Effect of Catchment Forest Cover on Macroinvertebrate Community Structure in Streams of Fiji

EFFECT OF CATCHMENT FOREST COVER ON MACROINVERTEBRATE COMMUNITY STRUCTURE IN STREAMS OF FIJI

Bindiya Rashni

Cover photo: Acochlidium fijiense (Haynes & Kenchington, 1991), the endemic and threatened shell-less aquatic gastropod found in Vucinivola stream, Nakorotubu district of Ra, Viti Levu.

A thesis submitted in fulfillment of the requirements for the degree of Master of Science of Marine Science

School of Marine Studies Faculty of Science, Technology and Environment The University of the South Pacific July, 2014

DEDICATION

To my dearest parents (my dad, Late Mr. Sathanand and my mum, Manjula Wati)

My siblings Shyam Sajay Nand, Reshmi Sudha Nandni and Ambika Nand for their love and continuous support

ACKNOWLEDGEMENTS

“Every creation of God has a message for the human beings who were made in his image” Late Mr. Sathanand (my dad)

& his words are verified by the holy Bhagavad Gita

“As rivers flow into the ocean but cannot make the vast ocean overflow, so flow the streams of the sense-world into the sea of peace”

Bhagavad-Gita 2:70

I had just graduated with a Bachelors degree in Marine Science (BScMS) in 2009 and was in a dilemma of whether to work or get enrolled in a postgraduate program. As usual, I prayed to almighty God and asked for guidance. Out of the blue, I received an e-mail from Dr. Susanne Pohler saying that a man named Prof. William Aalbersberg from the Institute of Applied Science (IAS) wanted to talk to me. Surprisingly, during the three years at Marine Campus, I was not aware that IAS existed in the same campus and that Prof. Aalbersberg was the director of IAS. I clearly remember the first time I met Prof. Aalbersberg. I entered the main IAS office and introduced myself to him. While discussing about my employment at IAS, he bent down slightly and picked up something from the floor next to my feet. I thought he had dropped some valuable of his but it was just two pieces of crumpled paper probably left around carelessly by someone. I felt a bit embarrassed for having younger eyes and not noticing it but that day, I did witness the action behind one of the wise quotes by Mahatma Gandhi “be the change you wish to see in this world”. As I walked out of the IAS office, I was content I came to the right place and my journey here would be a productive and memorable one.

First and foremost I would like to thank all the lecturers (late Mr. Johnson Seeto, Dr. Susanne Pohler, Dr. Vina Ram Bidesi and Dr. Milika Naqasima-Sobey) who voted for me during the IAS meeting for selection of an Indo-Fijian student to expand FLMMA

iii work in the Indo-Fijian community of Korolevu-i-wai district. This has been a great learning experience.

To my supervisory team, Prof. Aalbersberg, Dr. Alison Haynes, James Comley, Marika Tuiwawa and Dr. Brodie, thank you very much for your guidance, assistance and support throughout my MSc journey. To Prof. Aalbersberg, sir I am highly grateful for the graduate assistantship opportunity with IAS and allowing me to work with two units, the Environment Unit and the South Pacific Regional Herbarium. If it wasn’t for this GAship offer, I would have never discovered my passion for freshwater macroinvertebrate taxonomy. Special thanks for trusting me with an entire new field and providing a research topic despite my marine background.

To Dr. Haynes, I salute you for the immense freshwater macroinvertebrate work that you have done locally and regionally. You have laid a strong foundation for these challenging little critters that reflect valuable message for the people of Oceania. Thank you so much for lending me your doctrines of freshwater macroinvertebrates. Being your apprentice was a wonderful experience. Like I always say, working with you in lab is like working with a living legend. Thank you also for your comments on my results and discussion chapters.

To James Comley, hats off to you the GIS and Stats guru for IAS, special thanks for taking out time from your daily work schedule as well as weekends to assist me with Permanova analysis. Thank you for continuously believing in me and monitoring my progress as well as reviewing my chapter drafts.

To Dr. Sarah Pene, my life guard, special thanks for reviewing all my chapters and making sense out of my crappy paragraphs. Thanks heaps for the words of encouragement and support and keeping me interested in my thesis.

As the Lao Tzu saying goes “nature does not hurry, yet everything is accomplished”, huge thanks to the father of Herbarium, Mr. Marika Tuiwawa and my Herbarium family (Alivereti Naikatini, Dr. Sarah Pene, Manoa Maiwaqa, Lekima Copeland, Tokasaya Cakacaka, Apaitia Liga, Hilda Waqa-Sakiti, Mereia Katafono and Siteri Tikoca) for a wonderful journey to some of the most heavenly places in mountainous terrains of Fiji.

The five year field experience has been a blessing indeed and molded me into a wiser and better researcher. Biodiversity team, you rock. To my field mate and brother, Lekima Copeland, special thanks for taking care of me during all the field trips.

To my Environment Unit family (Dr. Bale Tamata, Alifereti Tawake, Ron Vave, Semisi Meo, Rusiate Ratuniata, Ron Simpson, Hans Karl Wendt, Laitia Tamata, Joji Sivo, Fulori Nainoca, Leigh-Anne Buliruarua, Albert Whippy and Tomasi Delana) huge thanks for the support and encouragement all the way. To Isimeli Logan, huge thanks for assisting me during my field work in Kadavu and being my older brother. To Hans Karl Wendt, thank you so much for such awesome study site maps.

To Apisai Bogiva, Laitia and Hans and Rusiate, my brotherly circle, thank you for taking care of me and all your advices during the challenging times. This journey would have been tougher without you guys.

To the IAS administration team, Aisha Khan, Rina Segran, Reshma Prasad and Loata Qorovarua, thanks heaps for your assistance with administration of my research vote code. To our creative graphic artist, Mere Brown, special thanks for assisting with the poster design for my thesis research work.

To the analytical and microbiology team of IAS, huge thanks for analyzing and providing data for my field site water samples. Your assistance is highly appreciated.

My sincere thanks to the Institute of Applied Science for funding this research. Huge thanks to the Ra Provincial Office for allowing me to carry out my field work in Nakorotubu district. To the people of Naocobau, Saioko and Nabukadra villages, huge thanks for your hospitality and support. Special thanks to my field assistants, Josua Vunivesi, Samu Bale, Apenisa dau, Rai Malani and Laisiasa Cava. My sincere gratitude to Jona Nacayatani and Laisiasa for driving us to and from the study sites. Thanks for safe driving. Also thanks to Samu Copeland and Pita Koroi for all the laughter and fun despite the tiring hike.

Huge thanks to the Kadavu Provincial Office for allowing me to carry out my field work in Nakasaleka district. To the Tui Nakasa, my deepest appreciation for your

v support throughout my stay in the Nakasaleka district. To the people of the following villages of Nakasaleka, Kavala bay, Namajiu, Lomanikoro and Nakaugasele, huge thanks for your hospitality and support. To my special field troop ( Ratu Rabici, Radike Isoa,Tomasi Peckham, Jone Bulawa, Martin Liwaliwa, Vilikesa Bulewa, Wani Seru, Jovesa Rokona, Isoa Coalala, Apakuki Soata, Manoa Meli and Mosese Bete), thank you for such a fun-filled field work. Also thanks you for your assistance with carrying my field gears.

To the marine studies team at the University of the South Pacific, Shiv Sharma and Jone Lima, thanks heaps for your assistance with the laboratory equipments during the four horrific taxonomy months. To the late Mr. Johnson Seeto, special thanks for your guidance and support as well as for your assistance with identification of some crustaceans. To my genius and lovely crustacean specialist, Ms Laura Williams, special thanks for your brains on the confusing shrimp identification. To my CRISP room friends, (Epeli Loganimoce, Kelly Thomas Brown, Jerome Taoi and John Kaituu’, thanks for allowing me to crash at your desks, when I needed a change in write-up space. Also thanks for all the fun and laughter with your paddling team. To our genius Monal Lal, my life saver, special thanks for sending over all the journal articles on time for without it, I would not have achieved my discussion chapter and thesis submission deadline. To Yashika Nand, special thanks for helping me with fine-tuning my study conceptual model. Your assistance is highly appreciated.

