occurring in the catchment and impacting on the river health.

From the interview data, it was evident that the monitoring of river health has been often an ad-hoc activity that is responding to an immediate problem rather than with a longer-term view to understand, enhance, and sustain river health. At present, there are multiple government agencies responsible for the management of the HNR system, and as such, this has limited the effectiveness of most initiatives to improve river health. In order to create effective community partnership and increased awareness of the river health among different stakeholders, appropriate tools, and strategies need to be in place for communication and engagement at catchment level.

* * * * *

CHAPTER 5

MANAGING RIVER HEALTH: ISSUES AND

CHALLENGES

Summary

This chapter examines the aspects of community values and perspectives on goods and services they receive from a peri-urban river system and issues related to its sustainable management. The contemporary river values of community are mostly associated with recreational and leisure activities. The distance from the river and age of the participants appeared as the most influential social demographic factors affecting participants’ perspectives on river health. However, when the survey participants did not change their hometown for more than 10 years, their familiarity on the changing conditions of the river system increased considerably. This clearly indicates that people tend to develop a strong bond with the natural resources around them and the amount of time they spent at a location is directly related to the strength of this relationship. To assess the river health, a number of visual indicators (e.g., floating debris and algal blooms) were suggested. If properly standardised, calibrated and investigated, such indicators have the potential for developing cost effective monitoring tools for detecting seasonal and spatial changes in river health. The monitoring based on such indicators can possibly act as a precursor to a more detailed and relatively expensive monitoring involving laboratory analysis of water quality parameters. The results of this chapter also highlight the need for an effective communication strategy and community engagement in river health improvement initiatives.

5.1 General

Previous studies that investigate the community perspectives on environmental assets report on the qualitative aspects of the ecosystem in relation to participants’ social-demographic factors. For example, the cross cultural social acceptance of environmental services based on participants’ proximity to the natural resource, gender and age are well documented (Larson and Santelmann, 2007, Anderson et al., 2008, Hunter et al., 2004, Dunlap et al., 2000). Other similar studies explore

ones cognitive behaviour and self-attachment to water resources and physical landscape through a concept of ‘sense of place’ (Tucker et al., 2006, Relph, 1997, Wardell-Johnson, 2006, Williams and Stewart, 1998). However, studies that particularly examine the community values and perspectives of peri-urban river systems considering its unique geographical location influenced by surrounding urban and rural lands are scarce in the literature. At present, due to various logistic reasons, budget restrictions and the nature of the diffused borders of peri-urban regions, many state and local authorities have turned a blind eye on the role of community in managing river health. A clear understanding of the community’s role in given river health management will not only promote development of appropriate strategies for effective community engagement but also aid in implementing community-based management decisions for sustainable river systems.

The main objective of this chapter is to explore how the local communities living in a peri-urban river catchment perceive and appreciate a range of goods and services provided by a healthy river system. The HNR is used as a case study and relevant field data are obtained through a community survey. By empirically examining the respondents’ age, gender and distance from the river system, an attempt is made to develop an in-depth understanding of the community views on the present condition, different types of indicators and issues related to the sustainable management of peri-urban river health.

5.2 Data Collection and Analysis

The formulation of survey questionnaire and survey administration are similar to steps followed in Chaper 3.

5.2.1 Analysis of the Survey Data

The analysis of the survey data involved a range of quantitative and qualitative techniques. All participants were grouped based on distance by matching participants’ postcodes and how far they live away from the river using Google™ Earth map tool. They were called ‘proximate residents’ if they were living within 10km of river and ‘distant residents’ if living more than 10km away. Using this measure, the distance from the edge of the river to the centre of a suburb represented by each postcode, was calculated. The raw data were tabulated using the Pivot Table tool in Microsoft Excel™ package. Categorical regression analysis and categorical correspondence analysis were also conducted using SPSS™. Open-

ended questions in the survey instrument were coded, indexed and analysed for common themes. New themes were added and existing themes were altered progressively to obtain meaningful insights from the analysis.

5.3 Results

5.3.1 Participants’ Profile

A total of 302 participants returned the questionnaires. The valid responses for each question varied slightly as no question in the survey was compulsory for the participants (Table 8 & 9). Younger participants (41%) responded to the survey more than all other age groups and male participation (37%) was half the number of female participants (63%). One in two participants was a proximate resident and has been living in their current residence for more than 10 years when they responded to the survey questionnaire (Q 1-4). It is noted that there was some over- representation of younger participants in the survey. However, considering a large number of survey participants included the survey the data collected are expected to provide a wide cross section views from the community.

Table 8. Socio-demographic features of survey participants.

Percentage Demographic feature Parameter (%) 18-25 41 26-40 21 Age 41-60 27 Over 60 11 Male 37 Gender Female 63 10km of the river 53 Distance from the river More than 10km from the river 47 > 1 year 11 1-5 years 26 Years of residency 5-10 years 16 < 10 years 47

Table 9. Key questions used in the survey questionnaire.

Question Question number Q1 What is the name of the suburb and the postcode you currently live in? Q2 Are you a male or female? Q3 Which of the following best describes your age? Q4 How long have you been living in your current residence? Q5 Can you tell us why this river is important to you? In your view, how would you see the importance of the following values of Q6 Hawkesbury-Nepean River? (Historic value, Cultural / Spiritual Value, Recreational value, Aesthetic value, Economic value, Biodiversity value) Can you share your thoughts about the current health of the Hawkesbury- Nepean River compared to your any previous experiences with this river? Q7 (Please select one of the five responses) (It has improved, It has not improved, It is now deteriorating, Nothing has changed, Do not know) Q8 In your opinion, what are the most obvious signs of an ‘unhealthy river’ ? Q9 In your opinion, what river health indicators are suitable for the HNR? Are you satisfied with the current measures taken by the river management Q10 authorities to improve the health of the HNR for future generation? Based on your knowledge, can you tell us, two or three most practical actions Q11 that can be implemented to improve the health of the HNR for future generations?

5.3.2 What do Participants Value the most of a Peri-urban River System?

The river values were investigated through two questions, viz., one open-ended and the other based on Likert-rating scale (Figure 11). The former shows the importance of the HNR to the community in general and the latter is a pie-chart representation of the specific river values. Participants referred to recreational activities when asked about the importance of the HNR to them (Q5). When the other common values of water resource were revealed to them (Q6), they intrinsically chose biodiversity followed by aesthetic values and cultural and spiritual values as the key aspects of the river system (Figure 12). A fewer number of survey participants said there is a value associated with ‘human consumption’ aspect of the river as it provides drinking water for proximate communities.

160 141 140

120

100

60 54 43 46 47 47

Number participants of Number 35 38 38 40 29 20 23 20 14 6 9 0

Figure 11. Importance of the different functions of the Hawkesbury-Nepean River System to the survey participants.

14% Historic value 25% Cultural / Spiritual Value

16% Recreational value

Aesthetic value 12%

13% Economic value

20% Biodiversity value

Figure 12. Values of the Hawkesbury-Nepean River to the survey participants.

5.3.3 Effects of Participant’s Socio-demographic Characteristics

The community view on the present condition of the HNR is presented in Figure 13 (Q7). One in three participants (32%) indicated their lack of knowledge about the present status of the HNR while one in four participants (26%) said the river health was deteriorating at present. The categorical modelling revealed three socio- demographic characteristics of the participants’z., viyears of residency (β = 0.234), distance from the river (β = 0.188) and the age of participants (β = 0.154) as the most influential and significant (p < 0.05) factors affecting participants’ perspectives on the current state of the HNR (Table 10).

The correspondence analysis further indicated how each parameter is linked with the river condition. Figure 14a & 14b indicates symmetric plots of correspondence analysis representing the association between river health condition and years of residency and participants’ age. In Figure 14a, the first dimension separates people who lived at the current residence for less than a year from those who lived more than 5 years. The second dimension separates statements about the improvements in river health (i.e., improved, not improved) from statements about the changing condition over time scale (i.e., do not know, nothing has changed, deteriorating). Interestingly, when number of years participants occupied at a certain residence increased, their understanding on the current state of the river health also increased. It is noted that the distance from the river played a role in the participants’ values and perspectives on river health. In particular, distant participants often responded that they did not know about the state of the HNR while the proximate participants were of the view that river health was deteriorating. This could mean participants living farther from river may be less enthusiastic in their active contributions towards river health improvement initiatives.

In the Figure 14b, the first dimension separates two younger age groups (18-25 and 26-40) from two mature age groups (41-60 and over 60). The second dimension separates the statements related to degradation (not improved, deteriorating) from status related to mixed state of the river system (improved, nothing has changed, do not know). Clearly, the mature participants often perceived that the current health of the HNR is deteriorating while younger participants were not quite sure of the present status of the HNR. Participants in the middle age group (26-40) were more associated with the ‘nothing has changed’ aspect of the river health.

Do not know

26% 32% Improved

Not Improved

14% Nothing has changed 12%

16% Now deteriorating

Figure 13. Views of the survey participants for the present condition of river health.

Figure 14. Symmetric plot of correspondence analysis representing the present status of the HNR System with participant’s (a) years of residency and (b) age.

Table 10. Significance of socio-demographic factors on community view on the condition of river.

(N.B., Dependent Variable: Present river health condition)

Beta Degrees of F-value Significance coefficients freedom Age 0.159 3 7.159 0.000 Gender 0.104 1 3.129 0.078 Distance 0.188 1 8.849 0.003 Years at current residence 0.234 3 12.288 0.000

5.3.4 River Health Indicators

In Q8 and Q9, the suitability and options for indicators of river health were examined. Similar to river values, participants were requested to provide their views on river health indicators in general (Q8) and rank nine known indicators (Q9) from low to high importance. Figure 15 summarise the responses provided by the participants for the two questions. The decline in biota and change in water-colour were the two key indicators of river health from participants’ perspectives. In similar proportions, participants stated that the appearance of rubbish, weeds and surface pollutants such as oil and scum were the indicators of the river system changing towards an unhealthy status (see Plate 6). When this question was asked with pre- determined answers (Q9), participants said ‘presence of floating material’ (75%) followed by ‘algal blooms’ (73%) and ‘reduced flow’ (71%) as the three important river health indicators (Figure 15). These were followed by indicators associated with ‘deposition of sand’ (64%) followed by ‘bank erosion’ (48%) and ‘clarity of water’ (41%).

Plate 6. Oil scum on the surface water at Penrith.

Overall, the types of indicators provided by the participants to the open-ended question (Q8) are similar to the list of known indicators, which were provided to them in Q9. With a marginal difference in the order of key indicators, participants showed a strong preference towards visual indicators in both questions. For example, when asked about the best signs of an unhealthy river without any clues, they said declined biota, presence of floating materials and changes to water colour as the best indicators. However when the clues were given, participants’ first choice was presence of surface pollutants, followed by algal blooms and reduced river flow. The floating materials according to the survey participants are petrochemical slicks, food scrap and beverage containers. The declined biota mentioned above refers to reduced fish and prawn catch and reduced harvest in aquaculture.

200 172

160 148

121 120

74 80 65 54 41 47 40

Number participants of Number 24

Figure 15. Indicators of general river health proposed by the survey participants.

5.3.5 River Health Improvement Programs and Community Satisfaction

Nearly half of the survey participants were not satisfied with the current measures, initiated by river management authorities to improve the river health (Q10). About 40% of participants said they did not know about such measures while a small proportion (7%) of participants noted they are happy with the current government measures. The proximity from the river system and the number of years resided at the current residence were significant (p< 0.05) and was the most influential (β= 0.171, 0.122) social demographic factors affecting participants’ satisfaction for river health improvement programs (Table 11).

The distant participants indicated a considerable dissatisfaction towards the government measures than proximate participants (Figure 16). Further, it was observed the longer they lived in the river catchment the more likely they will be critical of river health improvement programs. In particular, participants who lived more than 10 years in the catchment are more likely to indicate dissatisfaction with river health improvement programs (Figure 17).

80 Satisfied Not satisfied Do not know

61 60

40 36 Counts

20 11 3 0 Proximate Distanant

Figure 16. Community satisfaction of river health improvement programs.

Figure 17. Symmetric plot of correspondence analysis representing the community satisfaction of river health improvement programs with participants residence time.

Table 11. Significance of socio-demographic factors influencing community satisfaction of river health improvement programs.

(N.B., Dependent Variable: Community satisfaction)

Beta Degrees of F-value Significance coefficients freedom Age 0.03 3 0.256 0.857 Gender 0.115 1 3.803 0.052 Distance 0.171 1 5.567 0.019 Years at current residence 0.122 3 5.423 0.001

5.3.6 Practical Actions and River Health Communication

Regardless of participants’ social-demographic characteristics, all participants strongly agreed that community education must be strengthened to assist in river health improvement (Figure 18) (Q11). They further suggested the need to update the current environmental regulations and make them more focused on practical actions that will lead to less polluted runoff from the agricultural, industrial, and residential lands in peri-urban regions. The participants also suggested that measures that will enhance environmental flow, regulate recreational activities and improve community engagement could be critical for river health management in a longer-term.

Figure 19 indicates the methods that survey participants preferred for receiving regular updates on the condition of the river health. One in three participants indicated local newspaper as the best tool for this purpose. One in five participants indicated preference for a newsletter and local radio announcements. The use of media such as television, Short-Message-Service (SMS) and electronic mail were the least popular tools among the survey participants.

106 100 88

80 68 61 60 48 49

40 34 36

Number participants of Number 20 22 16 18 20 9 12 12

Figure 18. Practical actions to improve the river health proposed by the survey participants.

TV 5% 7% SMS & E mail alert 27% 10% Other

Display boards 11% Local Radio announcement

21% Newsletter 19% Newspaper

Figure 19. Participants’ view on communicating river health to wider community.

5.4 Discussion

5.4.1 Community Values of River Systems

Values of water resource in general and river in particular are difficult to comprehend and evaluate as they sustain a vast amount of social, environmental and economic activities within the catchment. Nevertheless, environmental scientists and anthropologist are constantly being urged to clearly articulate the ecological, social and environmental values in support of policy and management decisions (Syme et al., 1999). The recreational aspects appeared more important to community than biodiversity or spiritual values. An earlier study indicated the richness of recreational values possessed by the HNR, viz., picnicking, relaxing and walking along the river and the findings of the present survey further confirms this aspect (Tucker et al., 2006). The HNR system is one of the freely accessible recreational spots in the Western Sydney region for many peri-urban inhabitants whose employment is primarily based in urban townships. As such, their attachment to the river systems has developed mainly through numerous recreational activities during leisure time (see Plate 7). Other key values that are often assumed important in ecosystem assets, viz., life supporting values, and place meaning were of lower importance when compared with recreational values in the context of the HNR.

Plate 7. HNR is used by residents for recreational activities.

The acceptance of environmental values differs notably in urban, rural and indigenous communities. For example, based on urbanised townships of Jakarta, Ouagadougou, Brasilia, Madrid, Munich and Osaka, Biagi and Ferro (2011) observed how the community perceptions were associated with functional and aesthetic values of the water resources. Wardell-Johnson (2006) reported recreational and life-sustaining values followed by intrinsic values as important for rural communities based on the Condamine River system in North Queensland. On the other hand, when the indigenous communities in the Northern Territory, Australia are considered, the river values were more related to cultural and spiritual aspects (Strang, 2005, Zander and Straton, 2010). The indigenous communities are intimately attached to rivers through customary tenure, cosmological beliefs and vast amount of knowledge about the landscape accumulated within their communities over many years (Jackson et al., 2008).

The peri-urban inhabitants emphasise utilitarian values heavily (such as recreational activities) than intrinsic values of their local rivers. Further, the community’s preferences clearly showcased that their decisions on river health values depend on the availability of in-depth information. Without ample information, community believe that the river system only allow them to engage in recreational activities. For them to embrace a range of eco-centric values effectively, a clear mechanism on distribution of such information is required.

5.4.2 Factors Affecting the Understanding of River Health

The distance from the river and age of the participants were found to be the most influential social demographic factors affecting participants’ understanding of river health. However, gender also plays a significant role in social perceptions towards environmental attitudes and behaviour (Liere and Dunlap, 1978, Hines et al., 1987, Zelezny et al., 2000, Hunter et al., 2004). By reviewing studies published between 1988 and 1998, Zelezny et al., (2000) found that, females had strong environmental attitudes and behaviour patterns than males. One common reason for female dominance towards positive environmental attitudes is described through ‘socialisation’ theory which defines as a ‘behaviour that is predicted by the socialisation whereby individuals are shaped by gender expectations within the context of cultural norms’ (Zelezny et al., 2000). This may be also due to the women’s role as interdependent and compassionate caregivers (Beutel and Marini, 1995, Smith and Moore, 2011). On the other hand males are ‘socialised to be more independent and competitive’ (Zelezny et al., 2000). Gender was not observed as a

significant variable impacting the community perception of river health, in this chapter.

Further, we observed that the longer the participants’ residence time, the more knowledgeable they became over the fluctuating condition of the river system. When the survey participants did not change their hometown for more than 5 years, their familiarity of the changing condition of the river system increased considerably. This clearly indicates that participants tend to develop an in-depth knowledge about the natural system around them if they spend more time around the system. Baker and Palmer (2006, p. 400) described this view as when ‘a person’s length of time residing within a community increases, individuals tend to experience a higher level of community pride and tend to be better educated in terms of the availability of recreational and civic opportunities amiable to them in their communities’.

A majority of the participants were unaware of the present condition of the HNR and identified the present condition as deteriorated. A link between the community perception of river health and distance they lived away from the river system was further observed. The distant and younger participants were often unaware of the river health status while proximate and mature participants indicated a deteriorating condition of river heath. Many natural resource managers consider adjacent communities of water body as an integral part of the larger ecosystem and are often consulted prior to the implementation of new management decisions (Grumbine, 1994). Living closer to a natural water resource benefit the proximate communities immensely by providing them easy and free access to the resource to enjoy the solitude and at the same time engage in a number of recreational activities at a time convenient to them. Thus, proximate communities tend to have positive attitudes towards their recreational experience and a strong sense of stewardship and attachment to the water resource compared to distant residents (Larson and Santelmann, 2007). In particular, when the community was farther away from the HNR, their understanding of the river condition became poor regardless of their level of attachment to the river. It also does not appear that conventional community education addresses this issue adequately, because many younger participants were also not aware of the present condition of the river. The age of a person is usually negatively associated with environmental attitudes with younger individuals having more understanding about the ecological benefits of a natural resource and their pro-environmental attitudes are positively associated with ones level of

education (Dunlap et al., 2000). However, in the context of the HNR, younger participants were less aware of the river health condition.

To assess the deteriorating condition of the HNR, the participants suggested a range of indicators that are mostly visual in nature. In particular, they suggested indicators that daily river users can easily monitor. For example, number of individual species per catch by anglers and aquaculture farmers, the amount of floating material, noxious odours and water colour can be used for the purposes of river health assessment. Many odour-causing compounds (i.e., aliphatic hydrocarbons, sulphur-containing compounds, aldehydes, ketones and alicyclic alcohols) are created by the metabolic activities of microbiological organisms and blue-green algae species (Bao et al., 1997, Suffet et al., 1999). Hence, the odour levels may be a suitable surrogate indicator of accelerated biological activities in the river system.

It was further observed a considerable difference between the visual indicators suggested in the present study and the single-perspective and integrated indicators proposed by previous researchers for this purpose (Harris, 1995, Chessman, 1995, Whitton and Kelly, 1995, Turner and Rabalais, 2003). The single-perspective indicators involve a wide array of faunal and floral groups such as macroinvertebrates, fish, photosynthetic organisms and diatoms. Each species reflects the environmental stress as a ‘response-indication’ that can be monitored over different spatio-temporal scales. Integrated indicator tools on the other hand are available as a package of different indicators or as a measure of a particular ecosystem function viz., gross primary production, respiration rate (House, 1989, Bunn et al., 1999). Most visual indicators suggested by the survey participants indirectly integrate a variety of standard indicators and can be easily observed without an aid of a scientific instrument. If properly standardised, calibrated and investigated these visual indicators can be used as grassroots based inexpensive and easy to monitor ‘river health indicators’ as opposed to laboratory based indicators.

5.4.3 Community Engagement and River Health Management

Community engagement and consultation is an integral part of the management of natural assets such as rivers in Australia (EP&A Act 1979). The practical actions suggested by the participants are timely for the management of river health for future generations. In number of occasions, the participants indicated that they were

unaware of the present river health condition or river health improvement programs initiated by the river management authorities. Participants started to think broadly about the eco-centric river values in comparison to anthropo-centric river values when they were given detailed information about this aspect. At present, awareness programs are not run for this purpose, particularly for the interested cohorts in the community. However, smaller scale community education programs are intermittently conducted in isolated pockets of the HNR catchment for primary and high school students and for some anglers (through monthly meetings). In order to engage community effectively into river health management program, catchment- wide community education schemes need to be designed and run on regular basis.

A number of other practical actions suggested by the survey participants towards the improvement of river health are currently being attempted by the key administrative stakeholders of the HNR. For example, a Priority Sewerage Program is underway linking the households with old septic tank systems to the main sewer system in the HNR catchment (Sydney Water, 2011). This aims to properly dispose of domestic wastewater which would otherwise discharge into the river systems during technical failures in the septic tank systems (Thomas, 2000). Similarly, a number of sewage disposal facilities have been upgraded to produce superior quality discharge with significantly low amounts of plant nutrients (Sydney Water, 2007). At present, a number of initiatives to improve river health by increasing catchment health are also underway viz., Nutrient Smart Management, Water Smart Farms and Buy-back of Irrigation Licences under the HNR Recovery Program. As a result, positive changes to the present condition of the HNR will be expected in the coming years regardless of on-going urbanisation in the catchment.

5.4.4 Communicating about River Health

In order to engage the community in various river health management opportunities, the need for a proper communication tool is vital. Regular communication also plays a key role in ‘proximity empowerment’ where people develop a strong sense of attachment with the landscape and involve in the management of environment. This further enhances the effectiveness of river health improvement programs. Conventionally, the communication between river management authorities and the community was maintained through publically available government documents (i.e., annual reports, report cards) and in some occasions through display boards. The former has been the most common method attempted for various river systems in Australia (Peter et al., 2008, Victoria River Health Card, 2009, D.E.C.C.W., 2010). In

the context of the HNR, the communication about river conditions occurs mainly through display boards erected on key locations indicating the algae hazards and through the state of the environment reports. The traditional types of communication (i.e., newspapers and newsletters) are found to be more popular among the survey participants compared with the contemporary electronic methods such as SMS. In particular, mature aged participants who are 41 years old and over preferred the newspaper method and the least favourable methods of communication for them were the television, SMS and electronic mail.

Interestingly, the community preferences on the effectiveness of different media types observed in the present survey are different from observations in the literature. Yoon and Kim (2001) compared community preferences on different types of media for the purpose of advertising commercial products. The participants perceived internet as the most preferred type of media over television for commercial advertising. A recent study further confirms that internet has displaced the traditional media (i.e., newspapers, television and radio) among community participants in Ohio (Ha and Fang, 2011). However, the survey participants strongly believed electronic media are less preferred tools of communication of river health over other traditional methods. In general, the management of peri-urban river system must be a two way process with on-going community engagement and education through an effective communication instrument. This survey clearly indicates the importance of exchanging such information on river health matters with different community groups and how such activity will promote positive feedback loop in the peri-urban landscapes.

5.5 Concluding Remarks

The community views and values for peri-urban river system emerged were similar to those reported in published literature for both urban and rural river systems. In particular, the community views reveals that the key contemporary river values of the community are associated with recreational and leisure activities. Three demographic factors, viz., distance from the river, age of participants and residence time, appeared important as to how the survey participants appreciated a range of social and environmental benefits of a healthy river system. When compared with the participants residing close to the river (<10km), those residing at farther distance do value a healthy river in their landscape but tend to be less enthusiastic in contributing in river improvement programs.

A number of visual river health indicators (e.g., floating debris and odour) were proposed by the survey participants but further research is needed to explore whether such indicators can be incorporated into existing river health assessment methodologies. If properly standardised, calibrated and investigated, such visual indicators have the potential for developing cost effective monitoring tool. Further, the monitoring based on such indicators can possibly act as a precursor to a detailed and relatively more expensive monitoring based on laboratory testing for water quality and other parameters. The survey highlighted the need for mechanisms that will help in effective engagement of community in river health improvement program. Further, there is a need for a properly designed catchment- wide community education program to secure support of local communities.

* * * * *

CHAPTER 6

SPATIAL AND TEMPORAL TRENDS IN RIVER

WATER QUALITY

Pinto, U., Maheshwari, B. & Ollerton, R.L. (2012). Analysis of long-term water quality 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.

Summary

In this chapter, the application of Factor Analysis (FA), Hierarchical Agglomerative Cluster Analysis (HACA) and Trend Analysis (TA) techniques to evaluate long-term historic data sets was demonstrated. Six water quality parameters, viz., temperature, Chlorophyll a, dissolved oxygen, oxides of nitrogen, suspended solids and reactive silicates, measured at weekly intervals between 1985 and 2008 at 12 monitoring stations located along the 300 km length of the HNR system were evaluated to understand the human and natural influences on the river system in a peri-urban landscape. The application of FA extracted three latent factors which explained more than 70% of the total variance of the data and related to the ’bio- geographical’, ‘natural’ and ‘nutrient pollutant’ dimensions of the HNR system. The bio-geographical and nutrient pollution factors more likely related to the direct influence of changes and activities of peri-urban natures and accounted for about 50% of variability in water quality. However, the usefulness of the FA was limited due to the incomplete set of data for some variables. The application of HACA indicated two major clusters representing clean and polluted zones of the river for the years considered in the analysis. On the spatial scale, one cluster was represented by the upper and lower sections of the river (clean zone) and accounted for approximately 158 km of the river. The other cluster was represented by the middle section (polluted zone) with a length of approximately 98 km. Taking into account the various effects of nutrient loads, sewerage effluents, agricultural and other pollutants originating from point and non-point-sources along the river and via the major tributaries of the HNR, TA indicated how the point sources influence the river water quality on spatio-temporal scales. Water temperature has significantly increased while suspended solids have significantly decreased (p < 0.05) over the past 26 years. The analysis of water quality data through FA, HACA and TA helped to characterise the key sections and cluster the key water quality variables of the

HNR system. The insights gained from this chapter have the potential to contribute to improving the effectiveness of river health monitoring programs in terms of cost, time and effort, particularly in a peri-urban context.

6.1 General

Worldwide, rivers draining from extensive residential and agricultural areas have become highly polluted over the past few decades and remain a sensitive issue for river management authorities (Simeonov et al., 2003). The condition of a river system depends upon its hydrology and complex interactions between biotic and abiotic interactions, which are greatly influenced by the lithological aspects of catchment, climatic variability, land-use changes and other activities of anthropogenic nature. In turn, these factors affect river’s surface water quality and quantity in such way that a spatio-temporal variation can be observed in a ‘variation gradient’ influencing the health of river systems. To understand the spatial and temporal variability of local rivers, many river management authorities focus on measuring water quality parameters as surrogate measures to assess river health (Nnane et al., 2011). In order to manage river water quality, it is critical that we develop a firm knowledge on the complexity in measured water quality parameters due to pollution originating from anthropogenic factors (spatial scale) and natural factors associated with climate variability (temporal scale) (Alberto et al., 2001, Dillon and Kirchner, 1975). For example, nitrogen and mercury are naturally deposited from the atmosphere in river basins through precipitation. Similarly, alkalinity and acidity in surface waters are strongly affected by weathering of bedrocks without human intervention (Ouyang et al., 2006, St-Hilaire et al., 2004).

Rivers often carry a significant amount of pollutant loads originating from natural and anthropogenic sources via one-way flow to the ocean. In particular, rivers in peri- urban landscape have become increasingly vulnerable to a higher level of degradation due to rapid population growth, damming of freshwater for drinking, building of STPs to treat municipal effluent originating mostly from urban townships and extraction of freshwater to grow vegetables over the past decade (Ford, 1999, Buxton et al., 2006). In recent years the peri-urbanisation and increased area of impervious surfaces in urban landscapes have led to the generation of large volumes of municipal runoff during rainy periods which end up in rivers, negatively affecting river biota (Pejaman et al., 2009). Treated and untreated sewage effluent, groundwater seepage carrying toxic leachates and nutrient-rich runoff originating

from agricultural activities have profoundly altered water quality and the wellbeing of biotic communities in many river basins. As such, water resources in peri-urban landscape have failed to efficiently provide a number of social and environmental services they used to provide when they were at pristine conditions. A possible way to alleviate river pollution and manage river water quality is to clearly understand these interactions of anthropogenic and natural factors, particularly through a well- designed and cost effective monitoring program. The collected data can then be analysed to evaluate temporal and spatial variations after which informed water management decisions may be implemented.

For this purpose, it is a common practice that government agencies collect a number of representative biological, physicochemical and hydro-morphological parameters through a series of monitoring stations along a river at regular intervals. This often yields a large volume of data that is expensive in terms of collection, handling, storage and analysis. The intensity of monitoring has often been questioned when there are budget cuts. Therefore, it is not surprising that in many instances long-term data are not available or contain significant gaps. Further, most of such historic data are not collected for the purpose of specific statistical analysis. To make the monitoring sustainable and useful it is timely to devise analytical methodologies to identify the key water quality parameters which describe the health of a river system and key sections of a river that describe the highest temporal variations in water quality. Previous studies have also highlighted the need for maintaining long-term records on benchmark stations for the better management of river systems (Askey-Doran et al., 2009).

For the analysis of long-term datasets originating from river monitoring programs, multivariate statistical techniques such as Principal Component Analysis (PCA), Factor Analysis (FA), Hierarchical Agglomerative Cluster Analysis (HACA) and Discriminant Function Analysis (DA) have been proposed (Simeonov et al., 2003, Vega et al., 1998, Alberto et al., 2001). Of these techniques, PCA and FA have gained more acceptance over other techniques due to their ability to handle large datasets and reduce the dimensionality to a manageable size while keeping as much of the original information as possible. In particular, FA helps to understand the structure of the data set by allowing researchers to identify latent factors (combination of variables that influence condition of river, e.g., pH) unique to a given river system by extracting the most useful groups of measured variables (Fields, 2009).

The long-term monitoring of river water quality is likely to suffer from data gaps due to funding cuts and other reasons, nevertheless we need to assess river health based on the available information. Therefore, the two main objectives of this chapter are, (i) to obtain a snapshot of water quality degradation along different sections of the river over the past decade by dealing with limited data collected by the river managing authorities and (ii) to understand the major trends in water quality variables and relate them with land-use and other changes in the catchment. The present study differs from many other similar studies due to a number of reasons. Firstly, it identifies the major impacts on river waters in rural-urban fringes of the landscape where anthropogenic impacts are quite different to a purely urban river system. Secondly, it takes long-term data variability of the river system into consideration to better understand the magnitude of impacts on river water quality due to peri-urbanisation. Almost all similar studies, which follow the same statistical approach to understand river water quality, tend to examine short-term data set collected mostly between 1-8 years (Ouyang, 2005, Shrestha and Kazama, 2007, Ouyang et al., 2006, Filik Iscen et al., 2008, Nnane et al., 2011). Finally, the findings of this chapter, proposes the usefulness of multivariate techniques, which were previously applied mostly to urban river systems, to reduce dimensionality of data set and understanding the key sections of a peri-urban river for monitoring purposes.

6.2 Data Analyses

6.2.1 Screening the Data set for Analysis

Water quality data at weekly intervals were obtained from the SCA for a total period of 23 years (1985-2008). Due to changes in priorities of the organisation, funding arrangements and other logistic reasons at different times during this period, the monitoring at some sites was discontinued temporarily or permanently and some parameters were excluded from the monitoring (i.e., river flow). Therefore, six water quality variables, viz., Temperature (TEMP), Chlorophyll a (CHL), Dissolved Oxygen (DO), Oxides of Nitrogen (NOx), Suspended Solids (SS) and Reactive Silicate (SIL) were chosen due to their continuity of monitoring. The total number of monitoring sites included in the analysis was 12. Five sites were located on the Nepean River (Maldon Weir, Sharpes Weir, Wallacia, Penrith Weir and Yarramundi) and seven on the Hawkesbury River (North Richmond, Wilberforce, Sackville, Lower Portland, Wisemans Ferry, Gunderman, Peats Ferry) (Figure 20 and Table 12). Maldon Weir

is the most upstream station and Peats Ferry station is the most downstream station, located close to the mouth of the HNR system. It is noted that short-term (daily or weekly) low flow or high flow events are important but including such flow events in monthly values will result in masking of their influence through averaging effect at the monthly level. Further, this analysis includes stations from headwaters to the river mouth (highly impacted by tide) thus, inclusion of river flow in general was not appropriate.

An initial examination of the full dataset revealed that over 23 years of monitoring across 12 stations the water quality data were often collected on different days of the month. This was probably due to the time required for field sampling and other logistic reasons. It was also noted that the distributions of some parameters were positively skewed while others were negatively skewed. In addition, there were statistically extreme readings at various monitoring stations. To overcome these problems, following the approach of Ouyang (2005) and St-Hilaire et al. (2004), the annual median values of the parameters TEMP, CHL, DO, NOx, SS and SIL were calculated and used these values as the basis for analysis. Variables were also log- transformed as necessary and subsequently normalised for FA and HACA. This reduced the effects of different units attached to the variables and misclassification that could arise due to differences in magnitude of numerical values. However, untransformed data were used for TA.

Figure 20. Schematic representation of the Hawkesbury-Nepean River and its tributaries.

Table 12. Distances of the monitoring stations from ocean.

(* Distance calculated from Google Maps)

Distance Site Name (km) Maldon Weir 256 Sharpes Weir 216 Nepean Wallacia Bridge 183 River Penrith Weir 162 Yarramundi Bridge 143

North Richmond 140 Wilberforce 122 Sackville Ferry 97 Hawkesbury Lower Portland 83 River Wisemans Ferry *64 Gunderman *49 Peats Ferry Bridge 16

6.2.2 Factor Analysis

Factor analysis is a tool, which usually follows a PCA reducing the dimensionality of the water quality data set without loss of embedded information. In FA, 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).

The difference between Principal Components (PCs) obtained in PCA and VFs obtained in FA is that PC are linear combinations of observable water quality parameters but VF are unobservable, hypothetical and latent variables (Alberto et al., 2001, Shrestha and Kazama, 2007, Chapman, 1992). Because of FA, a small number of factors will usually account for approximately the same amount of information as with the much larger set of original observations. To retain the factor loadings that are important, past studies have suggested different cut off values of significance. For example, Alberto et al., (2001) and Simeonov et al., (2003) suggested 0.7, Ouyang et al., (2005) and Pejaman et al., (2009) suggested 0.75 as the cut-off value of significance to retain the factors. Rather low value of 0.6 and extremely high value of 0.95 was used by Mazlum et al., (1999) and Ouyang et al., (2006) in their studies. However, Liu et al., (2003) provided an easy-to-follow scale- based methodology for this purpose. According to this scale, factor loadings above 0.7 are considered as ‘strong effects’, factor loadings between 0.7 and 0.5 are

considered as ‘moderate effects’ and factor loadings between 0.5 - 0.3 are considered as ‘weak effects’ (Liu et al., 2003). The factor loadings in this chapter were interpreted based on guidelines provided by Liu et al., (2003).

In the past, FA has been used in conjunction with Analysis of Variance (ANOVA) methods to determine the significance of two simultaneously occurring environmental factors (Simeonov et al., 2003). Similarly, they have been used with multiple regression analysis to understand the contribution of identified sources (by PCA) to the concentration of each parameter, a modelling approach commonly known as ‘source apportionment’, and for the development of indices of sediment quality (Vega et al., 1998, Simeonov et al., 2003, Shin and Lam, 2001). In water research, PCA has been widely used to link multivariate community structure and abiotic gradients in marine environments, to assess seasonal and temporal variation in surface freshwater and to understand groundwater dynamics in multiple locations (Clarke and Ainsworth, 1993, Shrestha and Kazama, 2007, Winter et al., 2000). The PCA alone can only indicate possible variables related to the highest variance of the data set. To increase the clarity of the analysis, FA should also be used identifying similar variables responsible for differences in overall variances.

The annual median values (1994 - 2008) of each monitoring station were evaluated using the FA. Sackville, Gunderman and Peats Ferry sites were removed from the analysis due to large data gaps. Annual median datasets were first examined for assumptions of multivariate normality using Anderson-Darling test and Draftsman’s plots prior to FA. Keiser-Meyer-Olkin (KMO) and Bartlett’s tests were used to check whether the data were suitable for the FA analysis (Shrestha and Kazama, 2007). The KMO test is a measure of sample adequacy in which a value of zero indicates diffusion in the correlation pattern and one indicates relative compactness of correlations. As a rule of thumb, values greater than 0.5 are considered acceptable. If the values are less than 0.5, it is then necessary to consider collecting more data and/or reassessing which variables to include. A significant Bartlett’s test (p < 0.05) indicates that the correlation matrix is not an identity matrix. Furthermore, it indicates that there are significant relationships among the variables used in the matrix and hence the dataset is appropriate for FA (Fields, 2009). Factors were extracted using PCA method and were mathematically rotated using varimax rotation in order to reduce the contribution of variables with minor significance and then factor loadings below 0.1 were suppressed for ease of reading and interpretation in the result (Mazlum et al., 1999). The selection of varimax rotation was done in accordance

with previously published work involving spatial and temporal chemometric evaluations (Alberto et al., 2001, Vega et al., 1998). A schematic diagram of the main steps followed during FA is summarised in Figure 21.

Check the Producing a covariance / Raw field Median Data matrix covariance / correlation matrix data calculation (cases x variable) correlation for sample matrix adequacy, test for identity matrix PCA

PCs are extracted with Eigenvalues above 1 FA

Set a suitable Extracted Interpret the cut-off value PCs are significance of rotated of varifactors significance (Varimax rotation)

Figure 21. PCA and FA methodology.

6.2.3 Cluster Analysis

Cluster analysis is potentially a powerful technique to analyse water quality data monitored along river systems. There are two types of cluster analysis available, hierarchical and non-hierarchical. The most widely used method is the HACA - a simple multivariate technique that separates variables into distinct groups based on their natural characteristics. The HACA classification is a non-parametric, unsupervised method and does not depend upon assumptions of normality (Vega et al., 1998). The HACA has been applied to large sets of river data in many parts of the world, e.g., Japan (Fuji River), Argentina (Suquia River) and Iran (Haraz River) for visualizing the differences in spatial and temporal scales and as the first explanatory method (Zhou et al., 2007, Shrestha and Kazama, 2007, Alberto et al., 2001, Pejaman et al., 2009).

The HACA were employed using the squared Euclidean distance as distance measure and Ward’s linkage method to understand similarities among monitoring stations established along the HNR system. The HACA requires a data set without

missing values, thus annual data collected in 1985, 1986, 1996, 1997, 2000 and 2001 (Gunderman site excluded) were only included in the analysis. The use of squared Euclidean distance measure has been widely applied for multivariate surface water quality classifications as it is capable of progressively placing a greater weight on variables which are further apart (Alberto et al., 2001, Shrestha and Kazama, 2007). Similarly, the Ward’s linkage is suitable for this type of work because it uses ANOVA approach to evaluate the distances between and within cluster variances in an attempt to minimize the sum of squares of any two clusters that can be formed at each step (Shrestha and Kazama, 2007).

6.2.4 Trend Analysis

Trend Analysis is an exploratory data visualisation technique to identify an emerging pattern on either a temporal or a spatial scale. This application is useful, especially when dealing with a large set of historic data records. Burt et al., (2010) extensively used different types of TA plots to understand seasonal cycles, episodic responses and long-term trends of nitrate concentration in UK rivers. To test the statistical significance of trends in the time-series data, the Mann-Kendall tool is often used because in reality, the long term water quality data originating from large river systems do not follow conventional probability distributions (i.e., normal or lognormal distributions) and contains missing data points (Mann, 1945, Lettenmaier et al., 1991, Yue et al., 2002). The null hypothesis for the MK statistic is that there is no significant trend in each variable considered in the time-series.

For the TA annual median values of all sites between 1991 and 2008 (NOx-total of 18 years, CHL-total of 17 years) were used. However, we included a slightly larger data records in the MK test, collected between 1985 and 2009. The serial independence was checked with the Dublin-Watson test statistic prior to conducting the MK test. Dublin-Watson test (value = 2) indicated that serial correlation is not a problem to yield a valid MK test result (Deepesh and Madan, 2012). The TA plots of variables were created using Origin™ 8.1 to visualize overall fluctuations of river water quality variables on spatial and temporal scales and the MK statistic was calculated using an Excel™ macro developed for this purpose (Anders, 2010).