Special thanks to Dr. Satish Choy, Prof. Stuart Bunn, Dr. Martin Haase and Prof. Peter Lin for e-mailing me their publications on freshwater ecology. Your assistance is highly appreciated. To the late Dr. Stephen Moore, special thanks for your assistance with macroinvertebrate specimen identification. To Mr. Nick Carter, my friend and mentor, huge thanks for your support and guidance during the process of my thesis data analysis. Thanks also for the writing tips and the journal articles. To Mr. Karuna Reddy, the statistician at USP research office thanks heaps for assisting me understanding of statistical functions and applications. To my friends abroad, Brian Weeks and Qistina Azman and Setoki Tuiteci, thanks for your well wishes and the quick responses in e-

vi mailing journal articles. To my freshwater macroinvertebrate taxonomy buddy, Ilaitia Finau, thanks for my first exposure to invertebrate sampling in Savura creek.

To my seven comrades-in-arms of the MSc troop, Valerie Waqanivavalagi, Payal Maharaj, Siteri Tikoca, Mereia Katafono, Surava Elaisa, Sharon Raj and Ahilya Singh, my deepest thanks for keeping me sane, for all the fun and laughter, for support and encouragement and making sure that I didn’t turn into a lab geek. To my superhero, Mr. Senikarawa Jale Mar, thanks heaps for always going an extra mile when I was not able to.

Lastly, my deepest gratitude to my family, especially my parents for their continuous words of encouragement and assistance in making wise decisions and for keeping me in their prayers daily. To both my brothers and my sister, thank you for your love and support and understanding. To my aunty, Sita Chennaiya, thank you for your support, encouragement and appreciation throughout my MSc journey. To the Waqanivavalagi family, my sincere thanks for your spiritual assistance. This journey would not have been same without you all.

धयवादं! Vinaka vakalevu! Thank you all. Stay blessed and beautiful!

vii

ABBREVIATIONS & ACRONYMS

ANZECC Australian and New Zealand Environment and Conservation Council APHA American Public Health Association BHP Broken Hill Propriety Company CBAM Community Based Adaptive Management CPOM Coarse Particulate Organic Matter COWRIE Coastal and Watershed Restoration for the Integrity of Island Environments CRISP Coral Reef Initiative in the South Pacific DistLM Distance Based Linear Model FFG Functional Feeding Group FSM Federated States of Micronesia HIES Household Income and Economic Survey IAS Institute of Applied Sciences IBM SPSS International Business Machines Corporation Statistical Package for the Social Sciences IUCN International Union for the Conservation of Nature MCI Macroinvertebrate Community Index MDS Multi-Dimensional Scaling PCO Principal Coordinate Analysis PERMANOVA Permutational Multivariate Analysis of Variance QMCI Semi-quantitative Macroinvertebrate Community Index SIGNAL Stream Invertebrate Grade Number-Average Level TDS Total Dissolved Solids USP University of the South Pacific WANI Water and Nature Initiative

viii

TABLE OF CONTENTS

DECLARATION ...... I

DEDICATION...... ii

ACKNOWLEDGEMENTS ...... III

ABBREVIATIONS & ACRONYMS ...... VIII

LIST OF FIGURES ...... XIII

LIST OF TABLES...... XV

ABSTRACT XVI

CHAPTER 1. INTRODUCTION & LITERATURE REVIEW ...... 1

1.1 Freshwater: an essential resource ...... 1 1.1.1 Global overview ...... 1 1.1.2 Biogeochemical process ...... 1 1.1.3 Biodiversity ...... 1 1.1.4 Human use of freshwater ...... 4 1.1.5 Food security ...... 5 1.1.6 Energy production ...... 6

1.2 Freshwater: A focus on status and threats ...... 6 1.2.1 Pacific freshwater vulnerability ...... 7 1.2.2 Threats to Fiji freshwater systems ...... 8

1.3 Landuse and stream ecosystems: a focus on anthropogenic impacts on stream bio-integrity ...... 10 1.3.1 Logging ...... 10 1.3.2 Dams ...... 11 1.3.3 Road construction ...... 11 1.3.4 Urbanisation ...... 11 1.3.5 Mining ...... 12 1.3.6 Invasive species ...... 13 1.3.7 Agricultural practices ...... 13 1.3.8 Landuse in Fiji: a focus on stream biotic community ...... 14

1.4 Natural effects on biological indicators of stream health: a focus on distribution and species richness ...... 18

1.5 Measuring stream health: a focus on biophysical monitoring ...... 21

1.6 Rationale for this study ...... 24

1.7 Research aims and objectives ...... 26

1.8 Structure of this thesis ...... 26

1.9 Significance of the study ...... 27

CHAPTER 2. METHODOLOGY ...... 28

2.1 Overview ...... 28

2.2 Catchment selection criteria ...... 28

2.3 Study site description ...... 29 2.3.1 Nakasaleka, Kadavu ...... 29 2.3.2 Ra ...... 31

2.4 General description of streams ...... 34

2.5 Sampling period and weather ...... 35

2.6 Data collection ...... 35

2.7 Habitat characteristics ...... 36

2.8 Physicochemical sampling and analysis ...... 38

2.9 Macroinvertebrate sampling and processing ...... 38 2.9.1 Sampling technique one: kick-net sampler ...... 39 2.9.2 Sampling technique two: Surber sampler ...... 40 2.9.3 Specimen identification ...... 41

2.10 GIS ...... 41 2.10.1 Forest Cover analysis ...... 41

2.11 Statistical analysis ...... 42 2.11.1 Macroinvertebrate Metrics ...... 42 2.11.2 Multivariate relationships ...... 43

CHAPTER 3. RESULTS & DISCUSSION OF BIOTIC (MACROINVERTEBRATE COMMUNITY STRUCTURE) VARIABLES .... 46

3.1 Sampling Method: Surber sampler and kick-net macroinvertebrate community composition comparison ...... 46

3.2 Biological communities: macroinvertebrate assemblages ...... 50 3.2.1 Taxonomic Metrics- proportion of species at Group level ...... 50 3.2.2 New additions to freshwater macroinvertebrate taxonomy of Fiji ...... 52 3.2.3 Functional Metrics- proportion of macroinvertebrate in Functional Feeding Groups ...... 54 3.2.4 Macroinvertebrate community variables-diversity measures ...... 58 3.2.5 Principal Coordinate Analysis (PCO) of macroinvertebrate assemblages: both Ra and Kadavu ...... 66 3.2.6 Principal Coordinate Analysis (PCO) of macroinvertebrate assemblages- Kadavu streams ...... 68 3.2.7 Principal Coordinate Analysis (PCO) of macroinvertebrate assemblages- Ra streams ...... 70 3.2.8 General discussion on PCO analysis: macroinvertebrate association with catchment forest cover ...... 72 3.2.9 Patterns in macroinvertebrate assemblage composition and variability ...... 78

CHAPTER 4. RESULTS & DISCUSSION OF ABIOTIC (ENVIRONMENTAL) VARIABLES 82

4.1 Forest Cover analysis ...... 85

4.2 Linking abiotic variables and forest cover to biological communities ...... 88 4.2.1 Abiotic correlation with macroinvertebrate assemblage: both Kadavu and Ra streams ...... 88 4.2.2 Abiotic variable correlation with macroinvertebrate assemblage: Kadavu streams ...... 90 4.2.3 Abiotic variable correlation with macroinvertebrate assemblage: Ra streams 91 4.2.4 Distance based linear models (DistLM) results ...... 93 4.2.5 Best predictory abiotic variable for variation in macroinvertebrate community structure ...... 94

4.3 Methodological Limitations ...... 95

CHAPTER 5. CONCLUSIONS ...... 96

5.1 Recommendations ...... 101 xi

CHAPTER 6. REFERENCES ...... 103

CHAPTER 7. APPENDICES ...... 126

7.1 Habitat Assessment form used during the study ...... 126

7.2 Environmental variable graphs...... 128

7.3 Functional Feeding groups for macroinvertebrates recorded during the study...... 144

7.4 PERMANOVA analyses on macroinvertebrate species diversity measures...... 145

7.5 Macroinvertebrate species and corresponding abundances recorded across Kadavu sampling stations (Surber sampler data) ...... 148

7.6 Macroinvertebrate species and corresponding abundances recorded across Ra sampling stations (Surber sampler data) ...... 150