6.3 Results and Discussion

6.3.1 Descriptive Statistics of the Data Set

The summary statistics of the original data set indicating the number of data points, range, minimum, maximum, mean, standard error, standard deviation, and variance are shown in Table 13.

The mean temperature gradually increases across the stations towards the mouth with the highest recorded at Gunderman site (20.40C). The mean values of CHL levels gradually increase between Penrith Weir and Sackville Ferry, but then decrease towards the river mouth. The mean value of CHL was highest at Sackville Ferry (31.5 mg/L) and lowest at Peats Ferry (2.88 mg/L). The mean DO remained relatively constant across the stations. Interestingly, the mean SS steadily increased towards the mouth with the lowest recorded at the most upstream station (Maldon Weir) and the highest at most downstream station (Peats Ferry Bridge). Mean values for NOx indicated two clear peaks, one at Sharpes Weir and other at Wilberforce. From Lower Portland onwards, the mean NOx levels gradually declined towards the mouth. For SIL, the mean values followed a similar trend with two peaks at Penrith Weir and North Richmond.

Table 13. Descriptive statistics of river data.

Descriptive Statistics N Range Min. Max. Mean Std. Err. Std. Div. Variance N Range Min. Max. Mean Std. Err. Std. Div. Variance TEMP(0C) 363 20.90 10.30 31.20 19.67 0.27 5.14 26.40 TEMP(0C) 650 27.08 4.42 31.50 17.83 0.20 5.04 25.36 CHL (ug/L) 383 81.53 0.79 82.32 20.60 0.70 13.79 190.04 CHL (ug/L) 668 77.20 0.00 77.20 6.63 0.35 9.13 83.35 DO (mg/L) 359 10.20 3.90 14.10 8.31 0.10 1.89 3.56 DO (mg/L) 642 15.40 3.40 18.80 8.87 0.08 2.08 4.34 SS (mg/L) 337 241.00 1.00 242.00 17.73 1.02 18.67 348.60 SS (mg/L) 538 121.00 1.00 122.00 3.17 0.27 6.38 40.65 NOx (mg/L) 366 4.29 0.01 4.30 1.15 0.05 0.87 0.76 NOx (mg/L) 666 1.63 0.00 1.63 0.28 0.01 0.26 0.07 Wilberforce SIL (mg/L) 346 6.18 0.02 6.20 1.90 0.10 1.78 3.15

Maldon WeirMaldon SIL (mg/L) 643 4.90 0.10 5.00 1.76 0.04 0.93 0.86 Total Years 24 Missing Years 0 Total Years 24 Missing Years 0 N Range Min. Max. Mean Std. Err. Std. Div. Variance N Range Min. Max. Mean Std. Err. Std. Div. Variance TEMP(0C) 439 18.50 10.80 29.30 20.05 0.24 5.02 25.18 TEMP(0C) 665 22.80 7.00 29.80 19.15 0.20 5.20 27.01 CHL (ug/L) 458 191.20 1.00 192.20 31.51 1.30 27.78 771.64 CHL (ug/L) 707 101.40 0.20 101.60 13.36 0.55 14.54 211.33 DO (mg/L) 431 12.50 1.80 14.30 8.68 0.09 1.84 3.37 DO (mg/L) 651 14.30 3.20 17.50 9.05 0.07 1.85 3.43 SS (mg/L) 266 174.00 1.00 175.00 17.03 1.17 19.01 361.23 SS (mg/L) 559 87.50 0.50 88.00 5.28 0.33 7.84 61.41 NOx (mg/L) 436 3.18 0.00 3.18 0.67 0.03 0.57 0.32 SIL (mg/L) 377 6.00 0.00 6.00 1.36 0.09 1.70 2.88

NOx (mg/L) 693 5.89 0.01 5.90 1.50 0.04 1.14 1.31 Sackville Ferry 17 2002, 2003, 2004, 2005, 2006, Sharpes Weir SIL (mg/L) 666 6.14 0.10 6.24 1.92 0.04 1.08 1.16 Total Years Missing Years 2007, 2008 Total Years 24 Missing Years 0 N Range Min. Max. Mean Std. Err. Std. Div. Variance N Range Min. Max. Mean Std. Err. Std. Div. Variance TEMP(0C) 319 18.40 10.60 29.00 20.27 0.28 5.02 25.18 TEMP(0C) 1013 24.00 7.50 31.50 19.53 0.17 5.38 29.00 CHL (ug/L) 337 252.60 0.50 253.10 19.59 1.00 18.35 336.68 CHL (ug/L) 1037 155.43 0.20 155.63 10.65 0.37 11.93 142.24 DO (mg/L) 317 9.40 4.00 13.40 8.49 0.09 1.62 2.63 DO (mg/L) 989 12.10 3.80 15.90 8.57 0.06 1.77 3.14 SS (mg/L) 304 100.00 2.00 102.00 9.99 0.54 9.34 87.29 SS (mg/L) 586 488.00 1.00 489.00 9.65 1.33 32.21 1037.51 NOx (mg/L) 334 1.99 0.01 2.00 0.25 0.02 0.28 0.08 NOx (mg/L) 1036 3.42 0.00 3.42 0.33 0.01 0.34 0.11

Lower Portland SIL (mg/L) 286 5.90 0.10 6.00 1.63 0.09 1.51 2.28

Wallacia BridgeWallacia SIL (mg/L) 895 23.90 0.10 24.00 1.78 0.04 1.34 1.80 Total Years 24 Missing Years 0 Total Years 24 Missing Years 0 N Range Min. Max. Mean Std. Err. Std. Div. Variance N Range Min. Max. Mean Std. Err. Std. Div. Variance TEMP(0C) 425 17.80 11.00 28.80 20.01 0.23 4.79 22.91 0 TEMP( C) 813 25.70 8.00 33.70 19.18 0.18 5.05 25.53 CHL (ug/L) 433 52.68 0.50 53.18 7.92 0.35 7.38 54.42 CHL (ug/L) 821 35.98 0.00 35.98 4.63 0.13 3.86 14.92 DO (mg/L) 420 10.80 4.30 15.10 7.90 0.07 1.47 2.15 DO (mg/L) 806 12.10 1.50 13.60 8.84 0.06 1.60 2.56 SS (mg/L) 401 149.00 1.00 150.00 15.51 0.85 16.96 287.60 SS (mg/L) 569 259.00 0.00 259.00 7.86 1.09 26.12 682.22 NOx (mg/L) 424 1.07 0.01 1.08 0.22 0.01 0.19 0.04 NOx (mg/L) 854 1.80 0.00 1.80 0.16 0.01 0.17 0.03 SIL (mg/L) 390 6.20 0.10 6.30 1.62 0.08 1.48 2.20 Ferry Wiseman's Penrith WeirPenrith SIL (mg/L) 796 5.45 0.10 5.55 2.49 0.04 1.13 1.28 Total Years 24 Missing Years 0 Total Years 24 Missing Years 0 N Range Min. Max. Mean Std. Err. Std. Div. Variance N Range Min. Max. Mean Std. Err. Std. Div. Variance TEMP(0C) 255 16.40 12.00 28.40 20.41 0.28 4.55 20.67 0 TEMP( C) 764 20.70 8.90 29.60 19.23 0.18 4.98 24.78 CHL (ug/L) 246 40.38 0.30 40.68 4.86 0.33 5.13 26.27 CHL (ug/L) 805 111.62 0.40 112.02 11.56 0.48 13.55 183.68 DO (mg/L) 247 8.30 4.10 12.40 7.33 0.08 1.31 1.72 DO (mg/L) 756 12.40 2.50 14.90 8.26 0.07 1.85 3.41 SS (mg/L) 235 177.00 1.00 178.00 18.74 1.59 24.37 594.09 SS (mg/L) 572 120.00 1.00 121.00 5.49 0.36 8.62 74.30 NOx (mg/L) 261 0.90 0.01 0.91 0.20 0.01 0.16 0.03 NOx (mg/L) 803 2.29 0.01 2.30 0.50 0.01 0.36 0.13 Gunderman SIL (mg/L) 218 5.40 0.10 5.50 1.63 0.09 1.40 1.96 2002, 2003, 2004, 2005, SIL (mg/L) 725 5.39 0.01 5.40 1.88 0.05 1.28 1.65 16 Total Years Missing Years Yarramundi Bridge Yarramundi 2006,2007,2008 Total Years 24 Missing Years 0 N Range Min. Max. Mean Std. Err. Std. Div. Variance N Range Min. Max. Mean Std. Err. Std. Div. Variance TEMP(0C) 225 16.50 11.00 27.50 19.85 0.27 4.03 16.23 0 978 21.90 8.30 30.20 19.56 0.17 5.36 28.74 TEMP( C) CHL (ug/L) 230 12.13 0.06 12.19 2.88 0.13 1.98 3.92 CHL (ug/L) 1025 118.98 0.24 119.22 13.62 0.45 14.51 210.57 DO (mg/L) 221 9.70 4.80 14.50 7.11 0.08 1.18 1.38 DO (mg/L) 961 10.20 4.40 14.60 8.98 0.05 1.64 2.68 SS (mg/L) 211 400.00 1.00 401.00 32.74 4.06 58.97 3477.24 SS (mg/L) 551 189.00 1.00 190.00 5.40 0.51 11.94 142.53 NOx (mg/L) 225 0.66 0.00 0.66 0.06 0.01 0.08 0.01 NOx (mg/L) 970 4.22 0.00 4.22 0.32 0.01 0.27 0.07 SIL (mg/L) 204 3.00 0.10 3.10 0.77 0.03 0.48 0.23

SIL (mg/L) 933 6.27 0.03 6.30 2.60 0.04 1.28 1.65 Bridge Ferry Peats

North Richmond 16 2002, 2003, 2004, 2005, Total Years 24 Missing Years 0 Total Years Missing Years 2006,2007,2008

6.3.2 Factor Analysis

The FA was employed to understand the compositional patterns between environmental variables in the HNR system and possibly identify factors affecting each variable. The KMO was 0.473 and Bartlett’s test was significant (p < 0.05). The

low value of KMO probably resulted due to a low number of variables being collected over the years. The low values indicated by KMO have taken into consideration, when interpreting our results. The correlation matrix of the water quality variables is provided in Table 14. Most correlations between parameters tended to be weak with the highest correlation coefficient between SS and CHL (0.661). A low correlation coefficient (0.207) was recorded for NOx and CHL, indicating that CHL did not fully depend upon dissolved nutrients in the HNR system. This finding is similar to research conducted on the effects of nitrogenous compounds on Chlorophyll levels under laboratory conditions (Buapeta et al., 2008).

- -3 Buapeta et al., 2008 found the presence of NO3 and PO4 has a higher impact on + - the Chlorophyll content of Ulva reticulata than the presence NH4 and NO3 .

The PCA extracted three components with eigenvalues above 1 capturing 74% of the total (Table 15) and the FA yielded three VFs with weak, moderate and strong factor loadings (Table 16). Only the strong factor loadings are reported from the analysis. For the data analysed, VF1 explained 29% of the total variance and indicates strong positive loadings on CHL (0.794) and SS (0.702) and a negative loading on SIL (-0.785). VF2 accounted for 25% of the total variance and recorded loadings of -0.793 on DO and of 0.750 on TEMP. VF3 accounted for 19% of the total variance and had the highest positive loading on NOx (0.894). It is also interesting to note that VF1 represents the bio-geographical factor of water quality, VF2 the natural factor and VF3 the nutrient pollution factor. The bio-geographical and nutrient pollution factors more likely related to the direct influence of changes and activities of peri-urban natures and account for about 50% of variability in water quality.

The three variables CHL, SIL, and SS picked up by VF1 were considered further. Here, VF1 is indicative of CHL, which is a biological measure of water quality, SIL which is a measure of catchment geology and SS which is a measure of total organic and inorganic suspended matter. Suspended solids usually consist of particulates with a diameter less than 62µm (Waters, 1995). In many instances, suspended solids are related to soil leaching and soil erosion (Simeonov et al., 2003, Shrestha and Kazama, 2007). On the other hand, SIL are derived from weathering of silicate rocks (Semhi et al., 2000). However, the factor loadings are opposite for SIL and SS suggesting the origin and existence of SIL in the HNR system are not similar to those of SS. Further, SIL did not appear to contribute significantly to the increase in CHL levels. The relationship between CHL and SS contradicts the fact that increased suspended solids reduce the light penetration into the water column thereby reducing the growth of Chlorophyll a producing flora. Instead, suspended solids appear to have had an additive effect on Chlorophyll a producing flora in the HNR system. Nutrients flocculated around suspended solid particles may be supporting the increase in Chlorophyll a producing flora (Heathwaite, 1994). Further to this the negative loading of SIL in connection with a positive loading of CHL may be an indication of silicate uptake by diatoms during phytoplankton blooms. However, further research is warranted to understand these processes.

Two variables (DO and TEMP) with high absolute factor loadings on VF2 are indicative of oxygen de-saturation processes in the water column. The inverse relationship obtained by VF2 reflects warmer water holding less oxygen as it saturates with gas faster than colder water. Shrestha and Kazama (2007) reported similar results (0.889 for temperature and -0.885 for DO under VF2) for the Fuji River in Japan. This is a well-known natural phenomenon in all waterways and the HNR follows the same pattern.

Variables with a large loading on VF3 relate to effects of nutrient pollutants in the river. This factor has been reported in several previous studies for highly urbanized river systems including Aliakmon, Axios, Gallikos, Loudias and Strymon Rivers (Simeonov et al., 2003) and Fuji River (Shrestha and Kazama, 2007). The raw data for NOx included in the analysis contained both nitrate and nitrite oxides. The increased levels of NOx were probably due to non-point-source pollution from agricultural areas and orchards in the catchment, as well as, from STP effluent discharges to the river. Nitrogenous fertilizers applied for crops undergo nitrification process and NO3-N reaches river waters through groundwater seepage and runoff (Shrestha and Kazama, 2007). The nutrient factor only accounts for 19% of the total variance and was captured by the third PC. However, compared to the first (29%) and second components (25%) the effect of the third component (19%) is relatively low in the total variance. Nevertheless, it indicates the prevalence of nutrient pollution in the HNR.

The FA seemingly suffered from the limited number of complete data sets, resulting from short lengths of continuous monitoring and reduced numbers of parameters in a number of occasions. As such the data shortage attributed to the time scale, special distances and number of water quality variables considered in the analysis has some impact on the outcome of FA. This problem was also reflected in the KMO pre-test. More frequent and systematic collection of river data is required to obtain additional VFs that would aid in identifying other variables affecting the health of the HNR system. Factor analysis in evaluating the surface water quality usually achieves two targets. Firstly, it reduces the number of variables to a manageable size and secondly it indicates variables important for a larger proportion of water quality variability in the river system. The application of FA did not significantly reduce the number of variables. Nevertheless, it revealed three underlying factors governed the health of the HNR system in the past decade.

Table 14. The values of the correlation matrix of the six physico-chemical variables.

TEMP CHL DO SS NOx SIL 1 TEMP 0.289 1 CHL -0.296 -0.038 1 DO 0.290 0.661 -0.442 1 SS -0.08 0.204 0.065 0.051 1 NOx SIL -0.042 -0.316 0.119 -0.313 0.117 1

Table 15. Eigenvalues for the first three PCs for the water quality data set of HNR.

Rotation Sums of Squared Loadings Component Total % of Variance Cumulative % 1 1.756 29.27 29.27

2 1.523 25.384 54.654 3 1.147 19.124 73.778

Table 16. Rotated factor loadings for extracted principal components.

Variables Factors VF1 VF2 VF3 CHL 0.794 (Strong) 0.191 0.379 SIL -0.785 (Strong) 0.120 0.395 SS 0.702 (Strong) 0.518 0.147 DO -0.793 (Strong) 0.158 TEMP 0.750 (Strong) NOx -0.111 0.894 (strong)

6.3.3 Hierarchical Agglomerative Cluster Analysis

The HACA was employed to understand the similarities and dissimilarities in monitoring sites giving particular emphasis to spatial variability of the river. For increased clarity, the distinct clusters obtained from the dendrograms are relisted, and presented in (Table 17). Gunderman site was removed for the year 2001 data column because of a lack of continuity in available data. For the six-year period (1985-1986, 1996-1997, 2000-2001) considered in this analysis HACA identified two broad clusters. Most upstream and downstream stations were included in one cluster while stations belonging to the middle reach were included in the second cluster. There is a slight mismatch in this pattern in 1985 and 1996 however the broad clusters remained the same.

HACA clustered the 12 monitoring stations into two distinct groups based on six water quality variables. Overall, this pattern was consistent for the six-year (non- consecutive) period. The clusters observed, did not belong to broader areas of the HNR system predicted by previous studies (Markich and Brown, 1998, Diamond, 2004). The geology of catchment often influences the quality of the runoff and groundwater that contributes to the flow in the river. Diamond (2004) identified three main types of landscapes of the HNR catchment - the upper one-third having many poorly accessible gorges, the mid one-third situating in the midst of agricultural farmlands and the lower one-third ending with tidal slopes and alluvial soil pockets. According to Markich and Brown (1998), the HNR catchment has three major geological formations - the Narrabeen Group, the Hawkesbury Sandstone and the Wianamatta Group. Much of the catchment of the Nepean River and its tributaries is underlain by shale yielding high salinity in groundwater and resulting in high conductivity of river water while the headwaters lie mostly on sandstone (Growns et al., 1995). If the catchment geology and different landscape formations had a considerable influence on river waters, a clustering pattern that reflects these geographical differences would be expected. However, the clusters do not seem to match the above divisions of the catchment based on geological formations.

Alternatively, one might expect a clustering pattern based on tidal influence. For example, the saline intrusion into the HNR system (due to tidal-inflow) is usually limited to Colo River junction near Sackville station. Tidal effects in some instances extend up to Grose River junction near Yarramundi station further upstream of the river (SPCC, 1983). If saline water had a profound effect on the six water quality variables, the analysis would have grouped the monitoring stations into two clusters each representing upstream and downstream sites. This pattern was also not observed in any dendrograms. The clusters were indicative of pollutant loads entering the river. The middle reach identified by the second cluster (highlighted cells in Table 17 includes Yarramundi, North Richmond, and Wiseman’s Ferry stations. In 1997 and 2000, the polluted reach was indicated by three sites, Wilberforce, Sackville, and Lower Portland (Figure 22). Between these two monitoring stations the HNR receives pollutant loads from three major tributaries (McDonald River, Colo River and Grose River) and four creeks (Eastern Creek, Cattai Creek, Ropes Creek and South Creek) which transport substantial amounts of treated effluent from Quakers Hill, Riverstone and St. Mary’s sewage treatment facilities. It has been reported that 95% of dry weather flow at South Creek consists of domestic or treated effluent (Thoms et al., 2000). Therefore, the discharge from

STPs has a greater influence on the water quality of major creeks in HNR. The middle stream monitoring stations are also situated in highly peri-urbanised areas of the catchment that drain large quantities of urban runoff during rainfall events. If all tributaries are considered, about 90% of flows through various tributaries are effluent (Thoms et al., 2000). If the hypothesis that most upstream reaches of a river are relatively less polluted is true then the inclusion of most downstream sites together with most up-stream sites in one cluster, as shown by HACA, is indicative of self-purification and an assimilative capacity of the HNR. The river tends to improve its water quality close to the mouth due to tidal influence, less pollutant inflows and water purification effect of dense mangrove vegetation. Thus, the two clusters in this analysis may be considered to represent a ‘clean-zone’ and a ‘polluted-zone’ of the HNR system.

The results of this analysis are consistent with other peri-urban rivers in the world reported by a number of researchers (e.g., Zhou et al. 2007; Pejaman et al., 2009; Simeonov et al., 2003). They report on three clusters representing less polluted areas of the river, areas polluted with sewage effluent from STPs, industrial wastewater and domestic wastewater, and areas polluted with agricultural runoff. For example, using 23 variables measured over a 12-month period, Zhou et al. (2007) obtained three clusters of a watercourse in Hong Kong representing a near pristine water quality area, a moderately polluted area and a highly polluted area. Further, Alberto et al. (2001) obtained two clusters belongs to upstream and downstream sampling stations based on 22 parameters monitored over a 2-year period in the Suquia River basin in Argentina. A recent cluster analysis done in Iran for the Haraz River basin which used 10 water quality parameters, (DO, Faecal

Coliform, pH, Temperature, BOD, NO3, TP, TS and Discharge) revealed three clusters with similar water quality which were related to low, moderate and high polluted areas of the basin (Pejaman et al., 2009). Similar to the HNR system, water quality of the Haraz River was affected by pollution originating from agricultural activities, sand mining activities, and untreated sewage effluent. Interestingly, HACA was able to differentiate these spatial changes. Simeonov et al. (2003) reported similar results based on the Aliakmon, Axios, Gallikos, Loudias and Strymon River systems in Northern Greece where the effects of irrigation activities along the river and stream discharges gave rise to two distinct clusters representing clean and polluted areas of the river.

Figure 22. Dendrograms showing clustering of monitoring stations based on water quality in year 1997 and 2000.

Table 17. Two clusters obtained from dendrograms at 25-rescaled distance cluster combine level (Non highlighted-cluster 1, Highlighted- cluster 2).

1 9 8 5 1 9 8 6 1 9 9 6 1 9 9 7 2 0 0 0 2001 Maldon Weir Maldon Weir Maldon Weir Maldon Weir Maldon Weir Maldon Weir Sharpes Sharpes Sharpes Sharpes Sharpes Sharpes Weir Weir Weir Weir Weir Weir Wallacia Wallacia Wallacia Wallacia Wallacia Wallacia Penrith Weir Penrith Weir Penrith Weir Penrith Weir Penrith Weir Penrith Weir Yarramundi Yarramundi Yarramundi Yarramundi Yarramundi Yarramundi N. Richmond N. Richmond N. Richmond N. Richmond N. Richmond N. Richmond Wilberforce Wilberforce Wilberforce Wilberforce Wilberforce Wilberforce Sackville Sackville Sackville Sackville Sackville Sackville L. Portland L. Portland L. Portland L. Portland L. Portland L. Portland Wisemans Wisemans Wisemans Wisemans Wisemans Wisemans Ferry Ferry Ferry Ferry Ferry Ferry Gunderman Gunderman Gunderman Gunderman Gunderman Peats Ferry Peats Ferry Peats Ferry Peats Ferry Peats Ferry Peats Ferry

6.3.4 Trend Analysis

The MK test revealed a significantly increasing trend for annual median TEMP (p < 0.05) and significantly decreasing trend for SS (p < 0.05) over the last two decades (Figure 23). Trends for both CHL and SIL variables decreased over time and were not significant. Minor increasing trends were also observed for DO and NOx on temporal scale. The temperature of river waters plays a considerable role in the energy utilization process, cooling capacity, dissolving capacity, and most importantly maintaining healthy biological communities within an aquatic ecosystem. However, water temperature is greatly affected by the atmospheric temperatures and to a lesser extent by discharge rates, although such increases are marginal (Pekárová et al., 2011, Moatar and Gailhard, 2006). Pekárová et al. (2011) observed a slight increase (0.12oC) in water temperature due to an increase of 1.5oC air temperature over a 50 year period for the river Bela basin in Slovakia (Pekárová et al., 2011). Similarly, for the river Loir (France), the mean annual water temperature increased by 0.8 oC over 95 years (Moatar and Gailhard, 2006). In line with past observations, an increasing trend for the water temperature was observed and this trend may be attributed to the consequences of global warming. As to the variable SS, the significantly decreasing trend observed is an interesting observation from

the analysis. In natural waters, presence of SS increases the nitrification process and this leads to an increase in NOx levels (Xia et al., 2004). The SS indicates the opposite of this relationship on temporal scale. This may be due to the effects of tide, which has not accounted for in this analysis.

The trends for variation of NOx and CHL are shown in Figure 24 and Figure 25. There are clear highs and lows in both plots. For NOx variation, there are three clear peaks over Sharpes Weir, Wilberforce and Yarramundi in descending order (Figure 24). All sites from Lower Portland to the mouth of the river have relatively low levels of NOx for all years. The graph for CHL indicated only two peaks, one at Sharpes Weir and other between Wilberforce and Lower Portland (Figure 25). Chlorophyll a for all other sites remained relatively low and may be due to dominance by attached aquatic plants in those sections. The decreasing spatial trend in NOx is due to the reduced impacts of STP discharge and dilution effect from salt-water intrusion towards the mouth of the river system.

Two steep peaks at Sharpes Weir and Wilberforce were noted. Increased NOx levels observed over consecutive years at Sharpes Weir are most likely due to effluent discharge from adjacent West Camden and Picton STPs. The West Camden STP releases 3112 ML of effluent annually into the environment (Howard, 2009). On the other hand, the peak at Wilberforce probably represents nutrients entering the HNR system just upstream through South Creek, one of the most polluted tributaries of the HNR. A detailed report by Krogh et al. (2008) on the analysis of water quality of the HNR system suggested that regardless of minor trends in decreasing nitrogen levels, NOx levels remained above the Australian and New Zealand Environment and Conservation Council and Agricultural and Resource Management Council of Australia and New Zealand (ANZECC and ARMCANZ) guidelines particularly near Sharpes Weir monitoring station. The results obtained in the TA further agree with a modelling exercise by Qin et al., (1995). They showed that the effect of the West Camden STP on total nitrogen levels had dropped considerably by 10 km downstream of the discharge point (Qin et al., 1995). Further, the Wilberforce site exceeded the ANZECC and ARMCANZ guidelines in more than 90% of the sampling sessions (SCA, 2010). In 1990, about 75% nitrogen loads in Wilberforce reach was related to treated effluent originating from two nearby STPs (James, 1997). Therefore, it was assumed that the peak observed at Wilberforce was strongly influenced by McGrath Hill, South Windsor, Riverstone, St Mary’s, Quakers Hill and Rouse Hill STPs attached to Cattai and South creeks. However,

there was no peak visible at the Penrith monitoring site although it is located near Penrith STP. This is because the effluents from the Penrith STP are discharged via Boundary Creek below Penrith Weir and monitoring is undertaken just above the discharge point (see Plate 8). It is assumed that the peaks recorded on consecutive years at Yarramundi are indicative of the effects of effluent discharged by the Penrith STP. Overall, the heights of peaks reduced around 2000, due to major upgrades made to the STP.

HNR

Boundary Creek

Plate 8. Boundary Creek STP discharge point at Penrith.

The variation of CHL was relatively high at Sharpes Weir and between Wilberforce & Wisemans Ferry compared to other sites on the river (Figure 25). Rahman and Salbe (1995) found that South Creek was a major carrier of diffused source pollution into the main river. The two creeks carry runoff, stormwater and effluent from sub- catchments that are highly urbanised but include agricultural areas. Wilberforce and Sackville monitoring stations are situated downstream of Cattai Creek confluence and South Creek confluence. Thus, effects of nutrient loads from these creeks have probably influenced the increased CHL levels. Change in river morphology is another possible factor for increased CHL levels. Previous river depth reports indicate that the depth of the HNR reach between Cattai Creek and the Colo River confluences ranges from 4 m and 6.5 m and is deeper than the depths in both Cattai

Creek and the Colo River (SPCC, 1984). Sackville station is located in this deeper reach of the river. Slow moving water resulting from increased depth and high nutrient loads from point and non-point-sources favour native and exotic floating macrophyte species in the HNR (Thiebaud and Williams, 2007). Similarly, Krogh et al., (2008) reported that increased nutrients (Nitrogen and Phosphorus) were related to STP discharges in the HNR.

The DO variable exhibited a slight decrease towards the mouth of the river but no obvious corresponding pattern of SIL or TEMP was observed for the years considered in this analysis. However, the median temperature was consistent along the HNR with slight fluctuations due to local weather conditions.

Predominantly the peaks observed at some monitoring stations were directly associated with treated effluent discharged by STPs and pollutant loads released by major creeks. South Creek and Cattai Creeks are the two most influential pollutant carriers into the HNR. However, in recent years agricultural land contributed more (Nitrogen 64%) to the total nutrients in the river than STPs (Nitrogen 27%) (EPA, 2002). The patterns of highs and lows for a particular variable were similar when they occurred for both annual median values and overall mean values calculated using all raw data (see Table 13, Figure 24 & 25). This further confirms the trends observed for each variable. The peaks were clearly noted for the monitoring site located downstream of the pollutant source, and they were almost invisible for corresponding years for the monitoring site located upstream of the pollutant source. Nevertheless, these effects do not seem to present along the river at the same level. These results further confirm the zonation of river obtained by the HACA to some degree as the peaks obtained for NOx at Yarramundi and Wilberforce sites and peaks obtained for CHL between Wilberforce and Lower Portland were within the polluted zone identified by the HACA.

The analysis presented in this chapter serves as a window to an understanding of the major impacts and challenges to peri-urban rivers in the present time. The findings based on long-term historic trends are particularly applicable to all major rivers that flow through rural-urban fringes of the terrain and impacted by agricultural activities and effluent discharges. While the usual suspects of pollutant sources for urban rivers are common for peri-urban rivers, we found that the trends in water quality variables are more complex and affected by other factors such as runoff from agricultural sources.

100

MK statistic-Temporal scale 1985-2009

200 157

150

100 53 33

0 MK statistic TEMP CHL DO SS NOx SIL

-50 - - 48 50

-100

-150 -

-200 185

-250

Figure 23. Multivariate MK test results for temporal trends.

(Note that for variables CHL and SS a decreasing trend and for other variables an increasing trend are observed. Also, these trends are significant (α = 0.05) for variables TEMP and SS.)

101

Figure 24. Long-term median NOx variation along HNR.

Figure 25. Long-term median Chlorophyll a variation along HNR.

102

6.4 Concluding Remarks

The long-term monitoring of river water quality is likely to suffer from data gap due to funding cuts and other reasons, nevertheless we need to assess river health based on the available information. This chapter demonstrated how the FA, HACA and TA techniques can be applied to evaluate long-term historic data sets. The FA was useful in grouping variables based on some meaningful interactions within the river system. In particular, it was useful in accounting for the interactions of wastewater inputs, nutrient inputs, effluent discharge and geomorphologic effects and helped to group the variables based on the factor loadings. Using the HNR system as a case study, the application of FA indicated that there are three main factors, viz., bio- geographical, natural and nutrient pollutant, affecting the overall condition of river system in a peri-urban context. The three components extracted with the application of PCA explained more than 70% of the total variance of the data. The bio- geographical and nutrient pollution factors more likely related to the direct influence of changes and activities of peri-urban natures and accounted for about 50% of variability in water quality.

The application of HACA helped in identifying similarities and dissimilarities in monitoring stations based on the six water quality parameters, viz., TEMP, CHL, DO, NOx, SS and SIL. This analysis indicated that the monitoring stations of the HNR could be divided into two distinct zones, ‘clean’ and ‘polluted’. Further, this chapter demonstrated that HACA can be employed for objective classification of river waters for future monitoring programs by government agencies, particularly to design cost effective monitoring programs. The TA indicated considerable temporal and spatial variations in NOx and CHL in places where the HNR system is impacted by point-source discharges from sewage treatment plants. The TEMP and SS variables showed significantly increasing and decreasing trends in spatio-temporal scales for the last 26 years. Direct impact by the diffused source of pollution is not very clear from the results, although TA supported the findings of HACA.

* * * * *

103

CHAPTER 7

BIOTIC ASSEMBLAGES AS RIVER HEALTH

INDICATORS

Summary

In this chapter, the analysis of water quality and biological (phytoplankton and benthic-macroinvertebrates) data collected from the HNR during three sampling surveys conducted in February, May and August 2011 is reported. The variability in water quality parameters clearly indicated a complex pattern, depending on the season (Interaction p = 0.001), which highlighted how the river condition is stressed at multiple points due to human intervention. In particular, the downstream locations indicated an accumulation of nutrients, the presence of increased sediments and phytoplankton related variables such as total counts, bio-volumes, Chlorophyll a and total phosphorus. The patterns of phytoplankton communities varied in a complex way depending on the season (Interaction p = 0.001). Abundances of phytoplankton were also found in less concentration where the water column is calm. However, when the water clarity drops due to tidal cycles, inflows from tributaries and intense boating activities the phytoplankton abundances also increased considerably. On the other hand, benthic-macroinvertebrates compositions were significantly different between locations (p = 0.001) with increased abundances associated with upstream sites. Aphanocapsa holsatica and Chironomid larvae appeared as the important indicators for upstream and downstream site differences in water quality. Water temperature influenced the phytoplankton communitypattern (ρw = 0.408) while pH influenced the benthic-macroinvertebrate community pattern w=0.4 (ρ 37). The findings in this chapter provide some valuable insights into the interactions of patterns of water quality on peri-urban river systems and biotic assemblages as to the suitability of benthic-macroinvertebrates and phytoplankton assemblages as indicators for monitoring and assessing river health.

104

7.1 General

The structure of a river is formed with multiple gradients which are related to hydrology, geography and anthropogenic inputs (González-Oreja and Saiz-Salinas, 1998). When river systems are impacted the level of deterioration is reflected in water quality as a measurable change and in aquatic faunal and floral species as a shift in their distribution and abundance patterns. For the past decade the interrelationships among qualitative and quantitative aspects of water quality, faunal and floral species have been extensively documented for large river systems around the globe (Whitton and Kelly, 1995, Alberto et al., 2001, Bergey and Ward, 1989, Biggs, 1996, Buapeta et al., 2008, Hoelzl, 2007, Jansen et al., 2004, Harris, 1995). Similarly, a variety of approaches based on ecological functions and integrated- indices have been proposed to assess the condition of urban river systems (Bunn et al., 1999, Young et al., 2004, Wright, 1995, Karr, 1991). While such studies have advanced our current knowledge on how freshwater streams respond to persistent environmental stress, studies that holistically investigate the perturbation through benthic-macroinvertebrates, phytoplankton species, and water quality indicators in a peri-urban context are scarce in the literature. As such, an investigation of this type is timely in view of increasing pressure from peri-urban activities on many Australian and world rivers.

Understanding the patterns of distribution and abundance of phytoplankton species and benthic-macroinvertebrates as surrogates for perturbation has many advantages. In the recent years, phytoplankton species have been collectively employed to understand surface water acidification, paleolimnological evaluations and nutrient dynamics (Stevenson et al., 1996, Smol, 1992, Battarbee et al., 1999, Stevenson and Smol, 2003, Kelly and Whitton, 1995, Lung and Paerl, 1988, Elliott et al., 2000). The phytoplankton assemblages are suitable indicators in freshwater systems because many persist before and after environmental stress and are highly sensitive to nutrient pollutants (Whitton and Kelly, 1995, Reavie et al., 2010). Further, the sample collection and identification of phytoplankton species is relatively easy and cost effective due to their mass distribution patterns throughout the course of streams. Thus, phytoplankton data complements the standard physio- chemical variables of a river system.

On the other hand, sedimentary macroinvertebrates are also often considered as surrogate measures of river health assessment as they play a key role in biological

105

processes such as nutrient recycling, metabolism of pollutants, dispersion of secondary products and transmission of energy through food webs (Covich et al., 1999, Lu, 2005). These species are suitable candidates for river health assessment purposes due to a number of reasons. Firstly, many of them are sedentary and often exposed to significant anthropogenic pollution such as solid waste, industrial effluent, discharge from boats and sewage treatment plants (Bilyard, 1987). Secondly, the benthic species indicate a high resolution of taxonomic diversity resulting in different response levels to various toxicants (Warwick and Clarke, 1991). Finally, they have a long life span and many have a patchy dispersion throughout the streambed. Such qualities are particularly important when using them to evaluate long term environmental stress rather than the static condition of aquatic systems (Bilyard, 1987, Underwood and Peterson, 1988).

The key objectives of this chapter are to understand the spatial and temporal trends of water quality variables in relation to the peri-urban impacts (peri-urban impacts are described in Chapter 1, 2 and 3), determine how benthic macroinvertebrate communities and phytoplankton communities are influenced by water quality over temporal and spatial scales and evaluate suitable phytoplankton and benthic macroinvertebrate species as indicators for river health assessment.

7.2 Materials and Methods

7.2.1 Location Selection and Characterisation

Samples were collected from four locations along the HNR, viz., Penrith (Lat. - 33.755303, Long. 150.671047), Yarramundi (Lat. -33.613105, Long. 150.700010), Cattai (Lat. -33.558361, Long. 150.889275) and Sackville (Lat. -33.501894, Long. 150.875187) over three different seasons (February, May and August) in 2011 (Figure 26). 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 the close vicinity of these locations providing access to existing water quality data sets if required. Thirdly, the locations were easily accessible by boat and were well separated.

All locations were sampled using a two-stage nested sampling design to minimise standard error due to the patchiness associated with dense growth of macrophytes in the river bed (Hurlbert, 1984). Within each location, there were a total of 16 sub-

106

sites in the sampling design. A grid including 16 squares (the dimensions were slightly adjusted to match the variable river widths at each location) was overlaid on a Google Earth™ map layer and four sub-sites were randomly chosen for sample collection during each month. To ensure that sampling within each location were different for the various seasons, coordinates of sample locations were obtained using a Garmin eTrex®-H handheld global positioning unit and those coordinates helped in the selection of appropriate sampling locations for the subsequent sampling sessions.

All locations indicated distinct habitat types influenced by human and natural factors. The downstream locations (Cattai and Sackville) in the study were predominantly covered with a sandy substrate while one upstream location (Penrith) had a muddy substrate and the other location (Yarramundi) was characterised by cobbles and pebbles (see Plate 9). Water quality and benthic species at the Cattai and Sackville locations were exposed to strong tides, boating and fishing activities. The Penrith location had calm water retained by a weir (built in 1909) and the Yarramundi location consisted of shallow water (1-5 m). Both sites were not affected by the tidal cycle. However, Cattai and Sackville locations were sampled at incoming tide between 8 AM and 12 noon.

Figure 26. Map with key locations on the Hawkesbury-Nepean River system.

( = sampling location)

107

Plate 9. Quality of the bed sediments collected from the HNR.

(a: Penrith, b: Yarramundi, c: Sackville)

7.2.2 Water Quality Variables

We used a digital water quality meter (AQUA Read™ - AM 100) to obtain in situ measurements of temperature (measured in degrees Celsius), pH, dissolved oxygen (DO measured in milligrams per litre), electrical conductivity (EC measured in micro Siemens per centimetre at 250C), total dissolved solids (TDS measured in measured in milligrams per litre), salinity (SAL measured in parts per thousand) and oxidation- reduction-potential (ORP in milli Volts) (Plate 10). From each sub-site, 1 L of water sample was collected in an acid rinsed, high-density polyethylene bottle (HDPE) for laboratory analysis. A separate sample was collected in an aluminium wrapped HDPE bottle (to avoid exposure to direct sunlight) for Chlorophyll a and phaeophytin analysis. All water samples were stored below 40C and analysed within 72 hours of collection. Manganese (Mn) (3111B), Chlorophyll a (10200H) and phaeophytin (10200H) were analysed using methods described by Standard Methods for the Examination of Water and Wastewater (Eaton and Franson, 2005). The Escherichia coli (organisms/100mL), total nitrogen (TN), total phosphorus (TP), suspended solids (SS) and rainfall data were supplemented by accessing data bases managed by the SCA and the Australian Bureau of Meteorology.

108

Plate 10. Water quality sampling at the HNR.

(a: Taking water quality readings at Yarramundi using AQUA Read™ - AM 100, b: Pre-labelled HDPE sample bottles for different water quality analysis, c: Unloading the boat at Cattai for river sampling)

7.2.3 Phytoplankton and Benthic-macroinvertebrate Sampling

Samples for phytoplankton identification, bio-volume (mm3/L) and total phytoplankton cell counts (cells/mL) were fixed using Lugol’s iodine solution and sent away for analysis within 5 hours of collection (Hötzel et al., 1999). The phytoplankton species identification results were presented under four major phyla, viz., blue-green algae (Cyanophyta), green algae (Chlorophyta), diatoms (Bacillariophyta) and dinoflagellates (Dinophyta). Species belonging to phylum Euglenomophyta and Cryptophyta were also noted in the results.

We collected benthic-macroinvertebrates using a Ponar grab sampler (15 cm x 15 cm x 15cm). From each sub-site two replicate grabs were obtained (approximately 1 kg per grab). Sediment samples were washed and filtered through a series of 0.5 μm, 1 μm and 2 μm sieves in the laboratory within 24 hours of collection. The species retained on the residue were handpicked, preserved in 75% ethanol and identified, using their key morphological characteristics, under a Leica™ MZ-12

109

stereomicroscope (Plate 11). The benthic-macroinvertebrates were collected in accordance with scientific collection permit no P10/0061-1.0 under the NSW Fisheries Management Act 1994.