7.7 Macroinvertebrate species and corresponding abundances recorded across Ra sampling stations (kick-net data) ...... 152

xii

LIST OF FIGURES Figure 1.0: Generic conceptual model for effects of forest cover on the macroinvertebrate community structure in Fiji rural streams. The major landuse types that affect forest cover which directly or indirectly affect habitat features and physiochemical parameters that support macroinvertebrate community structure are shown inside the boxes...... 25 Figure 2.1: Map showing the study sites in Kadavu and Ra province as part of Fiji archipelago ...... 28 Figure 2.2: Sampling stations in Kadavu province ...... 30 Figure 2.3: Sampling stations in Ra province ...... 32 Figure 2.4: kick-net sampling technique ...... 39 Figure 2.5: Surber Sampling technique ...... 40 Figure 3.1: Multi-Dimensional Scaling ordination plot showing macroinvertebrate assemblage similarity using the kick-net and Surber sampler...... 47 Figure 3.2: Total taxa number per Functional Feeding Group across all study sites ..... 54 Figure 3.3: Proportion of total abundance that each functional feeding group made at streams in Kadavu and Ra ...... 55 Figure 3.4: Mean number of species collected across six streams. Error bars represent standard error. Shared letters for each stream denote no significant difference (p>0.05) based on pairwise PERMANOVA tests. The first three bars represent streams from Kadavu province (Korolevu, Naqewa and Savumiri) and the last three bars represent streams from Ra province (Naikawaqa, Taveu and Vucinivola). The level of catchment forest cover (HF= highly forested, MF= moderately forested and LF= least forested) is indicated on each column...... 59 Figure 3.5: Mean number of individuals in the macroinvertebrate community collected across six streams. Error bars represent standard error. Shared letters for each stream denote no significant difference (p>0.05) based on pairwise PERMANOVA tests. The first three bars represent streams from Kadavu province (Korolevu, Naqewa and Savumiri) and the last three bars represent streams from Ra province (Naikawaqa, Taveu and Vucinivola). The level of catchment forest cover (HF= highly forested, MF= moderately forested, LF= least forested) is indicated on each column...... 62 Figure 3.6: Mean Shannon Weiner diversity Index measured across six streams. Error bars represent standard error. Shared letters for each stream denote no significant difference (p>0.05) based on pairwise PERMANOVA tests. The first three bars represent streams from Kadavu province (Korolevu, Naqewa and Savumiri) and the last three bars represent streams from Ra province (Naikawaqa, Taveu and Vucinivola). The level of catchment forest cover (HF= highly forested, MF= moderately forested, LF= least forested) is indicated on each column...... 64

xiii

Figure 3.7: PCO ordination plot showing macroinvertebrate assemblage similarities among streams of Ra and Kadavu province. The vectors in the plot represent the species most responsible for the dissimilarities between sampling sites...... 67 Figure 3.8: PCO ordination plot showing macroinvertebrate assemblage similarities among three streams in Kadavu. The vectors in the plot represent the species most responsible for the dissimilarities between sampling sites...... 69 Figure 3.9: PCO ordination plot showing macroinvertebrate assemblage similarities among streams in Ra. The vectors in the plot represent the species most responsible for the dissimilarities between sampling sites...... 71 Figure 4.1: Percentage forest cover of catchments studied in Kadavu and Ra province 85 Figure 4.2: Nakasaleka study site catchment boundaries with proportion of forested and non-forested land cover...... 86 Figure 4.3: Nakorotubu study site catchment boundaries with proportion of forested and non-forested land cover...... 87 Figure 4.4: PCO ordination plot showing the environmental variable associated with clusters among streams of Ra and Kadavu province. The vectors in the plot represent the abiotic variables most responsible for differences between all sampling sites...... 89 Figure 4.5: PCO ordination plot showing the environmental variable associated with clusters among streams in Kadavu. The vectors in the plot represent the abiotic factors most responsible for distinguishing differences between sampling sites in the Kadavu streams...... 90 Figure 4.6: PCO ordination plot showing the environmental variables associated with clusters among streams in Ra. The vectors in the plot represent the abiotic factors most responsible for distinguishing differences between sampling sites in the Ra streams. .. 92 Figure 7.1: (a-f) Streambed substrate composition across sites surveyed within Ra and Kadavu streams...... 143

xiv

LIST OF TABLES Table 2.2.2: General classification of Functional Feeding Group categories for aquatic macroinvertebrates ...... 42 Table 3.1: Number of macroinvertebrate taxa recorded in each of the taxonomic group across 18 sites ...... 51 Table 3.2: Number of individual macroinvertebrate taxa per stream associated with the PCO analysis...... 72 Table 3.3: Number of individual macroinvertebrate taxa per stream associated with the PCO analysis...... 74 Table 3.4: PERMANOVA results on the macroinvertebrate data examining the effects of season within streams...... 78 Table 4.1: Water quality results for the six streams in comparison to the Fiji freshwater and ANZECC Guidelines ...... 83 Table 4.2: Streams and associated percent forested catchments at study sites...... 85 Table 4.3: PERMANOVA (DistLM) results on the macroinvertebrate data examining the correlation between biotic community and the corresponding environmental variables...... 93 Table 7.1: (a-o): Bar graphs showing physical and chemical variables measured at wet and dry season across study sites ...... 128 Table 7.2: PERMANOVA results on the species richness (S) data within stream. Significant value is in bold...... 145 Table 7.3: PERMANOVA pair-wise test on the effect of stream on species richness (S). Significant values are in bold. P (perm) is p-value based on n where n = unique permutations...... 145 Table 7.4: PERMANOVA results on the total number of individuals (N) data within stream. Significant value is in bold...... 145 Table 7.5: PERMANOVA pair-wise test on the effect of stream on total number of individuals (N). Significant values are in bold. P (perm) is p-value based on n where n = unique permutations...... 146 Table 7.6: PERMANOVA results on the Shannon-Weiner diversity index (Loge) data within stream. Significant value is in bold...... 146 Table 7.7: PERMANOVA pair-wise test on the effect of stream on Shannon-Weiner diversity index (Loge). Significant values are in bold. P (perm) is p-value based on n where n = unique permutations...... 147

ABSTRACT This research aimed at investigating the health of streams draining catchments with varying percentage forest cover in order to better understand the effect that the alteration of catchment forest cover has on the stream macroinvertebrate fauna.

To achieve the aim of the research, three objectives were designed. These comprised of assessing (a) the biotic parameters (macroinvertebrate community structure) and (b) the abiotic parameters (water quality and in-stream habitat features) of stream health and then (c) examining the relationships between catchment forest cover and biotic and abiotic parameters. Six streams (three in each of two catchments) of Ra and Kadavu provinces were sampled for macroinvertebrates, water quality and habitat data during the dry and wet seasons. Macroinvertebrate samples were collected using a Surber sampler and a kick-net. Percentage forest cover of the catchment was determined using Berkeley Image Segmentation and ArcGIS 10.x program. The biotic and abiotic data were analyzed using both univariate and multivariate statistical methods.

A total of 35,334 individual macroinvertebrates were collected and identified to the lowest possible taxonomical level. A total of 140 distinct taxa in 53 families were collected and an additional new record of 14 aquatic families in the phylum Insecta for Fiji was documented. The macroinvertebrate abundance, species richness and diversity graphs did not show any apparent relationship with catchment forest cover.

The results of PCO (Principal Coordinate Analysis) analyses associating abiotic variables with macroinvertebrate community assemblage in Kadavu streams showed that ‘percent forested’ was the factor that explained the most variation in the macroinvertebrate community. The results of PCO analyses associating abiotic variables with macroinvertebrate community assemblage in Ra streams showed that the factors, ‘percent forested’, canopy cover, temperature and conductivity explained the most variation in the macroinvertebrate community.

The PCO analyses also identified several species that are strongly associated with either highly forested or poorly forested catchments, thus making them potential candidates for indicators of watershed health/intactness. The species associated with highly

xvi forested catchment streams were a finger-net caddisfly Chimarra sp. A, the atyid shrimps Atyoida pilipes and Atyopsis spinipes, and the neritid gastropods Neritilia rubida, Septaria sanguisuga, and Neritina squamaepicta. The species associated with the least forested catchment streams were a clinging mayfly nymph Pseudocloeon sp. and the net spinning caddisfly larvae Abacaria fijiana.

The Distance Based Linear Modeling (DistLM) results showed that the best solution single variable model highlighted that the predictor percent forested had the highest R2 correlation value (7.6197E-2) and was therefore the single factor that explained the greatest proportion of variance in the macroinvertebrate assemblage.

The combined results of the species richness graphs, the PCO analyses and the DistLM analysis when taken in their totality show that streams draining highly forested catchments have more ‘sediment sensitive’ species in comparison to streams draining less forested catchments. This is a pioneering study on a catchment-scale level effect and corresponding ecological response in Fiji streams. It is hoped that the information gathered through the current study will be used to facilitate development of appropriate stream health and watershed health indicators and contribute to watershed management in Fiji.

xvii

Chapter 1. INTRODUCTION & LITERATURE REVIEW

1.1 Freshwater: an essential resource

1.1.1 Global overview Freshwater bodies such as small springs, brooks, ephemeral streams, lakes and rivers comprise only 0.8% of the earths’ surface and contain 0.01% of water on earth (Dudgeon et al. 2006; Strayer 2006; Dobson & Frid 2009; Vörösmarty et al. 2010). Despite being small in comparison to oceans, rivers and associated tributaries play a vital role in global biogeochemical processes. Dobson and Frid (2009) define a river as a channel of water fed by converging tributaries whose movement is controlled gravitationally.

1.1.2 Biogeochemical process Rivers and streams are an indispensable component of the global hydrologic cycle as they transport excess water precipitated on land to ocean (Gutknecht et al. 2008). They are major conduits for global particle transport from land to oceans as they transport the majority (95 percent) of terrigenous sediments and particulate material (Chakrapani 2005). In addition to water and sediments, rivers transport nutrients. Those nutrients essential for biological processes such as nitrogen, phosphorous and sulphur once delivered into riverine system go through cycling processes (McClain et al. 2001; Pacini et al. 2008). Rivers also play a vital role in the global carbon cycle by delivering carbon from land to ocean. The annual riverine carbon export to oceans is approximately 900 x 1012 g (900 Tg) (Mckee 2003).

1.1.3 Biodiversity Freshwater bodies support a high proportion of the global biodiversity, supporting at least 100,000 species out of approximately 1.8 million of all species known to science (Dudgeon et al. 2006).

1.1.3.1 Fish biodiversity Inland waters are home to over 10,000 freshwater dispersant fish species. These comprise approximately 40% of global fish diversity and one quarter of global vertebrate diversity (Dudgeon et al. 2006). The freshwater bodies of the United States and Canada support nearly 800 native fish species in 36 families (Helfrich & Neves 2009). Indonesia (1,300 species) and Papua New Guinea (329 species) have the highest freshwater fish species richness in the Indo-Pacific region. Selected Fijian streams and rivers surveyed on the four main islands of Fiji (Viti Levu, Vanua Levu, Taveuni and Kadavu) are known to support over 90 freshwater fish species (Boseto et al. 2007); approximately 4% of the described fish species of Fiji (Seeto & Baldwin 2010).

The importance of freshwater systems is not limited to freshwater dispersant fish species but also includes a significant number of diadromous reef fishes and invertebrates. Dispersant species are those that are not able to cross saltwater barriers and thus their feeding and reproductive biome are limited to freshwater (Abell et al. 2008). Diadromous species undertake regular, seasonal and life-stage-consistent migrations between marine and freshwater environments (McDowall 2008). There are three subcategories of diadromous species; anadromous, catadromous and amphidromous. Anadromous species complete their feeding and growth stage at sea and migrate to freshwater to spawn. Amphidromous species migrate to freshwater during their juvenile stage when they are less than ca. 50 mm long. Thereafter they undertake prolonged feeding and complete their maturation to spawn. Catadromous species complete their feeding and growth stage in freshwater and the adult migrates to sea to reproduce (McDowall 1997, 2008).

Globally, freshwater systems support approximately 250 diadromous fish species; approximately one percent of the global fish diversity (McDowall 1997, 2008). These together with other freshwater related vertebrates (amphibians, aquatic reptiles, and mammals) combine to give roughly one third of all vertebrate species being confined to freshwater (Dudgeon et al. 2006). However, the contribution to the global biodiversity by freshwater fauna does not end here as new species continue to be described every year even in well known groups and well explored regions. For instance in the last five

2 years about 465 new freshwater fish species have been described for South America alone (Abell et al. 2008). According to Jenkins et al. (2010), a survey of 20 river basins of three main islands of Fiji within six years showed that the river systems of Fiji are dominated by diadromous fish (amphidromous gobiids) in contrast to tropical river systems of northern Australia and Papua New Guinea which contain a high proportion of resident fish.

1.1.3.2 Invertebrate biodiversity Freshwater systems also support a highly diverse group of resident invertebrates. These are invertebrates that have freshwater as their reproductive, feeding and growth habitat. Strayer (2006) stated that formally described freshwater invertebrates comprise about 90,000 species representing 17 phyla and approximately 570 families. However, the species list is continually being updated with new discoveries. Hershler (1998, 1999) described 2 new genera and 63 new mollusc species for North America while Kristensen and Funch (2000) discovered a new class of invertebrate, the Micrognathozoa, from freshwaters of Greenland. Strayer (2006) predicted that about 20,000 to 200,000 species of freshwater invertebrates are yet to be discovered. Aquatic insect fauna represent over nine percent of the total insect fauna of Ireland (Lucey & Dorris 2001). Tropical island streams not only support aquatic insect fauna but commonly support the non-insect diadromous invertebrates such as atyid shrimps, palaemonid prawns, and neritid snails (Resh 2005).

The freshwater macroinvertebrates of Fiji belong to seven recognized phyla and 45 families; namely Insecta (25 families), Crustacea (4 families), Mollusca (8 families), Nemotoda (2 families), Annelida (3 families), Platyhelminthes (1 family) and Porifera (2 families) (Haynes 1988a, 1999a, 2001a; Jeng et al. 2003; Haynes 2009; Department of Environment 2010). Mollusca are well described with 51 known snail and limpet species (Haynes 1985, 2001a, 2009) and a single freshwater bivalve species, a clam locally known as kai (Batissa violacea). Crustacea are also well described with 27 known shrimp and prawn species and one freshwater crab (Shokita et al. 1985;Choy 1991). However, many Fijian freshwater macroinvertebrates are yet to be fully described to genus or species level.

1.1.4 Human use of freshwater Streams and rivers provide many other essential ecosystem services for humans such as drinking water, subsistence and commercial fisheries, recreational (swimming, boating, fishing) and power generation (Kling et al. 2003). Water on earth is composed of 97.5% saltwater and 2.5% freshwater. Of the freshwater available, the majority is not readily accessible to humans for use as approximately 70% of it is frozen in the icecaps of Antarctica and Greenland and most of the remaining either lies in deep underground aquifers (groundwater) or is present as soil moisture (Postel et al. 1996; Engelman et al. 2008). The amount of water to which humans have access is less than one percent of the freshwater on earth (approximately 0.007% of all water on earth) for direct human uses (Engelman et al. 2008; Dobson & Frid 2009). This amount of water is contained in lakes, rivers, reservoirs and underground sources which are renewed regularly by rain and snowfall (Dobson & Frid 2009) and are shallow enough to be tapped at affordable costs (Postel et al. 1996).

1.1.4.1 Water consumption in Fiji In Fiji, The Water Authority of Fiji (established in 2010) is responsible for the water supply in accordance with the Fiji National Drinking Water Quality Standards (approved by cabinet in 2011) (Water Authority of Fiji 2012). However, the first round of these standards is limited to the centralized reticulated and metered water supply within urban areas while the rural area water supply is catered for by un-metered community-specific water catchments or household-roof catchments (Secretariat of the Pacific Community 2007). The Household Income and Expenditure Survey (HIES) showed that 93% of households in urban areas had access to an improved water supply with water piped directly to the household (Fiji Bureau of Statistics 2013). Water supply in rural areas is sourced largely either from community-specific water catchments; typically a concreted section of a creek or catchment feeding supply to the nearby village or through a communal well. In Fiji, there are a total of 1,174 un-metered village water supplies. These community water supply initiatives are installed by the Water Authority of Fiji (and previously the Public Works Department) and then the communities are responsible for maintenance, management and routine upkeep of these systems. The villages lacking this communal infrastructure acquire water at the

4 household level through rainwater harvest tanks or from rivers or creeks (Asia Development Bank 2003; Secretariat of the Pacific Community 2007). According to the ADB survey, one third of all rural households in Fiji lack access to an improved water source (Asia Development Bank 2003)

1.1.5 Food security Inland streams and rivers support subsistence fisheries which form an important component of food security, as well as generate local income for people in many countries. Freshwater catch consists largely of fish although crustaceans, molluscs and some reptiles comprise a minority of the catch (Welcomme et al. 2010). In 2008, global capture inland fisheries production was estimated to be 10.2 million tonnes. Two-thirds of this estimate was accounted for by Asia (66.4%) while the Oceania region contributed only 0.2 percent (Food and Agriculture Organization of the United Nations 2010).