Plate 11. Collection and identification of benthic-macroinvertebrates.

(a: Preparation of sediment samples, b: Hand picking the benthic species, c: Identification and imaging the benthic specimens, d: May fly (Family Baetidae), e: Caddis fly, f: Non-biting midge (Family Chironomidae))

7.2.4 Taxa Identification

It was considered that the family level identification is sufficient for the benthic- macroinvertebrate study following similar work by Chessman (1995) and Armitage et al., (1983). The family level taxonomic resolution is rapid, cost effective and yields a high accuracy in the absence of a taxonomic expert (Rosenberg and Resh, 1993, Mustow, 2002). However, the phytoplankton species were identified to the species level where possible. The taxonomic nomenclature in the study was thoroughly checked for consistency using Integrated Taxonomic Information System (I.T.I.S., 2011).

7.2.5 Data Analysis

The multivariate relationships among species compositions and water quality variables were further investigated using Permutational Multivariate Analysis of Variance (PERMANOVA), PCA and Similarity Percentage (SIMPER) where relevant

110

(Clarke and Warwick, 2001, Clarke and Gorley, 2006, Anderson, 2001). PERMANOVA was particularly chosen for this study because it computes responses of multiple variables on multiple factors using permutation methods or Monte Carlo asymptotic distributions, is flexible for unbalanced design, and has designs that lack replication. This test also minimizes the experiment wide statistical error compared to analysis of the same data set through numerous ANOVAs for each variable. The SIMPER on the other hand provides an average Bray-Curtis dissimilarity percentage for species that primarily account for the observed assemblage differences between location and seasons.

The water quality variables were assessed using PCA as this analysis transmutes large numbers of naturally correlated variables into a small number of uncorrelated variables and assists with identifying the most meaningful variables that describe the majority of the variance in the water quality data set. The PCs are the linear combinations of originally measured water quality parameters given as eigenvalues (indicating the variances accounted by the component) and eigenvectors (indicating the directions of the PCA axis) (Chapman, 1992). The PCA was performed on normalised water quality data sets. To understand the linkages between biotic assemblages and multivariate environmental patterns, the BIOENV (PRIMER test for linking biota to multivariate environmental patterns) routine was performed (Clarke and Ainsworth, 1993, Clarke and Warwick, 2001). This tool calculates a rank correlation (ρw) between a Bray-Curtis Similarity matrix of biotic assemblages and Euclidean distances of environmental variables. The rank correlations indicate a best combination of environmental variables that are closely related to the biotic community pattern observed in benthic-macroinvertebrate and phytoplankton communities.

All environmental variables were screened for inter-correlations using Draftsman’s plots. Three highly correlated variables viz., TDS (correlation with EC r = 1.0) and SAL (correlation with TDS r =0.9) were removed from the analysis. Subsequently, this data set was z-scale transformed and normalised prior to multivariate analysis. The multivariate analysis was conducted using PRIMER™ (v.6).

Following Amstrong (1969), the dominant species of phytoplankton species was calculated as a percentage of the total population for a particular location. We also calculated the Shannon diversity index for both species in addition to the species richness index for Ephemeroptera, Plecoptera and Trichoptera (EPT index).

111

7.3 Results

7.3.1 Water Quality

The water quality variables, both abiotic and biotic, varied with distance along the river but in complex patterns depending on the season (season x location interaction significant, p = 0.001, Table 18). For a number of variables, there was a pattern of gradual increase from upstream to downstream (Figure 27 & Figure 28). All these variables, including TN and TP, are also varied by the presence of a clear peak at Cattai for at least one season.

Water temperature demonstrated a seasonal pattern with high readings in February (summer) (29.20C) and low readings in August (winter) (10.90C) (Figure 27c, Table 19). The pH of river water ranged between 6.8-9.3 and the mean value reached the upper level of the Australian and New Zealand guidelines (ANZECC) for aquatic systems for lowland rivers (ANZECC, 2000). The mean concentrations of nutrients, TN (0.38 mg/L) and TP (0.02 mg/L) also slightly exceeded the ANZECC guidelines for fresh and marine water quality default trigger values of 0.35 mg/L and 0.025 mg/L (ANZECC, 2000). These two nutrient variables and turbidity are found in increased concentrations during August. The total counts of phytoplankton recorded its highest values at Sackville during May and February exceeding the ANZECC guidelines for recreational activities. The values of DO and pH followed a similar seasonal and spatial pattern, which decreased towards Cattai and then marginally increased at Sackville (Figure 27j, h). The E.coli was considerably low at Penrith during the warmer months and slightly increased towards the downstream sites (Figure 28m). This pattern inverted over colder months with the high readings recorded at Penrith exceeding the Australian drinking water quality guidelines (N.R.M.M.C., 2011). Turbidity reported a maximum of 53.3 NTU at Cattai in August, which exceeded the ANZECC guidelines for aquatic systems. Cattai and Sackville sites received the highest rainfall (58.2 mm) during May and August (Figure 29). In comparison with trends of rainfall data, turbidity, suspended solids and EC followed similar spatial and temporal trends as monthly rainfall patterns.

The PCA ordination of water quality further reveals the interrelationships between water quality on spatial and seasonal scale. Figure 30 indicates multi-dimensional relationships of 15 water quality variables in two-dimensional space. PC1 accounted for the much of the variability (30.3%) which increased to 50.9% when PC2 was

112

included and to 68% when PC3 was included (Table 20 & Table 21). The sediment- related variables (turbidity, SS, EC) were heavily weighted on PC1. Similarly, phytoplankton related variables (phaeophytin, bio-volume, Chlorophyll a, TP) and ORP scored the second largest absolute values on PC1 axes increasing from left to right. This is an indication of their presence in the river system at similar times and similar patterns of concentrations as turbidity, SS ad EC. The vector lengths indicate the relative contribution of each variable to PC1 and PC2. The vectors lengths were also similar for these variables towards the most downstream sites, indicating their high correlation with Cattai and Sackville for May and August. Thus, PC1 is an indication of the river clarity due to sediment and phytoplankton biomass. Temperature (0.503) and Mn (0.519) are approximately equally weighted in PC2 indicating their similar behaviour at lower reaches over February. Chlorophyll a, phytoplankton total counts and DO are nearly on a straight-line at opposite directions to each other, so they are negatively correlated. In particular, DO indicated a higher correlation with the most upstream sites Penrith and Yarramundi during May and August.

The vector line for E.coli pointed towards Penrith indicating its presence is more associated with Penrith (Figure 30). The E.coli was abundant at upstream site (Penrith) especially during May and February and dropped to very low levels in August (Figure 28m). Laboratory studies indicate low temperatures favour the survivability and abundance of E.coli in rivers (Bogosian et al., 1996, Sampson et al., 2006, Brettar and Hofle, 1992). However, our results are not consistent with this trend. The impoundment may provide a suitable habitat for E.coli organisms, which originated from the Australian Wood Duck (Chenonetta jubata) population, to proliferate near the weir.

TN indicated a large negative coefficient (-0.466) on PC2 and an almost perpendicular to the TP vector line. This showed a relatively small influence of TN on phytoplankton compared with that of TP. If the points for locations on the plot are close to each other, then the water quality of those sites are similar. The tidal and non-tidal sites are approximately on the opposite direction on PC1 axis indicating the influence of tide on the patterns of water quality variables. There is a clear separation between the February samples and the other two seasons on PC2 axis. In May and August, the water quality has been similar, but in February water quality was more variable for all matrices.

113

36 50 a 16 b c 30 40 12 30 24 8 20 18

10 SS (mg/L) 4

Temp (Deg.Temp cel.) 12 0 Turbidity (NTU) Pen Yar Cat Sack Pen Yar Cat Sack Pen Yar Cat Sack 450 0.06 0.15 d e f

300 0.04 0.10

TP (mg/ L) TP 0.02 0.05 EC (uS/EC cm)

150 (mg/L) Mn

Pen Yar Cat Sack Pen Yar Cat Sack Pen Yar Cat Sack 0.8 240 g 8.8 h i 0.6 8.4 200 0.4 8.0

pH 160

TN (mg/L) TN 0.2 7.6 ORP (mv) 0.0 7.2 120 Pen Yar Cat Sack Pen Yar Cat Sack Pen Yar Cat Sack 14 12 j 10 8 Feb 6 DO (mg/L) DO 4 May August Pen Yar Cat Sack

Figure 27. Spatial and temporal trends of abiotic water quality variables.

(N.B., X-axis represents the location. The distances between locations are not to scale. SS: Suspended solids, Temp: Temperature, EC: Electrical conductivity, TP: Total Phosphorus, Mn: Manganese, TN: Total Nitrogen, ORP: Oxidation reduction potential and DO: Dissolved oxygen.)

114

8.0x104 9 250 m k 8 l 200 6.0x104 7 6 150 4 4.0x10 5 100 4 4 2.0x10 3 50 Phaeophytin (ug/L) Total counts (Cells/mL)

2 mL) (org./100 E.coli 0.0 0 1 Pen Yar Cat Sack Pen Yar Cat Sack Pen Yar Cat Sack

20 n 20 o

15 15 a (ug/ L) 10 Feb 10 May 5 August 5 Bio volume (mm3 / L) (mm3 volume Bio

0 Chlorophyll- 0 Pen Yar Cat Sack Pen Yar Cat Sack

Figure 28. Spatial and temporal trends of biotic water quality variables.

(N.B., X-axis represents the location. The distances between locations are not to scale.)

70 Feb May Aug 60

Rainfall (mm) 40

Penrith Yarramundi Cattai Sackville Location

Figure 29. Seasonal rainfall patterns across sampling location.

115

4 Cat_Feb Location

Cat_Feb Sac_Feb Penrith Sac_FebSac_Feb Sac_Feb Yarramundi Cat_Feb Cat_Feb Cattai

Yar_Feb Sackville Yar_FebYar_Feb 2 Yar_Feb MnTemp Pen_Feb

Chl-a SS ORPTC Pen_Feb Pen_Feb Pen_Feb Sac_AugSac_Aug TP 0 Cat_AugSac_AugCat_Aug Cat_Aug

PC2 Sac_May TurbSac_MayECSac_May E.coli Sac_Aug Sac_MayPhaeo-aBioVol pH Pen_AugPen_Aug Cat_MayCat_MayCat_May Pen_Aug Pen_May Cat_May Pen_Aug Pen_MayPen_MayPen_May DO Yar_MayYar_MayYar_May TN Yar_AugYar_AugYar_May -2 Yar_AugYar_Aug

-4 -4 -2 0 2 4 PC1 Figure 30. PCA Eigen vector plot of water quality variables.

(The obscured labels are Turbidity, EC, Phaeophytin and bio-volume.)

Table 18. PERMANOVA results.

Significant differences between: Location Season Interaction df Pseudo.F p df Pseudo.F p df Pseudo.F p Water quality 3 40.37 0.001 2 3.71 0.004 6 12.75 0.001 Phytoplankton 3 28.59 0.001 2 3.17 0.006 6 10.48 0.001 3 4.84 0.001 2 1.22 ns 5 1.86 ns Benthos

116

Table 19. Descriptive statistics of water quality variables.

In the Table, * Denotes ANZECC guidelines for aquatic systems for lowland rivers, ** denotes ANZECC for recreational activities, *** denotes Australian drinking water guidelines.

ANZECC Variable Minimum Maximum guidelines Temperature (0C) 10.9 29.2 35C max** 6-50*, 0.0 53.7 Turbidity (NTU) 0.2-0.5*** pH 6.8 9.3 6.5-8* Dissolved Oxygen (mg/L) 3.6 15.5 - Electrical Conductivity (uS/cm, 250C) 126.8 417.3 - Oxidation Reduction Potential (mv) 109.1 230.9 - Total Nitrogen (mg/L) 0.02 0.38 0.35* Total phosphorus (mg/L) 0.01 0.06 0.025* Manganese (mg/L) 0.02 0.05 0.1-0.5*** Chlorophyll-a (ug/L) 0.4 21.5 0.35* Phaeophytin (ug/L) 0.2 17.0 - Phytoplankton bio-vol (mm3/L) 0.2 27.7 - Phytoplankton total counts (cells/mL) 1431.0 77571.0 15000-20000** E.coli (org/100mL) 3.0 210.0 0*** Suspended solids (mg/L) 1.3 17.0 -

Table 20. Properties of the PCA plot for water quality variables, showing the variation explained by the individual PCs and cumulative percentage of variation.

Principal Percentage of Cumulative Component variation (%) variation (%) 1 30.3 30.3 2 20.6 50.9 3 17.3 68.2 4 10.6 78.8

117

Table 21. Properties of PCA plot for water quality variables, showing variables and eigenvectors for four PCs.

Variable PC 1 PC 2 PC 3 Temperature 0.101 0.503 -0.034 Turbidity -0.436 -0.048 0.118 pH 0.012 -0.153 -0.253 Dissolved oxygen 0.173 -0.318 -0.313 Electrical conductivity -0.406 -0.05 -0.013 ORP -0.204 0.148 0.318 Total nitrogen -0.161 -0.466 0.232 Total phosphorus -0.303 0.029 0.364 Manganese 0.091 0.519 0.024 Suspended solids -0.424 0.182 -0.013 Chlorophyll a -0.25 0.188 -0.116 Phaeophytin -0.275 -0.058 -0.255 Bio-volume -0.293 -0.057 -0.399 Total counts -0.136 0.179 -0.478 E.coli 0.116 -0.022 0.262

7.3.2 The Composition of Phytoplankton Communities

A total of 116 phytoplankton species were identified in this study. Phytoplankton abundances increased towards the downstream sites but diversity fluctuated on spatial and temporal scales with increased diversity recorded for Yarramundi (February and August) and Cattai (May) (Figure 31). Phytoplankton species compositions indicated a complex pattern over locations influenced by season (season x location interaction significant, p = 0.001, Table 18).

The pair-wise comparisons revealed that within the location factor, all seasonal differences were significant for all locations except for Penrith. For Penrith, only May and August were significantly different from each other. Similarly, a consistent pattern of significant differences between upstream (Penrith, Yarramundi) and downstream (Cattai, Sackville) was observed for February, May and August. The average similarity between and within groups showed a gradual decrease in similarity between February and August samples for all locations. A gradual decrease in similarity between Penrith and the downstream location is consistent for all seasons. This implies a higher degree of complexity in the phytoplankton community patterns which are strongly affected by seasonal and spatial effects. Aphanocapsa holsatica was responsible for the largest contribution of site and

118

seasonal dissimilarities, except for Penrith-Yarramundi and February-August (Table 22). In particular, Sackville was dominated by A. holsatica regardless of the season. Thus, this species may be a suitable pollution indicator for peri-urban river systems due to increased turbidity, SS and EC.

The species dominance indicated marked differences over different seasons and locations. During the February sampling, blue-green algal species A. holsatica (Phylum: Cyanophyta) dominated at all the locations (Penrith-40%, Yarramundi- 25%, Cattai-28% and Sackville 62%) (Figure 32a). However, in May, green-algae species (Phylum: Chlorophyta) represented a total abundance of 37%, 51% and 21% at Penrith, Yarramundi and Cattai locations respectively (Figure 32b). In August, a mix of species belonging to the Phylum Cyanophyta, Chlorophyta and Bacillariophyta appeared dominant across four locations (Figure 32c). Throughout the sampling period Sackville was dominated by A. holsatica (February-62%, May- 66%, August-39%). Bacillariophyta appeared in large numbers only during August and indicated a clear declining pattern downstream of the river (Penrith-51%, Yarramundi-19%, Cattai-10%). The BIOENV analysis showed temperature as the best water quality variable that correlated highly with the seasonal and spatial patterns of phytoplankton communities w(ρ = 0.443) (Table 23).

119

5 3.0x10 Feb a. 2.5x105 May 2.0x105 Aug 1.5x105 1.0x105 Abundance 5.0x104 0.0

Penrith Yarramundi Cattai Sackville

3.0 b. 2.5

2.0

1.5

1.0

0.5 Mean Shannon DiversityMean Index Penrith Yarramundi Cattai Sackville

Location

Figure 31. The abundance and mean Shannon diversity of phytoplankton species across four locations in different seasons. Error bars represent ±SE.

120

a. February

Relative Abundance (%) Abundance Relative 0 Ulothrix CATTAI PENRITH SWC SACKVILLE Chroomonas type type type Crucigenia spp Crucigenia Crucigenia spp Crucigenia Crucigenia spp Crucigenia YARRAMUNDI Kirchneriella sp Aphanothece sp Coelosphaerium Aphanocapsa sp Aphanocapsa sp Cyanogranis type Cyanogranis type Cyanogranis type Cyanogranis type Scenedesmus sp 1 sp Scenedesmus Scenedesmus sp 3 sp Scenedesmus Merismopedia sp small Aphanocapsa holsatica Aphanocapsa Aphanocapsa holsatica Aphanocapsa Aphanocapsa holsatica Aphanocapsa

b. May

0 Relative (%) Abundance Relative CATTAI PENRITH (4.5um) SACKVILLE Chroomonas Cyanocatena Monad (6um) Type type type Diam Diam Crucigenia spp Crucigenia YARRAMUNDI Kirchneriella sp Dictyosphaerium Aphanocapsa sp Cyanogranis type Cyanogranis type Scenedesmus sp 3 sp Scenedesmus Scenedesmus sp 3 sp Scenedesmus Scenedesmus sp 3 sp Scenedesmus Anabaena circinalis Microcystis flos-aquae Ankistrodesmus small Ankistrodesmus small Aphanocapsa holsatica Aphanocapsa Aphanocapsa holsatica Aphanocapsa Dictyosphaerium 4.5um Dictyosphaerium 4.5um

c. August

0 Relative (%) Abundance Relative CATTAI PENRITH SACKVILLE Chroomonas Chroomonas Chroomonas Cryptomonas Aphanocapsa holsatica type Aphanocapsa holsatica type 4.5um Diam 4.5um Diam 4.5um Diam Crucigenia spp Crucigenia Crucigenia spp Crucigenia Crucigenia spp Crucigenia Crucigenia spp Crucigenia YARRAMUNDI Dictyosphaerium Dictyosphaerium Dictyosphaerium Scenedesmus sp 3 sp Scenedesmus Scenedesmus sp 3 sp Scenedesmus Scenedesmus sp 3 sp Scenedesmus Scenedesmus sp 3 sp Scenedesmus Asterionella formosa Asterionella Asterionella formosa Asterionella Aulacoseira granulate

Figure 32. Dominant phytoplankton species during February, May and August.

(Green : Chlorophyta, Blue : Cyanophyta, Red : Bacillariophyta, Black : Other Phyla)

121

Table 22. SIMPER results for phytoplankton species.

(Cumulative contribution up to 35% average dissimilarity. Site 1 and 2 refer to the first and second locations under group column)

Av. Ab. Av. Ab. Ave. Contrib. Cum. Location Species name site 1 site 2 Diss. % % Pen-Yar Dictyosphaerium sp. 30.5 13.99 4.54 8.92 8.92 Monad sp. 0.57 22.58 3.93 7.72 16.64 Cyanogranis type 29.47 7.51 3.47 6.83 23.47 Aphanocapsa holsatica 30.34 15.77 2.96 5.81 29.28 Chroomonas 21.7 20.27 2.58 5.06 34.34 Aphanothece sp. 15.61 0 2.3 4.52 38.86 Pen-Cat Aphanocapsa holsatica 30.34 22.29 5.18 8.74 8.74 Dictyosphaerium sp. 30.05 22.71 4.4 7.43 16.18 Crucigenia sp. 12.17 43.71 4.08 6.89 23.06 Chroomonas 21.7 25.02 2.24 3.78 26.84 Desmidium 0.55 14.62 2 3.38 30.23 Kirchneriella sp. 11.34 26.48 1.94 3.27 33.49 Aphanocapsa sp. 0 17.9 1.9 3.22 36.71 Yar-Cat Aphanocapsa holsatica 15.77 22.29 4.23 7.53 7.53 Crucigenia sp. 18.85 43.71 3.6 6.41 13.94 Monad sp. 22.58 0.33 3.12 5.56 19.5 Cyanogranis type 7.51 27.84 2.85 5.08 24.58 Scenedesmus sp. 22.69 18.66 2.63 4.68 29.26 Kirchneriella sp. 8.48 26.48 2.58 4.6 33.85 Dictyosphaerium sp. 13.99 22.71 2.58 4.59 38.44 Pen-Sac Aphanocapsa holsatica 30.34 208.52 13.55 19.96 19.96 Aphanocapsa sp. 0 57.82 3.97 5.85 25.81 Crucigenia sp. 12.17 48.16 2.9 4.27 30.09 Cyanodictyon sp. 0 42.6 2.86 4.21 34.3 Dictyosphaerium sp. 30.05 36.72 2.49 3.67 37.97 Yar-Sac Aphanocapsa holsatica 15.77 208.52 15.47 21.48 21.48 Aphanocapsa sp. 0 57.82 4.34 6.03 27.51 Cyanogranis type 7.51 46.21 3.08 4.28 31.78 Cyanodictyon sp. 0 42.6 2.91 4.04 35.82 Cat-Sac Aphanocapsa holsatica 22.29 208.52 13.06 26.99 26.99 Cyanodictyon sp. 0 42.6 2.58 5.33 32.32 Aphanocapsa sp. 17.9 57.82 2.5 5.16 37.48 Season Feb-May Aphanocapsa holsatica 84.32 87.77 5.6 9.85 9.85 Cyanogranis type 61.38 21.89 3.96 6.96 16.81 Desmidium 0.41 22.6 3.12 5.48 22.29 Monad sp. 0 14.75 2.78 4.89 27.18 Dictyosphaerium sp. 9.11 29.78 2.59 4.55 31.73 Scenedesmus sp. 12.53 26.69 2.19 3.85 35.58 Feb-Aug Cyanogranis type 61.38 0 6.6 10.2 10.2 Aphanocapsa holsatica 84.32 35.6 6.49 10.03 20.23 Dictyosphaerium sp. 9.11 38.71 3.66 5.65 25.88 Aphanocapsa sp. 32.06 0 2.85 4.41 30.3 Chroomonas 13.76 34.86 2.72 4.2 34.5 Asterionella formosa 0 19.7 2.68 3.95 38.45 May-Aug Aphanocapsa holsatica 87.77 35.6 1.01 7.64 7.64 Dictyosphaerium sp. 29.78 38.71 1.69 7.46 15.1 Desmidium 22.6 0 1.02 5.61 20.71 Cyanocatena 18.84 0 1.14 4.7 25.41 Asterionella formosa 0 19.7 2.45 4.67 30.08 Cyanogranis type 21.89 0 1.06 4.51 34.59 Chroomonas 18.56 34.86 0.93 4.35 38.94

122

Table 23. Summary of BIOENV routine connecting multivariate environmental patterns with biotic assemblages.

Environmental variable contributing to Correlation ( ) ρw maximum rank correlation variables Phytoplankton 0.443 Temperature Benthos 0.409 pH

7.3.3 The Composition of Benthic-macroinvertebrate Communities

A total of 25 Benthic-macroinvertebrate families and 475 individuals were obtained from all locations during the three sampling surveys in this study. The most diverse phylum was Arthropoda with 18 Families. The PERMDISP multivariate dispersion test was significant (p = 0.000) and there were a number of zero data cases for Penrith and this was removed from the data set and run as an unbalanced PERMANOVA test (Table 18). Type one, two and three sums of square methods were attempted, and they all produced similar results. Thus, type three was chosen for interpretation. The PERMANOVA test indicated that benthic species compositions were significantly different between locations (p = 0.001) but not significantly different between seasons. Interaction between season and location was also not significant. Pair-wise comparisons revealed the Yarramundi site was significantly different from Cattai, Sackville and Penrith.

A clear peak was observed at Yarramundi for species abundance, diversity and for EPT index (Figure 33c). The Epemeroptera, Plecoptera and Trichoptera are highly sensitive to pollutants and their abundance is accounted in the EPT index (Merritt and Cummins, 2009, Barbour et al., 1999, Lydy et al., 2000). SIMPER reveals the Family Chironomidae (Order: Diptera) was responsible for the maximum dissimilarity between most locations and seasons, except for Yarramundi-Cattai, thus acting as a suitable indicator species for location differences (Table 24). Towards the downstream locations such as Cattai and Sackville, the sediment samples indicated an increased average abundance of Chironomidae family. The average abundance of this family across survey locations was also high in August compared with May and February. BIOENV identified pH as the best variable that correlated highly with the observed assemblages’ data and that best explained the biological patterns with a rank correlation of ρw = 0.409 (Table 23).

123

180 160 Feb a. 140 May 120 Aug 100 80

Abundance 60 40 20 0 Penrith Yarramundi Cattai Sackville 2.0

b. 1.5

1.0

0.5

Mean Shannon DiversityMean 0.0

Penrith Yarramundi Cattai Sackville 60 c. 50 40 30 20

Mean EPT Index EPT Mean 10 0

Penrith Yarramundi Cattai Sackville Location

Figure 33. The abundance, mean Shannon diversity and EPT index of benthic macroinvertebrates. Error bars represent ±SE.

124

Table 24. SIMPER results for benthic species.

(Cumulative contribution up to 50% average dissimilarity. Site 1 and 2 refer to the first and second locations under group column)

Ave. Abu. Ave. Abu. Ave. Contrib. Cum. Group Family name site 1 site 2 Diss. % % Pen-Yar Chironomidae 1.57 4.44 14.88 17.57 17.57 Baetidae 0 3 13.76 16.24 33.82 Leptoceridae 0 2.78 13.25 15.65 49.46 Lumbriculidae 0 1.3 5.51 6.51 55.97 Pen-Cat Chironomidae 1.57 1.97 26.21 52.85 52.85 Yar-Cat Baetidae 3 0 13.81 17.43 17.43 Leptoceridae 2.78 0 13.28 16.75 34.18 Chironomidae 4.44 1.97 12.96 16.36 50.53 Pen-Sac Chironomidae 1.57 2.09 30.61 40.71 40.71 Physidae/Planoribidae 0.29 0 10.57 14.05 54.76 Yar-Sac Chironomidae 4.44 2.09 13.87 16.89 16.89 Leptoceridae 2.78 0 12.95 15.76 32.65 Baetidae 3 0.17 12.14 14.78 47.43 Lumbriculidae 1.3 0.17 5.5 6.7 54.13 Cat-Sak Chironomidae 1.97 2.09 23.09 46.81 46.81 Hymenosomatidae 0 0.4 9.1 18.44 65.25 Season Feb-May Chironomidae 1.5 1.87 23.08 35.37 35.37 Physidae/Planoribidae 0.22 0 7.4 11.35 46.72 Atyidae 0.13 0 5.71 8.76 55.47 Feb-Aug Chironomidae 1.5 4.19 35.18 56 56 May-Aug Chironomidae 1.87 4.19 28.38 50.5 50.5

7.4 Discussion

7.4.1 Spatial and Temporal Trends in the Water Quality

Through the PCA ordination plot, it is evident that the water quality of the HNR was strongly influenced by the interaction of both natural and anthropogenic effects associated with season and location such as tide, inflows from tributaries, rainfall, urban effluent discharges and impoundments. During February, recreational activities occur at a higher frequency in the river. However, in colder months, such activities are less common but rainfall events start to impact on the river water quality especially towards the downstream sites. This pattern is evident in a number of water quality variables. For example, spatial trends of monthly rainfall closely resembled the turbidity, suspended solids and EC values across four locations and this clearly indicates how the run-off from impervious surfaces and agricultural areas determine impacts on peri-urban river systems. The significance of agricultural runoff from market gardens into the HNR river system has previously been reported by Baginska et. al., (1998).

125

Another interesting result is the association of suspended and particulate matter related variables (turbidity, SS and EC) and phytoplankton related variables (biovolume, Chlorophyll a, Phaeophytin, TP, ORP) which became apparent at downstream sites. Historic data records also agree on increased Chlorophyll a levels towards downstream reaches of the HNR (See Chapter 6). The behaviour of turbidity and TP indicates their co-occurrence and thereby promotes phytoplankton assemblages (see Figure 30). Most particulate matter and TP entering the freshwater systems through run-off are readily available for the proliferation of phytoplankton species (Smith, 1987). However, the effect of TN on phytoplankton communities is marginal and in fact, a negative association of TN with phytoplankton- related variables was evident from the PCA ordination plot. This result is consistent with the previous literature that suggested the nutrient association with phytoplankton in surface waters. TP concentrations between 0.02 mg/L to 0.1 mg/L have more influence in promoting the phytoplankton assemblages compared with TN given other hydrological factors, viz., water temperature, depth, mixing patterns, are constant (Basu and Pick, 1996, Perkins and Underwood, 2000, Mainstone and Parr, 2002). However, a long standing debate exists about the relative importance of an ideal nutrient ratio (TN:TP) on phytoplankton abundance in surface waters initially reported by Smith (1983) based on lake environments. For example Reynolds (1988) argued that nutrient limitation is rare in large rivers and instead plant nutrients are present in excess for most phytoplankton. However, based on the present study, the influence of TP on the abundance of phytoplanktons is clearly evident regardless of the TN:TP ratio of the water column.

Increased levels of SS in water negatively affects the abundance of periphyton by limiting the light penetration and positively affects the phytoplankton by acting as a nutrient carrier (Bilotta and Brazier, 2008, Heathwaite, 1994). In the present study, TP increased with SS promoting the phytoplankton biomass towards the lower reaches (Cattai and Sackville). The low levels of plant nutrients observed at Penrith may be due to nutrient assimilation by the dense E. densa, Elodea sp. and Vallesneria sp. populations in the weir pool (see Plate 12).

126

Plate 12. Dense growth of E. densa in the Penrith weir pool.

(a: Grab sampler filled with E. densa, b: Penrith weir, and c: Dense growth of the weed)

The decline in DO and pH values at Cattai in August is interesting. One reason for the drop in DO is related to increased phytoplankton biomass due to metabolism of organic matter entering the river through South and Cattai Creeks. The drop in pH was probably due to the production of ammonia and other organic acids under anaerobic fermentation. The presence of anaerobic conditions resulting in an acidic pH has been previously reported for the HNR and other urban rivers (Fuji River, Japan and Pisuerga River, Spain) (Shrestha and Kazama, 2007, Vega et al., 1998, Pinto and Maheshwari, 2011). However, DO and alkalinity increased towards Sackville. Tidal inflows seem to have a greater role in this observation and warrant further research.

Electrical conductivity clearly followed the seasonal and spatial patterns of Chlorophyll a. The electrical conductivity is one of the key variables influencing the growth of aquatic flora although it has rarely been investigated in detail (Bunt et al.,

- + 3- 1982). Camargo et al., (2005) found high nutrients (NO 3, NH 4, PO 4) and high conductivity in four mountain river systems in Spain which showed a significant correlation with increased periphytic Chlorophyll a levels. Thus, in the absence of

127

nutrient analysis, EC can be used as a surrogate measure to predict Chlorophyll a levels.

The ORP is a measure of amount of oxidisers and reducers in the water column. Large positive values are related to elevated oxidising conditions created by oxidising agents. When the water column becomes deep, it leads to thermal stratification. As a result an oxycline is formed creating reducing conditions in the hypolimnion generating an influx of sediment bound nutrients and heavy metals (Turner and Erskine, 2005, Søndergaard, 2009). In the present study, the increased ORP towards downstream reaches may occur as a result of oxidising pollutants entering the river system through South and Cattai Creeks and increased river depth. It was noted that the increasing concentrations of strong oxidising agent Mn towards downstream sites may be the main cause for the above observation.

In summary the spatial and temporal water quality variability in the HNR did not follow the previously accepted ‘river continuum’ concept where physical water quality varies from the headwaters to the mouth in a gradual fashion (Vannote et al., 1980, Karr, 1999). Rather, the water quality was influenced at multiple points due to geographical, natural and anthropogenic factors. These changes are particularly prominent near confluences and impoundments and water quality tends to vary strongly over short distances in these areas.

7.4.2 Phytoplankton communities

The persistence of algae in freshwater has been an ongoing problem for many water management authorities in Australia and other parts of the world. The distribution and abundance of phytoplankton species are affected by a large number of physio- chemical variables and hence are considered as a natural indicator of the health of fresh and marine waters. However, the appearance of numerous genera depending on the season and location confounds the direct comparison of the similar taxonomic level findings with other studies and poses a greater difficulty in establishing one definitive indicator of water quality.

A clear pattern of phytoplankton succession in the present study was observed over February, May and August with different classes of species dominating in each season, viz., Cyanophyta, Chlorophyta and a mix of both species together with Bacillariophyta respectively. When the season changed, so did the dominant groups of phytoplankton appearing at each location. Similar patterns of succession have

128

been previously reported in the HNR (SPCC, 1983, Hawkins et al., 1994). This pattern initiated with a Bacillariophyta dominant system which gradually changed to a Chlorophyta dominant system and then to a Cyanophyta dominant system. In many Australian freshwater systems, Bacillariophyta appears in abundance during colder months (August) and Dinophyta and Cyanophyta become dominant during warmer months (February) (Australian Water Resources Council, 1991). However, the particular order of succession seems to be greatly influenced by climatic and physio-chemical factors. A seasonal succession is reported for German freshwater systems with Bacillariophyta being dominant in Spring and May while Chlorophyta appears more during February (Lemmermann, 1907 cited in(Swale, 1969). A phytoplankton succession influenced by physio-chemical variables was reported for American rivers where Bacillariophyta dominated at low nutrient conditions, a mix of Bacillariophyta, Chlorophyta and Cyanophyta dominated at high nutrient and turbid conditions promoting the proliferation of Cyanophyta (Heiskary and Markus, 2001).

The seasonal succession pattern among phytoplankton species in the HNR greatly reduces our ability to establish a particular genus as an indicator species for river health monitoring purposes. This difficulty is also explained in the context of lakes and river systems by previous researchers (Reynolds, 1990, Wehr and Descy, 1998). Nevertheless, A. holsatica found to be the key species responsible for the maximum dissimilarity between locations and seasons suggesting its suitability as a potential indicator. The average abundance of A. holsatica remained high at Sackville regardless of the season and may be related to the increased turbidity, suspended solids and EC recorded at downstream locations. Aphanocapsa holsatica is a blue-green algae that forms globular amorphous gelatinous colonies with an average colony volume of 6000 (μm3) (Bellinger and Sigee, 2010). It belongs to a low risk category for recreational activities in Australia and not reported as a toxin producing species (Kankaanpaa et al., 2005, ACT Health, 2010). Although Aphanocapsa appears promising for seasonal and location discrimination, in-depth monitoring and standardisation is required prior to its adaption for peri-urban river health assessments considering the naturally occurring succession.

During the present study, it was further observed how water temperature strongly influenced the composition of the phytoplankton community structure. The importance of temperature, as a suitable variable for the prediction of the eutrophication risk in the HNR system through a discriminant function model has been previously reported (unpublished). The temperature effect on phytoplankton is

129

mainly through their chemical and photosynthetic pathways (Raven, 1974). In particular, temperature plays a key role in the metabolic reactions of many phytoplankton families viz., nitrogen fixation and excretion of nitrogenous compounds in Cyanophyta is greatly dependent on the thermal effects (Jones and Stewart, 1969). A number of previous studies also agree on the temperature effects on phytoplankton communities although some report on the collective effects of the influence of nutrient variables. Brogueira et al., (2007) reported on a combination of variables (temperature, Salinity, Si(OH)4 and TP) that best described the phytoplankton community pattern of the Tagus estuary in Portugal. From a similar analysis, Ignatiades et al., (2009) reported P-PO4 followed by N-NO3, salinity, temperature as suitable variables that best correlated with phytoplankton patterns in the Mediterranean sea. The association of temperature with phytoplankton assemblages of estuarine waters is also reported in China and Taiwan (Shen et al., 2011, Tew et al., 2006). The phytoplankton abundances of large river systems respond dynamically to the physical condition of water quality driven by spatial and seasonal differences. Interestingly, none of the major plant nutrient-related variables appear as crucial to the phytoplankton community composition in the HNR. Thus, this could be the prime reason for the lack of correlation observed between phytoplankton communities and nutrinents in the HNR.

7.4.3 Benthic-macroinvertebrates

The ecology of benthic-macroinvertebrates in relation to environmental stress has been well documented (Sanders, 1958, Gray, 1974, Snelgrove and Butman, 1994, Rygg, 1985, Heip et al., 1988). In Australia, investigations were conducted in Port Jackson, Port Hacking, Port Pirie and Botany Bay estuaries to understand how macrobenthic species are impacted by strong pollution gradients (Stark, 1998, Ward and Hutching, 1996, Morrisey et al., 1996). In particular, many previous studies have been conducted on the HNR system investigating the community response in relation to the condition of the sediment bed in the estuarine reach between Broken Bay and the Colo River confluence (Jones et al., 1986, Jones, 1987, Jones, 1990, MacFarlane and Booth, 2001, Lasiak and Underwood, 2002). As such, they did not examine the effects of water quality on benthic assemblages in detail, but rather focused on only a few variables of the water column (i.e., water salinity, temperature and dissolved oxygen content).

The PERMANOVA test and EPT sensitive species index both agree the distinct nature of the Yarramundi site is characterised by the consistently high species

130

abundance and diversity throughout the sampling seasons. A number of natural (e.g., substrate type, oxygen availability, food availability), anthropogenic factors (e.g., wastewater, impoundments) and internal community dynamics could be responsible for such differences (de Haas and Kraak, 2008, Clarke and Warwick, 2001). When a waterway is polluted, a reduced abundance or shift in community composition from sensitive taxon to tolerant taxon is expected (Sprague et al., 1965, Chadwick et al., 1986). However, two theories describe the fluctuating state of the pollution sources and biotic species compositions in streams. One theory stipulates that species diversity remains highest at the intermediate levels of disturbance and decreases at a higher level of disturbance due to stress (Huston, 1979, Connell, 1974). The other postulates that species richness, abundance and biomass initially reduces with increasing pollution and then rises again with increasing pollution as pollutant-tolerant species increase (Pearson and Rosenberg, 1978, González-Oreja and Saiz-Salinas, 1998). Thus, our results agree with the first theory on pollution- diversity interaction. The observed decreasing trend for species abundance and diversity towards the downstream locations is also consistent with previous studies (Azrina et al., 2006, Varnosfaderany et al., 2010).

A few Chironomids, at all locations, throughout the sampling seasons with a consistently increased abundance were recorded mostly towards downstream sites such as Cattai and Sackville. Chironomids are tolerant to pollutants and their presence in abundance indicates deterioration in the aquatic environment. Our observation may be related to the accumulated pollutants at downstream locations which entered the river system mainly through South and Cattai Creeks and upstream tributaries. This observation is also in agreement with a number of similar studies that reported the dominance of Chironomid taxon in streams polluted with heavy metals and urban pollutants (Sprague et al., 1965, Chadwick et al., 1986, Clements, 1994, Walsh et al., 2005, Varnosfaderany et al., 2010). However, its potential as a bio-indicator, for site discrimination in peri-urban river systems, requires further evaluation.

Although the results were not significant, in general a very low abundance of benthic species at Penrith and Cattai was observed. It was hypothesised that these locations are affected by a number of localised stressors. For example, the sediment samples collected at Penrith were rich in organic matter. Due to the establishment of weirs, most sediments, nutrients and pollutants were trapped and accumulated in the river bed over time. As a result, the sediment bed now harbours a dense

131

population of E. densa. On the other hand, the sandy sediment cover at Cattai was exposed due to heavy loads of treated effluent entering from South and Cattai Creeks and consequently to a thorough mix with tidal inflows. Further, the Cattai site was deeper (more than 20 m deep) and wider (100 m wide) compared to the other locations thereby creating greater practical difficulty in obtaining a representative sample of the benthic-macroinvertebrates. Benthic-macroinvertebrates can spread over a long spatial scale and such mobility response creates practical difficulties when collecting them from the sediment bed (Covich et al., 1999).