In Fiji, the inland communities that have no direct access to marine resources depend heavily on the freshwater finfish and invertebrates as an important source of dietary protein (Haynes 1999b; Boseto 2006). The most commonly harvested finfish and invertebrates include freshwater mussel (Batissa violacea), eels, decapod crustaceans such as freshwater prawns (Macrobrachium spp.), Atyid shrimps and crabs, and tilapia and grass carps (Food and Agriculture Organization of the United Nations 2009). Freshwater eels, Flagtails (Kulia species) and a number of goby species are reported to be an important source of protein for the interior villages (Gillett 2009). Out of these consumed inland fishery species, the freshwater mussel (Batissa violacea) and decapod crustaceans are the only freshwater commodities that hold commercial importance (Food and Agriculture Organization of the United Nations 2009; Gillett 2009). It has been reported that in 2004, 2,526 tonnes of Batissa worth approximately F$2.2 million and 500 tonnes of decapod crustaceans valued at approximately F$6 million were sold in municipal and nonmunicipal markets. The figures from 2007 shows that the freshwater catch was 4,146 tonnes worth $3,858,750 though this was not broken down by commodity. This figure contributed to approximately four percent of the total fishery production of Fiji (Gillett 2009).

Globally freshwater supports a recreational fishery worth US$38 billion in the United States of America and US$8 billion in Europe though there is no defined recreational fishery in the high Pacific Islands (United Nations Environment Programme 2010; Welcomme et al. 2010).

1.1.6 Energy production Streams and rivers also provide power generation worldwide. The electricity generated by hydropower is termed as hydroelectricity (National Geographic Society 2012). It is a widely used form of renewable energy; used by 150 countries. In 2010, hydroelectricity accounted for 16 percent of global electricity consumption; 3,427 terawatt-hours of electricity. The Asia-Pacific region generated 32 percent of global hydropower (World Watch Institute 2012). In Fiji, hydropower accounts for 62.1% of national electricity production. There are four (Monasavu Hydro, Nadarivatu Hydro, Wainikasou Hydro and Nagado Hydro) hydroelectric dams in Fiji (Fiji Department of Energy c.2010).

1.2 Freshwater: A focus on status and threats It has been well established that globally the health of freshwater bodies is in decline (Jurado et al. 2008; Vörösmarty et al. 2010; Darwall et al. 2012) resulting in declines in biodiversity far greater than those in the most affected terrestrial ecosystems (Dudgeon et al. 2006). The decline in health of freshwater bodies is driven by many factors; human alteration of the environment being a crucial one (Engelman et al. 2008; Jurado et al. 2008; Darwall et al. 2012). The majority of freshwater ecosystems are inclusive of catchments from which water and materials are drawn by humans (Dudgeon et al. 2006; Burcher et al. 2007). According to recent studies, more than 40% of the human populations are able to access the 263 geographically and temporally available international rivers draining 45% of earth’s land surface (Dudgeon et al. 2006; Vörösmarty et al. 2010). The natural position of freshwater bodies within landscapes increases their vulnerability to anthropogenic impacts. Rivers and streams lie low within landscapes. This poses a risk of making them a “sink” for wastes, sediments and pollutant runoffs. The limited small volume feature of these freshwater bodies limits their capacity for diluting contaminants or mitigating impacts (Dudgeon et al. 2006).

Continuous human expansion and development coupled by demands for water by humans over the past hundred years have led to immense transformation of land and water systems (Dudgeon et al. 2006; Vörösmarty et al. 2010). Several pieces of recent research indicate that landuse has been identified as a major disturbance to freshwater bodies worldwide (Allan 2004; DeFries & Eshleman 2004; Dudgeon et al. 2006; Burcher et al. 2007; Engelman et al. 2008; Vörösmarty et al. 2010). The intensity of the transformation of land within watersheds containing freshwater bodies (Allan 2004; Burcher et al. 2007) today poses threats to global freshwater biodiversity. This is because changes in land cover lead to accelerated erosion, degraded water quality and loss of critical watershed ecosystem services and processes which are detrimental to freshwater biodiversity (Roth et al. 1996; Strayer et al. 2003; Allan 2004; Crowe & Hay 2004; Chadwick et al. 2006; Burcher et al. 2007). According to Vörösmarty et al. (2010) approximately 80% of the human population (4.8 billion) in the year 2000 lived in areas associated with high threat to either human water security or biodiversity. Vörösmarty et al. (2010) found that more than 30 of the 47 leading rivers collectively discharging half of global runoff to the oceans are evidences of threatened human water security and biodiversity. This leaves a remarkable tiny fraction of the rivers unaffected by humans.

1.2.1 Pacific freshwater vulnerability Freshwater resources of Pacific islands range from large rivers in high volcanic islands to low lying countries where no surface water resources exist; potable and economic needs are catered for by rainwater. The larger high volcanic islands such as Fiji and Papua New Guinea have adequate freshwater resources on the wet-side compared to the dry-side where water sources can be over-used. Tonga and Niue have significant groundwater sources while the low lying islands; Nauru, Niue, Tonga, Kiribati, Tuvalu and Republic of Marshall Islands lack significant freshwater resources (Duncan 2012; Secretariat of the Pacific Community 2012).

Freshwater resources of the Pacific islands are fragile due to the small sizes of these islands, lack of natural storage, vulnerability to natural hazards and competing land use. Regionally identified human induced pressures on freshwater resource or associated

7 catchments include expanding populations, poorly planned developments, progressive deforestation, introduced freshwater species, pressures of tourism, mining discharges, industrial wastewater, pollution from sanitation systems and agricultural chemicals (Watling & Chape 1992; Ellison 2009; Duncan 2012; Secretariat of the Pacific Community 2012). Agricultural chemical seepage and sediment loads arising from deforestation, mining and agricultural activities have been identified as significant threats to freshwater ecosystems. High volcanic islands are known for the largest and most varied freshwater biodiversity and greater range of habitats while the low-lying islands are often known for island endemic speciation (Ellison 2009; Duncan 2012; Secretariat of the Pacific Community 2012). Vulnerability of the island biodiversity means the associated ecosystems are also amongst the globally most endangered. Thus the status of the freshwater ecosystems of the Pacific islands is accepted as highly threatened with species extinction rate amongst the one of the highest (Ellison 2009; Duncan 2012).

1.2.2 Threats to Fiji freshwater systems As developmental pressures give rise to freshwater ecological insecurities in Pacific islands, Fiji is not an exception. River systems of Fiji range from small mountain narrow streams, steep torrents to very wide and large mature rivers in lowlands (Secretariat of the Pacific Community 2007; Duncan 2012). In Fiji, the most common threats to rivers include urban sprawl, pollution from untreated domestic discharges, logging, progressive deforestation and alteration to the floodplain for agriculture (Watling & Chape 1992; Haynes 1994, 1999b; Falkland 2003; Ellison 2009; Duncan 2012). Viti Levu is the largest island and contains approximately 79% of the country’s population (Fiji Bureau of Statistics 2012). A recent study indicates significant deforestation in Western Viti Levu; Sigatoka, the Ba river valleys and the Wainibuka catchments (Ellison 2009). These areas are known for poor agricultural practices such as planting on slopes or close to stream banks which leads to massive soil erosion, siltation and eutrophication. Inappropriate logging practices in high forest areas which compromise forest regeneration and increases soil erosion leads to sediment loads (Watling & Chape 1992; Ellison 2009). The high rainfall range of 2,000mm to 6,000mm falling on mountain catchments of Viti Levu provides the island with

8 abundant water resources. Yet these high flows are often accompanied by high sediment loads, exacerbated by deforestation and poor agricultural practices.

Soil losses of 35 tonnes/ha/year were estimated for the eastern side of Viti Levu; in the Waibau area including Rewa River Basin (Nunn 1990). The sediment load in the Waimanu River in Naitasiri indicated that an average loss of silt for the agriculturally associated catchment was 53 tonnes/ha/yr corresponding to a loss of 2.0-2.5mm of soil per year (Lovell & Sykes 2007). Increase in pressure on catchments from increasing population coupled with competing landuse and water demands in Fiji threatens the already vulnerable freshwater ecosystems by affecting water chemistry and faunal habitats.