It was found that pH is the key water quality variable influencing the community patterns of benthic-macroinvertebrates. The pH of the HNR water mostly remained above 7.2 with high alkaline readings relating to Penrith and Yarramundi in May and August. The pH of water depends on episodic rainfall events, surface runoff, extended drought periods and acidity or alkalinity of the bed sediment (Hämäläinen and Huttunen, 1996, Likert, 1932). The pH effect on benthic species also reported through metal toxicity (Kullberg, 1992). Previous studies report on the influence of climatic events such as floods, drought conditions and salinity gradients on benthic species in the HNR but the influence of pH is not reported (Jones et al., 1986, Jones, 1987, Jones, 1990, Lasiak and Underwood, 2002). Although our data does not determine the causal relationship between acidity/alkalinity of water and the benthic macroinvertebrates, Costa (1967) and Sutclifte & Carrick (1973) (cited in(Wade et al., 1989) provided laboratory evidence as to how Gammarus pulex (Phylum: Arthropoda) avoided acidic water (pH <6.2) for living and Baetis rhodani (Phylum: Arthropoda) avoided acidic reaches when laying eggs. However, from the current study it is neither clear as to how alkaline pH of the water column is associated with the benthic species nor where it originated from. Perhaps this may be due to the nature of the run-off, sediment geology and the composition of treated effluent in the peri-urban landscape. Nevertheless, water pH would be a potential indicator for river acidification and will be useful for future management decisions on protection of benthic macroinvertebrates and other commercially important aquatic species.

7.5 Concluding Remarks

The field sites monitored during this study were impacted by multiple stressors of an anthropogenic nature due to urbanisation of the catchment. The natural flow of the river was blocked at Penrith weir and for this reason the water quality at the weir

132

was considerably different to that of the Cattai and Sackville sites which were influenced by tidal inflows and pollutants entering the river system from tributaries. Yarramundi was minimally disturbed compared to the other three sites. Therefore, a peri-urban river system such as the HNR system does not agree with the characteristics of a ‘river continuum’ concept often applicable to rivers in rural landscapes. Instead, the peri-urban river system is impacted over multiple locations due to geographical, natural and anthropogenic factors.

The study indicated that downstream sites in peri-urban river systems are likely to be impacted more than upstream sites due to an accumulation of nutrients, increased sediment levels and phytoplankton related variables such as total counts, bio-volumes, Chlorophyll a levels as well as Phosphorus. Further, these impacts can be influenced by seasonal rainfall events and tidal inflows. Also, the fluctuation in river water quality tends to be influenced by the season of the year - it is fairly even during warmer months and large variations during cooler months.

There was some relationship observed between the clarity of river water and phytoplankton abundance however the phytoplankton diversity fluctuated on a seasonal basis. A seasonal succession of phytoplanktons indicated summer, autumn and winter samples were dominated by Cyanophyta, Chlorophyta and a mix of species respectively. The study identified that A. holsatica and Chironomid larvae may be useful indicators for assessing river health, both at seasonal and spatial scales, and are worth examining further for possible development of biotic indicators. It was observed that the water temperature was the most influential water quality variable for phytoplankton species compositions. However, pH appears as the most influential variable for benthic-macroinvertebrates. Overall, these findings are important for future river management strategies such as developing predictive tools to assess algal blooms, routine assessment of the river condition through biotic assemblages and management of pollutant sources that alter the acidity and alkalinity of water.

* * * * *

133

CHAPTER 8

IMPACTS OF WATER QUALITY ON AQUATIC LIFE IN

RIVER SYSTEM

Pinto, U. & Maheshwari, B. (2012). Impacts of water quality on the harvest of school prawn (Metapenaeus macleayi) in a peri-urban river system. Journal of Shellfish Research, 31, 1-7.

Summary

In this chapter, using the harvest of school prawn (Metapenaeus macleayi) will be evaluated using selected water quality and weather parameters. The school prawn is important among commercial prawn trawler operators, but its harvest is affected in a complex way by a number of interacting water quality and other variables. Using the HNR system as a case study, Pearson correlation and HACA analysis were employed to assess the influence of the selected water quality (n = 7), quantity (n = 1) and weather (n = 2) parameters on the prawn harvest. Using data records (n = 104) collected over a nine year period, it was found that water temperature (r = 0.63, p < 0.01), dissolved oxygen (r = -0.59, p < 0.01) and rainfall (r = 0.26, p < 0.01) as significantly correlated variables with the prawn harvest. The HACA produced three distinct clusters of variables - nutrient availability for prawns (the total nitrogen, the total phosphorus, reactive silicate, turbidity and suspended solids), the physical river environment (temperature, rainfall and river flow) and the biochemical river environment (dissolved oxygen and Chlorophyll a). The study revealed that two key variables (temperature and rainfall) representing the physical river environment are statistically significant to affect prawn harvest in the study area, and therefore from fishing industry point of view the future river management need to focus on strategies that will improve the physical river environment, particularly to cope with the impacts of future peri-urban developments and climate change scenarios.

8.1 General

Fishing industry based on prawns and shrimps is important in many tropical and subtropical regions of the world. In Australia, school prawn (Metapenaeus macleayi; Haswell) occurs along the east coast of Australia. It is one of the key species that

134

inhabits estuarine environments from Tin Can Bay, Queensland to Corner Inlet, Victoria (Ruello, 1973b, Glaister, 1978b). Estuarine prawn trawling is restricted to three regions in NSW viz., the Clarence, the Hunter and the Hawkesbury Rivers (Broadhurst and Kennelly, 1994, Montgomery et al., 2012). Of the numerous other aquatic species caught in the Hawkesbury River, M. macleayi has been the most popular type of seafood among Sydney fish consumers and their catch area falls mostly within the lower estuarine reaches of the HNR system (Hawkesbury Trawl Association, 2001, Howard and Howard, 2005).

Many prawn species seasonally migrate from shallow coastal waters into the deeper freshwaters to complete their life cycle. The spawning of M. macleayi mainly occurs at the estuaries, post-larval stages then move upstream where salinity is less than 20% and mature individuals return to the ocean for breeding (Ruello, 1973b, Rowling et al., 2008-2009). Based on a number of studies, Ruello (1973b) suggests that the main food source of M. macleayi includes chitinous remains of crustaceans, annelid worms, algae and diatoms. Opportunistic school prawns bury themselves slightly under the sediments and use chelipeds to transfer food particles into the mouthparts (Ruello, 1973a). The adult individuals are markedly more present in turbid coastal waters than in clean waters (Ruello, 1973a). In the Hawkesbury River, the prawn trawling is allowed in reaches downstream of Lower Portland, near the confluence of Colo River (Figure 34). The prawn catches predominantly consist of M. macleayi (40%) and Eastern King prawn (Penaeus plebejus) (2%) (Howard and Howard, 2005).

Over the last decade, the peri-urban zones of the Hawkesbury-Nepean catchment have been significantly affected by land-use changes, largely related to housing development, to accommodate the rising population. The current Metropolitan Plan for Sydney proposes 770,000 homes on top of the existing 1.68 million households to accommodate a forecast of 6 million extra residents in Sydney by 2036 (NSW Department of Planning, 2010) (see Plate 13). Commercial fishers in the region are concerned that future urbanisation will result in changes to the natural pattern of river flow, presence of excessive algae growth in upstream reaches, increased discharge of treated effluent and will have negative consequences on their fish catch (Hawkesbury Trawl Association, 2001). Although the effects of rainfall (Ruello, 1973b), ecological interactions with sediment bed (Ruello, 1973a), migration habits (Ruello, 1977), effects of river discharge (Glaister, 1978a) and mortality patterns (Montgomery et al., 2012) of school prawns have been well studied in some detail

135

there has been less attention given to understanding the influence of water quality on the prawn harvest.

Plate 13. New housing developments in Western Sydney region.

The main objective of this chapter is to examine how the water quality and weather variables influence the harvest of school prawns (M. macleayi) (hereafter simply referred to prawns) in a peri-urban river system by using Pearson correlation and HACA. A reach along the HNR system is used as a field site for this study. Considering the overwhelming land-use changes due to urbanisation in the Hawkesbury - Nepean catchment over the last three decades, this study will not only increase our present understanding of the prawn harvest, but will also help in prioritising the river management decisions to cope with future impacts of urbanisation and balancing the interests of different river users.

8.2 Materials and Methods

8.2.1 Data Collection

For the purpose of this study, prawn catch, water quality, and quantity and weather data of the study area were obtained from three government agencies viz., NSW

136

Department of Primary Industries, SCA, and Australian of Bureau of Meteorology. The prawn harvest (kg) was based on the information provided by registered anglers and prawn trawler operators to the NSW Department of Primary Industries. The date and location where the prawns were caught were the month of the year and the river reach between Lower Portland and the river mouth (Figure 34 & Figure 35). The prawn harvest weight is the sum of all reported gross landings (kg) in a month under different fishing gear used for harvest. The water quality data, Chlorophyll a (µg/L), temperature (0C) dissolved oxygen (DO), total nitrogen (mg/L), total phosphorus (mg/L), turbidity (NTU) and suspended solids (mg/L) were obtained from a routine government water quality monitoring program conducted by the SCA. Collection and analysis of this data has been in accordance with the American Public Health Association (APHA) Standard Methods for the Examination of Water and Wastewater (Eaton and Franson, 2005).

Total monthly flows (ML) recorded at Penrith weir were included in this study, because these flows have a major influence on the downstream macrophyte community, fish life and water quality (Healthy Rivers Commission, 1998). Monthly rainfall (mm) data at Richmond was selected due to the continuity in data collection. This information was obtained from the Australian Bureau of Metrology (Bureau of Meteorology, 2011).

137

Figure 34. Map of the Hawkesbury-Nepean River system indicating the main tributaries and data collection points.

PACIFIC OCEAN

PENRITH WEIR RICHMOND LOWER PORTLAND Hawkesbury-Nepean River system

Flow data Rainfall data Water quality data Prawn harvest data 0 km 83 km 160 km 140 km

Figure 35. Schematic diagram of the distribution of data collection points along the HNR.

(Distances shown are not to the scale)

138

8.2.2 Statistical Analysis

The raw data set (n = 104) includes monthly water quality, quantity and rainfall data collected between 2000 and 2009. The Pearson Correlation Coefficients among the variables were calculated using raw data records while z-scale transformed data was used for the HACA and line graphs. Squared Euclidean Distance was used as a distance measure and Ward’s method as a linkage method in HACA. The use of Squared Euclidean Distance measure has been widely adopted for multivariate surface water quality classifications as it is capable of progressively placing a greater weight on variables which are further apart (Alberto et al., 2001, Shrestha and Kazama, 2007). The actual distances among variables were further rescaled and presented as a number between 0 and 25 in the dendrogram.

8.3 Results

The descriptive statistics of the data set used in this study are shown in (Table 25) The water temperature varied from 110C to 290C depending on the season and indicated the highest significant correlation with the prawn harvest (r = 0.63, p < 0.01) (Table 26). Dissolved oxygen also indicated a significantly negative correlation with prawn harvests between 2000 and 2009 (r = -0.59, p < 0.01). The correlation between rainfall and prawn harvest was low although significant (r = 0.26, p < 0.01). All other variables indicated a considerably low correlation with the prawn harvest (Table 25). The maximum rainfall in the study area was 239 mm in March 2002.

The dendrograms indicate the variables considered in the study can be divided into three distinguishable clusters of variables based on their characteristics at a rescaled distance unit of 15 (Figure 36). The first cluster combines variables related to nutrient availability for prawns, viz. the total nitrogen, the total phosphorus, reactive silicate, turbidity and suspended solids. The second cluster relates to the physical river environment for prawns and includes three variables, viz., temperature, rainfall and river flow. The final cluster relates to the biochemical river environment as indirectly indicated by Chlorophyll a and dissolved oxygen level of the river water. It should be noted that the variable prawn harvest is part of the second cluster, i.e., the physical river environment, and suggest that temporal variation of prawn harvest over the years tend to follow the trends observed for temperature, rainfall and flow. The prawn harvest is associated with trends in water

139

temperature the value of r = 0.63 and p < 0.01, dissolved oxygen r = -0.59 and p < 0.01 and rainfall r = 0.26 and p < 0.01.

The fluctuation of significantly correlated variables, viz., water temperature and rainfall with the prawn harvest in standardised units are presented in Figure 37a and b. The prawn season of the HNR is between September and June and often an increased harvest is recorded around December. Thus, the consistent low prawn harvest readings coincide with the small number of prawn trawler operators reporting the catch data between June and August. During the prawn season, the catch weights closely follow the seasonal temperature fluctuations (Figure 37a). The dissolved oxygen is always high in colder months (i.e., July-August; see Figure 37c) and this is due to the effects of water temperature and increased oxygen levels produced by algae. Rainfall indicated at least one distinct peak during each prawn season. High rainfall events are rare during colder months when prawn trawling is not commonly occurring.

Cluster 1

Cluster 2

Cluster 3

Figure 36. Dendrogram of Hierarchical Agglomerative Cluster Analysis indicating the similarities among water quality variables, prawn harvest and rainfall using Ward linkage measure and Squad Euclidean Distances.

(The similarity measure values were proportionally rescaled to facilitate reading)

140

a. Prawn harvest weight Vs. Water temperature Prawns Temperature 3 3

2 2

1 1

0 0

-1 -1 Water Temperature Prawn harvest weight -2 -2

-3 -3 Apr-06 Apr-05 Oct-05 Apr-04 Oct-04 Apr-03 Oct-03 Apr-02 Oct-02 Apr-01 Oct-01 Oct-00 Apr-09 Apr-08 Oct-08 Apr-07 Oct-07 Oct-06 Jun-02 Jun-01 Jun-08 Jun-07 Jun-06 Jun-05 Jun-04 Jun-03 Feb-02 Feb-01 Feb-08 Feb-07 Feb-06 Feb-05 Feb-04 Feb-03 Feb-09 Dec-00 Dec-07 Dec-06 Dec-05 Dec-04 Dec-03 Dec-02 Dec-01 Dec-08 Aug-04 Aug-03 Aug-02 Aug-01 Aug-00 Aug-08 Aug-07 Aug-06 Aug-05

b. Prawn harvest weight Vs. Rainfall Prawns Rainfall-Richmond 3 4

2 3 2 1 1 0 Richmond 0 - -1 -1

-2 -2 Rainfall Prawn harvest weight -3 -3 Apr-09 Apr-08 Oct-08 Apr-07 Oct-07 Apr-06 Oct-06 Apr-05 Oct-05 Apr-04 Oct-04 Apr-03 Oct-03 Apr-02 Oct-02 Apr-01 Oct-01 Oct-00 Jun-08 Jun-07 Jun-06 Jun-05 Jun-04 Jun-03 Jun-02 Jun-01 Feb-09 Feb-08 Feb-07 Feb-06 Feb-05 Feb-04 Feb-03 Feb-02 Feb-01 Dec-08 Dec-07 Dec-06 Dec-05 Dec-04 Dec-03 Dec-02 Dec-01 Dec-00 Aug-08 Aug-07 Aug-06 Aug-05 Aug-04 Aug-03 Aug-02 Aug-01 Aug-00

c. Temperature Vs. Dissolved Oxygen

Temperature Dissolved Oxygen 3 3

2 2 1 1 0 0

Temperature -1 -1

-2 -2 Dissolved oxygen -3 -3 Apr-09 Apr-08 Oct-08 Apr-07 Oct-07 Apr-06 Oct-06 Apr-05 Oct-05 Apr-04 Oct-04 Apr-03 Oct-03 Apr-02 Oct-02 Apr-01 Oct-01 Oct-00 Jun-08 Jun-07 Jun-06 Jun-05 Jun-04 Jun-03 Jun-02 Jun-01 Feb-09 Feb-08 Feb-07 Feb-06 Feb-05 Feb-04 Feb-03 Feb-02 Feb-01 Dec-08 Dec-07 Dec-06 Dec-05 Dec-04 Dec-03 Dec-02 Dec-01 Dec-00 Aug-08 Aug-07 Aug-06 Aug-05 Aug-04 Aug-03 Aug-02 Aug-01 Aug-00

Figure 37. The pattern of variation of (a) prawn harvest and water temperature, (b) prawn harvest and rainfall and (c) temperature and dissolved oxygen.

(N.B., The units of all data are z-scale standardised, Highlighted months from Septembers-June indicate the prawn season)

141

Table 25. The descriptive statistics of the data set.

Std. Variables Min. Max. Mean Deviation Prawn, School (kg) 91 23172.2 6981.85 5796.1 Flows at Penrith Weir (ML) 880.9 150065.1 9579.62 20932.7 Rainfall-Richmond 0.4 255.6 60.21 55.8 Chlorophyll-a (µg/L) 1.3 67 20.16 11.52 Temperature (o C) 11.4 29 20.41 5.06 Nitrogen Total (mg/L) 0.1 1.56 0.52 0.23 Phosphorus Total (mg/L) 0.01 0.087 0.03 0.02 Turbidity (NTU) 3 51.1 10.63 7.23 Dissolved Oxygen (mg/L) 4.1 12.6 8.74 1.71 Suspended Solids (mg/L) 3 35 9.92 5.94

Silicate Reactive (SiO2 mg/L) 0.08 6 1.57 1.46

Table 26. Pearson correlations between prawn harvest, flow, rainfall, Chlorophyll a (Chl), Temperature (Temp), total nitrogen (TN), total Phosphorus (TP), turbidity (Turb), dissolved oxygen (DO), suspended solids (SS) and Reactive Silicates (SIL).

Prawns Flow Rainfall Chl Temp TN TP Turb DO SS Sil Prawns 1 Flow 0.03 1 Rainfall 0.26** 0.63** 1 Chl 0.11 -0.12 0.13 1 Temp 0.63** 0.08 0.35** 0.34** 1 TN -0.12 0.15 0.1 -0.07 -0.16 1 TP 0.06 0.22* 0.21* -0.03 0.15 0.61** 1 Turb 0.14 0.41** 0.25* 0.02 0.19 0.49** 0.65** 1 DO -0.59** -0.24* -0.39** 0.11 -0.77** -0.07 -0.33** -0.30** 1 SS 0.07 -0.03 0.05 0.38** 0.17 0.20* 0.48** 0.51** -0.08 1 SIL 0.08 0.29** 0.19 -0.44** -0.12 0.53** 0.48** 0.43** -0.28** 0.02 1 **Correlation is significant at the 0.01 level (two-tailed). *Correlation is significant at the 0.05 level (two-tailed).

8.4 Discussion

8.4.1 Water Temperature and Prawn Harvest

Commercially important prawn species inhabit, for a major proportion of their life cycle, estuarine regions of river systems which are often prone to temperature and salinity fluctuations that occur due to tidal and freshwater mixing (Aziz and Greenwood, 1981). In this study, temperature was observed as an important variable determining the catch of prawns in the HNR. A significant correlation was observed in the present study agrees with the previously published temperature effects on Penaeus vannamei (Pacific white shrimp) and M. bennettae (Wyban et al., 1995, Aziz and Greenwood, 1981). Based on outdoor ponds in Hawaii, water

142

temperature was observed to be significantly correlated with the growth of P. vannamei suggesting that shrimps less than 10g are more likely to grow faster in waters up to 300C (Wyban et al., 1995). Leung and Hochman Lawrence (1990) incorporated this relationship into an economic model for aquaculture farmers. Based on laboratory experiments, Aziz and Greenwood (1981) reported the survivability of juvenile M. bennettae in a wide temperature range of 8.10C to 32.90C and how their lower and upper lethal temperature levels could be altered by exposing them to varying acclimation temperatures. The current climatic predictions for the Western Sydney region implies a rise in the atmospheric temperature (170C- 290C) by 1.60C by 2030 and 4.80C by 2070 (C.S.I.R.O., 2007). If the water temperature increases in future, as being predicted, the present study indicates that the climate change may impact on the prawn harvest while prawns adapt to new temperature regimes and cope with changes in DO levels and other factors. As such, the future change in temperature regime will be an important management consideration for many prawn trawler operators to sustain their livelihood.

Due to a high correlation between the prawn harvest and water temperature as shown in this study, it may be possible to develop a rapid assessment tool to predict prawn harvest using water temperature and some easy to measure water quality variables. The availability of such a tool could help the prawn trawler operators in the region to plan and manage their business more objectively. At present, some prawn trawler operators use the clarity of water to obtain some indication of their prawn harvest (Howard, M 2011, pers. comm.) but it is often too subjective.

8.4.2 Dissolved Oxygen and Prawn Harvest

The DO is the next important variable after water temperature, and it particularly plays a key role in the life of prawns. However, the value of DO is affected by temperature, shade, nutrient richness, turbulence, decomposition of large volumes of algal blooms and organic matter (Kramer, 1987, Davis, 1975). A significantly negative correlation (r = 0.77, p < 0.01) was observed between the prawn harvest and DO levels. Further, DO is affected by water temperature and Chlorophyll a bearing phytoplanktons as evident from cluster 3 (Figure 36). The prawn harvest season is usually in the warmer months where the river system is naturally low in dissolved oxygen levels, and DO could be further reduced when phytoplankton biomasses are present in abundance either in live or dead form. This makes the interaction of the variables affecting prawn harvest along with DO quite complex.

143

Pinto and Maheshwari (2011) reported on anaerobic fermentation occurring in the lower reach of the HNR system which led to the depletion of the available oxygen for fish life, mostly due to peri-urban development occurring in the region. The lowest value recorded for dissolved oxygen in the data set is 4 mg/L, which is well above the lethal level reported for most Penaeids prawn species (0.2-1.0 mg/L) (Allan and Maguire, 1991). This indicates that there will be a considerable reduction in prawn catches in downstream reaches if the phytoplankton blooms prevails in the upstream reaches and organic matter continues to accumulate in the lower reaches of the river system.

8.4.3 Prawn Harvest and other Variables

A low but statistically significant positive correlation between prawn harvest and rainfall (r = 0.26, p < 0.01) was observed. The prawn season of the HNR is usually the wet season of the year for this region and so an increase in river flows due to rainfall events is likely to be beneficial in terms of prawn harvest. Although not significant, the HACA suggested that the patterns of variation of river flow and rainfall tend to be similar to prawn catches (Figure 37b). This observation is consistent with some published literature relating to hydrological regimes and their influence on the growth of prawns and other aquatic living beings (Hildebrand and Gunter, 1953, Subrahmanyam, 1964, Glaister, 1978a, Gillson, 2011). Heavy rainfall events often carry large amounts of plant detritus, sediment and organic food scraps that immensely benefit the survival and growth of juvenile M. macleayi. Hildebrand and Gunter (1953) observed how the fluctuations in Penaeus setiferus is related to the salinity levels influenced by rainfall, while Subrahmanyam (1964) showed how the larval recruitment of Penaeus monodon is stimulated by rainfall in the Godavari estuarine although the reason for the observed pattern was not stipulated.

In Australia, significant emphasis has also been given to studies relating to the impacts of river flow on the downstream fisheries productivity in coastal systems. A study conducted by the Logan River system in Queensland indicated how the increased seasonal pattern of river flow discharges positively influences the production of commercial and recreational coastal fisheries (Gillson, 2011, Loneragan, 1999). In particular, Glaister (1978a), reported a direct relationship between river discharge and the adult M. macleayi caught in the Clarence River, NSW. The present study suggests the catch of M. macleayi is affected by upstream flows regardless of their origin (i.e., rainfall or flow releases) in the Hawkesbury River. Thus, this is an important finding for prawn trawler operators as the river

144

system is often subject to variable flow regimes by river management authorities due to heavy rainfall events (Ball and Keane, 2006, Krogh et al., 2008). It is also evident from this study that natural rainfall events are more beneficial yielding an increased harvest compared to the human controlled flow events.

The nutrient dynamics in the diet of M. macleayi in Hawkesbury River has not been previously reported. A considerably low correlation between the total phosphorus and prawn harvest (r = 0.06) was noted. The low value of the correlation may be due to the feeding habits and nutrient requirements of prawns and shrimps and somewhat different to the findings of previous studies (Kitabayashi et al., (1971); Ambasankar et al., (2011). However, further research is warranted to verify this fact and assess the degree of beneficial effects of TP for the phytoplankton and fish communities.

8.4.5 Implications for Prawn Industry

The study revealed that two key variables (temperature and rainfall) representing the physical river environment are statistically significant to affect prawn harvest in the study area. It is now increasingly believed that rapid urbanisation is causing warming of urban areas along with global warming (Kataoka et al., 2009). This is because urbanisation results in substantial changes to the land surfaces, especially increase in areas with concrete, roads and tiled roofs. These changes perturb the local environment through changes in heat absorption, storage and radiation and thus leading to increase in local temperature. This warming is referred to ‘urban heat island’ (UHI) effect. Therefore, from fishing industry point of view, the future river management need to focus on strategies that will improve the physical river environment, particularly the aspects that affect temperature and rainfall through future peri-urban development and climate change.

The increased water demand for drinking and industrial uses and production of large volumes of nutrient rich surface run off and domestic effluent from peri-urban landscape are also important for river management, especially for the protection of aquatic life. The study highlighted that prawn harvest is impacted in a complex manner by combination of interrelated variables and factors (including the possibility of climate change) which need careful attention and in-depth understanding to manage the river system in the long-run to sustain fishing industry.

145

It is important to note that the results of this study could have been improved by incorporating the confounded flow effects, tidal inflows at downstream reaches, lunar periodicity and inter-species competition that may possibly influence the prawn harvest data. However, the determination of the river flow at lower reaches of the HNR is an extremely difficult task as the river system receives variable inflows (considerably impacted by sewage treatment plants) from tributaries and tidal inflows. While practically it is difficult to conduct large fisheries surveys, this data (even at monthly intervals) provided sufficient evidence to understand relationships between prawn harvest and water quality and weather variables and design appropriate river management strategies.

8.5 Concluding Remarks

In this study, the effects of water quality and quantity and weather variables on the harvest of M. macleayi in a peri-urban river system were analysed. The analysis indicated that water temperature, dissolved oxygen and rainfall were the variables that were significantly correlated with prawn harvest. Further, the analysis indicated that the variables considered can be grouped into three distinct clusters – nutrient availability for prawns (the total nitrogen, the total phosphorus, reactive silicate, turbidity and suspended solids), the physical river environment (temperature, rainfall and river flow) and the biochemical river environment (dissolved oxygen and Chlorophyll a). The study revealed that two key variables (temperature and rainfall) representing the physical river environment are statistically significant to affect prawn harvest, and therefore from fishing industry point of view the future river management need to focus on strategies that will improve the physical river environment, particularly the aspects of future peri-urban development and climate change.

* * * * *

146

CHAPTER 9

KEY INDICATORS FOR RIVER HEALTH

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

Summary

An array of river health assessment approaches and water quality variables have been suggested in the past for assessing the level of river health. However, the selection of suitable variables to be monitored for the assessment remains ambiguous and often it is not practical to monitor all the suggested variables. In this study, a multivariate data reduction technique, called FA was employed, to identify the key river health variables for a peri-urban river system, viz., the HNR system in NSW, Australia. Out of 40 water quality variables included in the analysis, the FA identified nine key variables, under three varifactors (VFs), explaining 50% of the variance in the river water quality. Variables in the first, second and third VFs revealed anaerobic conditions, microbial quality and effects of eutrophication in the HNR. Thus, the present work shows a notable reduction in the number of variables and the application of FA for identification of key variables was found promising. The finding of this study has potential application in designing a cost-effective river health-monitoring program by reducing the number of variables to be monitored in a peri-urban situation. It can also assist in partitioning variables according to their unique contribution to the total variance......

9.1 Issues in River Health Assessment

Major problems associated with river health assessment approaches are related to the rationale for variable selection and our limited understanding of the environmental complexity. In the past, recommendations from a panel of experts who interact remotely and anonymously provide feedback, a method commonly known as ‘DELPHI’ forecasting was considered as the best way to select variables to be included in water quality indices (Dalkey and Helmer, 1963). This has historically gained much acceptance although it suffers from lack of human understanding of complex interactions of ecosystems. Another river health

147

assessment approach considers a hierarchical model, which combines catchment, habitat and biota, assuming that the components at higher levels of hierarchy affect components at lower levels (i.e., catchment health affects river health). In other words, this methodology chooses variables on the assumption that ecological integrity is represented by all the major components of the environment that comprise an ecosystem (Norris et al., 2007). Thus, a package of representative variables on catchment disturbance, hydrological changes, water quality, soil, physical form, fringing zone and aquatic biota are included to assess the overall health of river. Similarly, some researchers proposed to measure riparian habitat conditions and argue that habitats close to the river waters have a stronger relationship between their physical and biological composition and the river health (Jansen and Robertson, 2001). However, the above argument proves the lack of coherence in variable selection procedures for the purpose of reliable river health assessment.

Secondly, it is becoming extremely difficult to find sites that are ‘pristine’ for use as reference sites to establish indicators and subsequently develop composite indices. The river health assessment approaches which are heavily dependent on condition of reference sites (Observed versus Expected commonly known as O/E ratio) have a tendency to misinterpret the true state of rivers. In Australia, many composite river health approaches such as AUSRIVAS (National River Health Assessment Program), FARWH (National Framework for the Assessment of River and Wetland Health), IRC (Tasmanian Index of River Condition) and SRA (Sustainable River Audit, systematic river health assessment for the Murray Darling Basin) consider the state of pristine condition or pre-European reference conditions to assess the health of river systems at state and national levels (Barmuta et al., 2002, Norris et al., 2007, Askey-Doran et al., 2009, Peter et al., 2008). However, if the reference sites are already impacted, the river health assessment becomes erroneous. Similarly, selecting reference sites which are ‘minimally-disturbed’ may become difficult due to large environmental variability inherited in individual sites (Underwood, 1994).

Thirdly, there are complex interactions occurring in natural environment, and for this reason incorporating ecological interactions into river health assessment approaches becomes difficult. For example, we still do not know the full extent of response thresholds to a given stressor by ecological functions or biological organisms (i.e., leaf-litter decomposition is increased by nutrient enrichment but decreased by acidic pH water) (Young et al., 2004). As such, the interpretation of

148

ecological function based methodologies has become cumbersome due to multiple factors affecting a chosen ecological function. Similarly, the distribution and abundance of species and communities have been widely used to assess river health in response to two distinct stressor levels. However, distribution and abundance of biological species are concurrently affected by predators and internal community dynamics. Thus, it is extremely difficult to predict whether the change in species distribution and abundance is due to an external impact or community dynamics. Past studies indicated that diversity remains highest at the intermediate levels of disturbance and decreases at higher levels of disturbance due to stress (Huston, 1979, Connell, 1974). Therefore, methods based on community interactions and ecological functions require an in-depth understanding of the community dynamics before a meaningful assessment of river system health is made.

9.2 Identifying Key Variables

In order to assess river health comprehensively, there is a need to establish a clear rationale to identify key variables of river health. There are numerous problems associated with specific water quality variables monitored for river health assessment approaches as well as variables selected for routine assessment of river condition. Often, ongoing river monitoring schemes yield a large volume of data that is expensive in terms of collection and storage. The intensity of monitoring is also questioned when budget cuts occur. Clearly, there is a need to make monitoring more effective in terms of cost and effort involved. Considering that river health is analogous to human health (Fairweather, 1999b, Schofield and Davies, 1996), in this study an attempt is made to understand the key variables of peri- urban river systems to include in river health assessment similar to a physician assessing patient health based on major symptoms. Identifying key variables of river health has many advantages. Firstly, with a small number of representative variables, government agencies could significantly reduce their monitoring cost. Secondly, key variables can be examined for a specific river management purpose such as developing a risk assessment framework for recreation activities or identifying an emerging problem in the river health. Therefore, using the HNR system in NSW, Australia as a case study, the main objective of this study is to employ a multivariate approach to identify key variables that have a significant relationship with overall changes in river water quality in peri-urban contexts.

149

9.3 Materials and Methods

9.3.1 Application of Multivariate Techniques

In this study, a multivariate approach was employed because multivariate tools can handle interacting water quality variables on a spatio-temporal scale and guide in developing management strategies for complex water resources within a region (Pejaman et al., 2009, Vega et al., 1998). Based on matrix algebra, FA has been selected because it can handle large sets of data points often present in routine water quality monitoring programs. The analysis can help to reduce the dimensionality of the data set to a manageable size and keeps as much of the original information as possible. In water research, FA has been employed to link marine community interaction with environmental variables to assess seasonal and temporal variation in surface water bodies and to understand groundwater interactions in multiple locations (Clarke and Ainsworth, 1993, Shrestha and Kazama, 2007, Winter et al., 2000).

For the application of FA, it is a necessary requirement to remove the highly correlated parameters prior to the analysis because singularity (variables that are highly correlated) prevents the determination of unique contribution of a particular variable to a factor (Fields, 2009). This is a crucial step because in reality agencies measure a vast range of water quality parameters that are similar in nature. Avoiding the overlapping variables or variables that explain the same dimension of water quality greatly saves a monitoring budget, especially when there are limited funds available. By extracting the most useful groups of variables, FA helps to understand the structure of a data set in terms of latent factors (i.e., the factors that are not directly measured but reflect the interactions of several individual variables) and sources of pollution unique to a chosen river system (Fields, 2009). In the present study, FA was employed to considerably reduce the number of variables obtained in a routine monitoring program in the HNR and identify the latent factors relative to river health in peri-urban landscapes.

9.3.2 Data Analysis

Forty parameters measured on monthly basis for four adjacent sites (Wiseman’s Ferry, Lower Portland, Wilberforce, and North Richmond) on the Hawkesbury River, part of the HNR was obtained (Figure 38). The data supplied by the SCA, were for two consecutive years viz., 2008 and 2009. The Wiseman’s Ferry and Lower

150

Portland sites mostly contained brackish water. These sites were selected due to the continuity in the range of water quality parameters collected between 2008 and 2009. Data reported as below detection level (

All variables were assessed for multivariate normality using the Anderson-Darling test and Draftsman’s plots. Necessary variables were log, square root and reciprocal square root transformed to obtain normality. Variables that were highly correlated with each other and variables, which could not be transformed to normality, were removed from the list. All variables were z-scale standardised prior to the analysis. Factors were extracted using the principal component method and rotated using varimax rotation on a correlation matrix. During the analysis, factors indicating Eigenvalues ≥ 1 were retained. All absolute factorloadings <0.5 were suppressed for ease of interpretation. Microsoft ExcelTM 2003, MinitabTM 15 and SPSSTM 18 were used for statistical computations throughout the analysis.

Figure 38. Map with key locations on the Hawkesbury-Nepean River system

( = sampling location).

151

9.4 Results and Discussion

9.4.1 Variability in Water Quality Data

Table 27 and Figure 39 show the highly correlated variables observed in the initial data set. The descriptive statistics of the monitored data for 2008 and 2009 are summarised in Table 28 and Table 29. The pre-test results of FA are shown in Table 30. Table 31 and Table 32 show the results of total variance explained by the varifactors and rotated factor loading on individual variables respectively. The determinant value for the variables used in this study was zero (Table 30). In FA, existence of extreme multi-collinearity (many correlated variables) and singularity (perfectly correlated variables) causes problems. When there are too many variables in the original data, which are correlated, it becomes difficult to determine their unique contribution to a particular factor. Past studies suggest to keep the determinant value <0.00001 to avoid such difficulties (Fields, 2009). The KMO test statistic for sampling adequacy was 0.692 and Bartlett’s test of sphericity was significant (p < 0.05) (Table 30). The results of pre-tests indicate that correlations of the data were relatively compact to yield distinct and reliable factors and the R- matrix was not an identity matrix. Hence, FA was appropriate for the data. The input data matrix contained 16 variables for four sites and therefore the input data matrix for FA comprised of 16 variables and 104 cases for 2008 and 2009.

Five PCs which had Eigenvalues >1 were retained, summarizing almost 74% of the total variance in the water quality data (Table 31). The first PC accounted for 17% of the total variance while the second and the third PCs accounted for similar variances of 16%. Altogether, first three components reflected 50% of the total variance in the entire data set. Corresponding VFs and factor loadings are described in Table 32. Only factor loading above 0.5 have been considered for interpretation as they have a moderate to high effect on the relevant factor (Zhou et al., 2007).

152

Al Total

1 Fe Total 0.160

0.08 Mn Total 0.0016

8 Algal BioVol 80000

4000 ASU Cyano 0 2000 1000 ASU Toxic 0 10000

5000 ASU Algae 0 300000

150000 Algal TC 0 300000

150000 Cyano TC 0 2 1 Toxic Cyano BioVol 0 2000 1000 E. coli 0

0 1 2 0 8 6 0 8 6 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 2 0.0 0.5 1.0 .0 . 0 .1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 40 80 10 20 50 00 00 00 00 00 1 15 30 15 30

Figure 39. A sample of plots showing the extent correlation among the selected variables.

(Refer to Table 28 and 29 for the units of the variable on X and Y - axis.)

153

Table 27. Highly correlated variables in the study.

Variable 1 Variable 2 Pearson Corr. coefficient Al Total NTU 0.83 Al Total Fe Total 0.91 Mn Total Fe Total 0.72 Algal BioVol ASU algae 0.92 ASU Cyano Algal BioVol 0.80 Coli. Therm E.coli 0.92 Coli TC E.coli 0.79 Algal TC Cyano TC 0.99 Cyano BioVol ASU Cyano 0.97 Cyano BioVol Toxic Cyano BioVol 0.92 Toxic Cyano BioVol ASU Toxic 0.99 Cyano BioVol ASU Toxic 0.91 N-TN NOx 0.82 Loren Chl-a 0.99 UV NTU 0.84 UV Colour 0.93 SS NTU 0.71

154

Table 28. Statistical descriptive of original water quality variables (2008).

(N-Number of data points, STD-Standard Deviation)

Descriptive Statistics-2008 Variable and Code N Min Max Mean STD Statistic Std. Error Algal Total Count (cells/mL)-'Algal TC' 46 770 337954 33944.54 8582.82 58211.51 Algal biovolume (mm3/L)-'Algal BioVol' 46 0.214 7.146 1.68 0.21 1.44 Areal Standard Unit (Cyano)-'ASU Cyano' 46 0 3275 296.73 85.76 581.67 Areal Standard Unit (Toxic)-'ASU-Toxic' 46 0 1248 119.20 38.80 263.14 Areal Standard Unit (algae)-'ASU Algae' 46 144 5523 1552.76 186.05 1261.87 Chlorophyll-a (ug/L)-'Chl-a' 52 2.1 59 15.94 1.73 12.45 Cyanobacteria Total Count (cells/mL)-Cyano TC' 46 0 325699 27525.36 8272.59 56107.44 Cyanobacterial biovolume (mm3/L) 46 0 1.812 0.19 0.05 0.37 Lorenzen (ug/L)-'Loren' 52 0.5 53 14.36 1.61 11.61 Phaeophytin (ug/L)-'Phaeo' 52 0.2 9.1 1.80 0.25 1.83 Toxic Cyanobacterial Count (cells/mL)-'Toxic Cyano TC' 46 0 35970 2385.86 948.30 6431.67 Toxic Cyanobacterial biovolume (mm3/L)-'Toxic Cyano BioVol' 46 0 1.008 0.12 0.04 0.26 Clostridium perfringens (CFU/100mL) 52 1 180 11.31 4.17 30.05 Coliforms Thermotolerant (cfu/100mL)-'Coli Therm' 52 1 3000 91.96 58.58 422.41 Coliforms Total (cfu/100mL)-'Coli TC' 52 46 98000 5739.37 2575.59 18572.87 E. coli (orgs/100mL)-'Ecoli' 52 1 2400 91.48 49.95 360.21 Enterococci (cfu/100ml)-'Entero' 52 1 1200 80.31 32.45 234.00 Manganese Filtered (mg/L)-"MnFilt' 52 0.001 0.123 0.02 0.00 0.03 Manganese Total (mg/L)-'MnTot' 52 0.003 0.144 0.06 0.00 0.03 Aluminium Filtered (mg/L)-'AlFilt' 52 0.01 0.2 0.02 0.00 0.03 Aluminium Total (mg/L)-'AlTot' 52 0.01 1.17 0.26 0.03 0.23 Iron Filtered (mg/L)-'FeFilt' 52 0.05 0.52 0.13 0.01 0.10 Iron Total (mg/L)-'FeTot' 52 0.05 2.23 0.62 0.06 0.45 Dissolved Organic Carbon (mg/L)-'DOC' 52 2 11 5.06 0.28 1.99 UV Absorbing constituents (organic)-'UV' 52 0.08 0.78 0.19 0.02 0.13 Nitrogen Ammoniacal (mg/L)-'N-Amm' 52 0.01 0.087 0.02 0.00 0.02 Nitrogen Oxidised (mg/L)-'NOx' 52 0.01 1.03 0.33 0.04 0.26 Nitrogen TKN (mg/L)-'NTKN' 52 0.1 0.7 0.28 0.02 0.13 Nitrogen Total (mg/L)-'NTN' 52 0.3 1.3 0.61 0.04 0.26 Phosphorus Filterable (mg/L)-'PFilt' 52 0.005 0.037 0.01 0.00 0.01 Phosphorus Total (mg/L)-'TP' 52 0.01 0.15 0.02 0.00 0.02 Silicate Reactive (SiO2 mg/L)-'Sil' 52 0.1 5.87 2.69 0.24 1.72 Alkalinity (mgCaCO3/L)-'Alk' 52 18 59 34.02 1.51 10.91 Conductivity Field (mS/cm)-'EC' 52 0.178 17.38 2.06 0.63 4.53 Dissolved Oxygen (%Sat)-'DO' 52 51.88 126.55 89.12 2.12 15.26 Suspended Solids (mg/L)-'SS' 52 1 41 9.58 1.07 7.70 Temperature (Deg C)-'Temp' 52 10.2 26.7 19.60 0.70 5.02 True Colour at 400nm-'Col' 52 5 50 15.65 1.35 9.70 Turbidity Lab/Field (NTU)-'NTU' 52 1.7 89.1 12.44 1.99 14.37 pH (Lab/Field)-'pH' 52 6.62 9.03 7.46 0.06 0.43

155

Table 29. Statistical descriptive of original water quality variables (2009).