Influences on the aquatic systems are not straightforward. Strayer et al. (2003) explained four major reasons why the specific relationships between land-use conversion and ecological responses are difficult to establish: (1) The landuse features differ among catchments and regions and across political boundaries such as landuse types, rates of conversion, and spatial distribution (2) changes in land use are capable of driving river hydrology and channel morphology into a state of instability which may take several decades to stabilise, (3) ecological responses may delay physical habitat modifications and duration of such delay and the associated effects are not always identified, and (4) there is lack a of information concerning the effectiveness of the management actions introduced for mediating the effects of development on streams. Thus, understanding and foreseeing the effects of landuse change on riverine ecosystems are difficult scientific issues and major challenges for present day ecology (Strayer et al. 2003).

However, in addressing the contemporary freshwater ecology challenges in his research, Allan (2004) states that recent developments have given rise to numerous studies that seek to establish relationships between land use and state of streams: (a) worldwide recognition of the extent and significance of changes in land use and land cover (b) conceptual and methodological advancement in landscape ecology together with the readily accessible land use/land cover data and (c) increasing applications of stream health indicators for assessing status and trends of riverine systems.

1.3 Landuse and stream ecosystems: a focus on anthropogenic impacts on stream bio-integrity Rivers have acted as magnets for human settlement over centuries leaving only a few river catchments unaffected by humans in some way. Rivers and streams play a vital role in the livelihood of inland communities. They not only produce resources that are consumed in rural areas but also provide much for commercial markets. Despite their importance, these ecosystems are becoming vulnerable due to human induced impacts. Freshwater ecosystems define the structure and functioning of the watershed they drain (Allan 2004; Bailey et al. 2004; Bunn et al. 2010). This is because stream ecosystems integrate the landscapes they drain in terms of water transport and respond to changing terrestrial conditions within catchments influencing stream structure and function (Burcher et al. 2007). Therefore, a delicate balance exists between streams and their surrounding ecosystem. However, this balance can be significantly altered through land cover changes and watershed disturbances allied to various human activities such as logging, agriculture, urbanization, construction of roads and dams, point source inputs, impoundment, channelization mining, and species introduction (Haynes 1994; Harding et al. 1998; Haynes 1999b; Fuchs et al. 2003; Alberti et al. 2007; Jenkins et al. 2010; Bernhadt & Palmer 2011; Nejadhashemi et al. 2011), all of which has led to alteration in factors related to stream health including water quality, habitat structure, flow regime, energy sources and stream biodiversity.

1.3.1 Logging Logging changes a number of aspects of a watershed that can have multiple effects on stream habitats. The removal of riparian forests including streamside trees causes alterations in stream hydrology and allocthonous inputs resulting in higher sediment and debris inputs, loss of physical in-stream macro-habitats and shallower and wider stream channels and eventually higher sediment discharge to sea (Fuchs et al. 2003). In addition to that, catchment cover clearance affects the stream water chemistry and associated processes. For instance removal of streamside trees (providing bank stability, shade and leaf litter deposition) and riparian forest (prevents soil erosion) may lead to increase in water temperature (Poole & Berman 2001) which in turn affects hydrologic

10 regimes and flow pathways, primary production and organic matter dynamics (Hindara et al. 2003) and nutrient uptake by plants (Sabater et al. 2000; Gravelle et al. 2009).

1.3.2 Dams Dams create problems for the freshwater community in many ways. Dams tend to trap most of the nutrients carried by rivers/streams and in doing so they alter natural in- stream nutrient concentration and flow (Kondolf et al. 2002; Strayer 2006). Damming lowers high flows in rivers which are essential to maintain distribution, abundance and diversity of stream biota and successional phases of riparian vegetation (Kondolf et al. 2002). Additionally, changes in flow and nutrient concentration can result in differences in total nutrient export relative to unimpacted catchments as well as affect the migratory pathways of running water species (Larinier 2000). Change in natural flow due to dam construction in California was one of the reasons for establishment of exotic species that prey upon the native salmon below dams (Kondolf et al. 2002).

1.3.3 Road construction The road construction activities in mountainous watersheds can increase the amount of sediment entering streams. An increase in suspended sediments reduces feeding ability of stream fauna by affecting respiration through gills (Gravelle et al. 2009). Toxicants and adsorbed nutrients on suspended sediments alter algal growth rate and biomass; toxic compounds incorporated into periphyton tend to have impacts throughout the food chain. If the quantity of fine sediments is great it gets deposited on the river bed which may have multiple effects on river biota. These include: (1) covering streambed substratum surface biofilms (food for grazing invertebrates); reducing its organic content and nutritional value, (2) covering gravel riffles and depleting invertebrate habitat (3), directly smothering aquatic plants (macrophytes and periphyton) and reducing their biomass, (4) infiltrating into fish spawning gravels and reducing incubation success and (5) eliminating juvenile fish habitat via filling interstices of cobble beds (Kondolf et al. 2002; Crowe & Hay 2004).

1.3.4 Urbanisation Urban development leads to an increase in impervious surfaces which decrease the amount of water that infiltrates into the soil. This increases overland flow and storm

11 water conveyance intensity. To cater for the increased flow volume, streams undergo geomorphic change resulting in increased channel erosion (Schoonover et al. 2005). Along with the former may come associated problems such as high sediment loads, heavy metals, nutrients and bacteria loadings. Increases in overland flow and associated inputs causes alteration of stream hydrology and water chemistry that in turn, result in degraded water quality, poor biological habitats and conditions necessary for survival of stream biota; eventually loss of associated biotic richness (Schoonover et al. 2005; Chadwick et al. 2006). A three year study in the Provo River, Utah U.S.A showed that urban runoff caused drastic changes in riverine substrate and macroinvertebrate community composition. Increasing urbanization caused an increase in river substrate compaction and decrease in macroinvertebrate density and species diversity. Macroinvertebrate communities in urban reaches were dominated by ‘tolerant’ species such as snails and leeches compared to non-urban reaches, dominated by ‘sensitive’ species (Gray 2004).

1.3.5 Mining Mining leads to extensive chemical and hydrological alterations in streams. Continuous deposition of tailings in particular results in a stream becoming shallower. Deposition of increasing loads of sediments and debris leads to reduction in feeding ability of fauna, instream-habitat loss, decline in fish incubation and juvenile survival success and eventually decline in fish population (Kondolf et al. 2002; Crowe & Hay 2004; Bernhadt & Palmer 2011). A well-documented example of this occurred on the Ok Tedi River of Papua New Guinea. For two decades, Australian-British mining giant BHP daily dumped 80,000 tons of tailings (rock waste) containing copper, zinc, cadmium and lead directly into the Fly and Ok Tedi Rivers (Hettler et al. 1997). Chemicals from the tailings killed or contaminated fish while massive dumping into the river exceeded its carrying capacity resulting in the river bed being raised by 10 m. This changed the hydrology of the river; a relatively deep and slow river became shallower and developed rapids which altered fish habitats and thus severely depleted fish stock (Burton 1999; Marshall 2002). Although two mining companies (bauxite mine in Bua and gold mine in Vatukoula) are currently operating in Fiji, their impacts on Fiji riverine fauna are yet to be documented.

1.3.6 Invasive species Introduction of invasive species has been documented to cause extinction or worsen range reductions of native species in many parts of the world. These species often invade and colonize new habitats resulting in devastating consequences for the native flora and fauna such as reduction in numbers (Ricciardi et al. 1998), forced extinction, or hybridization by alien fish species (Strayer et al. 1999; Taylor et al. 1984). The zebra mussel invasion in eastern North America altered physical, biological and chemical characteristics of many rivers and lakes as well as extirpating populations of native unionid mussels (Ricciardi et al. 1998; Strayer et al. 1999). Exotic freshwater fish species have altered the ecology of aquatic ecosystems. In the United States, the grass carp (Ctenopharyngodon idella) is known for reducing natural aquatic vegetation while the common carp (Cyprinus carpio) reduces water quality by increasing turbidity. These ecological alterations have caused extinction of some native fish species (Taylor et al. 1984). In Florida, exotic aquatic plants such as hydrilla (Hydrilla verticillata), water hyacinth (Eichhornia crassipes), and water lettuce (Pistia straiotes) are commonly known for choking waterways and altering nutrient cycles, causing alteration in natural aquatic species community assemblages (Office of Technology Assessment 1993). In Fiji, the Chinese grass carp (Ctenopharyngodon idella) was introduced to control the excessive growth of Hydrilla verticillata (Murty & Cavuilati 1977) which aggravated silting up of lowland Rewa river channel (Hughes 1977).