(N-Number of data points, STD-Standard Deviation)

Descriptive Statistics 2009 Variable and Code N Min Max Mean STD Statistic Std. Error Algal Total Count (cells/mL)-'Algal TC' 41 643 233506 52845.10 9180.42 58783.38 Algal biovolume (mm3/L)-'Algal BioVol' 41 0.231 15.07 1.70 0.36 2.33 Areal Standard Unit (Cyano)-'ASU Cyano' 41 0 7158 555.46 182.66 1169.60 Areal Standard Unit (Toxic)-'ASU-Toxic' 41 0 2502 179.30 182.66 419.86 Areal Standard Unit (algae)-'ASU Algae' 41 217 10650 1619.45 264.95 1696.50 Chlorophyll-a (ug/L)-'Chl-a' 52 2.2 45.4 14.02 1.44 10.41 Cyanobacteria Total Count (cells/mL)-Cyano TC' 41 0 229567 47171.90 9003.68 57651.70 Cyanobacterial biovolume (mm3/L) 41 0 8.227 0.48 0.21 1.31 Lorenzen (ug/L)-'Loren' 52 1.8 41.5 12.58 1.34 9.64 Phaeophytin (ug/L)-'Phaeo' 52 0.2 8.2 1.75 0.22 1.59 Toxic Cyanobacterial Count (cells/mL)-'Toxic Cyano TC' 41 0 36222 2602.79 933.77 5979.08 Toxic Cyanobacterial biovolume (mm3/L)-'Toxic Cyano BioVol' 41 0 2.653 0.20 0.07 0.45 Clostridium perfringens (CFU/100mL) 52 1 340 15.60 6.72 48.46 Coliforms Thermotolerant (cfu/100mL)-'Coli Therm' 52 1 990 48.81 20.67 149.08 Coliforms Total (cfu/100mL)-'Coli TC' 52 70 24000 2230.08 491.92 3547.27 E. coli (orgs/100mL)-'Ecoli' 52 1 2000 68.12 38.78 279.66 Enterococci (cfu/100ml)-'Entero' 52 1 1600 81.65 32.85 236.86 Manganese Filtered (mg/L)-"MnFilt' 52 0.001 0.054 0.01 0.00 0.01 Manganese Total (mg/L)-'MnTot' 52 0.011 0.142 0.05 0.00 0.02 Aluminium Filtered (mg/L)-'AlFilt' 52 0.01 0.08 0.02 0.00 0.01 Aluminium Total (mg/L)-'AlTot' 52 0.01 0.9 0.19 0.03 0.19 Iron Filtered (mg/L)-'FeFilt' 52 0.05 0.36 0.10 0.01 0.07 Iron Total (mg/L)-'FeTot' 52 0.05 1.6 0.43 0.04 0.32 Dissolved Organic Carbon (mg/L)-'DOC' 52 2 19 4.28 0.35 2.54 UV Absorbing constituents (organic)-'UV' 52 0.06 0.3 0.14 0.01 0.07 Nitrogen Ammoniacal (mg/L)-'N-Amm' 52 0.005 0.088 0.01 0.00 0.02 Nitrogen Oxidised (mg/L)-'NOx' 52 0.002 0.85 0.20 0.03 0.21 Nitrogen TKN (mg/L)-'NTKN' 52 0.05 0.92 0.33 0.02 0.15 Nitrogen Total (mg/L)-'NTN' 52 0.27 1.24 0.53 0.03 0.21 Phosphorus Filterable (mg/L)-'PFilt' 52 0.001 0.037 0.00 0.00 0.01 Phosphorus Total (mg/L)-'TP' 52 0.005 0.122 0.02 0.00 0.02 Silicate Reactive (SiO2 mg/L)-'Sil' 52 0.05 4.81 1.47 0.19 1.36 Alkalinity (mgCaCO3/L)-'Alk' 52 7 71 40.42 1.96 14.15 Conductivity Field (mS/cm)-'EC' 52 0.016 17.9 1.78 0.50 3.62 Dissolved Oxygen (%Sat)-'DO' 52 61.9 120.84 90.44 1.85 13.34 Suspended Solids (mg/L)-'SS' 52 1 25 8.22 0.81 5.83 Temperature (Deg C)-'Temp' 52 11.1 30.4 21.05 0.77 5.55 True Colour at 400nm-'Col' 52 4 26 10.82 0.75 5.44 Turbidity Lab/Field (NTU)-'NTU' 52 1.23 24.8 9.17 0.88 6.32 pH (Lab/Field)-'pH' 52 7.12 9.2 7.64 0.05 0.35

Table 30. Results of KMO and Bartlett's tests.

Kaiser-Meyer-Olkin Measure of Sampling Adequacy 0.692 Bartlett's Approx. Chi-Square 838.834 Test of df 120 Sphericity Sig. 0.000

156

Table 31. Total variance explained by the varifactors.

Rotation Sums of Squared Loadings Component Total % of Variance Cumulative % 1 2.80 17.49 17.49 2 2.68 16.77 34.27 3 2.68 16.73 51.00 4 2.00 12.48 63.47 5 1.67 10.44 73.91

Table 32. Rotated Component Matrix.

Variable Factors

VF1 VF2 VF3 VF4 VF5 pH -0.863 NTU 0.766 DO -0.699 DOC Entero 0.86 E.coli 0.734 NTKN Chl-a 0.843 AlgalBioVol 0.824 MnFilt -0.642 Phaeo 0.536 Temp -0.867 NOx 0.82 Alk 0.878 EC -0.692 Sil -0.504

9.4.2 Identification of Water quality factors

Water, biota and river geography are the ultimate endpoints of human induced pollution. It was assumed that the water chemistry is fundamental to river health assessments as it has multiple stressors embedded into it, such as the composition and quality of catchment and atmospheric and human influence (Markich and Brown, 1998). As such, subtle changes in water quality can be identified quickly and efficiently even before they appear in a biological community. Water chemistry being the basis for river health and using the human health analogy it was attempted to understand the health of HNR system through a few key variables. For this, a popular dimension reduction multivariate approach FA was employed (Fields, 2009). This tool has previously been used for many purposes but its potential for identifying

157

overall river condition based on a few variables for river health assessment purposes has not been attempted in detail.

The first VF, the most important of all five VFs, contained three variables with both positive and negative factor loadings (pH -0.863, NTU 0.766, DO -0.699). The values of pH and DO variables were negatively correlated with VF1, while NTU was positively correlated. Past studies suggested that the factors containing pH and DO with negative factor loadings show the presence of anaerobic conditions in the river (Shrestha and Kazama, 2007, Vega et al., 1998). Shrestha and Kazama (2007) obtained Biological Oxygen Demand, Chemical Oxygen Demand and Ammonia with pH and DO while Vega et al. (1998) obtained discharge with pH and DO. Nevertheless, they both agree that this factor is a clear indication of organic matter entering the waterway producing ammonia and other organic acids under anaerobic fermentation, resulting in low pH of water. In the present study, neither discharge nor BOD was used as an input variable in the analysis. However, turbidity, which is some way related to organic pollutant loads, is grouped in the first factor. Thus, the first factor gives an indication of anaerobic conditions in the river. These results further explain the anaerobic fermentation, once known as a common problem for many urban rivers, is now appearing in peri-urban river system. At present, these three variables are listed as the key variables to be measured in regional streams in the community water quality monitoring networks in Australia. These networks encourage ordinary citizens to become involved and be active in the protection and management of their waterways and catchments (Waterwatch, 2004). The present analysis further confirms the importance of the three variables to obtain a quick snapshot of the river water quality.

The second VF indicates the microbiological quality of the river water. It consists of Enterococci and E. coli variables with strong positive loadings (0.860 and 0.734) on the factor. The microbial species belonging to genus Enterococcus are gram positive, facultative anaerobes (ability to survive with or without the presence of Oxygen) mostly present in the human faeces with two dominant species, viz., E. faecalis and E. faecium (Murray, 1990). On the other hand, E. coli is a gram negative bacteria found in the intestines of warm blooded animals and has been historically used as an indicator organism of faecal contamination of water. It is now accepted the ability of Enterococci spp. to act as a more stable and reliable indicator than E.coli or faecal coliforms in brackish water (Jin et al., 2004). The EU Water Framework Directive still insists using E.coli and Enterococci species as the most

158

robust indicator species to assess bathing water (European Environment Agency, 2009) In the present study, Enterococci have a higher correlation with the VF2 while E.coli have a slightly low correlation with VF2. Two previous studies include faecal and total coliform counts in FA while analysing water quality data obtained from Suquia River (Argentina) and Fuji River (Japan) (Alberto et al., 2001, Shrestha and Kazama, 2007). The former found coliforms as important under third VF, while the latter did not obtain coliforms as an important variable for water quality variance. Peri-urban rivers are particularly prone to microbiological contamination due to activities of large scale farms operating in the riparian lands. However, microbiological contamination is less or confounded by other highly variable water quality parameters in urban river systems draining from sewered areas. Results of the present study agree with work by Alberto et al., (2001). The original data set contained five types of microbiological indicators, viz., Clostridium perfringens, Clostridium thermo-tolerant, Coliform total, E.coli and Enterococci spp. The multivariate analysis carried out in the present study helped to reduce the five microbiological variables to two microbiological variables that explained 16% of the total variance in peri-urban river system. This factor further indicated the importance of Enterococci spp. as a potential indicator for river health assessment in brackish water sections of this peri-urban river system.

The third VF contains two Chlorophyll related variables (Chlorophyll a and Phaeophytin), one algal related (algal bio-volume) variable and one plant micronutrient (Manganese). This factor can be viewed as representing the eutrophication effects of the river system. Chlorophyll a is a pigment found in all plant cells, including eukaryotic algae and prokaryotic blue-green algae cyanobacteria (Carlson and Simpson, 1996). When Chlorophyll is degraded due to light, it could occur through either losing the centre Magnesium ion or the phytol tail. The former produces the pigment Phaeophytin, while the latter produces a molecule called chlorophyllide (Yentsch, 1965). Eutrophication has recently become significant threat to nutrient equilibrium in peri-urban river systems due to increased agricultural activities and discharge of municipal water from treatment facilities. Based on the peri-urban section of Yangtze River system China, (Zhang et al. (2007)) reported higher levels of nitrogen and phosphorus levels in surface river waters than in surface water reported from other districts of China. On the other hand, algal bio volumes are now being used to quantitatively measure the volume of algae cells in a sample to determine the risks associated with mixed species of toxin producing algae species for recreational activities in Australia (ACT Health, 2010).

159

Thus, the amount of Chlorophyll a, Phaeophytin and algal bio-volumes can be viewed as suitable indicators of eutrophication and nuisance algal blooms in recreational waters of peri-urban rivers. The presence of Manganese in this factor is indicative of its association with the growth of macrophytes in the river waters as a micronutrient. It does not seem related to the effects of eutrophication or algal blooms.

9.4.3 Key Water Quality Variables for Monitoring

During the analysis, three groups of variables were obtained explaining 50% of variance in the river quality. By measuring nine variables listed under three factors, 50% of the variance in the river water quality was understood. Therefore, depending on the time and resource availability, river management authorities can first focus on variables described in the first, second and third VFs to understand the half of the variance occurring in the river system. Variables grouped under fourth and fifth VFs can then be measured to account for up to 74% of the total variance. Nevertheless, in descending order of importance, the variables identified in the first, second and third varifactors were related to both apparent and latent dimensions (i.e., anaerobic condition, microbial quality and eutrophication) of river water quality.

The foregoing discussion has indicated the possibility of identifying the most appropriate water quality variables for routine monitoring and inclusion in river health assessment approaches in peri-urban situation, where the effects of urbanisation is a dominating factor. By selecting key variables for monitoring, it is possible to greatly reduce the data collection cost over the long term. The FA is also a valuable tool in isolating key variables to be used in a composite river health index. By doing this, the whole approach expects to avoid the use of reference sites, provision of equal degree of importance to each environmental component in the river system and use the view of an expert panel to assess river health.

In general, the present study has indicated that river health in peri-urban landscapes is prone to higher degrees of degradation. River health is usually assessed through regular monitoring of water quality and specific river health approaches. Although water quality variables have a significant influence on river health, there are other aspects in river ecosystems that also play a role in determining the health. As a result, the latter has gained much acceptance in assessing river health compared to routine monitoring in recent years. However, there are a number of discrepancies in either method in terms of variable selection and application of river assessment to

160

an aquatic system. For example, ecological function based river health assessment approaches suffer from a lack of human understanding on complex ecological interactions when interpreting the results while the rationale for selecting key variables remains ambiguous for other river health assessment approaches. Furthermore, we are losing sites which were once considered ‘pristine’ and as a result the accuracy and reliability of river health assessment approaches based on reference site conditions are subject to a higher degree of scrutiny.

A number of restrictions of this approach and areas for future research should be mentioned. It was assumed that the observed parameters important for majority of the variances occurring in the HNR system in this study are subjected to a certain degree of spatio-temporal variation and require a standardisation with long-term data. There is also some uncertainty associated with the ecological relationships between chosen water quality parameters and biotic communities. As such, further research is warranted to unearth the biotic interactions with selected variables.

9.5 Concluding Remarks

In this study, a bottom-to-top approach was followed for understanding the key variables and factors of river health conditions assuming that river waters are fundamental to the health of an entire river system. The FA approach employed in the present study clearly identified three groups of water quality variables that explains majority of the river health conditions using a few key variables that can easily be monitored. Anaerobic fermentation, microbial pollution and eutrophication are three key environmental problems faced by peri-urban rivers. Variables in the first, second and third VFs revealed anaerobic conditions, microbial quality and effects of eutrophication in the HNR system. Thus, the present work found nine variables are sufficient to explain up to 50% of the river variance. The application of FA for identification of key variables is promising. Its major advantages are considerable cost reduction in river health monitoring and assessment programs. The FA approach can provide guidance in variable selection for river health assessment in a peri-urban context and helping in the partition of variables according to their unique contribution to the total variance.

* * * * *

161

CHAPTER 10

DEVELOPING RIVER HEALTH ASSESSMENT

TOOLS

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

Summary

The methodology currently available to river managers for assessment of river conditions for eutrophication and microbial risks is often time consuming and costly. There is a need for efficient predictive tools based on easily measured variables for implementing appropriate management strategies and providing advice to local river users on river health and associated risks. In this chapter the HNR system as case study, a stepwise discriminant function analysis was employed to develop two predictive models, one for river eutrophication risk and the other for microbial risk. The models are intended for a preliminary assessment of a river reach, particularly to assess the level of risk (high or low) for algal bloom and whether the river water is suitable for primary contact activities such as swimming. The input variables for both models included saturated dissolved oxygen and turbidity, while the eutrophication risk model included temperature as an additional variable. When validated with an independent data set, both models predicted the observed risk category accurately in two out of three instances. Since the models developed in this study use only two or three easy to measure variables, their application can help in rapid assessment of river conditions, result in potential cost saving in river monitoring programs and assist in providing timely advice to community and other users for a particular aspect of river use..

10.1 General

Prediction of eutrophication and bacterial water quality is important for river health management and issuing short and long-term advisories on the suitability of water quality to a wide range of river users. Often in peri-urban river systems, the persistence of algal blooms and the presence of pathogenic bacteria exceeding the accepted trigger guidelines are key challenges for river managers. Such issues also diminish the aesthetic appeal and recreational opportunities of the river system. So

162

far the prediction of algal blooms due to eutrophication and microbial water quality has been difficult due to the influence of multiple factors, and the data collection and analysis for prediction is often time consuming, costly, laborious and may not adequately facilitate timely advice to river users.

Management decisions can be quickly implemented if a reliable prediction tool is available that uses measurements which can be carried out in a short time (in few hours rather than days). Therefore, the main aim of this study is to develop two predictive models, one for river eutrophication risk and the other for microbial risk. Predictive Discriminant Function Analysis (PDFA) was used, for assessing the likelihood of a river system exceeding Chlorophyll a for predicting eutrophication risk and that or Enterococci microbial risk in a peri-urban river system. In particular, the focus of this chapter is to examine the applicability of PDFA methodology for developing rapid, site-specific predictive tools that will help assess the risks of eutrophication and the microbial water quality changes. Using observed data set from the study area, the models were validated and provide assurance for any user who prefers to adapt this technique for a local river system. It should be noted that the term ‘risk’ throughout in this study refers to the likelihood of an event occurring at n% of the time and does not take ‘exposure’ into account as in a conventional risk assessment process.

10.1.1 Complexity of Assessing Eutrophication

Human induced eutrophication is one of the most prevalent concerns among water authorities around the world (Moss et al., 1989, Brezonik et al., 1999, Billen et al., 1994). 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 have altered nutrients and sediment-rich stormwater run-off during high rainfall events. As such, the urbanisation has dramatically increased nitrogen and phosphorus loadings in lotic ecosystems (Cole et al., 1993, Neal et al., 2006).

When a freshwater system is under persistent eutrophication, it loses many services and amenities it generally provides to society and the biota (Postel and Carpenter, 1997). Although the direct impacts of eutrophication on human health are marginal, it greatly reduces the aesthetic appeal of a water body, impacts on the fishing industry and significantly increases water treatment costs (Hilton et al., 2006). Many drinking water reservoirs that are rich in plant nutrients cost government agencies

163

considerably as the water quality has been severely degraded in terms of taste, colour, odour and chemical nature (Smith, 2003). In the context of marine and estuarine environments, toxins produced by phytoplankton blooms cause large scale harm to shellfish, finfish and shrimp species (Corrales and Maclean, 1995). On the other hand, toxin-producing cyano-bacteria in nutrient rich water resources produce potentially deadly toxins (i.e., hepato-toxins, cyto-toxins and neurotoxins) that are hazardous to humans and animals (Smith, 2003).

Until recently much emphasis has been given to better understand and model the factors affecting eutrophication in lake environments whilst rivers have received less attention on the prediction of nutrient dynamics (Hilton et al., 2006). For river systems, trophic state indices (Carlson, 1977), regression based methods (Lohman et al., 1992, Dodds et al., 1997, Jassby, 2005), spatial analytical methods (Xu et al., 2001, Wang and Liu, 2005) and multivariate statistical methods (Los and Wijsman, 2007, Primpas et al., 2010) have been attempted. In other studies, the emphasis was on exploring relationships between Chlorophyll a levels and abiotic variables such as total and reactive forms of plant nutrients, local flood regimes, residence time, channel retention, distribution and timing of discharge velocities and channel length that may be attributed to spatial and temporal patterns of eutrophication (Dodds et al., 1997, Reynolds, 2000, Biggs, 2000).

10.1.2 Enterococci as an Indicator of Microbial Water Quality

Faecal pollution in peri-urban rivers, regardless of whether they originated from humans or animals, is a serious concern for many river users from a health perspective. It greatly restricts the use of recreational water for primary contact leisure activities and aquaculture of shellfish (Collins and Rutherford, 2004). Enterococcus and Escherichia coli are two key indicator species that have been extensively utilised to assess the microbiological health of recreational waters (Kinzelman et al., 2003, Eaton and Franson, 2005). Enterococci organisms are gram positive, facultative anaerobic bacteria (ability to survive with or without the presence of oxygen) and mostly present in the human faeces with two dominant species, viz., Enterococcus faecalis and Enterococcus faecium (Murray, 1990). E. coli on the other hand is a gram negative, rod-shaped, non-spore forming bacteria that belongs to a diverse group of enteric bacteria called coliforms (Atlas, 1998). Whilst both species commonly appear in the faecal matter of warm-blooded animals, their presence in large quantities strongly indicates pollution by faecal matter and

164

thus the potential to present increased numbers of other pathogenic organisms (Ahmed et al., 2007).

The Enterococci and E.coli have interchangeably been used as key indicators of microbial health of recreational waters. However, there has been some uncertainty as to which species is more satisfactory for assessing fresh and marine waters. Wade et al., (2003) reviewed 27 potential studies and suggested the suitability of Enterococci to assess marine water, and to a lesser extent E.coli for freshwater systems, as indicators of gastro-intestinal-illnesses. In brackish water, Enterococci survive as a stable and reliable indicator than E.coli (Jin et al., 2004). In marine environments, the survival of E.coli is greatly weaken due to a number of abiotic factors compared to intestinal Enterococci (Sinton et al., 2002). The increased mortality of E.coli in marine environments occurs mainly due to temperature, sedimentation and Ultraviolet light (Orlob, 1956, Rittenberg et al., 1958, Gameson and Saxon, 1967).

Many epidemiological studies indicate Enterococci as a suitable indicator for anthropogenic sources of faecal pollution. In general, Enterococci counts correlate well with the health outcomes in both marine and freshwater environments while E.coli does well only in freshwater systems (NHMRC, 2008). Further, Enterococci counts have indicated a clear dose-response relationship for disease outcomes in marine waters while dose-response for E.coli is negligible (WHO, 2003, NHMRC, 2008). As such, World Health Organisation (WHO) and Australian National Medical Council (NHMRC) recommend using Enterococci as the ‘single-preferred’ compliance indicator for assessing the faecal contamination of recreational waters for primary contact recreation where a whole or part of the body immersion occurs or where there is a risk of swallowing water (WHO, 2003, NHMRC, 2008). A study based on seven New Zealand swimming beaches also concluded that illness had better correlated with Enterococci numbers than with faecal coliform counts (Bandaranayake et al., 1995). Therefore, Enterococcus is a reliable indicator for assessing risk associated with recreational waters and this parameter has been used in this study. It is especially so in the context of HNR where nearly a 50% of the river has some influence from the marine water and inorganic salts that enter the river through surface run-off over the agricultural lands.

Regardless of the type of bacterial organism, in order to predict the bacterial numbers in water, a robust modelling tool that accounts for multiple factors is required. Many traditional bacterial concentration models available for this purpose

165

are based on a time-dependent-decay relationship, a distance-dependent-decay relationship or a combination of thereof (Auer and Niehaus, 1993, Kay and McDonald, 1980, Collins and Rutherford, 2004). However the utility of such models are hindered when adapting them to assess surface waters in peri-urban landscape due to its unique geo-hydrological characteristics. For example, many decaying coefficients are developed under laboratory conditions and have an increased potential to produce inaccurate predictions if applied in field conditions (Jenkins et al., 1984). Secondly the prediction of travel time down a stream becomes difficult and errorneous in distance-dependent decay models when the flow is highly irregular (Jenkins et al., 1984). This is particularly true in peri-urban river systems where the main river is blocked by many impoundments and water is extracted and treated effluent is discharged through multiple locations. A significat proportion of environmental variables also impact on the efficiency of these models.

Kay and McDonald (1983) predicted total coliforms and E.coli using a multiple regression model in two impoundments in the UK. By using a range of predictor variables related to timing and magnitude of hydrological events (i.e., distance to main stream input, time since previous flood and weekly and daily rainfall) and water quality variables (i.e., water temperature, DO saturation) they showed how short- term hydro-meteorological variables are related to bacterial concentrations. Similarly, the source of the pathogenic organisms also plays a critical role in prediction models that determine the risk level to the humans (Ahmed et al., 2007, Parajuli et al., 2009, Soller et al., 2010). Work by Soller et al., (2010) explained how the risks to humans from recreational waters polluted by cattle faeces was substantially different to human faeces and if the faeces originated from chicken, pig or gulls, how this risk was lowered considerably.

10.2 Material and Methods

10.2.1 Discriminant Function Analysis Approach

Discriminant function analysis is a multivariate, pattern recognition technique to determine the ability of a suite of continuous explanatory variables to predict membership of a naturally occurring categorical variable. Basic discriminant function analysis has two categories; the first is the PDFA and the second is Descriptive Discriminant Analysis. The former predicts a particular group which an observation

166

(based on measured variables) belongs to, while the latter separates a priori group based on linear transformations of predictor variables (Williams, 1983).

A stepwise PDFA was employed in this study. The PDFA has been extensively applied in a wide variety of disciplines including medicine (e.g., determination of gender from bone characteristics) (Dixit et al., 2007), ecology (e.g., determination of impact of forest management on woodpecker nesting habitat) (Conner and Adkisson, 1976), business (e.g., prediction of corporate bankruptcy) (Altman, 1968) and linguistics (e.g., categorising text into pre-determined text genre categories) (Karlgren and Cutting, 1994). In the field of water research, the application of this tool has been to predict trace elements in small rivers (Herrmann et al., 1977), understand key variables important for seasonal and temporal variation and reduce dimensionality of water quality data generated from routine monitoring programs (Alberto et al., 2001, Shrestha and Kazama, 2007). However, its application for predicting a particular health attribute in water quality has been over looked.

When predicting a group membership, PDFA generates a composite variate named discriminant function (DF) which is a linear combination of the explanatory variables (Huberty and Barton, 1989). This function divides the data space into clear regions (i.e., usually into a dichotomy) which are subsequently used to separate samples with common properties into the same groups (Alberto et al., 2001). The DF equation is expressed as,

DF = a + b1X1 + b2X2 + b3X3 + … bnXn Eqn. (1)

Where,

DF = discriminant function,

b i = discriminant coefficients,

Xi = score of the explanatory variable,

a = constant, and

n = the number of the predictor variable.

The discriminant coefficients of explanatory variables in PDFA are similar to coefficients obtained in multiple-regression and can be assessed for statistical significance for their relative importance. The prediction of a new group is a three-

167

step process. First, a discriminant score is calculated to a given data case based on the coefficients of explanatory variables that have a significant discriminatory power. A cut-off value is then calculated based on the group means of the explanatory variables called a group centroid. Finally, a new case is predicted as belonging to a particular group based on discriminant function score and group centroid. When there are two groups, only one DF is generated.

10.3 Data Collection and Analysis

10.3.1 Data Collection

Eight water quality variables measured on a monthly basis at three adjacent sites (North Richmond, Wilberforce, and Lower Portland) were obtained on the HNR (Figure 40). The Lower Portland site mostly contained brackish water while the other sites contained mainly freshwater and tertiary treated effluent throughout the year. Nevertheless, all sites are moderately influenced by salt-water intrusion to a certain degree depending on the tidal cycle. The monthly data supplied by the SCA were mainly for three consecutive years viz., 2008, 2009 and 2010. This data has been collected and analysed in accordance with the standard methods for the examination of water and wastewater by the American Public Health Association (APHA) (Eaton and Franson, 2005).

Figure 40. Map with key locations on the Hawkesbury-Nepean River system.

( indicates a sampling location, clear area indicates the HNR catchment, grey area indicates the land not a part of the HNR catchment).

168

10.3.2 Model Development

Models were developed using 2008 and 2009 data sets for selected sites along the river and validated using the 2010 data for the same sites. The Chlorophyll a and Enterococci variables were chosen due to their ability to act as key indicators for eutrophication and microbial risk for primary contact recreational activities. A primary contact recreational activity includes any leisure or commercial activity where humans are directly exposed to river water. In this study, data reported as being below detection level were replaced with the minimum value reported for that data point.

The PDFA is highly sensitive to extreme values and to increase the model accuracy, thus outliers were removed from each variable. Outliers were identified by exploring variables using box plot function in MiniTabTM 16. The total number of outliers for each variable was less than 5%. The accuracy and usefulness of PDFA greatly depends on the normality of the data set. Therefore, all variables were assessed for multivariate normality using the Anderson-Darling test. Necessary variables were log and square root transformed to achieve normality (suspended solids was square root transformed and all other variables were log-transformed) and data were subsequently z-scale standardised prior to the analysis. The aim of the latter transformation was to make all variables in the analysis unit-less. We also looked at the multi-collinearity between variables using draftsman’s plots (Clarke and Ainsworth, 1993). A linear pattern was noticed in the draftsman’s plots between SS and turbidity. However, only turbidity appears as a significant variable in the discriminant function and therefore models were not affected by the issue of multi- collinearity.

To determine whether samples from different sites can be discriminated by easy-to- measure water quality variables, each sample was categorised in relation to their existing Chlorophyll a and Enterococci levels. A desktop review was carried out to obtain a suitable guideline for eutrophication risk and Chlorophyll a and 10 μg/L was chosen for this study (Table 33). Thus, the presence of Chlorophyll a concentration greater than 10 μg/L is considered as an indication of a high risk eutrophication event. Similarly, a guideline for Enterococcus was chosen as 35 cfu/100mL based on Australian and New Zealand Environment Conservation Council guideline (ANZECC, 2000). A number of previous studies also agree with similar threshold values as presented in Table 34 (Deely et al., 1997, Wade et al., 2003, Gannon and Busse, 1989). Each data point of Chlorophyll a was assigned a categorical code viz.,

169

equal or above 10 μg/L = 1 (high-risk blooms) and below 10 μg/L = 2 (low-risk blooms). Similarly, the values equal and above 35 cfu/100mL were given a categorical group number ‘one’ and values below that guideline was given number ‘two’.

During the stepwise PDFA process, Wilk’s lambda minimization method was used. At each step, the F-threshold criterion for inclusion and exclusion of variables were kept at the default values (minimum partial F to enter = 3.84, maximum partial F to remove = 2.71) in the SPSS statistical package (Ferreira et al., 2009). With this method, three and two explanatory variables were identified as discriminators between high and low risk associated with eutrophication and microbial quality of river water. For the eutrophication risk model and microbial risk model, the calculated cut-off value from group centroid was -0.216 and -0.202 respectively (Figure 41 & Figure 42). During the model validation process the DF calculated from each data case was matched with the above cut-off scores and assigned a risk category (i.e., high or low). The validation process was automated using LOGIC functions in MS Excel™ 2007.

We adapted a two-level ‘risk-based’ indication system as the output style for this model due to a number of reasons. Firstly, the word ‘risk’ means to the general public ‘something that should not be proceeded with’. Therefore, a high-risk situation is indicative of a state where further attention or testing is required to assess the condition. Secondly, it does not provide a clear cut-off point as to whether or not a river system is saturated with blooms or is not suitable for primary contact recreational activities (i.e., ‘yes’ or ‘no’ type output). The model output goes beyond the traditional ‘guideline’ approach to assess river health and provides a risk category, which is more perceivable to the public, than following a particular cut-off value. In this study, we assigned only two types of risk categories, high and low, to minimise the complexity of the model. Further, we avoided creating a middle risk group (i.e., moderate risk) as it can create problems particularly with the microbial risk model for river users who are immunologically compromised.

170

Low Risk High Risk

-0.5 -0.4 -0.3 -0.2 -0.1 0 0.1 0.2 0.3 0.4 0.5

Cut off score -0.216

Figure 41. Calculation of cut-off score for Eutrophication risk model.

High Risk Low Risk

-1.0 -0.8 -0.6 -0.4 -0.2 0.2 0.4 0.6 0.8 1.0 0

Cut off score -0.202

Figure 42. Calculation of cut-off score for Microbial risk model.

Table 33. Published Chlorophyll a guidelines.

Researchers Environment Chlorophyll a levels at different trophic status Oligotrophic : 0-0.66 ug/L (Xu et al., 2001) Lake Mesotrophic: 1.6-10 ug/L Eutrophic : Above 20 ug/L Oligotrophic : Below 2.5 ug/L (Wang and Liu, 2005) Lake Mesotrophic: 2.5-8.0 ug/L Eutrophic : Above 25 ug/L Aesthetically displeasing : Above (Dodds et al., 1997) River 100-150 ug/L Oligotrophic : Below 2.5 ug/L (Neal et al., 2006) River Mesotrophic: 2.5-8.0 ug/L Eutrophic : Above 25 ug/L (Muylaert et al., 2009) River Algal blooms: Above 16 ug/L Based on EU Urban Waste water Persistent algal blooms :above Directive (O'Higgins and Wilson, 2005) Estuarine 10ug/L

171

Table 34. Published Enterococci guidelines.

Researchers Enterococci guidelines Marine ecosystem = 35 cfu/100mL (Deely et al., 1997) Freshwater ecosystem = 33 cfu/100mL Marine ecosystem = 35 cfu/100mL (Wade et al., 2003) Freshwater ecosystem = 33 cfu/100mL (Gannon and Busse, 1989) Freshwater ecosystem = 33 cfu/100mL (ANZECC, 2000) Recreational waters = 35 cfu/100mL

10.3.3 Model Validation

Using the 2010 data set (36 data cases collected over 12 months), a discriminant score was calculated for a given data case based on relevant explanatory variables and determined whether the data case belongs to a high risk or low risk group based on cut-off values (see Figure 41 & Figure 42). All observed Chlorophyll a and Enterococci values were then converted into high or low risk groups based on the guidelines. Using the predicted and observed results, the sensitivity, specificity and the overall predictive accuracy of the models were computed as,

Sensitivity = x 100 Eqn. (2)

True positives True positive+False negatives

Specificity = x 100 Eqn. (3)

True negatives True negatives+False positives

Overall predictive accuracy = x 100 Eqn. (4)

True negatives + True positives Total data cases It should be noted that the term sensitivity in this study refers to the proportion of high risk cases predicted correctly by model while specificity refers to the proportion of low risk cases predicted correctly by model (Lalkhen and McCluskey, 2008).

172

10.4 Results and Discussion

10.4.1 General

The descriptive statistics in the original data set for the three sites between 2008 and 2010 obtained from SCA are given in Table 35. The eutrophication risk model consists of three explanatory variables viz., dissolved oxygen, temperature and turbidity (Eqn. 5; also see Table 36 and Table 37). The eutrophication risk model developed in the present study can be represented by Eqn. 5:

DFe =0.772 DO + 0.694 Temp + 1.098 Turb Eqn. (5)

Where,

DFe = discriminant function for the eutrophication risk model, DO = Dissolved oxygen, (Saturated %) Temp = Temperature, (Degrees Celsius) and, Turb = Turbidity (Nephelometric Turbidity Units)

When validated with 2010 data set, the model accurately predicted the eutrophication risk with 72% accuracy for the duration considered (Table 38).

The microbial risk model retained two explanatory variables, viz., dissolved oxygen and turbidity (Eqn. 6; also see Table 36 and 37). The microbial risk model developed in the present study can be represented by Eqn. 6:

DFmi = 1.083 DO + 0.71 Turb Eqn. (6)

Where,

DFmi = discriminant function for the microbial risk model

Box’s M indicates that the assumption of equality of covariance matrices was violated in both models (eutrophication risk model: p = 0.003, microbial risk model: p = 0.017) (Table 37). However, given the large sample size (n = 60) a significant result for Box’s M can be ignored. The PDFA is a robust analytical technique and withstand modest violations of the assumptions (Lachenbruch and Goldstein, 1979). Discriminant functions of both models were highly significant (eutrophication risk

173

model: λ = 0.544, χ2(3) = 34.3 p = 0.000, microbial risk model: λ = 0.747, χ2(2) = 16.6, p = 0.000).

The canonical correlation represents the strength of the overall relationships between the linear composites for the independent and dependent variables.

Considering two sets of variables (x1..xn and y1..yn) canonical correlation is the maximum linear combination between x’ and y’ (Hair et al., 1998). In the eutrophication and microbial risk models, the canonical correlations were 0.675 and 0.503 indicating that each model explained 45.6% and 25.3% of the variation in explanatory variables respectively. Similar to beta-weights in multiple-regression analysis, PDFA provides a set of weights called Standardised Canonical Discriminant Function Coefficients (SCDFC) (Burns and Burns, 2008). The SCDFC is an indicator of the discrimination ability of the independent variable. Thus, larger absolute values refer to a greater discriminatory power of the independent variables. In the eutrophication risk model turbidity appeared as stronger discriminatory variable (SCDFC = 1.098) while DO (SCDFC = 1.083) indicated a higher discriminatory power in microbial risk model. However, the strength of discrimination is marginal compared with the remaining variables within each model.

The model sensitivity, specificity, and overall accuracy are given in Table 38. The high risk data cases were well predicted by the eutrophication risk model (83%) while low risk cases were better predicted by the microbial risk model (74%). The microbial risk model indicated considerably low false positive cases (26%) than eutrophication risk model (39%). The overall prediction accuracy of the models was 72% for the eutrophication risk model and 69% for the microbial risk model. This means, in two out three cases, the models will be able to predict accurately the risk of eutrophication and microbial water quality degradation.

174

Table 35. Descriptive statistics of the data set used to develop the model (2008-2009).

(N-Number of data points, STD-Standard Deviation)

Variable N Mean Median Min Max STD Statistic Std. error Chlorophyll-a (ug/L) 60 16.53 1.59 17.4 2.2 59 12.3 Enterococci 60 67.47 28.26 15 1 1600 218.92 (org/100mL) EC (mS/cm) 60 0.36 0.02 0.32 0.15 0.8 0.15 DO (%) 60 92.76 1.63 92.23 61.9 120.84 12.59 Suspended solids 60 7.18 0.67 6.5 1 20 5.21 (mg/L) Temperature (Deg. C) 60 19.39 0.71 19 10.2 29.3 5.53 Turbidity (NTU) 60 8.46 0.74 7.1 1.23 21.9 5.7 pH 60 7.63 0.04 7.58 7.1 8.49 0.27

Table 36. The stepwise statistics of the discriminant function analysis.

Wilks' Lambda Variables Model Step Exact F entered Statistic df1 df2 df3 Statistic df1 df2 Sig. 1 NTU 0.703 1 1 58 24.465 1 58 0.000 Eutrophication 2 DO 0.653 2 1 58 15.154 2 57 0.000 risk Model 3 Temp 0.544 3 1 58 15.622 3 56 0.000 Microbial risk 1 DO 0.833 1 1 58 11.65 1 58 0.001 Model 2 Turb 0.747 2 1 58 9.639 2 57 0.000

Table 37. Model development statistics.

Canonical Stand. Canonical Model Data set Box's M Wilks' λ Correlation DF Coefficients

2008-2009 20.9 0.54 0.675 DO = 0.772 Eutrophication Risk Model (n = 60) (p < 0.05) (p < 0.05) (r = 45.6%) Temp = 0.694 Turb = 1.098 Microbial Risk 2008-2009 10.7 0.75 0.503 DO = 1.083 Model (n = 60) (p < 0.05) (p < 0.05) (r = 25.3%) Turb = 0.71

175

Table 38. Model validation statistics.

Variables in Overall Model Data set the DF Sensitivity Specificity accuracy equation (%) DO, Temp, Eutrophication 2010 Turb 83% 61% 72% Risk Model (n = 36)

Microbial Risk 2010 DO, Turb 56% 74% 69% Model (n = 36)

10.4.2 Role of Key Variables in the Eutrophication Risk Model

Human induced eutrophication is one of the most prevalent concerns among water authorities around the world (Moss et al., 1989, Brezonik et al., 1999, Billen et al., 1994). For the assessment of trophic status in lotic and lentic systems, two main types of indicator tools have been suggested, viz., tools based on biological factors and tools based on physio-chemical factors (Pesson 1980 cited in Parinet et al., 2004). The eutrophication risk model in the present study belongs to the latter type and predicts the risk of increased algal blooms levels in a river system from three easy-to-measure variables viz., dissolved oxygen, temperature, and turbidity (Eqn. 2). The adoption of Chlorophyll a as a response variable in this study was due to its higher correlation with phytoplankton biomass (Cloot and Roos, 1996, Forsberg and Ryding, 1980). The three explanatory variables in the DF model are related to the eutrophication process in multiple ways with varying levels of influence. If considered in isolation, the values may not describe the state of the trophic condition accurately and relationships between explanatory variables have been accounted for in PDFA to generate better discrimination. This has been clearly articulated by Parinet et al., (2004) when they investigated the cause and effect relationship at the Ivory Coast lake system in France.