1.3.7 Agricultural practices The agricultural transformation of land can lead to an alteration in water hydrology as well as an increase in nutrient input in streams which alters water chemistry and eventually affects the nutrient cycle (Poff et al. 2006). Increased nutrients in streams impact aquatic diversity and have been observed to result in a change in freshwater community assemblage. According to Peterson et al, (1993) phosphorus fertilization of a pristine tundra river led to increased algal biomass and productivity. This triggered an increase in populations of grazing invertebrates and invertebrate growth rate. An increase in population density of multiple species led to an increase in competition and eventually a decline in density of some groups of invertebrates (Peterson et al. 1993).

Diversion or modification of streams or rivers for irrigation of crops greatly affects stream hydrology and sediment regimes which cause alteration in instream habitats (Poff et al. 2006). These alterations in physical parameters and associated stream water chemistry in return may have significant impact on stream ecosystem processes and biodiversity and eventually affect overall stream health.

Transformation of land within the Pacific Island countries has led to destruction of natural habitats of many island endemic freshwater fish species resulting in severe decline in their population (Pippard 2012). These fish species are now listed on the IUCN Red List of threatened species. The freshwater goby (Akihito futuna) from the island of Futuna has been severely impacted by the construction of dams. Land clearance, pesticide use and dam construction in French Polynesia has led to an intense decline in the population of freshwater goby (Stiphodon discotorquatus). Agricultural run-off in the Federated States of Micronesia (FSM) has negatively affected the population of freshwater goby (Sicyopterus eudentatus) (Pippard 2012). Experiences suggest declines in population of a few freshwater macroinvertebrate species due to land transformations but their status is yet to be documented (Haynes 2012, pers. comm).

1.3.8 Landuse in Fiji: a focus on stream biotic community

While stream health is globally threatened, Fiji is not an exception with landuse being the main driving factor for declining stream health (Atherton et al. 2006; Boseto 2006). Recent studies for example Atherton et al. (2006) and Olson et al. (2009) have shown that watersheds are amongst the most threatened ecosystems in Fiji. The rapid shifts in landuse patterns over the past hundred years have resulted in Fiji's land surface today being less than 40% covered in forest (Olson et al. 2009). Over the years, the land surface has been cleared either by logging commercially for timber or by clearing to make way for agricultural land (Atherton et al. 2006). Burning is employed as the main method for both clearing land and maintaining land once cleared. The catchments of the main populated islands of Viti Levu, Vanua Levu, Taveuni, Ovalau and Kadavu are the

14 most threatened (Atherton et al. 2006; Olson et al. 2009). In many catchments on the dry leeward side of these islands, the native vegetation has now been entirely replaced with talasiga grassland and/or shrubby vegetation on land that has been allowed to become fallow (Keppel & Tuiwawa 2007). In modified forested catchments of Fiji, ecosystem services and species diversity have been altered (Olson et al. 2009).

Ecologically, many Fijian stream systems have undergone modifications, degradation and loss of species assemblages (Haynes 1999b). The local land cover changes for purposes such as commercial forestry (Haynes 1999b) on the catchment slope, cattle grazing and horticulture on the floodplains, human activity in clearing the forest vegetation for timber and conventional farming (Carpenter & Lawedrau 2002), constructions such as dams (Haynes 1994), roads, development of settlements, urban sprawl (Japan International Cooperation Agency & Ministry of Agriculture Fisheries and Forests 1998) and agriculture (Ram et al. 2007) has led to accelerated erosion and sedimentation (Printemps 2008) which end up in streams and changes habitat conditions. The changes in habitat conditions influence the “freshwater community composition, instream energy production and food web interactions that redirect energy flow towards higher instream trophic levels including predatory macroinvertebrates” (Arkle & Pilliod 2010) and vertebrates (Harvey & Hill 1991; Boseto 2006).

Loss of catchment forest cover has affected freshwater fish diversity in Fijian streams. In a six year study of 20 river basins within three main islands of Fiji (Viti Levu, Vanua Levu and Taveuni), Jenkins et al. (2010) noted a marked decline in freshwater fish diversity in mid reaches of streams where catchment forest cover was reduced below 50%. Freshwater macroinvertebrates are likely to be similarly affected. Haynes (1999) compared stream biodiversity adjacent to logged and unlogged areas of catchments in Viti Levu over a three year period. The results showed consistently lower diversity in the stream (Nabukavesi stream) adjacent to a logged area compared to an unlogged area (Wainikovu stream) and significantly high species diversity. She hypothesized that the low abundance of neritid gastropods in streams of a logged catchment was due to generated sediment loads covering the biofilm on which they grazed (Haynes 1999b).

The decline in these prey species may strongly affect populations of the predator species such as gudgeons. Gudgeons feed preferentially on benthic macroinvertebrates invertebrates and are important local freshwater fisheries resources for inland communities of Fiji (Jenkins et al. 2010).

Forestry activities over the past years in the Rewa watershed have led to lowland channel sedimentation and growth of exotic aquatic plants. The Rewa River is the largest riverine system of Fiji and drains approximately one third of Viti Levu; the main island of Fiji. Activities in the Rewa catchment such as commercial forestry, small-scale cultivation, fires and demand for firewood has led to deforestation while large scale developments such as urbanization, development of settlements, conventional farming and building of access roads has led to forest clearance (Japan International Cooperation Agency & Ministry of Agriculture Fisheries and Forests 1998; Carpenter & Lawedrau 2002). While these human activities are the main cause of catchment erosion, high rainfall (6,000-8,000mm annually) intensity and runoff bring high volumes of sediments and associated nutrients into lowland waterways (Carpenter & Lawedrau 2002). Over the years, the buildup of sediments and the high concentration of nutrients has led to the growth of exotic aquatic weeds in the Rewa River (Hughes 1977; Sundaresan & Reddy 1979; Carpenter & Lawedrau 2002). Severe infestation by Salvinia molesta, a highly competitive and aggressive floating aquatic fern choked Rewa river waterways in the 1970s. This aquatic plant propagates both horizontally and vertically forming thick mats and growing on its own dead materials (Hughes 1977; Sundaresan & Reddy 1979). The water column below the Salvinia mats has been found to have low oxygen levels; showing the possibility of suffocating other associated flora and fauna (Sundaresan & Reddy 1979). Other submerged exotic aquatic plants present in the river include Monochoria hastata, Potamogeton crispus and Hydrilla verticillata. Excessive growth of Hydrilla verticillata has aggravated silting up of lowland Rewa river channel (Hughes 1977). Over a seven year study, Hughes (1977) hypothesized that excess growth of H. verticillata might cause undue development of noxious algae responsible for fish mortality or liberation of toxic substances.

Faulty farming practices at the subsistence level in certain areas of Fiji has had negative effects on freshwater gastropod population. A study carried out on the island of Ovalau showed a depletion of freshwater gastropods associated with sedimentation effects caused by erosion due to faulty farming practices. Grazing neritid gastropods in Naisogo stream were reduced by 64% within a year upon washing away of a layer of mud into streams from village gardens during wet season heavy rainfall. The layer of mud (approx. 20mm thick) covered the streambed substrate and algal biofilms which served as food for the gastropods and hence starvation may have led to decline in population (Haynes 1989b). Suspended sediments affect filter feeding invertebrates holding local commercial value such as clam (Batissa violacea). Due to high suspended solids in water these clams remain closed and are unable to feed and thus their growth is retarded (Fuller 1974).

A study by Haynes (1994) showed that the construction of Monasavu dam on the headwaters of the Rewa River and impounding of the stream wiped out a natural community of freshwater invertebrates. Three years after damming only a few stress- tolerant species were present in the lake. She hypothesized that this was probably due to eutrophication taking place. Once the decomposition of vegetation decreased, more species returned but disappeared upon a 20m fall in water level (Haynes, 1994).

Exotic freshwater fish have been documented to cause reductions of native fish in Fiji. In an assessment of impacts of introduced freshwater fish species on native ones, Jenkins et al. (2010) found that on average, stream networks with established (tilapia) Oreochromis spp. populations had fewer species of native fish compared to systems with tilapia absent. Sites with tilapia had nine native species absent comprising eleotrids and gobiids which have important dietary and economic value; five of these fish species were endemic. The disappearance of native fishes was hypothesized to be due to predatory effects of tilapia which prey upon juvenile fish and fish larvae (Jenkins et al. 2010). The freshwater gastropod Viviparous japonicus was accidently introduced into fish ponds in the late 1980s accompanying the Macrobrachium prawn from Japan. It was first noticed in 1989 and within two years time their numbers had increased to the

17 extent that they clogged two fish ponds and made them inoperable (Haynes 1994). Impacts of Viviparous japonicus on natural stream biota at present are conjectural.