Two out of the three explanatory variables identified by the stepwise PDFA predicting the eutrophication risk viz., temperature and light attenuation (measured using turbidity as a surrogate indicator), are the key factors affecting the primary productivity of all phytoplankton species (Vymazal, 1995). One reason for wide adoption of temperature to understand phytoplankton ecology is the simplicity in measurement and influence on the chemical pathways within the photosynthetic cycles of phytoplankton species (Vogel, 1996, Raven, 1974). For example, the

176

growth rate of marine and freshwater algae species increases with rising temperature to an optimum of 30-350C and rapidly drops with further increase in temperature (Eppley, 1972, Smayda, 1969). A number of recent studies from Taiwan, Portugal and China also agree on the effects of temperature as the most influential variable for phytoplankton community structure (Tew et al., 2006, Brogueira et al., 2007, Shen et al., 2011).

Turbidity on the other hand is a measure of amount of light attenuation in the water column and greatly influenced by the presence of organic and inorganic matter. Most of the suspended matter can be characterised of having a particle size less than 62 μm and has a considerable ociation ass with the proliferation of phytoplankton, periphyton and macrophyte communities by maintaining the availability of light in the water column (Waters, 1995, Bilotta and Brazier, 2008). For example, Schanz (1985) showed how the light attenuation in vertical depth profiles at Lake Zurich is related to phytoplankton densities and how well the seasonal phytoplankton blooms can be interpreted using the light intensity. As such, turbidity plays a critical role in the light intensity (visible range of 400 nm-700 nm) influencing the phytoplankton productivity (Hill and Knight, 1988, Minshall, 1978, South and Whittick, 1987). In the present study, turbidity (SCDFC = 1.098) indicated a marginally increased role in discriminating the high and low risk algal blooms events in the HNR compared with DO and temperature (SCDFC = 0.772, 0.694).

The DO levels in river system are quite complex and mostly dependent on water depth, temperature, salinity, turbulence, photosynthesis and respiration rates of aquatic plants and breakdown of organic matter (Weiss, 1970, Badran, 2001, Wheeler et al., 2003, Young et al., 2004, Manasrah et al., 2006). In the present study, the DO indicated a positive discriminatory ability similar to water temperature and this contradicts the conventional DO-temperature and DO-phytoplankton relationships in water bodies reported by previous researchers (Ibanez et al., 2008, Shrestha and Kazama, 2007, Manasrah et al., 2006). Perhaps, DO in peri-urban river systems may be governed by a combination of above mentioned factors influenced by seasonality. Further studies are therefore required to find reasons for the positive relationship of DO in the equation.

The importance of measuring oxygen and temperature to understand the behaviour of algal biomass as surrogate measures of photosynthesis and respiration is also highlighted (Hornberger et al., 1977). In addition, at low DO levels (due to increased temperature), phosphorus is rapidly released from the sediment bed promoting

177

phytoplankton biomass (Lovley, 1993). Overall variables included in the eutrophication risk model represent linkages to a majority of the requirements for the persistence and proliferation of aquatic floral species and can be effectively used to predict the eutrophication risk to a water body.

10.4.3 Challenges of Modelling Chlorophyll a

The notion of nutrient enrichment as a precursor for an eutrophication event is commonly accepted. Many previous models developed for this purpose are based on the firm mechanistic understanding that nitrogen and phosphorus act as growth limiting nutrients for a variety of aquatic floral species (Primpas et al., 2010, Biggs, 2000, Dodds et al., 1998). The work by Dodds et al. (1998) is based on benthic Chlorophyll levels while Primpas et al., 2010 presents a trophic level index based on phytoplankton biomass. However, Biggs (2000) claims that many earlier attempts to model benthic algal biomass from dissolved nutrient with increased explanatory power has not always been successful. This occurred mainly due to our lack of understanding on the nutrient sources (i.e., soluble nutrients, total nutrients, and sediment bound nutrients) for different types of algae species (i.e., macrophytes, benthic and periphytic-algal species) responsible for eutrophication. Whilst dissolved form of nutrients found to have a notably high relationship with benthic algae in streams (Dodds et al., 1998, Dodds, 2006), it may not necessarily be similar to other type of algae such as phytoplanktons in large rivers. We observed the difficulty involved when attempting to predict planktonic Chlorophyll a from dissolved total nitrogen (r = 0.00) and total phosphorus (r = 0.00) in the HNR (see Figures 43 and 44) was observed. Further, the routine collection of benthic Chlorophyll a becomes extremely difficult in deep, murky rivers (such as the HNR).

A great amount of uncertainty also exists when models are developed for rivers which directly relate nutrient levels to benthic, suspended and macrophyte algal communities because, quite often, availability of suspended and benthic nutrients is affected by stream flow and REDOX potential of the water body. For example, in a lake environment, if the oxygen consumption rate is high at the sediment-water interface, it leads to changes in REDOX potential (Søndergaard, 2009). Previous studies suggest that the oxycline created as a result of thermal stratification developed on the HNR led to the formation of deoxygenated and reducing conditions in the hypolimnion. This generated an influx of phosphorus from the sediment with the existing reducing conditions, resulting in a cyano-bacterial bloom (Turner and Erskine, 2005). Therefore, models based on plant nutrients need to be

178

evaluated with caution and only interpreted in conjunction with a particular type of aquatic floral species (i.e., macrophyte, benthic or periphytic algal species). Keeping this aspect in mind, the inclusion of plant nutrients into the models were avoided. However, DO variable in the present study acts as a surrogate for REDOX potential.

In the present study, a considerable proportion of the variation in Chlorophyll a was unexplained (54.5%). This is largely due to abiotic (i.e., river velocity, river depth) and biotic (i.e., macroinvertebrates) determinants which are affecting Chlorophyll a on spatial and temporal scales. Biggs (1996) reported a similar problem when developing a model for benthic Chlorophyll a, without including important bio- physical factors such as stream flow, light penetration and impacts by the grazers.

For example if the flushing rate is faster than the algal development rate, algal communities would be unable to consume all nutrients and grow up to a maximum biomass (Hilton et al., 2006). Alternatively, aquatic plants would not establish in rivers with a high frequency of flood flows and unstable bed sediment (Biggs, 1996). Similarly, riparian shading could significantly reduce the benthic algal biomass, especially when the stream width is restricted to less than 4.5-5.5 m (Davies-Colley and Quinn, 1998). The growth of excessive macrophytes as a result of eutrophication is also influenced by ecological factors. Some researchers have particularly looked at the ecology of macrophytes in great detail viz., interaction between macrophyte biodiversity and species succession and interaction between macrophytes and algae as ecological interactions between species causing competitive exclusion of one species (Willby, 2001, Flynn et al., 2002). Further, macroinvertebrate grazing communities such as may-flies, caddis-flies, snails, chironomids and oligochaetes impose a significant grazing pressure on the algal communities during hydrological benign periods (Biggs, 2000). The phytoplankton communities are also subjected to a grazing by a variety zooplankton assemblages and filter feeding fishes (Malone, 1980). For example, Harvey et al., (1935) reported that in marine waters, phytoplankton communities dropped rapidly during summer (regardless of increased temperature and non-nutrient limiting condition) due to rapid increase in zooplankton assemblages. Practically, it is difficult to incorporate the pressure of grazers into prediction models.

179

0.07

0.06 r = 0.00 0.05

0.04 a (mg/L) - 0.03 Chl 0.02

0.01

0 0 0.5 1 1.5 2 TN (mg/L)

Figure 43. Relationship between TN and Chlorophyll a.

0.07

0.06

0.05 r = 0.00

0.04

0.03 a (mg/L) -

Chl 0.02

0.01

0 0 0.05 0.1 0.15 0.2 TP (mg/L)

Figure 44. Relationship between TP and Chlorophyll a.

10.4.4 Role of Key Variables in the Microbial Risk Model

An ideal microbial indicator of water positively associates with health risks, demonstrates a higher degree of resilience to environmental stress and can be

180

detected by a simple method (Kinzelman et al., 2003). Enterococci undoubtedly qualifies with the first two conditions as an ideal microbial indicator (WHO, 2003, Wade et al., 2003). In this chapter, the prediction is further simplified using two easy-to-measure variables. The explanatory variables (DO and turbidity) are similar to that of the eutrophication risk model without temperature as an additional variable.

While the underlying reason for the direction of this relationship is not clear, evidence from previous studies suggest how the clarity of water is associated with enteric bacterial populations in moving waters. Excessive volumes of suspended particulates are one of the major causes of turbidity in many streams. Thus, turbidity is used as a surrogate measure of suspended particulates in ANZECC guidelines and US EPA water quality criteria (ANZECC, 2000, US EPA, 2011). Bacterial species when combined with suspended particles form flocs and settle on the sediment bed (Marshall, 1980, Weiss, 1951, Jenkins et al., 1984). The process of flocculation is augmented when freshwater mixes with marine water (Weiss, 1951). When agitated by the continuous turbulence especially in river systems due to inflows of water from tributaries and boating activities, the re-suspension of bacterial-flocs causes water quality deterioration. The particular section of the HNR river system which the data set was obtained, has moderate influence of saline water, considerable influence from freshwater inflows from tributaries and impacted by all types of boating activities (i.e., leisure, commercial, prawn trawling). As a result, the water column is frequently agitated and turbidity becomes persistent.

It has also been clearly demonstrated how light penetration controlled by turbidity influences the bacterial numbers in rivers. Kay et al., (1994) found solar irradiance as the principal driver for survival of Enterococci based on the Seven Estuary and

Bristol Channel in the UK. Under high turbidity conditions, T90 (time taken for Enterococci concentration to decay by 90%) was 39.5 hours while under low turbidity conditions it was only 6.6 hours. They further obtained significant relationships (r > 0.8, p < 0.001) between the decay of Enterococci and salinity, suspended solids and turbidity. Confirming this relationship between enteric bacteria and suspended solids and turbidity, Hendricks (1971) found how easily such bacteria can multiply in oligotrophic stream beds due to increased sediments attached to the plant nutrients. As such, turbidity acts as a surrogate indicator for multiple water quality variables viz., Enterococci, salinity and plant nutrients.

181

Therefore, the role of turbidity as a surrogate indicator of Enterococci in the HNR and its inclusion in this model is justified.

The evidence related to direct effects of DO and presence of Enterococci in the literature is scarce. However, the inclusion of DO in our model suggests, it has accounted a number of synergistic effects from other physio-chemical variables that influence the bacterial numbers. For example, based on the River Burton in Spain, Barcina et al., (1986) found that DO levels had an inverse relationship with faecal and total coliforms respectively influenced by seasonality. Similar study found a higher survival of Enterococci organisms (64%) under deoxygenated conditions compared with oxygenated conditions (36%) over 7 days (Roslev et al., 2004). In Australia, the sudden change in DO levels and Enterococci has been included in a risk management framework implemented by NHMR Council as suitable indicators for drinking water (Stevens et al., 2003). While the exact reason for the positive association of DO evaluating the microbial risk is not clear from this study, the microbial risk model includes two surrogate indicators that could be used to predict high and low Enterococci numbers in the HNR. The complexity between and among variables reduces the predictability of Enterococci counts if each was considered in isolation and the only way to overcome this problem is through a combined model.

It should be noted that regulated peri-urban river systems such as the HNR indicate distinct water quality changes from headwater to the river mouth, and for this reason this model may not be suitable for reaches characterised by high salinity or nutrient level. Although the models presented here are site-specific, the methodology developed in the present study can be adapted for river systems elsewhere.

10.5 Concluding Remarks

Using the PDFA approach, two risk prediction models were developed, viz., the river eutrophication risk model and the microbial risk model, in the context of peri-urban river system. The DO and turbidity was found as common explanatory variables for discriminating the risk level of Chlorophyll a concentrations and Enterococci counts in peri-urban situation of the HNR. Further, for the eutrophication risk, water temperature was also an important variable. Both models predicted the observed risk category to a reasonable level of accuracy in two out of three instances; this being 72% for the eutrophication risk model and 69% for the microbial risk model. As such, this is an important finding of this study considering that the models use only two or three easy to measure variables. Further, the models developed in this

182

study use only two or three variables that can be measured easily and quickly. For this reason, their application can help in rapid assessment of river conditions, offer potential cost saving in river monitoring programs and assist in providing timely advice to community and other users for a particular aspect of river use.

The input variables were restricted to those measured by a multi-parameter water quality probe in this study. However, further research is warranted to explore if there are other easy to measure variables that could be included to improve the accuracy of model predictions. In addition, a detailed analysis is needed to understand the temporal and spatial differences in the predictions due to hydro-meteorological factors such as river flow, rainfall, dam releases, and catchment runoff. In general, it is important to understand the cause and effect relationship between the chosen abiotic variables with Chlorophyll a and Enterococci levels in water, and as such, this understanding will indirectly assist in identifying situations under which the models will perform better. Also, it should be noted that Chlorophyll a does not reflect the origin of the pigment and could have arisen from macrophytes, benthic or periphytic-algal species in the river. Validating such a model with longer-term data records will also provide confidence to the future applicability of the model. (Muylaert et al., 2009, O'Higgins and Wilson, 2005, Deely et al., 1997, Gannon and Busse, 1989)

* * * * *

183

CHAPTER 11

BRINGING IT TOGETHER : A FRAMEWORK FOR

ASSESSING RIVER HEALTH

Summary

The findings of the previous chapters of this thesis (Chapters 3-9) highlight complexity and the uncertainty involved in assessing the condition of river health. In particular, the meaning of river health varied with the role and interest of particular individuals or groups and river health indicators appeared highly vulnerable to spatial and temporal impacts such as rainfall, tidal inflows, impoundments and urban effluent discharge. Such complexity and uncertainty can be addressed through a river health assessment framework with step-by-step guidance to help river health management authorities develop site-specific tools suitable for their river systems by taking into account the local river hydrology, water quality aspects and insights from river users. The river health assessment based on such an approach is more appropriate than the application of a generic river health assessment tool as a given river system may have unique hydrological characteristics and scio-economic aspects that require careful attention. The present chapter proposes a river health assessment framework based on the key outcomes of this research project and showcases the role of each step in the framework. However, a comprehensive validation and application of the framework is beyond the scope of this thesis. Nevertheless, the predictive tools developed and knowledge gained using the HNR as a case study provide opportunities to tests the applicability of this framework for other world rivers.

11.1 General

A framework refers to a hypothetical or analytical construction that simplifies a complex process. For example, a risk management framework could be used to educate workers about foreseeable risks in the workplace, solutions to assess risks and methods to minimize risks in a number of different scenarios (Jones et al., 1990). Frameworks are used in almost every field today, viz., health, education, information technology, business, natural resource management, where multiple

184

competing interactions and interests influence key outcomes of a process (Harris and Chapman, 1997, Ichinose, 2003, Huang et al., 2009, Solomon and Miller, 2007, Painuly et al., 2007).

Some frameworks are theoretical and open, while others are analytical and often closed providing a step-by-step guidance towards achieving a desirable outcome. Former is usually based on a number of generic theoretical knowledge in the literature and accepted standards by institutions (Zhou et al., 2004, Peterson, 2003). Frameworks of this nature remain mostly open by providing broadly described steps without indicating linkages between the steps. However, the latter guides the user through a step-by-step process considering most possible interactions involved in the process (Trenberth, 2004, Narumi et al., 2009). They are also flexible, frequently provide the opportunity to link with previous steps and important in assessment of natural resources which are often subject to competing interest by stakeholders and internal biological, chemical and physical interactions.

Given the complexity in the meaning of river health and number of factors influencing the condition of the river system, a four-step framework is proposed to assess peri-urban river health, in this chapter. This framework guides the development of relevant tools for river health advisory information for common uses (e.g., swimming and fishing) and can help river authorities to detect trends in river health parameters and initiate management strategies to sustain a healthy river condition. The proposed framework does not advocate using pristine reference sites as the basis for choosing river health indicators. Instead, the framework considers multiple lines of evidence emerging from a number of sources, viz. stakeholders, general community and the river system itself for this purpose. The four key steps proposed in the river health assessment framework are (Figure 45):

i. Understand the river system from social and environmental perspectives, ii. Identify a suite of indicators for river health monitoring and assessment, iii. Develop predictive tools based on key river health indicators, and iv. Apply the tools for river health management and timely river health advisory.

185

UNDERSTAND IDENTIFY DEVELOP APPLY

Anthropogenic activities and Key indicators landscape-river Management for river health Advisory for interactions strategies for specific river longer-term uses river health

Predictive tools Stakeholder for river health views and indication expectations

Historical trends in water quality and aquatic life

Tools are NO YES satisfactory

Figure 45. The framework for assessing peri-urban river health condition.

11.1.1 Understand the River System

This is the first step in the river health assessment framework. As an initial step towards assessing river health, the emphasis here is to establish an in-depth understanding of major anthropogenic activities in the catchment, which have considerable influence on the river water quality and flow, aquatic life and stakeholder expectations. The three components within this step allocate equal weights on learning the past and present anthropogenic and environmental factors that have significantly influenced the river system.

Understanding the origin and ongoing impacts of anthropogenic activities on river health plays a crucial role in this framework. In many peri-urban settlements, the human population is on the rise, thus creating enormous pressures on river systems. The natural landscapes in peri-urban catchments are being altered to construct residential and industrial buildings. To cater for the ongoing potable water needs, peri-urban river systems are dammed, water is extracted and treated effluent is discharged into rivers over multiple locations. Establishing a clear understanding of the extent of such changes in hydrological regimes can assist in characterising the pressures on the river system. To further strengthen this understanding,

186

knowledge of local experts and key river users can be evaluated. A range of key stakeholders have an interest in a peri-urban river system including passive river users, farmers, recreational water users, fishers, river management authorities and municipal-water-treatment authorities. The stakeholders’ interest and expectations from the river system can vary depending on their level of involvement. The stakeholders can provide a wealth of firsthand information, particularly in characterising the management, policy and social aspects of the river system. Often their views and the level of expectations can vary and conflict. For example, a passive river user wants a clean looking river to enjoy the tranquillity and serenity of nature while downstream fishers want adequate quality and quantity of water flow in the river for a sustainable prawn harvest. Similarly, when upstream farmers extract more water, there will be less water to sustain biotic life in downstream reaches. To better understand such conflicting stakeholder expectations, a range of qualitative studies (e.g., one-to-one interview and field surveys,) can be conducted with the key stakeholders (including general community) in the catchment.

There are temporal and spatial water quality patterns, which can only be understood through an analysis of a time-series data set. As the last component of the first ‘understand’ step of this framework, a detailed analysis of historic water quality records can be conducted. This will help to identify the behaviour and key hydrologic events (i.e., floods, droughts, commissioning of a sewage treatment plant) that influence the present condition of the river. By using historical data, a user may further extend this investigation by exploring different reaches of the river system. Often, larger river systems indicate inherent differences in water quality for different river sections. Such differences are largely associated with pollutant loads entering via tributaries, nature of the adjacent land-use patterns, extent of recreational activities and geological formation of the sediment bed. If such reaches can be identified in advance, the river manager may be able to focus on distinct reaches when developing management strategies. Conventionally, attempts were made to identify different river reaches based on a priori criterion (i.e., distance from the river mouth, distance between two tributaries, freshwater versus salt-water reaches and expert opinion). However, the division of the main river system based on its historic water quality data will not only allow users to better focus river health improvement efforts on the condition of a specific section, but also support a range of river health management decisions to be effectively implemented based on key pollutant sources specific to a chosen section.

187

By successfully addressing each component in this step of the assessment, a higher level of understanding will emerge about the complexity of the given river system, its general patterns of water quality trends, major pollutant sources, key problems that need extra attention and the types of tailor-made tools that are required to satisfy community expectations of the river system.

11.1.2 Identify the River Health Indicators

Indicators are important measures in the decision making process. The second step in the river health assessment framework instigates the need for sound statistical techniques in identifying key indicators within a given reach of the river system. The purpose of the key indicators identified during this step is two-fold. They can be used to either develop tools for predicting future river health scenarios or assess the current river health condition. A suitable key indicator for river health assessment is one that:

• is cost effective in monitoring, • does not follow the trend of another variable, • has strong links with biotic life and key river functions such as primary production or respiration, and

• is sensitive to a wide range of anthropogenic disturbances within catchments and river systems.

For the selection of key indicators to be meaningful, multivariate statistical and pattern recognition techniques are suggested (Pinto and Maheshwari, 2011). A number of positive attributes make multivariate statistical tools the most suitable candidate for this purpose. Firstly, the multivariate tools can easily process a range of different data types originating from river systems (i.e., water chemistry, land-use patterns, sediment health) simultaneously. Secondly, multivariate tools can be used to distinguish the relationships between and among variables. Finally, the multivariate tools present results in a easy to read format making visualisation and interpretation user-friendly.

Often indicators are interrelated. However, a single indicator may not have the ability to indicate a crisis. In reality, a suite of interrelated indicators are therefore required. For example, one may suggest measuring nitrogen or phosphorus by considering the causal relationship between Chlorophyll a levels and plant nutrients to understand a phytoplankton bloom. However, a different set of water quality

188

variables (which are partially related to the behaviour of nutrient indicators) could suggest a better relationship. Factor analysis, methods based on Pearson product- moment Correlation and stepwise-multiple regression analysis are particularly suitable when a great deal of interrelationships occur between variables. Such analysis will also tell us the most influential suite of variables as performance indicators for a particular purpose. The selection of key indicators resembles the analogy of a physician who recommends a few simple pathology tests to diagnose a disease based on key symptoms. By successfully completing this step, one can identify a suite of key indicators that can be further evaluated for longer term river health monitoring and assessment through predictive tools.

11.1.3 Develop and Apply Predictive Tools

The focus of this step is to develop, validate, and apply specific predictive tools to generate advisory warnings for river users and support informed decision making for the management of longer-term river health.

During the ‘develop’ step, various combinations of indicators are thoroughly screened prior to them being included in the predictive tools. The screening involves understanding the nature of the indicators (predictor and response variables) and the purpose of the monitoring and distinguishing the strengths and weaknesses of the indicators. This can also be done by setting up a selection criteria to assess each key performance indicator. For example, one may want to include only the on- the-spot measurements (i.e., pH, EC, temperature) that do not require laboratory analysis rather than measurements (i.e., respiration rates) that require laboratory analysis over 24 hours or more. The interrelationships of the predictor and response variables are converted into a mathematical relationship that would predict a desired river health response (i.e., discoloration, algal bloom, reduced fish catch). Recommending a specific statistical tool for this purpose is beyond the scope of this study.

The predictive tools can be aimed at addressing a variety of common concerns of the community (e.g., whether someone can swim in the river) and rapidly identifying any changes in the water quality due to anthropogenic impacts. The preferred output style of the predictive tool can be risk-based rather than a ‘yes’ or ‘no’ result. An ‘intermediate’ result is not allowed, as river health cannot be expressed as ‘moderately’ suitable for any river use where humans have primary contact with the

189

water. However, when assessing the river health for secondary contact activity such as irrigation, a ‘moderately’ healthy condition may be acceptable.

If the tools fail to predict the river health condition with acceptable accuracy due to major hydrological alterations (i.e., establishment of impoundments, water diversion), the framework will not proceed to the application step. Instead, the user must repeat the previous steps. If the tools indicate acceptable accuracy, they can be used to issue timely alerts for daily river users. If the tools developed for management strategies indicate a noticeable difference compared to a set threshold, in-depth testing can be initiated to find the root cause of the extreme readings.

11.2 River Health Assessment Framework in Practice – An

Example of Practical Context

This framework was developed based on a range of outcomes from each chapter described in this thesis. Below is a brief description on how the framework steps performed in assessing the river health of the HNR system. In particular, the role of understand, identify, develop and apply steps of the framework are described in relation to Chapters 3-6, 7-9 and 10 respectively. Chapter 7 relates to both the Understand and Identify steps of the framework.

11.2.1 Understand

As the first step, the community aspect of the river health was investigated using an in-depth online survey (n = 302) and through a series of one-to-one interviews (n = 14) involving river users (active and passive), river researchers, river managers and river enthusiasts (Chapters 3, 4 and 5). Both studies gathered information on river activities, problems, management issues and community expectations from stakeholders who had different levels of attachment with the river system. One key aspect observed during both studies is the difference in how the general community and key informants perceived the meaning of a healthy river. A considerable proportion of the general community surveyed related river health to a number of visual cues while key informants’ emphasis revolved around the in-stream health (i.e., health of macrophyte, fish, bivalves, algae and water quality) and community satisfaction with the river (i.e., provision of environmental and social services). Primary contact recreational activities such as swimming followed by fishing and

190

boating appeared as the most common types of activities in which the majority of community members engaged with the HNR. These river activities were targeted during the third step of the framework.

In the ‘Understand’ step of the framework, historic water quality records held at SCA were accessed. The availability of the long term monitoring data set suffered from large data gaps due to funding cuts and other logistics reasons. This restricted the use of all data available in the database and only six water quality parameters, viz., temperature, Chlorophyll a, dissolved oxygen, oxides of nitrogen, suspended solids and reactive silicates, measured at weekly intervals between 1985 and 2008 at 12 monitoring stations were used in the analysis (Chapter 6). Taking into account the various effects of nutrient loads, sewage effluents, agricultural and other pollutants originating from point and non-point-sources along the river and via the major tributaries of the HNR, this analysis indicated how the point sources greatly influence the river water quality over the past decade. These finding further confirmed the community concerns on key pollutant sources of the HNR. When attempting to partition the river system based on its water quality, two major clusters representing clean and polluted zones of the river were observed. One cluster represented the upper and lower sections of the river (clean zone) and accounted for approximately 158 km of the river. The other cluster represented the middle section (polluted zone) with a length of approximately 98 km. Due to unavailability of biotic data, a detailed study was not conducted on aquatic species. Nevertheless, the understand step provided enough evidence to support the types of community expectations and landscape/river activities that greatly influenced the water quality. The two cluster partitioning of the river system was successfully used in later steps of the framework.

The historical data from at SCA did not contain regular data records on the phytoplanktons and aquatic species. In addressing this data gap, four locations along the HNR (Penrith, Yarramundi, Cattai and Sackville) were chosen to understand the distribution and abundance patterns of biotic assemblages (Chapter 7). Phytoplankton species revealed an increase in abundance towards the downstream sites and indicated a seasonal succession with Cyanophyta, Chlorophyta and mix of species dominating during summer, autumn and winter. On the other hand, benthic-macroinvertebrates indicated a patchy distribution across the sites with an increase in abundance of the pollutant tolerant Chironomid larvae towards downstream sites.

191

11.2.2 Identify

During the identify step, FA was employed to investigate the key indicators (Chapter 9). The data used in this analysis originated from the polluted zone of the river system. Out of 40 initial water quality variables, FA identified 9 key variables, explaining 50% of the variance in the river water quality. These performance indicators include pH, turbidity, dissolved oxygen, Enterococci, E.coli, Chlorophyll a, algal bio-volume, Manganese and Phaeophytin. The analysis further provided weights for factor loadings, assisting the variables to be segregated into meaningful groups (i.e., anaerobic conditions, microbial quality, and effects of eutrophication). Five other indicators (i.e., temperature, NOx, alkalinity, electrical conductivity, and silicates) also appeared in the analysis although their weights were below the cut-off threshold. The reduction in the number of variables occurred at several levels. The highest number of exclusions occurred when screening the original variables for multi-collinearity, (many correlated variables) and singularity (perfectly correlated variables). A few variables were also removed because they did not follow the multivariate normality and consisted of many data gaps. One interesting point to note during this step is that the key performance indicators identified by the factor analysis had already gained acceptance by international and national water quality monitoring programs. For example, Enterococci and E.coli appeared in the EU Water framework directive as key indicators of microbial contamination while chlorophyll related variables (i.e., Chlorophyll a and algal bio volume) were strongly suggested for river health assessment for recreational purposes by the ACT Government Health Directorate in Canberra, Australia (ACT Health, 2010, European Environment Agency, 2009).

The above water quality variables were further evaluated to see their influence on the harvest of economically important school prawns, community structure of phytoplankton and benthic-macroinvertebrate communities (Chapters 7 and 8). Water pH was identified as the most influential variable for the benthic macroinvertebrates (ρw = 0.437) with temperature as an influential variable for phytoplanktons (ρw = 0.408) and the harvest of school prawns (r = 0.63, p < 0.01). This step not only identified variables with the highest influence on biotic assemblages it also demonstrated how regular monitoring could be improved by reducing the number of water quality variables.

192

11.2.3 Develop and Apply

During this step, two predictive tools were developed to assess eutrophication and Enterococci abundances using the key performance indicators identified in the previous step (Chapters 7-9). These tools were designed to predict whether the river system is heading towards a high or low risk of algal blooms and whether it is suitable for primary contact recreational activities such as swimming. The development and validation of the tools were done in Chapter 6 using the data collected from a polluted zone of the river. Discriminant function analysis is the statistical tool used to develop the predictive algorithm. Linking with the first step of the framework, particular attention was given to addressing the two major concerns of many community members and key informants in terms of using the river waters. Both tools are important for people who want to have direct contact with the river system and river managers who want to implement long term management decisions to improve the quality of the water (Chapter 10). When developing the models, only predictor variables that can be measured through a digital multi-probe were included. The predictor variables for both models included saturated dissolved oxygen and turbidity, while the eutrophication model included temperature as an additional variable. The output style was kept as ‘high-risk’ and ‘low-risk’ so that the general community can self-assess the risks involved in the river activity. These models were also validated with independent data sets and the accuracy remained over 50%.

11.3 Concluding Remarks

The framework outlined in the chapter has the potential to comprehensively assess the health of peri-urban river systems. Local knowledge and history of water quality trends in the river system played a significant role in characterising the river stressors and its major uses. A single indicator may not be appropriate for river health assessment and a suite of variables need to be evaluated for a specific river use. The robustness of this framework can be further evaluated by applying the steps outlined in this chapter to evaluate river systems in other landscapes.

* * * * *

193

CHAPTER 12

CONCLUSIONS AND FUTURE RESEARCH

Summary

Specific conclusions of this thesis have been outlined at the end of Chapters 3-10. The aim of this chapter is to provide some overall conclusions summarised under key themes of the study and indicate suggestions for future research.

12.1 Meaning of River Health

It is evident from this study that the meaning of peri-urban river health remains as complex as the river system itself and the meaning is greatly influenced by personal attitude and one’s level of attachment with the river system. Key informants and general community perceived the meaning of river health slightly different. For example, community emphasised a healthy river based on a number of visual cues while key informants elaborated on biotic species and abiotic factors when defining the meaning of river health. Both groups however mentioned that community satisfaction and ecological integrity play a significant role in the meaning of river health. Therefore, the meaning itself is dynamic and can be adapted depending on the purpose which it serves. A single meaning of river health is not suitable to describe river health in all situations.

The description related to the meaning of river health was not affected by the participants’ age, gender and distance they live from the river. However, the distance from the river and age of the participants were the most influential social demographic factors affecting participants’ current view on the health of the HNR. This highlights the need to implement educational strategies that are tailored to suit different age groups to educate them about the importance of improved river health and their role as a resident to maintain a healthy river system.

The river system appeared quite important to the community as it plays a considerable role in the region through its visual appeal, sustaining ecological integrity, maintaining a hydrologic balance and providing potable water for life. These key river functions are directly associated with river services such as

194

recreational, fishing and agricultural activities. This was reflected in the community survey as well as in key informants’ interviews. Major impacts on peri-urban river health are the discharge of treated municipal effluent, surface run-off mainly containing nutrients from agricultural lands and the presence of exotic weeds (i.e., E. densa).

12.2 Views and Expectations of Stakeholders

The two groups, key informants and the community held opposing views about the sustainability of peri-urban river health. The key informants insisted the management of river health is a shared responsibility between the community and the government. However, the community suggested a majority of this responsibility is vested with the government and agencies responsible for managing the river. In reality, the management of a natural asset is a shared responsibility between its direct users and a regulating body. If the community is given regular information illustrating their role in managing the local river health and what can be done at a domestic level to improve the river health they are more likely to be engaged in the initiatives for sustainable river health management.

The presence of multiple government agencies responsible for the management of the HNR system has also limited the effectiveness of initiatives to improve river health. Key informants were of the view that many government initiatives were often operated on a small scale and did not continue until desired targets are achieved. As a result, community perceived that most river health improvement programs were unsatisfactory or had failed in the past.

Due to the competing interests, views and uses associated with peri-urban rivers, the need for a suitable river health assessment framework appeared as a timely resolution. This framework needs to integrate multiple indicators (i.e., biological, ecological, hydrological, chemical depending on the river use) for various river users and account for community needs and aspirations of the river system.

12.3 The Key River Health Indicators

A large variability was confronted in extracting key indicators for the assessment of river health due to seasonal, spatial, and anthropogenic influences. This was evident when exploring different data sets (i.e., historic and current) to find indicator variables. Rather than selecting single indicators, it is appropriate and more effective

195

to propose a suite of indicators for key river uses such as fishing, agriculture and drinking.

During the first attempt, 9 variables from a list of 40 were retained for the purpose of river health monitoring and assessment. The river health indicators found during this study are related to the anaerobic fermentation (pH, turbidity, and dissolved oxygen), microbial pollution (Enterococci and E.coli) and eutrophication (Chlorophyll a, algal bio-volume, Manganese, Phaeophytin) aspects of the river system. This indicator package accounted for 50% of the variance in river water quality. Further evaluation of these variables revealed dissolved oxygen and turbidity as key variables to determine eutrophication risk and Enterococci risk in the water. An in- depth evaluation of these water quality variables with weather parameters further revealed water temperature as a significantly influential variable for the prawn harvest and phytoplankton community structures. The pH also appeared as the most influential variable for benthic-macroinvertebrate community compositions.

12.4 River Health Assessment Framework and Tools

A simple four-step framework (steps: understand, identify, develop, and apply) was developed to assess peri-urban river systems. One of the important attributes of this framework approach is that it does not depend on pristine reference sites as a basis for choosing river health indicators. Instead, the framework considers multiple lines of evidence emerging from a number of sources, viz. stakeholders, general community and the river system itself for this purpose.

Using stepwise discriminant function analysis, two predictive tools one for eutrophication and the other for Enterococci abundances, were developed. The models are intended to assess whether a particular site in the river is heading to a high or low risk of algal blooms and whether it is suitable for primary contact recreational activities such as swimming. When validated with an independent data set both models predicted the observed risk category with a reasonable accuracy (above 50%).

12.5 Suggestions for Future Research

I. During this study, the river flow was not considered as a key variable due to lack of a reliable flow data set for the HNR. Further, it was assumed that the influence of antecedent flow condition is already embedded in the values of

196

water quality variables. However, if flow data is available, the impact of flow is worth examining especially at tidal mixing zones and also evaluation of its potential as a key variable. By accounting for the river flow, a greater proportion of error associated with the variable selection process and prediction tools can be minimised.

II. Factor analysis, Pearson product-moment correlation based approaches and linear-multiple regression analysis was used to understand the indicator variables. However, a number of other statistical tools such as logistic regression and multivariate regression can be evaluated to identify the variables that may be more effective as river health indicators.

III. The river health assessment framework in this study was developed using social and environmental data collected from a peri-urban river catchment. Thus, the flexibility of this framework can be further evaluated through its application to rivers in different geographical locations and comparing the results to understand whether the framework can consistently produce desirable outcomes.

IV. Due to time and budget constraints, this study only developed and validated predictive tools for selected river uses. Future researchers are encouraged to apply this framework as a basis to develop tools for other key uses of the river system such as drinking water supply, irrigation and visual amenities.

V. The performance of the framework can also be tested against different climatic scenarios and landuse patterns, since such aspects are of critical importance to local councils for decision-making purposes and development approvals. The framework was applied to assess neither climate nor financial scenarios in relation to river health. Further research is warranted to unearth the full spectrum of the social, economic and environmental capabilities of the framework.

VI. The predictive tools have the potential to be incorporated into a smart phone and web based application for routine access by river users and managers.

* * * * *

197

REFERENCES

ACT HEALTH. 2010. ACT Guidelines for the Management of Blue-Green Algae in Recreational Water [Online]. Department of Health, Canberra. Available: http://www.health.act.gov.au/c/health?a=sendfile&ft=p&fid=1281404492&sid = [Accessed 15 November 2010].

ADELL, G. 1999. Theories and models of the peri-urban interface: a changing conceptual landscape [Online]. London: Development Planning Unit, UCL. Available: http://discovery.ucl.ac.uk/43/1/DPU_PUI_Adell_THEORIES_MODELS.pdf [Accessed 12 December 2011].

AHMED, W., STEWART, J., GARDNER, T., POWELL, D., BROOKS, P., SULLIVAN, D. & TINDALE, N. 2007. Sourcing faecal pollution: A combination of library-dependent and library-independent methods to identify human faecal pollution in non-sewered catchments. Water Research, 41, 3771-3779.

ALBERTO, W., DIAZ M.P, AME M.V, PESCE S.B, HUED A.C & BISTONI M.A 2001. Pattern recognition techniques for the evaluation of spatial and temporal variations in water quality. A Case Study: Suqu a River Basin (Córdoba– Argentina). Water Research, 35, 2881-2894.

ALEM, A., JACOBSSON, L., ARAYA, M., KEBEDE, D. & KULLGREN, G. 1999. How are mental disorders seen and where is help sought in a rural Ethiopian community? A key informant study in Butajira, Ethiopia. Acta Psychiatrica Scandinavica, 100, 40-47.

ALLAN, G. L. & MAGUIRE, G. B. 1991. Lethal levels of low dissolved oxygen and effects of short-term oxygen stress on subsequent growth of juvenile Penaeus monodon. Aquaculture, 94, 27-37.

ALTMAN, E. I. 1968. Financial ratios, discriminant analysis and the prediction of corporate bankruptcy. The Journal of finance, 23, 589-609.

AMBASANKAR, K., ALI, S. A. & DAYAL, J. S. 2011. Effect of dietary supplementation of phosphorus on growth and phosphorus excretion in Indian white shrimp, Fenneropenaeus indicus (Milne Edwards). Indian Journal of Fisheries, 54, 305-310.

AMSTRONG, N. 1969. Ecological-Water Quality Models, Engineering Science, Inc. New York.

ANDERS, G. 2010. Multitest Version 5.2. Sweden: Linkopings University, Sweden.

ANDERSON, D., SCHNEIDER, I., WILHELM, S. & LEAHY, J. 2008. Proximate and distant visitors: Differences in importance ratings of beneficial experiences. Journal of Park & Recreation Administration, 26, 47-65.

ANDERSON, M. J. 2001. A new method for non-parametric multivariate analysis of variance. Austral Ecology, 26, 32-46.

ANZECC, A. 2000. Australian and New Zealand guidelines for fresh and marine water quality. Australian and New Zealand Environment and Conservation

198

Council and Agriculture and Resource Management Council of Australia and New Zealand, Canberra, 1-103.

ARAKEL, A. 1995. Towards developing sediment quality assessment guidelines for aquatic systems: an Australian perspective. Australian Journal of Earth Sciences, 42, 335-369.

ARMITAGE, P., MOSS, D., WRIGHT, J. & FURSE, M. 1983. The performance of a new biological water quality score system based on macroinvertebrates over a wide range of unpolluted running-water sites. Water Research, 17, 333- 347.

ASKEY-DORAN, M., HART, B., LADSON, T. & READ, M. 2009. Tasmanian River Condition Index [Online]. South Horbart, Tasmania: NRM South. Available: http://www.stors.tas.gov.au/au-7-0054-00463 [Accessed 18 June 2012].

ATLAS, R. M. 1998. Microbial ecology: fundamentals and applications (4th edition), Menlo Park, California, Benjamin-Cummings Pub. Co. 694p.

AUER, M. T. & NIEHAUS, S. L. 1993. Modeling fecal coliform bacteria--I. Field and laboratory determination of loss kinetics. Water Research, 27, 693-701.

AUSTRALIAN WATER RESOURCES COUNCIL 1991. Freshwater algal blooms- Occasional paper prepared by the Water Resources Management Committee of the Australian Water Resources Council.

AZIZ, K. & GREENWOOD, J. 1981. A laboratory investigation of temperature and salinity tolerances of juvenile Metapenaeus bennettae Racek and Dall (Crustacea: Penaeidae). Journal of Experimental Marine Biology and Ecology, 54, 137-147.

AZRINA, M., YAP, C., RAHIM ISMAIL, A., ISMAIL, A. & TAN, S. 2006. Anthropogenic impacts on the distribution and biodiversity of benthic macroinvertebrates and water quality of the Langat River, Peninsular Malaysia. Ecotoxicology and Environmental Safety, 64, 337-347.

BADRAN, M. I. 2001. Dissolved Oxygen, Chlorophyll a and Nutrients: Seasonal Cycles in Waters of the Gulf of Aquaba, Red Sea. Aquatic Ecosystem Health & Management, 4, 139-150.

BAGINSKA, B., CORNISH, P. S., HOLLINGER, E., KUCZERA, G. & JONES, D. 1998. Nutrient export from rural land in the Hawkesbury-Nepean catchment [Online]. Wagga Wagga: Australian Society of Agronomy. Available: http://www.regional.org.au/au/asa/1998/8/068baginska.htm [Accessed 28 May 2012 1998].

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, 39-46.

BAKER, D. A. & PALMER, R. J. 2006. Examining the effects of perceptions of community and recreation participation on quality of life. Social indicators research, 75, 395-418.

BALL, J. & KEANE, P. 2006. Assessment of impacts of the Replacement Flows Project on the Water cycle, water quality and aquatic ecology [Online].

199

Sydney: Sinclair Knight Merz and Sydney Water. Available: http://www.sydneywater.com.au/MajorProjects/pdf/appendixC1.pdf [Accessed 23 January 2012].

BANDARANAYAKE, D., SALMOND, C., TURNER, S. J., MCBRIDE, G. B., LEWIS, G. D. & TILL, D. G. 1995. Health effects of bathing at selected New Zealand marine beaches, Auckland, New Zealand, Ministry for the Environment Report, 98 p.

BAO, M. L., BARBIERI, K., BURRINI, D., GRIFFINI, O. & PANTANI, F. 1997. Determination of trace levels of taste and odor compounds in water by microextraction and gas chromatography-ion-trap detection-mass spectrometry. Water Research, 31, 1719-1727.

BARBOUR, M., GERRITSEN, J., SNYDER, B. & STRIBLING, J. 1999. Rapid Bioassessment Protocols for Use in Streams and Wadeable Rivers: Periphyton, Benthic Macroinvertebrates and Fish, Second Edition. [Online]. Washington, D.C.: U.S. Environmental Protection Agency; Office of Water. Available: http://water.epa.gov/scitech/monitoring/rsl/bioassessment/index.cfm [Accessed 02 January 2011].

BARCINA, I., ARANA, I., IRIBERRI, J. & EGEA, L. 1986. Factors affecting the survival of E. coli in a river. Hydrobiologia, 141, 249-253.

BARKER, K. K., BOSCO, C. & OANDASAN, I. F. 2005. Factors in implementing interprofessional education and collaborative practice initiatives: Findings from key informant interviews. Journal of Interprofessional Care, 19, 166- 176.

BARMUTA, L. A., CHESSMAN, B. C. & HART, B. T. 2002. Australian River Assessment System: Interpretation of the Outputs from AusRivAS-(Milestone Report) [Online]. Land and Water Resources Research and Development Corporation. Available: http://www.environment.gov.au/water/publications/environmental/rivers/nrhp/ outputs.html [Accessed 11 June 2012].

BARR, N. 2003. Future Agricultural Landscapes. Australian Planner, 40, 123-127.

BASU, B. K. & PICK, F. R. 1996. Factors regulating phytoplankton and zooplankton biomass in temperate rivers. Limnology and Oceanography, 1572-1577.

BATTARBEE, R. W., CHARLES, D. F., DIXIT, S. S. & RENBERG, I. 1999. Diatoms as indicators of surface water acidity. The Diatoms: Application for the Environmental and Earth Sciences, 85-127.

BELLINGER, E. & SIGEE, D. 2010. Freshwater algae: identification and use as bioindicators, Chichester, UK, John Wiley & Sons, Ltd, 284 p.

BERGEY, E. A. & WARD, J. 1989. Upstream-downstream movements of aquatic invertebrates in a Rocky Mountain stream. Hydrobiologia, 185, 71-82.

BEUTEL, A. M. & MARINI, M. M. 1995. Gender and values. American Sociological Review, 436-448.

200

BIAGI, M. & FERRO, M. 2011. Ecological Citizenship and Social Representation of Water. SAGE Open.

BIGGS, B. J. F. 1996. Hydraulic habitat of plants in streams. Regulated Rivers: Research & Management, 12, 131-144.

BIGGS, B. J. F. 2000. Eutrophication of Streams and Rivers: Dissolved Nutrient- Chlorophyll Relationships for Benthic Algae. Journal of the North American Benthological Society, 19, 17-31.

BILLEN, G., GARNIER, J. & HANSET, P. 1994. Modelling phytoplankton development in whole drainage networks: The RIVERSTRAHLER model applied to the Seine river system. Hydrobiologia, 289, 119-137.

BILOTTA, G. & BRAZIER, R. 2008. Understanding the influence of suspended solids on water quality and aquatic biota. Water Research, 42, 2849-2861.

BILYARD, G. R. 1987. The value of benthic infauna in marine pollution monitoring studies. Marine Pollution Bulletin, 18, 581-585.

BOGOSIAN, G., SAMMONS, L. E., MORRIS, P., O'NEIL, J. P., HEITKAMP, M. A. & WEBER, D. B. 1996. Death of the Escherichia coli K-12 strain W3110 in soil and water. Applied and environmental microbiology, 62, 4114.

BOULTON, A. J. 1999. An overview of river health assessment: philosophies, practice, problems and prognosis. Freshwater Biology, 41, 469-479.

BRETTAR, I. & HOFLE, M. G. 1992. Influence of ecosystematic factors on survival of Escherichia coli after large-scale release into lake water mesocosms. Applied and environmental microbiology, 58, 2201.

BREZONIK, P. L., BIERMAN JR, V. J., ALEXANDER, R., ANDERSON, J., BARKO, J., DORTCH, M., HATCH, L., HITCHCOCK, G. L., KEENEY, D. & MULLA, D. 1999. Effects of reducing nutrient loads to surface waters within the Mississippi River Basin and the Gulf of Mexico : Topic 4 Report for the Integrated Assessment of Hypoxia in the Gulf of Mexico Silver Springs, Maryland, NOAA Coastal Ocean Program, 130 p.

BRIGGS, J. & MWAMFUPE, D. 2000. Peri-urban development in an era of structural adjustment in Africa: the city of Dar es Salaam, Tanzania. Urban Studies, 37, 797.

BRITANNICA. 2010. Encyclopaedia Britannica Academic Edition [Online]. Available: http://www.britannica.com/EBchecked/topic/258178/health [Accessed 12 April 2012].

BROADHURST, M. & KENNELLY, S. 1994. Reducing the by-catch of juvenile fish (mulloway Argyrosomus hololepidotus) using square-mesh panels in codends in the Hawkesbury River prawn-trawl fishery, Australia. Fisheries Research, 19, 321-331.

BROGUEIRA, M. J., OLIVEIRA, M. D. R. & CABEÇADAS, G. 2007. Phytoplankton community structure defined by key environmental variables in Tagus estuary, Portugal. Marine Environmental Research, 64, 616-628.

201

BUAPETA, P., HIRANPANB, R., RITCHIEC, R. & PRATHEPA, A. 2008. Effect of nutrient inputs on growth, chlorophyll, and tissue nutrient concentration of ulva reticulata from a tropical habitat. ScienceAsia, 34, 245 - 252.

BUDDS, J. & MINAYA, A. 1999. Overview of initiatives regarding the management of the peri-urban interface [Online]. Development Planning Unit, UCL. Available: http://discovery.ucl.ac.uk/170/1/DPU_PUI_Budds_Minaya_OVERVIEW.pdf [Accessed 23 January 2012].

BUNN, S., DAVIES, P. & MOSISCH, T. 1999. Ecosystem measures of river health and their response to riparian and catchment degradation. Freshwater Biology, 41, 333-345.

BUNT, J., WILLIAMS, W. & CLAY, H. 1982. River water salinity and the distribution of mangrove species along several rivers in north Queensland. Australian Journal of Botany, 30, 401-412.

BUREAU OF METEOROLOGY. 2011. Climate data online [Online]. Available: http://www.bom.gov.au/climate/data/?ftr [Accessed 12 December 2011].

BURNS, R. B. & BURNS, R. A. 2008. Business research methods and statistics using SPSS, Thousand Oaks, California, SAGE Publications Ltd. 531 p.

BURT, T., HOWDEN, N., WORRALL, F. & WHELAN, M. 2010. Long-term monitoring of river water nitrate: how much data do we need? Journal of Environmental Monitoring, 12, 71-79.

BUXTON, M. & CHOY, D. L. 2008. Knowledge for managing Australian landscape- Change and continuity in peri-urban Australia [Online]. Austalian Government. Available: http://lwa.gov.au/files/products/social-and- institutional-research-program/pn22026/pn22026.pdf [Accessed 26 January 2012].

BUXTON, M., TIEMAN, G., BEKESSY, S., BUDGE, T., MERCER, D., COOTE, M. & MORCOMBE, J. 2006. Change and continuity in peri-urban Australia - State of the Peri-urban Regions: A Review of the Literature [Online]. Melbourne: RMIT University. Available: http://www.periurban.org.au/references/LandandWaterSubjectBibliography.p df [Accessed 11 June 2012].

C.S.I.R.O. 2007. Climate Change in the Hawkesbury-Nepean Catchment [Online]. Prepared for the New South Wales Government by the CSIRO. Available: http://www.penrithcity.nsw.gov.au/uploadedFiles/Website/Sustainability/Gree nhouse/Climate%20Change%20in%20the%20Hawkesbury- Nepean%20Catchment.pdf [Accessed 17 February 2012].

CAMARGO, J. A., ALONSO, A. & PUENTE, M. 2005. Eutrophication downstream from small reservoirs in mountain rivers of Central Spain. Water Research, 39, 3376-3384.

CARLSON, R. E. 1977. A trophic state index for lakes. Limnology and Oceanography, 22, 361-369.

202

CARLSON, R. E. & SIMPSON, J. 1996. A Coordinator’s Guide to Volunteer Lake Monitoring Methods, Madison, Wisconsin, North American Lake Management Society, 96 p.

CHADWICK, J. W., CANTON, S. P. & DENT, R. L. 1986. Recovery of benthic invertebrate communities in Silver Bow Creek, Montana, following improved metal mine wastewater treatment. Water, Air, & Soil Pollution, 28, 427-438.

CHAPMAN, P. 1992. Ecosystem health synthesis: can we get there from here? Journal of Aquatic Ecosystem Stress and Recovery, 1, 69-79.

CHESSMAN, B. 1995. Rapid assessment of rivers using macroinvertebrates: a procedure based on habitat-specific sampling, family level identification and a biotic index. Austral Ecology, 20, 122-129.

CLARKE, K. & AINSWORTH, M. 1993. A method of linking multivariate community structure to environmental variables. Marine Ecology-Progress Series, 92, 205-205.

CLARKE, K. & GORLEY, R. 2006. PRIMER-E v6: User Manual/Tutorial. Plymouth, UK: Plymouth Marine Laboratory, 192 p.

CLARKE, K. & WARWICK, R. 2001. Change in marine communities: an approach to statistical analysis and interpretation, 2nd edition, Plymouth, UK, PRIMER-E: Plymouth Marine Laboratory, 172 p.

CLARKE, M. 2006. Quantifying the Economic Value of Activities Dependent on the Hawkesbury-Nepean River [Online]. AgEconPlus Consulting. Available: http://www.seabeesboating.com/Downloads/AgEconPlus%20Hawkesbury%2 0NepeanEconomic%20Values%20Final%20Report.pdf [Accessed 20 April 2012].

CLEMENTS, W. H. 1994. Benthic invertebrate community responses to heavy metals in the Upper Arkansas River Basin, Colorado. Journal of the North American Benthological Society, 13, 30-44.

CLOOT, A. & ROOS, J. 1996. Modelling a relationship between phosphorus, pH, calcium and chlorophyll-a concentrations. Water SA, 22.

COLE, J. J., PEIERLS, B. L., CARACO, N. F. & PACE, M. L. 1993. Nitrogen loading of rivers as a human-driven process. Humans as components of ecosystems: the ecology of subtle human effects and populated areas, 141– 157.

COLLINS, R. & RUTHERFORD, K. 2004. Modelling bacterial water quality in streams draining pastoral land. Water Research, 38, 700-712.

CONNELL, J. 1974. Ecology: field experiments in marine ecology. Experimental Marine Biology, 21-54.

CONNER, R. N. & ADKISSON, C. S. 1976. Discriminant function analysis: a possible aid in determining the impact of forest management on woodpecker nesting habitat. Forest Science, 22, 122-127.

203

CORRALES, R. A. & MACLEAN, J. L. 1995. Impacts of harmful algae on seafarming in the Asia-Pacific areas. Journal of Applied Phycology, 7, 151- 162.

COSTANZA, R., NORTON, B. & HASKELL, B. 1992. Ecosystem health: New goals for environmental management, Washington DC, Island Press, 269 p.

COVICH, A. P., PALMER, M. A. & CROWL, T. A. 1999. The role of benthic invertebrate species in freshwater ecosystems. BioScience, 49, 119-127.

CUDE, C. 2001. Oregon wate quality index a tool for evaluating water quality management effectiveness. JAWRA Journal of the American Water Resources Association, 37, 125-137.

D.E.C.C.W. 2010. State of the Catchment 2010-Riverine Eco System:Hawkesbury Nepean Region [Online]. Sydney: Department of Environment, Climate Change and Water. Available: http://www.environment.nsw.gov.au/resources/soc/hawkesburynepean/1045 0HAWKNEPriver.pdf [Accessed 28 May 2012].

DA GAMA TORRES, H. 2011. Environmental Implications of Peri-urban Sprawl and the Urbanization of Secondary Cities in Latin America [Online]. Inter- American Development Bank. Available: http://idbdocs.iadb.org/wsdocs/getdocument.aspx?docnum=35769834 [Accessed 23 January 2012].

DALKEY, N. & HELMER, O. 1963. An experimental application of the Delphi method to the use of experts. Management Science, 9, 458-467.

DAVIES-COLLEY, R. J. & QUINN, J. M. 1998. Stream lighting in five regions of North Island, New Zealand: control by channel size and riparian vegetation. New Zealand Journal of Marine and Freshwater Research, 32, 591-605.

DAVIS, J. C. 1975. Minimal dissolved oxygen requirements of aquatic life with emphasis on Canadian species: a review. Journal of the Fisheries Board of Canada, 32, 2295-2332.

DE HAAS, E. M. & KRAAK, M. H. S. 2008. Species-specific responses of two benthic invertebrates explain their distribution along environmental gradients in freshwater habitats. Science of the Total Environment, 406, 430-435.

DE TOLEDO PIZA PELUSO, É. & BLAY, S. L. 2004. Community perception of mental disorders. Social psychiatry and psychiatric epidemiology, 39, 955- 961.

DEELY, J., HODGES, S., MCINTOSH, J. & BASSETT, D. 1997. Enterococcal numbers measured in waters of marine, lake, and river swimming sites of the Bay of Plenty, New Zealand. New Zealand Journal of Marine and Freshwater Research, 31, 89-101.

DEEPESH, M. & MADAN, K. J. 2012. Hydrologic Time Series Analysis: Theory and Practice, New York, Springer, 280 p.

DELANEY, A. E. 2010. Incorporation the human dimension into ecological-based management: the good, the bad and the ugly. Bulletin of the British Ecological Society, 41, 27-28.

204

DIAMOND, R. 2004. Water and Sydney's Future- Balancing the value of our rivers and economy [Online]. Sydney, NSW: Department of Infrastructure, Planning and Natural Resources. Available: http://www.water.nsw.gov.au/ [Accessed 11 June 2012].

DILLON, P. & KIRCHNER, W. 1975. The effects of geology and land use on the export of phosphorus from watersheds. Water Research, 9, 135-148.

DIXIT, S., KAKAR, S., AGARWAL, S. & CHOUDHRY, R. 2007. Sexing of human hip bones of Indian origin by discriminant function analysis. Journal of Forensic and Legal Medicine, 14, 429-435.

DODDS, W., SMITH, V. & ZANDER, B. 1997. Developing nutrient targets to control benthic chlorophyll levels in streams: a case study of the Clark Fork River. Water Research, 31, 1738-1750.

DODDS, W. K. 2006. Eutrophication and trophic state in rivers and streams. Limnology and Oceanography, 671-680.

DODDS, W. K., JONES, J. R. & WELCH, E. B. 1998. Suggested classification of stream trophic state: distributions of temperate stream types by chlorophyll, total nitrogen, and phosphorus. Water Research, 32, 1455-1462.

DUNLAP, R., VAN LIERE, K., MERTIG, A. & JONES, R. 2000. Measuring endorsement of the new environmental paradigm: a revised NEP scale. Journal of Social Issues, 56, 425-442.

EATON, A. D. & FRANSON, M. H. 2005. Standard Methods for the Examination of Water and Wastewater, American Public Health Association and American Water Works Association and Water Environment Federation, Washington, DC, 1200 pp.

ECK, D. 1982. The Goddess Ganges in Hindu Sacred Geography. In: HAWLEY, J. S. & WULFF, D. M. (eds.) Devi: Goddesses of India. Boston, Mass: Beacon Press, 137-53 p.

ELLIOTT, J. A., IRISH, A. E., REYNOLDS, C. S. & TETT, P. 2000. Modelling freshwater phytoplankton communities: an exercise in validation. Ecological Modelling, 128, 19-26.

ENVIRONMENTAL PROTECTION AGENCY (EPA). 2002. Green offset for sustainable development. Concept paper. [Online]. Available: http://www.environment.nsw.gov.au/resources/greenoffsets/greenoffsets.pdf [Accessed 01 December 2010].

EP & A. 1979. Environmental Planning and Assessment Act 1979 No 203 [Online]. NSW Government. Available: http://www.legislation.nsw.gov.au [Accessed 19 January 2012].

EPPLEY, R. W. 1972. Temperature and phytoplankton growth in the sea. Fish. Bull, 70, 1063-1085.

ERRINGTON, A. 1994. The peri-urban fringe: Europe's forgotten rural areas. Journal of Rural Studies, 10, 367-375.

205

EUROPEAN ENVIRONMENT AGENCY. 2009. Annual report, 2009 bathing season [Online]. Copenhagen, Denmark: Office for Official Publications of the European Union. Available: http://ec.europa.eu/environment/water/water- bathing/report2010/EU-wide%20report.pdf [Accessed 10 Feb 2011].

FAIRWEATHER, P. G. 1999a. Determining the 'health' of estuaries: Priorities for ecological research. Australian Journal of Ecology, 24, 441-451.

FAIRWEATHER, P. G. 1999b. State of environment indicators of 'river health': exploring the metaphor. Freshwater Biology, 41, 211-220.

FERREIRA, I. C. F. R., AIRES, E., BARREIRA, J. & ESTEVINHO, L. M. 2009. Antioxidant activity of Portuguese honey samples: Different contributions of the entire honey and phenolic extract. Food Chemistry, 114, 1438-1443.

FIELDS, A. 2009. Discovering Statistics Using SPSS (and sex,drugs and rock 'n' roll), London, SAGE Publications, 856 p.

FILIK ISCEN, C., EMIROGLU, Ö., ILHAN, S., ARSLAN, N., YILMAZ, V. & AHISKA, S. 2008. Application of multivariate statistical techniques in the assessment of surface water quality in Uluabat Lake, Turkey. Environmental Monitoring and Assessment, 144, 269-276.

FIREY, W. 1946. Ecological considerations in planning for rurban fringes. American Sociological Review, 11, 411-423.

FLYNN, N., SNOOK, D., WADE, A. & JARVIE, H. 2002. Macrophyte and periphyton dynamics in a UK Cretaceous chalk stream: the River Kennet, a tributary of the Thames. Science of The Total Environment, 282, 143-157.

FORD, T. 1999. Understanding population growth in the peri-urban region. International Journal of Population Geography, 5, 297-311.

FORSBERG, C. & RYDING, S. O. 1980. Eutrophication parameters and trophic state indices in 30 Swedish waste-receiving lakes. Arch. Hydrobiol, 89, 189- 207.

GAMESON, A. & SAXON, J. 1967. Field studies on effect of daylight on mortality of coliform bacteria. Water Research, 1, 279-295.

GANNON, J. J. & BUSSE, M. K. 1989. E. coli and enterococci levels in urban stormwater, river water and chlorinated treatment plant effluent. Water Research, 23, 1167-1176.

GAVIN, B., NICOLE, S. & PIETER, S. 1998. The Environmental Status of Hawkesbury River Sediments. Australian Geographical Studies, 36, 37-57.

GILLSON, J. 2011. Freshwater Flow and Fisheries Production in Estuarine and Coastal Systems: Where a Drop of Rain Is Not Lost. Reviews in Fisheries Science, 19, 168-186.

GLAISTER, J. 1978a. The Impact of River Discharge on Distribution and Production of the School Prawn Metapenaeus macleayi (Haswell)(Crustacea: Penaeidae) in the Clarence River Region, Northern New South Wales. Marine and Freshwater Research, 29, 311-323.

206

GLAISTER, J. 1978b. Movement and growth of tagged school prawns, Metapenaeus macleayi (Haswell)(Crustacea: Penaeidae), in the Clarence River region of northern New South Wales. Marine and Freshwater Research, 29, 645-657.

GONZÁLEZ-OREJA, J. A. & SAIZ-SALINAS, J. I. 1998. Exploring the relationships between abiotic variables and benthic community structure in a polluted estuarine system. Water Research, 32, 3799-3807.

GRAY, J. S. 1974. Animal-sediment relationships. Oceanography and Marine Biology An Annual Review 13, 223-261.

GREEN, D. & ERSKINE, W. 2000. Geomorphic effects of extractive industries and their Implications for river management: the case of the Hawkesbury-Nepean River, NSW, Chichester, UK, John Wiley, 124-149 p.

GROWNS, I., GEHRKE, P. C., ASTLES, K. L. & POLLARD, D. A. 2003. A comparison of fish assemblages associated with different riparian vegetation types in the Hawkesbury–Nepean River system. Fisheries Management & Ecology, 10, 209-220.

GROWNS, I. G. & GROWNS, J. E. 2001. Ecological effects of flow regulation on macroinvertebrate and periphytic diatom assemblages in the Hawkesbury - Nepean River, Australia. Regulated Rivers:Research & Management, 17, 275-293.

GROWNS, J. E., CHESSMAN, C., MCEVOY, P. K. & WRIGHT, I. A. 1995. Rapid assessment of rivers using macroinvertebrates: Case studies in the Nepean River and Blue Mountains, NSW. Austral Ecology, 20, 130-141.

GRUMBINE, R. E. 1994. What is ecosystem management? Conservation Biology, 8, 27-38.

HA, L. & FANG, L. 2011. Internet experience and time displacement of traditional news media use: An application of the theory of the niche. Telematics and Informatics, In Press, Corrected Proof.

HAEVEY, J. 2001. The natural economy. Nature, 413-463.

HAIR, J. F., ANDERSON, R. E., TATHAM, R. L. & WILLIAM, C. 1998. Multivariate data analysis. 5th ed. Upper Saddle River, New Jersey: Prentice Hall, 730 p.

HÄMÄLÄINEN, H. & HUTTUNEN, P. 1996. Inferring the minimum pH of streams from macroinvertebrates using weighted averaging regression and calibration. Freshwater Biology, 36, 697-709.

HARRIS, J. H. 1995. The use of fish in ecological assessments. Austral Ecology, 20, 65-80.

HARRIS, R. N. & CHAPMAN, D. S. 1997. Borehole temperatures and a baseline for 20th-century global warming estimates. Science, 275, 1618-1621.

HART, B., MAHER, B. & LAWRENCE, I. 1999. New generation water quality guidelines for ecosystem protection. Freshwater Biology, 41, 347-359.

207

HARVEY, H., COOPER, L., LEBOUR, M. V. & RUSSELL, F. 1935. Plankton production and its control. J. Mar. Biol. Assoc., 20, 407-441.

HAWKESBURY TRAWL ASSOCIATION. 2001. Environmental Action Plan [Online]. Available: http://www.hawkesburyharvest.com.au/imagesDB/members/Trawlers.pdf [Accessed 15 November 2011].

HAWKINS, P., HILL, B., SAUNDERS, J., KOBAYASHI, T. & WINDER, J. 1994. Algal bloom dynamics in the Hawkesbury-Nepean River (October 1992- March 1993), West Ryde, NSW, Australian Water Technologies.

HEALTHY RIVERS COMMISSION 1998. Independent inquiry into the Hawkesbury Nepean River system : final report August 1998 / Healthy Rivers Commission of New South Wales, Healthy Rivers Commission of New South Wales, Sydney.

HEALTHY RIVERS COMMISSION OF NSW. 2000. Securing Healthy Coastal Rivers: A Strategic Perspective [Online]. Sydney: NSW Department of Environment, Climate Change & Water. Available: http://www.shop.nsw.gov.au/pubdetails.jsp?publication=5693 [Accessed 11 June 2012].

HEARNSHAW, E., CULLEN, R. & HUGHEY, K. 2005. Ecosystem health demystified. 2nd National Workshop of the ANU's Economics and Environment Network (EEN).

HEATHWAITE, L. 1994. Eutrophication. Geographical Review, 7, 31-37.

HEIP, C., WARWICK, R., CARR, M., HERMAN, P., HUYS, R., SMOL, N. & VAN HOLSBEKE, K. 1988. Analysis of community attributes of the benthic meiofauna of Frierfjord/Langesundfjord. 46, 171-180.

HEISKARY, S. & MARKUS, H. 2001. Establishing Relationships Among Nutrient Concentrations, Phytoplankton Abundance, and Biochemical Oxygen Demand in Minnesota, USA, Rivers. Lake and reservoir management, 17, 251-262

HENDRICKS, C. W. 1971. Enteric bacterial metabolism of stream sediment eluates. Canadian Jounal of Microbiology, 17, 551-556.

HERBERS, J. 1989. A New Heartland. Country Journal, 16, 67-69.

HERRMANN, R., BOLZ, U., SYMADER, W. & RUMP, H. 1977. Interpretation and prediction of spatial variation in trace metals in small rivers by canonical and discriminant analyses. International Hydrologic Symposium. Fort Collins, Colarado.

HILDEBRAND, H. H. & GUNTER, G. 1953. Correlation of rainfall with the Texas catch of white shrimp, Penaeus setiferus (Linnaeus). Transactions of the American Fisheries Society, 82, 151-155.

HILL, W. R. & KNIGHT, A. W. 1988. Nutrient and light limitation of algae in northern California streams. Journal of Phycology, 24, 125-132.

208

HILTON, J., O'HARE, M., BOWES, M. J. & JONES, J. I. 2006. How green is my river? A new paradigm of eutrophication in rivers. Science of the Total Environment, 365, 66-83.

HINES, J. M., HUNGERFORD, H. R. & TOMERA, A. N. 1987. Analysis and synthesis of research on responsible environmental behavior: A meta- analysis. The Journal of Environmental Education, 18, 1-8.

HNCMA 2007. Hawkesbury Nepean River Health Strategy,. Goulburn NSW Australia: Hawkesbury Nepean Catchment Management Authority.

HOELZL, U. 2007. River habitat monitoring and assessment in Germany. Environmental Monitoring and Assessment, 127.

HOLDWAY, D., BRENNAN, S. & AHOKAS, J. 1995. Short review of selected fish biomarkers of xenobiotic exposure with an example using fish hepatic mixed‐ function oxidase. Australian Journal of Ecology, 20, 34-44.

HORNBERGER, G., KELLY, M. & COSBY, B. 1977. Evaluating eutrophication potential from river community productivity. Water Research, 11, 65-69.

HÖTZEL, G., CROOME, R., LAND, RESEARCH, W. R. & CORPORATION, D. 1999. A phytoplankton methods manual for Australian freshwaters, Land and Water Resources Research and Development Corp.

HOUSE, M. 1989. A water quality index for river management. Water and Environment Journal, 3, 336-344.

HOWARD, M. 2009. Aquatic Ecosystems productivity 'IS' reliant on water managers and systainable cities. 12th International River Symposium. Proceedings of 12th International River Symposium, 2009, September 21-24, Brisbane, Australia: International Watercentre.

HOWARD, M. & HOWARD, G. 2005. Investigation into Water and Wastewater Service Provision in the Greater Sydney Region [Online]. Available: http://www.ipart.nsw.gov.au/files/ [Accessed 18 November 2011].

HOWELL, J. & BENSON, D. 2000. Predicting potential impacts of environmental flows on weedy riparian vegetation of the Hawkesbury-Nepean River, south- eastern Australia. Austral Ecology, 25, 463-475.

HUANG, S., TANIGUCHI, M., YAMANO, M. & WANG, C. 2009. Detecting urbanization effects on surface and subsurface thermal environment - A case study of Osaka. Science of The Total Environment, 407, 3142-3152.

HUBERTY, C. J. & BARTON, R. M. 1989. An Introduction to Discriminant Analysis. Measurement and Evaluation in Counseling and Development, 22, 158-68.

HUNSAKER, C. T. 1990. Environmental monitoring and assessment program: ecological indicators, Atmospheric Research and Exposure Assessment Laboratory, Office of Research and Development, US Environmental Protection Agency.

HUNTER, L. M., HATCH, A. & JOHNSON, A. 2004. Cross National Gender Variation in Environmental Behaviors. Social Science Quarterly, 85, 677-694.

209

HURLBERT, S. H. 1984. Pseudoreplication and the Design of Ecological Field Experiments. Ecological Monographs, 54, 187-211.

HUSTON, M. 1979. A general hypothesis of species diversity. American Naturalist, 113, 81.

I.T.I.S. 2011. Intergrated Taxonomic Information System [Online]. Available: http://www.itis.gov/index.html [Accessed 31 October 2011].

IBANEZ, J. G., HERNANDEZ-ESPARZA, M., DORIA-SERRANO, C., FREGOSO- INFANTE, A. & SINGH, M. M. 2008. Dissolved Oxygen in Water. Environmental Chemistry, 16-27.

ICHINOSE, T. 2003. Regional warming related to land use change during recent 135 years in Japan. Journal of Global Environment Engineering, 9, 19-39.

IGNATIADES, L., GOTSIS-SKRETAS, O., PAGOU, K. & KRASAKOPOULOU, E. 2009. Diversification of phytoplankton community structure and related parameters along a large-scale longitudinal east–west transect of the Mediterranean Sea. Journal of Plankton Research, 31, 411.

ISC. 2006. Index of Stream Condition : User’s Manual (2nd edition) [Online]. Victoria: Department of Sustainability and Environment. Available: http://www.water.vic.gov.au/data/assets/pdf_file/0003/9921/ISCUsersManual 2ndEdition01.pdf [Accessed 2012 14 November].

JACKSON, S., STOECKL, N., STRATON, A. & STANLEY, O. 2008. The changing value of Australian tropical rivers. Geographical Research, 46, 275-290.

JAMES, D. 1997. Environmental incentives: Australian experience with economic instruments for environmental management, Environmental Economics Research Paper 5, Environment Australia, Commonwealth of Australia, Canberra, Environment Australia.

JANSEN, A., LAND & AUSTRALIA, W. 2004. Development and application of a method for the rapid appraisal of riparian condition [Online]. Land & Water Australia. Available: http://lwa.gov.au/files/products/river- landscapes/pr040656/pr040656.pdf [Accessed 2 October 2011].

JANSEN, A. & ROBERTSON, A. 2001. Relationships between livestock management and the ecological condition of riparian habitats along an Australian floodplain river. Journal of Applied Ecology, 38, 63-75.

JASSBY, A. D. 2005. Phytoplankton regulation in a eutrophic tidal river (San Joaquin River, California). San Francisco Estuary and Watershed Science, 3, 1-22.

JENKINS, A., KIRKBY, M., MCDONALD, A., NADEN, P. & KAY, D. 1984. A process based model of faecal bacterial levels in upland catchments. Water Science & Technology, 16, 453-462.

JIN, G., ENGLANDE, A., BRADFORD, H. & JENG, H. 2004. Comparison of E. coli, enterococci, and fecal coliform as indicators for brackish water quality assessment. Water environment research, 76, 245-255.

210

JOHN, C., REBEKAH, G. & DUNCAN, M. 1996. Intergrated catchment management in the Hawkesbury Nepean: A discussion paper. Richmond: Land and Water Resources Research and Development Corporation, University of Western Sydney.

JONES, A. 1987. Temporal patterns in the macrobenthic communities of the Hawkesbury Estuary. New South Wales. Australian Journal of Marine and Freshwater Research, 38, 607-624.

JONES, A. 1990. Zoobenthic variability associated with a flood and drought in the Hawkesbury estuary, New South Wales: Some consequences for environmental monitoring. Environmental Monitoring and Assessment, 14, 185-195.

JONES, A., WATSON-RUSSELL, C. & MURRAY, A. 1986. Spatial patterns in the macrobenthic communities of the Hawkesbury Estuary, New South Wales. Marine and Freshwater Research, 37, 521-543.

JONES, K. & STEWART, W. 1969. Nitrogen turnover in marine and brackish habitats. 3. The production of extracellular nitrogen by Clothrix scopulorum. Journal of the Marine Biological Association of the United Kingdom, 49, 475- 488.

JONES, P., GROISMAN, P. Y., COUGHLAN, M., PLUMMER, N., WANG, W. & KARL, T. 1990. Assessment of urbanization effects in time series of surface air temperature over land. Nature, 347, 169-172.

KANKAANPAA, H. T., HOLLIDAY, J., SCHRODER, H., GODDARD, T. J., VON FISTER, R. & CARMICHAEL, W. W. 2005. Cyanobacteria and prawn farming in northern New South Wales, Australia-A case study on cyanobacteria diversity and hepatotoxin bioaccumulation. Toxicology and applied pharmacology, 203, 243-256.

KARLGREN, J. & CUTTING, D. 1994. Recognizing text genres with simple metrics using discriminant analysis [Online]. Association for Computational Linguistics. Available: http://dl.acm.org/citation.cfm?id=991324 [Accessed 13 April 2010].

KARR, J. 1991. Biological integrity: a long-neglected aspect of water resource management. Ecological Applications, 66-84.

KARR, J. 1996. Ecological integrity and ecological health are not the same. In: PETER SCHULZE (ed.) Engineering within ecological constraints. The United States of America: National Academy of Engineers, 97-109 p.

KARR, J. 1999. Defining and measuring river health. Freshwater Biology, 41, 221- 234.

KARR, J. & THOMAS, T. 1996. Economics, ecology, and environmental quality. Ecological Applications, 6, 31-32.

KATAOKA, K., MATSUMOTO, F., ICHINOSE, T. & TANIGUCHI, M. 2009. Urban warming trends in several large Asian cities over the last 100 years. Science of The Total Environment, 407, 3112-3119.

211

KAY, D., JONES, F., WYER, M., FLEISHER, J., SALMON, R., GODFREE, A., ZELENAUCH-JACQUOTTE, A. & SHORE, R. 1994. Predicting likelihood of gastroenteritis from sea bathing: results from randomised exposure. The Lancet, 344, 905-909.

KAY, D. & MCDONALD, A. 1980. Reduction of coliform bacteria in two upland reservoirs: the significance of distance decay relationships. Water Research, 14, 305-318.

KAY, D. & MCDONALD, A. 1983. Predicting coliform concentrations in upland impoundments: design and calibration of a multivariate model. Applied and environmental microbiology, 46, 611.

KELLY, M. & WHITTON, B. 1995. The Trophic Diatom Index: a new index for monitoring eutrophication in rivers. Journal of Applied Phycology, 7, 433-444.

KIMMERIKONG 2005. Scoping study-Hawkesbury-Nepean River Estuary Management, Final Report. New South Walse Australia.

KINZELMAN, J., NG, C., JACKSON, E., GRADUS, S. & BAGLEY, R. 2003. Enterococci as indicators of Lake Michigan recreational water quality: Comparison of two methodologies and their impacts on public health regulatory events. Applied and environmental microbiology, 69, 92.

KITABAYASHI, K., KURATA, H., SHUDO, K., NAKAMURA, K. & ISHIKAWA, S. 1971. Studies on formula feed for kuruma prawn-I. On the relationship among glucosamine, phosphorus and calcium. Bull. Tokai Reg. Fish. Res. Lab, 65, 91-107.

KRAMER, D. L. 1987. Dissolved oxygen and fish behavior. Environmental Biology of Fishes, 18, 81-92.

KROGH, M., WRIGHT, A. & MILLER, J. 2008. Hawkesbury Nepean River Environmental Monitoring Program: Final Technical Report [Online]. Sydney: Department of Environment and Climate Change NSW and Sydney Catchment Authority. Available: http://www.environment.nsw.gov.au/resources/water/09112hnrempfintechrpt. pdf [Accessed 16 February 2012].

KULLBERG, A. 1992. Benthic macroinvertebrate community structure in 20 streams of varying pH and humic content. Environmental Pollution, 78, 103-106.

KUMAR, N., STERN, L. W. & ANDERSON, J. C. 1993. Conducting interorganizational research using key informants. Academy of Management Journal, 1633-1651.

LACHENBRUCH, P. A. & GOLDSTEIN, M. 1979. Discriminant analysis. Biometrics, 35, 69-85.

LALKHEN, A. G. & MCCLUSKEY, A. 2008. Clinical tests: sensitivity and specificity. Continuing Education in Anaesthesia, Critical Care & Pain, 8, 221-223.

LARSON, K. L. & SANTELMANN, M. V. 2007. An Analysis of the relationship between residents' proximity to water and attitudes about resource protection. The Professional Geographer, 59, 316-333.

212

LASIAK , T. & UNDERWOOD, A. 2002. Baseline survey of the benthic-invertibrate assemblages associated with fringing mangroves in the Brooklyn region of teh Hawkesbury River [Online]. Manly Vale, Autralia: Water Research Laboratory, The University of New South Walse. Available: http://www.hornsby.nsw.gov.au/media/documents/environment-and- waste/water-catchments/estuary-management/reports/brooklyn/Brooklyn- Estuary-Processes-Study-Vol2-2003.pdf [Accessed 18 June 2012].

LETTENMAIER, D. P., HOOPER, E. R., WAGONER, C. & FARIS, K. B. 1991. Trends in stream quality in the continental United States, 1978–1987. Water Resources Research, 27, 327-339.

LEUNG, P. S. & HOCHMANLAWRENCE, W. 1990. Modeling shrimp production and harvesting schedules. Agricultural Systems, 32, 233-249.

LIERE, K. D. & DUNLAP, R. E. 1978. Moral Norms and Environmental Behavior: An application of Schwartz's norm activation model to Yard Burning. Journal of Applied Social Psychology, 8, 174-188.

LIKERT, R. 1932. A technique for the measurement of attitudes. Archives of psychology, 140, 44-53.

LIU, C., LIN, K. & KUO, Y. 2003. Application of factor analysis in the assessment of groundwater quality in a blackfoot disease area in Taiwan. Science of The Total Environment, 313, 77-89.

LOHMAN, K., JONES, J. R. & PERKINS, B. D. 1992. Effects of nutrient enrichment and flood frequency on periphyton biomass in northern Ozark streams. Canadian Journal of Fisheries and Aquatic Sciences, 49, 1198-1205.

LONERAGAN, N. R. 1999. River flows and estuarine ecosystems: implications for coastal fisheries from a review and a case study of the Logan River, southeast Queensland. Australian Journal of Ecology, 24, 431-440.

LOS, F. & WIJSMAN, J. 2007. Application of a validated primary production model (BLOOM) as a screening tool for marine, coastal and transitional waters. Journal of Marine Systems, 64, 201-215.

LOVLEY, D. R. 1993. Anaerobes into heavy metal: dissimilatory metal reduction in anoxic environments. Trends in Ecology & Evolution, 8, 213-217.

LOW CHOY, D., SUTHERLAND, C., GLEESON, B., DODSON, J. & SIPE, N. 2008. Change and continuity in peri-urban Australia: peri-urban futures and sustainable development. Urban Research Program, Monograph, 4.

LU, L. 2005. The relationship between soft-bottom macrobenthic communities and environmental variables in Singaporean waters. Marine Pollution Bulletin, 51, 1034-1040.

LUNG, W. S. & PAERL, H. W. 1988. Modeling blue-green algal blooms in the lower Neuse River. Water Research, 22, 895-905.

LYDY, M., CRAWFORD, C. & FREY, J. 2000. A comparison of selected diversity, similarity, and biotic indices for detecting changes in benthic-invertebrate community structure and stream quality. Archives of environmental Contamination and Toxicology, 39, 469-479.

213

MACFARLANE, G. & BOOTH, D. 2001. Estuarine macrobenthic community structure in the Hawkesbury River, Australia: Relationships with sediment physicochemical and anthropogenic parameters. Environmental Monitoring and Assessment, 72, 51-78.

MAINSTONE, C. P. & PARR, W. 2002. Phosphorus in rivers--ecology and management. Science of The Total Environment, 282, 25-47.

MALONE, T. C. 1980. The physiological ecology of phytoplankton, Berkeley and New York, University of California Press, 625 p.

MANASRAH, R., RAHEED, M. & BADRAN, M. I. 2006. Relationships between water temperature, nutrients and dissolved oxygen in the northern Gulf of Aqaba, Red Sea. Oceanologia, 48, 237-253.

MANN, H. B. 1945. Nonparametric tests against trend. Econometrica: Journal of the Econometric Society, 245-259.

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, 201-230.

MARSHALL, K. 1980. Reactions of microorganisms, ions and macromolecules at interfaces. In: ELLWOOD, D., HEDGER, J., LATHAM, M., LYNCH, J. & SLATER, J. (eds.) Contemporary microbial ecology. London: Academic Press, 107-136.

MARSHALL, M. 1996. The key informant technique. Family Practice, 13, 92.

MAZLUM, N., OZER, A. & MAZLUM, S. 1999. Interpretation of water quality data by principal components analysis. Turkish Journal of Engineering & Environmental Sciences, 23, 19-26.

MERRIAM-WEBSTER. 2003. Merriam-Webster's Online Dictionary [Online]. Merriam-Webster. Available: http://www.merriam-webster.com/ [Accessed 12 April 2012].

MERRITT, R. W. & CUMMINS, K. W. 2009. An introduction to the aquatic insects of North America, 4th Revised Edition, US, Kendall/Hunt Publishing Company, 1214 p.

MEYER, J. 1997. Stream health: incorporating the human dimension to advance stream ecology. Journal of the North American Benthological Society, 16, 439-447.

MINSHALL, G. W. 1978. Autotrophy in stream ecosystems. BioScience, 767-771.

MOATAR, F. & GAILHARD, J. 2006. Water temperature behaviour in the River Loire since 1976 and 1881. Comptes Rendus Geosciences, 338, 319-328.

MONTGOMERY, S., BARCHIA, I. & WALSH, C. 2012. Estimating rates of mortality in stocks of Metapenaeus macleayi in estuaries of eastern Australia. Fisheries Research, 113, 55-67.

214

MORRISEY, D., UNDERWOOD, A. & HOWITT, L. 1996. Effects of copper on the faunas of marine soft-sediments: an experimental field study. Marine Biology, 125, 199-213.

MOSS, B., BOOKER, I., BALLS, H. & MANSON, K. 1989. Phytoplankton distribution in a temperate floodplain lake and river system. I. Hydrology, nutrient sources and phytoplankton biomass. Journal of Plankton Research, 11, 813.

MURRAY, B. 1990. The life and times of the Enterococcus. Clinical Microbiology Reviews, 3, 46.

MUSTOW, S. 2002. Biological monitoring of rivers in Thailand: use and adaptation of the BMWP score. Hydrobiologia, 479, 191-229.

MUYLAERT, K., SANCHEZ-PEREZ, J. M., TEISSIER, S., SAUVAGE, S., DAUTA, A. & VERVIER, P. 2009. Eutrophication and its effect on dissolved Si concentrations in the Garonne River (France). Journal of Limnology, 68, 368- 374.

N.R.M.M.C., N. H. M. R. C. 2011. Australian Drinking Water Guidelines Paper 6 National Water Quality Management Strategy [Online]. Canberra.: Commonwealth of Australia. Available: http://www.nhmrc.gov.au/_files_nhmrc/publications/attachments/eh52_aust_ drinking_water_guidelines_1.pdf [Accessed 28 May 2012].

NAIMAN, R., MAGNUSON, J., MCKNIGHT, D. & STANDFORD, J. 1995. The freshwater imperative: A research agenda, Washington, DC, Island Press, 161 p.

NARUMI, D., KONDO, A. & SHIMODA, Y. 2009. Effects of anthropogenic heat release upon the urban climate in a Japanese megacity. Environmental research, 109, 421-431.

NCOSS 2005. Submission to the department of planning on managing Sydney's growth centres. Surrey Hill: Available: http://www.ncoss.org.au/bookshelf/urban_development/submissions/growth- centres-Oct05.pdf [ 2 March 2011].

NEAL, C., HILTON, J., WADE, A. J., NEAL, M. & WICKHAM, H. 2006. Chlorophyll-a in the rivers of eastern England. Science of the Total Environment, 365, 84- 104.

NELSON, A. C. & DUEKER, K. J. 1990. The exurbanization of America and its planning policy implications. Journal of Planning Education and Research, 9, 91-100.

NHMRC. 2008. Guidelines for Managing Risks in Recreational Water [Online]. Canberra, ACT: Australian Government National Health and Medical Research Council. Available: http://www.nhmrc.gov.au/_files_nhmrc/publications/attachments/eh38.pdf [Accessed 12 January 2012].

NNANE, D. E., EBDON, J. E. & TAYLOR, H. D. 2011. Integrated analysis of water quality parameters for cost-effective faecal pollution management in river catchments. Water Research, 45, 2235–2246.

215

NORRIS, R. H., DYER, F., HAIRSINE, MARK KENNARD, SIMON LINKE, LINDA MERRIN, ARTHUR READ, WAYNE ROBINSON, CHRIS RYAN, SCOTT WILKINSON & WILLIAMS, D. 2007. Assessment of River and Wetland Health: A Framework for Comparative Assessment of the Ecological Condition of Australian Rivers and Wetlands. Canberra: National Water Commision.

NORRIS, R. H. & THOMAS, M. C. 1999. What is river health? Freshwater Biology, 41, 197-209.

NSW DEPARTMENT OF PLANNING. 2010. Metropolitan Plan for Sydney. Available: http://metroplansydney.nsw.gov.au/Portals/0/pdf/METRO2036_COMPLETE. pdf [Accessed 15 February 2012].

O'HIGGINS, T. & WILSON, J. 2005. Impact of the river Liffey discharge on nutrient and chlorophyll concentrations in the Liffey estuary and Dublin Bay (Irish Sea). Estuarine, Coastal and Shelf Science, 64, 323-334.

ORLOB, G. T. 1956. Viability of sewage bacteria in sea water. Sewage and Industrial Wastes, 28, 1147-1167.

OUYANG, Y. 2005. Evaluation of river water quality monitoring stations by principal component analysis. Water Research, 39, 2621-2635.

OUYANG, Y., NKEDI-KIZZA, P., WU, Q., SHINDE, D. & HUANG, C. 2006. Assessment of seasonal variations in surface water quality. Water Research, 40, 3800-3810.

PAINULY, A. S., SHRESTHA, S., HACKNEY, P. & KABBES, K. C. Distribution of metals and speciation of sediment grabs in Lake Burragorang in Sydney, Australia. 2007 Proceedings of the 2007 World Environmental and Water Resources Congress, 2007, May 15-19, Tampa, Florida. ASCE, 1-9.

PARAJULI, P. B., MANKIN, K. R. & BARNES, P. L. 2009. Source specific fecal bacteria modeling using soil and water assessment tool model. Bioresource Technology, 100, 953-963.

PARINET, B., LHOTE, A. & LEGUBE, B. 2004. Principal component analysis: an appropriate tool for water quality evaluation and management-application to a tropical lake system. Ecological Modelling, 178, 295-311.

PEARSON, M. & NORRIS, R. H. 1996. The effect of habitat-specific sampling on biological assessment of water quality. Freshwater Biology, 36, 419-434.

PEARSON, T. & ROSENBERG, R. 1978. Macrobenthic succession in relation to organic enrichment and pollution of the marine environment. Oceanography and Marine Biology: An Annual Review, 16.

PEJAMAN, A. H., BIDHENDI, N. G. R., KARBASSI, A. R., MEHRDADI, N. & BIDHENDI, E. M. 2009. Evaluation of spatial and seasonal variation in surface water quality using multivariate statistical techniques. International of Environmental Science and Technology, 6, 467-476.

PEKÁROVÁ, P., MIKLÁNEK, P., HALMOVÁ, D., ONDERKA, M., PEKÁR, J., KUCAROVA, K., LIOVA, S. & SKODA, P. 2011. Long-term trend and

216

multiannual variability of water temperature in the pristine Bela River basin (Slovakia). Journal of Hydrology, 400, 333–340.

PERKINS, R. & UNDERWOOD, G. 2000. Gradients of chlorophyll a and water chemistry along an eutrophic reservoir with determination of the limiting nutrient by in situ nutrient addition. Water Research, 34, 713-724.

PETER, D., HARRIS, J., HILLMAN, T. & WALKER, K. 2008. SRA Report 1: A Report on the Ecological Health of Rivers in the Murray-Darling Basin, 2004- 2007 [Online]. Canberra: Prepared by the Independent Sustainable Rivers Audit Group for the Murray-Darling Basin Ministerial Council. Available: http://www2.mdbc.gov.au/SRA/river_health_check_-_sra_report_one.html [Accessed 12 May 2012].

PETERSON, T. C. 2003. Assessment of urban versus rural in situ surface temperatures in the contiguous United States: No difference found. Journal of Climate, 16, 2941-2959.

PINTO, U. & MAHESHWARI, B. 2011. River Health Assessment in Peri-urban Landscapes: An Application of Multivariate Analysis to Identify the Key Variables. Water Research, 45, 3915-3924.

POSTEL, S. & CARPENTER, S. 1997. Freshwater ecosystem services. Nature’s Services: Societal Dependence on Natural Ecosystems, 195–214.

PRATI, L., PAVANELLO, R. & PESARIN, F. 1971. Assessment of surface water quality by a single index of pollution. Water Research, 5, 741-751.

PRIMPAS, I., TSIRTSIS, G., KARYDIS, M. & KOKKORIS, G. D. 2010. Principal component analysis: Development of a multivariate index for assessing eutrophication according to the European water framework directive. Ecological Indicators, 10, 178-183.

PRYOR, R. J. 1968. Defining the rural-urban fringe. Social Forces, 47, 202-215.

QIN, D., FISHER, I. & MAHESWARAN, S. 1995. Modelling Nepean River water quality due to proposed effluent and dam releases. Environment International, 21, 591-596.

QUEENBOROUGH, S. A. & COOKE, I. R. 2010. Do humans count in ecology? Quantitative methods can link socio-economics and ecology. Bulletin of the British Ecological Society, 41, 57-58.

RAHMAN, M. & SALBE, I. 1995. Modelling impacts of diffuse and point source nutrients on the water quality of South Creek catchment. Environment International, 21, 597-603.

RAPPORT, D. 1989. What constitutes ecosystem health ? Perspectives in Biology and Medicine, 33, 120-132.

RAPPORT, D., REGIER, H. & HUTCHINSON, T. 1985. Ecosystem behavior under stress. American Naturalist, 617-640.

RAVEN, J. 1974. Carbon dioxide fixation. In: STEWART, W. (ed.) Algal physiology and biochemistry. Oxford, UK: Blackwell Scientific, 434-449 p.

217

REAVIE, E. D., JICHA, T. M., ANGRADI, T. R., BOLGRIEN, D. W. & HILL, B. H. 2010. Algal assemblages for large river monitoring: Comparison among biovolume, absolute and relative abundance metrics. Ecological Indicators, 10, 167-177.

RELPH, E. 1997. Sense of place. Ten geographic ideas that changed the world, 205-226.

REYNOLDS, C. 1988. Potamoplnakton: paradigms, paradoxes, prognoses. In: ROUND, E. (ed.) Algae in the aquatic environment. Bristol, UK: Biopress Ltd., 285-311 p.

REYNOLDS, C. 2000. Hydroecology of river plankton: the role of variability in channel flow. Hydrological Processes, 14, 3119-3132.

REYNOLDS, C. S. 1990. The ecology of freshwater phytoplankton, Cambridge,UK, Cambridge University Press, 384 p.

RITTENBERG, S. C., MITTWER, T. & IVLER, D. 1958. Coliform Bacteria in Sediments Around Three Mine Sewage Outfalls. Limnology and Oceanography, 3, 101-108.

ROSENBERG, D. & RESH, V. 1993. Freshwater biomonitoring and benthic macroinvertebrates. In: ROSENBERG, D. & RESH, V. (eds.) Introduction to freshwater biomonitoring and benthic macroinvertebrates. New York, USA: Chapman and Hall, 1-9 p.

ROSLEV, P., BJERGBAEK, L. A. & HESSELSOE, M. 2004. Effect of oxygen on survival of faecal pollution indicators in drinking water. Journal of Applied Microbiology, 96, 938-945.

ROWLING, K., HEGARTY, A. M. & IVES, M. 2008-2009. Status of Fisheries Resources in NSW [Online]. Cronulla: Fisheries Research Centre of Excellence. Available: http://www.dpi.nsw.gov.au/__data/assets/pdf_file/0009/385389/WF_2011_O utput-1866_Rowling-and-Hegarty_Poster-re-Status-Report_POSTER.pdf [Accessed 7 May 2012].

RUELLO, N. V. 1973a. Burrowing, feeding, and spatial distribution of the school prawn Metapenaeus macleayi (Haswell) in the Hunter River region, Australia. Journal of Experimental Marine Biology and Ecology, 13, 189-206.

RUELLO, N. V. 1973b. The influence of rainfall on the distribution and abundance of the school prawn Metapenaeus macleayi in the Hunter River region (Australia). Marine Biology, 23, 221-228.

RUELLO, N. V. 1977. Migration and stock studies on the Australian school prawn Metapenaeus macleayi. Marine Biology, 41, 185-190.

RYGG, B. 1985. Effect of sediment copper on benthic fauna. Marine ecology progress series. Oldendorf, 25, 83-89.

SAMPSON, R. W., SWIATNICKI, S. A., OSINGA, V. L., SUPITA, J. L., MCDERMOTT, C. M. & KLEINHEINZ, G. 2006. Effects of temperature and sand on E. coli survival in a northern lake water microcosm. Journal of Water and Health, 4, 389-394.

218

SANDERS, H. L. 1958. Benthic studies in Buzzards Bay. I. Animal-sediment relationships. Limnology and Oceanography, 3, 245-258.

SCA. 2010. Water quality at Hawkesbury-Nepean River sites [Online]. Sydney Catchment Authority. Available: http://www.sca.nsw.gov.au/publications/awqmr08/stream/hnriver [Accessed 7 May 2012].

SCENIC QUALITY 1996. Hawkesbury-Nepean scenic quality study, Sydney, Department of Urban Affairs and Planning, 123 P.

SCHANZ, F. 1985. Vertical light attenuation and phytoplankton development in Lake Zurich. Limnology and Oceanography, 299-310.

SCHOFIELD, N. 2009. Australia wide assessment of river health. AWA Speciality Conference:Rivers and Reservoirs. 17-18 November, Canberra: Australian Water Association.

SCHOFIELD, N. & DAVIES, P. 1996. Measuring the health of our rivers. Water (Melbourne), 23, 39-43.

SEMHI, K., AMIOTTE SUCHET, P., CLAUER, N. & PROBST, J. 2000. Dissolved silica in the Garonne River waters: changes in the weathering dynamics. Environmental Geology, 40, 19-26.

SHEN, P. P., LI, G., HUANG, L. M., ZHANG, J. L. & TAN, Y. H. 2011. Spatio- temporal variability of phytoplankton assemblages in the Pearl River estuary, with special reference to the influence of turbidity and temperature. Continental Shelf Research, 31, 1672-1681.

SHIN, P. & LAM, W. 2001. Development of a marine sediment pollution index. Environmental Pollution, 113, 281-291.

SHRESTHA, S. & KAZAMA, F. 2007. Assessment of surface water quality using multivariate statistical techniques: A case study of the Fuji river basin, Japan. Environmental Modelling and Software, 22, 464-475.

SIMEONOV, V., STRATIS, J., SAMARA, C., ZACHARIADIS, G., VOUTSA, D., ANTHEMIDIS, A., SOFONIOU, M. & KOUIMTZIS, T. 2003. Assessment of the surface water quality in northern Greece. Water Research, 37, 4119– 4124.

SIMMONS, B. & SCOTT, J. 2006. The River has Recorded the Story, Living with the Hawkesbury River, Sydney, NSW, Australia. A History of Water: Water control and river biographies/edited by T. Tvedt and E. Jakobsson, 253.

SIMONOVSKI, J., OWENS, C. & BIRCH 2, G. 2003. Heavy metals in sediments of the Upper Hawkesbury-Nepean River. Australian Geographical Studies, 41, 196-207.

SINHA, S. P. 1993. Instant encyclopaedia of geography, Ned Delhi, India, Mittal Publications, 477 p.

SINTON, L. W., HALL, C. H., LYNCH, P. A. & DAVIES-COLLEY, R. J. 2002. Sunlight inactivation of fecal indicator bacteria and bacteriophages from

219

waste stabilization pond effluent in fresh and saline waters. Applied and environmental microbiology, 68, 1122.

SMAYDA, T. J. 1969. Experimental observations on the influence of temperature, light and salinity on cell division of the marine diatom Detonula confervacea. Journal of Phycology, 5, 150-157.

SMITH, C. M. 1987. Sediment, phosphorus, and nitrogen in channelised surface run off from a New Zealand pastoral catchment. New Zealand Journal of Marine and Freshwater Research, 21, 627-639.

SMITH, J. W. & MOORE, R. L. 2011. Perceptions of Community Benefits from Two Wild and Scenic Rivers. Environmental Management, 1-14.

SMITH, N. J. H. 1999. The Amazon River forest: a natural history of plants, animals, and people, New York, Oxford University Press, 313 p.

SMITH, V. H. 1983. Low nitrogen to phosphorus ratios favor dominance by blue- green algae in lake phytoplankton. Science, 221, 669.

SMITH, V. H. 2003. Eutrophication of freshwater and coastal marine ecosystems a global problem. Environmental Science and Pollution Research, 10, 126- 139.

SMOL, J. P. 1992. Paleolimnology: an important tool for effective ecosystem management. Journal of Aquatic Ecosystem Stress and Recovery 1, 49-58.

SNELGROVE, P. V. R. & BUTMAN, C. A. 1994. Animal sediment relationships revisited-cause versus effect. Oceanography and Marine Biology, 32, 111- 177.

SOLLER, J. A., SCHOEN, M. E., BARTRAND, T., RAVENSCROFT, J. E. & ASHBOLT, N. J. 2010. Estimated human health risks from exposure to recreational waters impacted by human and non-human sources of faecal contamination. Water Research, 44, 4674-4691.

SOLOMON, S., D. QIN, M. MANNING, Z. CHEN, M. MARQUIS, K.B. AVERYT, M. TIGNOR AND & MILLER, H. L. 2007. Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change, Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA, 996 p.

SØNDERGAARD, M. 2009. Redox Potential. In: GENE E. LIKENS (ed.) Encyclopedia of Inland Waters. Amsterdam; Boston: Elsevier, 703 p.

SOUTH, G. R. & WHITTICK, A. 1987. Introduction to phycology, Oxford, UK, Blackwell Scientific Publications, 341 p.

SPCC 1983. Water quality in the Hawkesbury-Nepean River- A study Recommendations. Sydney, NSW: State Pollution Control Commission, 207 p.

SPCC 1984. Sand and gravel extraction in the Upper Hawkesbury River. Sydney, NSW: State Pollution Control Commission, 18 p.

220

SPRAGUE, J., ELSON, P. & SAUNDERS, R. 1965. Sublethal copper-zinc pollution in a salmon river-a field and laboratory study. International Journal of Air and Water Pollution, 9, 531-543.

ST-HILAIRE, A., BRUN, G., COURTENAY, S., OUARDA, T., BOGHEN, A. & BOBÉE, B. 2004. Multivariate analysis of water quality in the Richibucto Drainage basin (New Brunswick, Canada). Journal of the American Water Resources Association, 40, 691-703.

STARK, J. S. 1998. Heavy metal pollution and macrobenthic assemblages in soft sediments in two Sydney estuaries, Australia. Marine and Freshwater Research, 49, 533-540.

STEVENS, M., ASHBOLT, N. & CUNLIFFE, D. 2003. Recommendations to change the use of colidorms as microbial indicators of drinking water quality [Online]. Canberra: National health and Medical Research Council, Australia. Available: http://www.nhmrc.gov.au/_files_nhmrc/publications/attachments/eh32.pdf [Accessed 1 June 2012].

STEVENSON, R. J., BOTHWELL, M. L. & LOWE, R. L. 1996. Algal ecology: freshwater benthic ecosystems, Amsterdam, Boston, Elsevier Inc., 753 p.

STEVENSON, R. J. & SMOL, J. P. 2003. Use of Algae in Environmental Assessments. In: JOHN, D. W. & ROBERT, G. S. (eds.) Freshwater Algae of North America. Burlington: Academic Press, 775-804 p.

STRANG, V. 2005. Water works: agency and creativity in the Mitchell River catchment. The Australian Journal of Anthropology, 16, 366-381.

SUBRAHMANYAM, M. 1964. Fluctuations in the prawn landings in the Godavari estuarine systems. Proceedings of. Indo-Pacific Fish. Coun., 11th Session, 11(2), 44-51.

SUFFET, I., KHIARI, D. & BRUCHET, A. 1999. The drinking water taste and odor wheel for the millennium: beyond geosmin and 2-methylisoborneol. Water Science & Technology, 40, 1-13.

SUTER II, G. 1993. A critique of ecosystem health concepts and indexes. Environmental Toxicology and Chemistry, 12, 1533-1539.

SWALE, E. 1969. Phytoplankton in two English rivers. The Journal of Ecology, 1-23.

SYDNEY WATER. 2007. Preffered Project Report for Replacement Flows Project [Online]. Sydney Water: Available: http://www.sydneywater.com.au/majorprojects/WesternSydney/Replacement FlowsProject/Environmentalassessment.cfm. [Accessed 20 May 2010].

SYDNEY WATER. 2011. Priority Sewerage Program [Online]. Sydney Water. Available: http://www.sydneywater.com.au/majorprojects/Wastewater/PrioritySewerage Program/ [Accessed 20 September 2011].

SYME, G., NANCARROW, B. & MCCREDDIN, J. 1999. Defining the components of fairness in the allocation of water to environmental and human uses. Journal of Environmental Management, 57, 51-70.

221

TEW, K. S., CHOU, W. R. & FANG, L. S. 2006. Phytoplankton Diversity and Community Structure in the Coastal Area of Chang-Hua Industrial Park during 2005. Platax, 31-41.

THIEBAUD, I. & WILLIAMS, R. 2007. Distribution of freshwater macrophytes in the Hawkesbury Nepean River from Warragamba Dam to Wisemans Ferry. Nelson Bay, Australia: NSW Department of Primary Industries Port Stephens Fisheries Centre.

THOMAS, A. 2000. The Easy Septic Guide. Developed by Social Change Media [Online]. Banstown: NSW Department of Local Government. Available: http://www.dlg.nsw.gov.au/DLG/Documents/information/ssguide.pdf [Accessed 23 August 2011].

THOMAS, D. 1974. The urban Fringe : Approacehs and Attitudes (Suburban growth: Geographical processes at the edge of the Western city), Aberdeen, Aberdeen University Press, 17–30 p.

THOMS, M., PARKER, C. & SIMONS, M. 2000. The dispersal and storage of trace metals in the Hawkesbury River valley. In: BRIZGA, S. & FINLAYSON, B. (eds.) River management: The Australasian experience. Sydney, Australia: John Wiley and Sons, 197-219 p.

TREMBLAY, M. 1957. The key informant technique: A nonethnographic application. American Anthropologist, 59, 688-701.

TRENBERTH, K. E. 2004. Climatology (communication arising): rural land-use change and climate. Nature, 427, 213-213.

TUCKER, D., JOHNSTON, C., LEVISTON, Z., JORGENSEN, B. & NANCARROW, B. 2006. Sense of Place: Towards a methodology to value externalities associated with urban water systems. Water for a Healthy Country National Research Flagship, Land and Water Perth, Australia: Commonwealth Scientific and Industrial Research Organisation.

TURAK, E., WADDELL, N. & JOHNSTONE, G. 2001. New South Wales (NSW) Australian River Assessment System (AUSRIVAS) Sampling and Processing Manual [Online]. Sydney, Australia: Environment Protection Authority. Available: http://ausrivas.canberra.edu.au/Bioassessment/Macroinvertebrates/Man/Sa mpling/NSW/NSW_Ausrivas_protocol_Version2_2004.pdf [Accessed 5 May 2011].

TURNER, E. R. & RABALAIS, N. N. 2003. Linking landscape and water quality in the Mississippi River Basin for 200 years. BioScience, 53, 563-572.

TURNER, L. & ERSKINE, W. 2005. Variability in the development, persistence and breakdown of thermal, oxygen and salt stratification on regulated rivers of south-eastern Australia. River Research and Applications, 21, 151-168.

UNDERWOOD, A. 1994. On beyond BACI: sampling designs that might reliably detect environmental disturbances. Ecological Applications, 4-15.

UNDERWOOD, A. & PETERSON, C. 1988. Towards an ecological framework for investigating pollution. Mar. Ecol. Prog. Ser., 46, 227-234.

222

US EPA. 2011. National recommended water quality criteria [Online]. Available: http://water.epa.gov/scitech/swguidance/standards/current/index.cfm#nonpri ority [Accessed 5 October 2011].

VANNOTE, R. L., MINSHALL, G. W., CUMMINS, K. W., SEDELL, J. R. & CUSHING, C. E. 1980. The river continuum concept. Canadian Journal of Fisheries and Aquatic Sciences, 37, 130-137.

VARLEY, I. 2002. Background to environmental flows and other factors affecting releases at dams. Sydney, Australia: Hawkesbury - Nepean River Management Forum.

VARNOSFADERANY, M. N., EBRAHIMI, E., MIRGHAFFARY, N. & SAFYANIAN, A. 2010. Biological assessment of the Zayandeh Rud River, Iran, using benthic macroinvertebrates. Limnologica-Ecology and Management of Inland Waters, 40, 226-232.

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, 3581-3592.

VICTORIA RIVER HEALTH CARD 2009. Securing our rivers for future generations:Victorian River Health Program Report Card 2002-2009, Victoria, Australia, The Department of Sustainability and Environment, 94 p.

VOGEL, S. 1996. Life in moving fluids: the physical biology of flow, New Jersey, USA, Princeton University Press, 484 p.

VUGTEVEEN, P., LEUVEN, R., HUIJBREGTS, M. & LENDERS, H. 2006. Redefinition and elaboration of river ecosystem health: perspective for river management. Hydrobiologia, 565, 289-308.

VYMAZAL, J. 1995. Algae and element cycling in wetlands, CRC Press, Florida, Lewis Publishers Inc., 689 p.

WADE, K., ORMEROD, S. & GEE, A. 1989. Classification and ordination of macroinvertebrate assemblages to predict stream acidity in upland Wales. Hydrobiologia, 171, 59-78.

WADE, T. J., PAI, N., EISENBERG, J. N. S. & COLFORD JR, J. M. 2003. Do US Environmental Protection Agency water quality guidelines for recreational waters prevent gastrointestinal illness? A systematic review and meta- analysis. Environmental Health Perspectives, 111, 1102.

WALSH, C. J., ALLISON, H. R., FEMINELLA, J. W., COTTINGHAM, P. D., GROFFMAN, P. M. & II, R. P. M. 2005. The Urban Stream Syndrome: Current Knowledge and the Search for a Cure. Journal of the North American Benthological Society, 24, 706-723.

WANG, X. & LIU, R. 2005. Spatial analysis and eutrophication assessment for chlorophyll a in Taihu Lake. Environmental Monitoring and Assessment, 101, 167-174.

WARD, T. J. & HUTCHING, P. A. 1996. Effects of trace metals on infaunal species composition in polluted intertidal and subtidal marine sediments near a lead smelter. Mar. Ecol. Prog. Ser., 135, 123-135.

223

WARDELL-JOHNSON, A. 2006. A Sense of Place: valuing landscapes in the Condamine Headwaters, Proceedings of APEN International Conference [Online]. Beechworth, Victoria, Australia: The Regional Institute. Available: http://regional.org.au/au/apen/2005/3/2770_wardelljohnson.htm [Accessed 5 March 2011].

WARNER, R. F. 1994. A theory of channel and floodplain responses to alternating regimes and its application to actual adjustments in the Hawkesbury River, Australia. In: KIRKBY, M. J. (ed.) Process Models and Theoretical Geomorphology. New York: John Wiley & Sons Ltd, 173-200 p.

WARWICK, R. & CLARKE, K. 1991. A comparison of some methods for analysing changes in benthic community structure. Journal of the Marine Biological Association of the United Kingdom, 71, 225-244.

WATERS, T. F. 1995. Sediment in streams. Sources, biological effects, and control. American Fisheries Society Monograph 7, 251.

WATERWATCH. 2004. Waterwatch Australia National Technical Manual [Online]. Canberra: Department of the Environment and Heritage. Available: http://www.waterwatch.org.au/ [Accessed 24 January 2012].

WEHR, J. D. & DESCY, J. P. 1998. Use of phytoplankton in large river management. Journal of Phycology, 34, 741-749.

WEISS, C. M. 1951. Adsorption of E. coli on river and estuarine silts. Sewage and Industrial Wastes, 23, 227-237.

WEISS, R. The solubility of nitrogen, oxygen and argon in water and seawater. 1970. Elsevier, 721-735.

WHEELER, P. A., HUYER, A. & FLEISCHBEIN, J. 2003. Cold halocline, increased nutrients and higher chlorophyll off Oregon in 2002. Geophys. Res. Lett, 30, 8021.

WHITTON, B. A. & KELLY, M. G. 1995. Use of algae and other plants for monitoring rivers. Austral Ecology, 20, 45-56.

WHO. 2003. Guidelines for Safe Recreational Water Environments: Vol -1 Coastal and Fresh waters [Online]. Geneva: World Health Organisation. Available: http://www.who.int/water_sanitation_health/bathing/srwe1/en/ [Accessed 12 January 2012].

WICKLUM, D. & DAVIES, R. 1995. Ecosystem health and integrity? Canadian Journal of Botany, 73, 997-1000.

WILLBY, N. J. 2001. Inter relationships between standing crop, biodiversity and trait attributes of hydrophytic vegetation in artificial waterways. Freshwater Biology, 46, 883-902.

WILLIAMS, B. K. 1983. Some observations of the use of discriminant analysis in ecology. Ecology, 64, 1283-1291.

WILLIAMS, D. R. & STEWART, S. I. 1998. Sense of place: An elusive concept that is finding a home in ecosystem management. Journal of Forestry, 96, 18-23.

224

WINTER, T., MALLORY, S., ALLEN, T. & ROSENBERRY, D. 2000. The use of principal component analysis for interpreting ground water hydrographs. Ground Water, 38, 234-246.

WISSINK, G. A. 1962. American cities in perspective: with special reference to the development of their fringe areas, Assen, USA, Van Gorcum, 320 p.

WRIGHT, J. 1995. Development and use of a system for predicting the macroinvertebrate fauna in flowing waters. Austral Ecology, 20, 181-197.

WYBAN, J., WALSH, W. A. & GODIN, D. M. 1995. Temperature effects on growth, feeding rate and feed conversion of the Pacific white shrimp (Penaeus vannamei). Aquaculture, 138, 267-279.

XIA, X., YANG, Z., HUANG, G., ZHANG, X., YU, H. & RONG, X. 2004. Nitrification in natural waters with high suspended-solid content: A study for the Yellow River. Chemosphere, 57, 1017-1029.

XU, F. L., TAO, S., DAWSON, R. & LI, B. G. 2001. A GIS-based method of lake eutrophication assessment. Ecological Modelling, 144, 231-244.

YENTSCH, C. 1965. Distribution of chlorophyll and phaeophytin in the open ocean. Deep Sea Research, 12, 653-666.

YOON, S. J. & KIM, J. H. 2001. Is the Internet more effective than traditional media? Factors affecting the choice of media. Journal of Advertising Research, 41, 53-60.

YOUNG, R., TOWNSEND, C. & MATTHAEI, C. 2004. Functional indicators of river ecosystem health–an interim guide for use in New Zealand. Prepared for Ministry for the Environment. Cawthron Report, 870, 54.

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, 254-271.

ZANDER, K. & STRATON, A. 2010. An economic assessement of the value of tropical river ecosystem services:Heterogenous preferences among Aboriginal and non-Aboriginal Australians. Ecological Economics, 69, 2417- 2426.

ZELEZNY, L. C., CHUA, P. P. & ALDRICH, C. 2000. New ways of thinking about environmentalism: Elaborating on gender differences in environmentalism. Journal of Social Issues, 56, 443-457.

ZHANG, Q., SHI, X., HUANG, B., YU, D., ÖBORN, I., BLOMBÄCK, K., WANG, H., PAGELLA, T. & SINCLAIR, F. 2007. Surface water quality of factory-based and vegetable-based peri-urban areas in the Yangtze River Delta region, China. Catena, 69, 57-64.

ZHOU, F., LIU, Y. & GUO, H. 2007. Application of multivariate statistical methods to water quality assessment of the watercourses in Northwestern new territories, Hong Kong. Environmental Monitoring and Assessment, 132, 1- 13.

225

ZHOU, L., DICKINSON, R. E., TIAN, Y., FANG, J., LI, Q., KAUFMANN, R. K., TUCKER, C. J. & MYNENI, R. B. 2004. Evidence for a significant urbanization effect on climate in China. Proceedings of the National Academy of Sciences of the United States of America, 101, 9540-9544.

226

APPENDICES

Appendix - A. Understanding the meaning of river health for a community: perspectives from the peri-urban region of the Hawkesbury–Nepean Catchment, Australia.

(PINTO, U., MAHESHWARI, B., SHRESTHA, S. & MORRIS, C. 2012. Understanding the meaning of river health for a community: perspectives from the peri-urban region of the Hawkesbury-Nepean Catchment, Australia. Water Policy, 1-18.)

Appendix - B. Analysis of long-term water quality for effective river health monitoring in peri-urban landscapes - A case study of the Hawkesbury-Nepean River System in NSW, Australia.

(PINTO, U., MAHESHWARI, B. & OLLERTON, R. 2012. Analysis of long-term water quality 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, 1-19.)

Appendix - C. Impacts of water quality on the harvest of school prawn (Metapenaeus macleayi) in a peri-urban river system.

(PINTO, U. & MAHESHWARI, B. 2012. Impacts of Water Quality on the Harvest of School Prawn (Metapenaeus macleayi) in a Peri-Urban River System. Journal of Shellfish Research, 31, 847-853.)

Appendix - D. River health assessment in peri-urban landscapes : An application of multivariate analysis to identify the key variables.

(PINTO, U. & MAHESHWARI, B. 2011. River Health Assessment in Peri-urban Landscapes: An Application of Multivariate Analysis to Identify the Key Variables. Water Research, 45, 3915-3924.)

Appendix - E. Modelling eutrophication and microbial risk in peri-urban river systems using discriminant function analysis.

(PINTO, U., MAHESHWARI, B., SHRESTHA, S. & MORRIS, C. 2012. Modelling eutrophication and microbial risks in peri-urban river systems using discriminant function analysis. Water Research, 46, 6476-88.)

Appendix - F. Poster presented for the Integrated Water Management Conference held in Germany 2011.

Appendix - G. Survey cover letter and questionnaire.

The Hawkesbury–Nepean River Health Survey

BACKGROUND

1. What is the name of the suburb and the postcode you currently live in?

Suburb: Postcode:

2. Are you

Male Female

3. Which of the following best describes your age?

18-25 40-60

25-40 Over 60

4. How long have you been living in your current residence?

Less than a year 5 -10 Years

1-5 Years More than 10 years

5. Can you tell us your residency status in Australia? (OPTIONAL)

I was born and raised in Australia

I have migrated to Australia

I am a temporary resident / visitor / student

6. Can you rank your knowledge about the following aspects of the Hawkesbury- Nepean River? (0-None, 3-Excellent)

0 1 2 3

Present condition of the river

The use / role of the river in the region

Problems / issues faced by the river

7. Can you tell us why this river is important to you?

This river is important to me because,

RIVER HEALTH & ACTIVITIES

8. In your own words, how would you describe a ‘healthy river’?

9. In your opinion, what are the most obvious signs of an ‘unhealthy river’?

Sign:1

Sign:2

Sign:3

10. Now, can you share your thoughts about the current health of the Hawkesbury- Nepean River compared to your any previous experiences with this river? (Please select one of the five responses)

It has improved It is now deteriorating Do not know

It has not improved Nothing has changed

11. If you want to inform ordinary citizens about the current status of river health, what would be the most effective way of doing it? (You can select more than one option)

Newsletter / Flyer Newspaper notice SMS / E-mail alert

Display boards Local radio Other announcement

12. Do you think any of your current activities are having an impact on the health of Hawkesbury-Nepean River? (For example, discharge of nutrient rich stormwater from your garden)

Yes No Do not know

13. Would you be willing to reduce that impact if you knew about it more?

Yes No

14. In your view, what river dependant activities of yours or others in the community will be most affected if the present river flow reduced to a very low level in the next 20 years? (Please list 3 major activities that will be affected in the order of increasing importance)

Activity:1

Activity:2

Activity:3

15. If the river health improves, what activities you would like to be involved more in the future?

Activity:1

Activity:2

Activity:3

RIVER VALUES & STRESSORS

16. In your view, how would you see the importance of the following values of Hawkesbury-Nepean River? (1-Low importance, 4-High importance)

1 2 3 4

Historic value (i.e. historic events & landmarks, European settlement)

Cultural / Spiritual Value (i.e. traditional Aboriginal cultural values, feeling of being close to nature)

Recreational value (i.e. water sports, amateur fishing)

Aesthetic value (i.e. a place to spend a day, visual amenity)

Economic value (i.e. prawns, sand extraction, water for market gardens and drinking)

Biodiversity value (i.e. habitat for various aquatic plants & animals breeding grounds for seasonal fish, refuge for migrant birds)

Others______

17. In your view, how much influence do the following factors have on the health of the Hawkesbury-Nepean River? Please respond to as many factors as you can. (1-Low influence, 4-High influence)

1 2 3 4

Discharge of nutrients with effluent from sewage treatment plants

Extraction of water for irrigation

Extraction of water for drinking

Recreational activities (e.g., boating, skiing, recreational fishing)

Run-off (e.g., residential areas, car parks, roads, catchment)

Sand and gravel extraction (e.g., quarrying)

Invasion by exotic animals and plants

Commercial fishing / prawn trawling / oyster aquaculture

Climate change

Others______

18. In your opinion, how important are the following indicators showing the health of the Hawkesbury-Nepean River?(1-Low importance, 4-High importance)

1 2 3 4

Algal blooms

Noxious odours

Not enough water flowing in the river

Weeds in the river and river banks

Clarity of water

Bank erosion

Deposition of sand

Loss of fish, prawns, oysters etc.

Presence of floating material (e.g., oil, scum, rubbish)

Other______

RIVER MANAGEMENT

19. Who would be the most capable in looking after the health of the Hawkesbury- Nepean River? (1-Least effective, 4-Most effective)

1 2 3 4

Nature (e.g., the river will keep healthy by itself)

Community groups (e.g., LandCare, StreamWatch)

Government agencies (e.g. local, state and federal government)

Individuals (e.g., like yourself)

Others______

20. Are you satisfied with the current measures taken by the river management authorities to improve the health of the Hawkesbury-Nepean River for future generation?

Yes No

I do not know

21. Based on your knowledge, can you tell us, two or three most practical actions that can be implemented to improve the health of the Hawkesbury- Nepean River for future generations?

Action:1

Action:2

Action:3

22. In your view, how would you contribute to improve the health of the Hawkesbury- Nepean River if an opportunity arises?

Not Yes No sure

Contribute financially (e.g., pay a levy with water bill)

Join a community group to improve river health

Take measures to reduce my water usage

Influence relevant authorities to work toward improving river health

Influence the state and federal members for some positive action

Others______

YOUR UNDERSTANDING

23. When completing this survey, my answers were based on, (You can select more than one option)

I live close to the river

My close association with the river (e.g., engaged in fishing, boating, picnicking, farming, travelling on ferry)

The knowledge and understanding gained from media (e.g., TV, news papers)

Other______

24. When completing this survey, which section(s) of the river did you have in mind? (Please use the schematic diagram below)

Section -1

Section -2

Section -3

Section -4

Section -5

Section -6

25. Do you have any other comments?

26. Would you like to receive an electronic or a printed copy of the survey results? (OPTIONAL)

YES

E-mail:

Postal Post Code Address

Thank You For Your Participation

Appendix - H. Newspaper announcements about the survey. Two examples from the Hawkesbury Gazette and Hornsby and upper North shore Advocate local newspapers.