Strayer et al. (2003) clearly outlines in his research that land cover alterations intended for various land-use purposes affects stream ecosystems by altering the timing, amount, and kind of inputs of water, light, organic matter, and other materials to the channel, which can have intense consequences for stream ecosystem functioning and threaten the beneficial services provided.

1.4 Natural effects on biological indicators of stream health: a focus on distribution and species richness Although human-induced pressure has had extensive impacts on the integrity of freshwater ecosystems, there are also natural factors that affect freshwater taxa distribution, trophic relationships and community assemblages. These natural factors include the natural geographic distribution of species, natural variations in habitats and differing modes and opportunities for dispersal of organisms.

Freshwater biodiversity is not spread uniformly across the globe. The diversity of species across the Pacific Island region decreases from west to east with increasing isolation and decreasing island size and age (Cowie & Holland 2006). The small size and isolation of islets in the eastern Pacific results in relatively small and restricted populations and low primary diversity and therefore freshwater biota community structure is unique in each Pacific island country. Over time, species have been able to cross the saltwater barrier successfully or get introduced accidently by humans (Haynes 1990; Paulay 1994; Resh & de Szalay 1995; Craig 2003; Cowie & Holland 2006; Abell et al. 2008; Ellison 2009).

Faunal assemblages in Pacific island streams are dictated by organisms’ dispersal abilities and life-history characteristics required for traversing oceanic expanses and successfully colonizing insular freshwaters (Covich 2006; Smith et al. 2003)., relatively few taxonomic groups of organisms (e.g. amphidromous fishes, gastropods and crustaceans) possess these traits and therefore, over time, similar faunal communities

18 have colonized the streams and rivers of isolated oceanic islands (Ryan 1991; Covich 2006). Hence, the freshwater fauna of the oceanic islands are characterised by a relatively low number of fishes, crustaceans, molluscs and aquatic insects that have complex life-cycles and differing mode and opportunity of dispersal (Covich 2006; Nelson 1999; Ryan 1991; Smith et al. 2003). The two modes of dispersal observed among the freshwater fauna are (a) passive, long-distance movement whereby the organisms traverse the broad geographic barrier of oceanic islands with assistance (e.g. the gastropod veligers that swim along with sea currents across oceanic islands) and (b) active, relatively short-distance movement whereby the organisms move across islands, more effectively within same archipelago, without assistance (e.g. the flying adults of aquatic insects such as moths and dragonflies/damselflies).

The stream invertebrates of oceanic islands comprise aquatic insects, crustaceans, gastropods, bivalves, sponges and a species of nudibranchs. Many macroinvertebrates (except aquatic insects) have a diadromous life cycle and therefore their veliger or planktonic stage allows passive dispersal via sea currents (Haynes 1990; Resh & de Szalay 1995; Craig 2003; Cowie & Holland 2006). Fijian streams possess invertebrate taxa that are related to those of New Zealand, Australia, PNG, Solomon Islands and Asia (Marinov & Waqa-Sakiti 2013; Marinov & Pikacha 2013; Kalkman & Orr 2013; Chase 1983; Choy 1991; Haynes 2009; Williams 1980; Winterbourn et al. 2006). Fiji freshwater insects are known to be closest to those of Australia and New Zealand, (Williams 1980; Winterbourn et al. 2006) they have many families in common although a few Odonata appeared to be more closely related to those of Papua New Guinea and Solomon Islands (Marinov & Waqa-Sakiti 2013; Marinov & Pikacha 2013; Kalkman & Orr 2013). Many of Fiji’s crustaceans (prawns, shrimps and crabs) are also present in Asia (Chase 1983; Choy 1991). The main Pacific Island gastropods, also present in Fiji, belong to the families Neritidae and Thiaridae originated in South East Asia (Indonesia, Philippines) where many of the same species are found. Exceptions are the rissooidean gastropods Fluviopupa spp. whose centre of origin is New Zealand (Zielske 2014).

Freshwater biotic community distribution is dictated by several geographical factors, including (1) island age - whereby older islands tend to accumulate more species

19 originating from the source continents and archipelagos coupled with additional endemic taxa as a result of speciation (Covich 2006), (2) island size, location and maximum height (Haynes 1990) (3) basin size and shape (Covich 2006) (4) drainage network pattern and density (5) and channel size (Covich 2006).

A study by Haynes (1990) on the number of freshwater gastropods present on 14 Pacific islands showed that the geographical factors determining the number of species on islands are: (1) island area and island maximum height which determine diversity and size of freshwater habitat availability and (2) island location, i.e. distance from a possible source of new immigrants.

A study on the pristine (uninhabited and completely forested) Yela River catchment on Kosrae in the Federated States of Micronesia showed a species-poor freshwater community. Six species each of fishes and shrimps and four snail species were found with Chironomid larvae representing 85% of the exceptionally low insect biomass; Ephemeroptera, Trichoptera, and Plecoptera were completely absent (Benstead et al. 2009). The authors suggested that the species-poor community was driven by biogeochemical process taking place on this remote Pacific Island.

Freshwater biota owes its distribution and species richness to features of a riverine system, such as the elevation of reaches within landscape, pool-run-cascades-chute- riffle dynamics, gradient and hydrological regime, amount of canopy cover and sunlight penetration, litter quality and diversity, number and size of streambed substrates, channel width and depth, current velocity, riparian width and complexity of plants, flood plain quality, bank stability and erosion (Gorman & Karr 1978; Allan 1995; Pires et al. 2000; Leroy & Marks 2006; Clarke et al. 2008; Petrie 2008). Variability of habitat types affects the distribution of freshwater communities. Niche specific taxa may not be present in all similarly pristine aquatic environments; thus direct comparisons are not useful. For example in a study comparing macroinvertebrate assemblages living on autumn-shed leaf litter in three streams in Arizona, the assemblages colonizing leaf litter mixtures differed from the assemblages colonising single tree leaf species (Leroy & Marks 2006). In another study, Mackay and Kalff (1973) found that distribution of two species of case caddisfly larvae was based on the organic substrate type in the

20 stream bed. Two species of Pycnopsyche (Limnephilidae) larvae of West Creek, Quebec inhabit allochthonous organic materials and are contemporaneous and similar in size. However Pycnopsyche gentilis prefers fallen leaves as food and to make cases while Pycnopsyche luculenta utilizes twigs and leaf material (Mackay & Kalff 1973).

The integrity of the aquatic food chain in a stream also affects the distribution of freshwater communities. Freshwater macroinvertebrates are essential food for other organisms within the stream ecosystem. Their diversity and abundance is essential to sustain fish population. The success of these macroinvertebrates depends on the abundance of high quality food, mostly algae and leaf litter (Cummins & Klug 1979). Any variation in habitat dynamics such as canopy cover, temperature, sediment and nutrient load that affects the food source tends to affect lower trophic levels which can also affect higher trophic levels; accounting for the patchy distribution of aquatic fauna along a natural gradient (Allan 1995). Experiences suggest this is also true for Fiji though it has not been documented as yet.

1.5 Measuring stream health: a focus on biophysical monitoring A number of studies internationally have shown the relationship between land cover alteration in watershed and ecological health of streams; in other words, with declining watershed health comes declining stream ecosystem health (Roth et al. 1996; Quinn et al. 1997; Sabater et al. 2000; McClain et al. 2001; Stewart et al. 2001; Hindara et al. 2003; Roy et al. 2003; Strayer et al. 2003; Allan 2004; DeFries & Eshleman 2004; Chadwick et al. 2006; Alberti et al. 2007; Burcher et al. 2008; Death & Collier 2010; Nejadhashemi et al. 2011; Magierowski et al. 2012). A case study from American Samoa (Wade et al. 2008) shows that there is a link between landuse, stream biophysical parameters and the ecological health of streams. Streams integrate the landscape they drain and any alterations (e.g removal of riparian forest) to the landscape may lead to alteration in factors related to ecological health of a stream including water quality, habitat structure, flow regime, energy sources and stream biodiversity.

To understand this cause-effect relationship, there is a need to assess ecological health of streams (hereinafter ‘ecological health of stream’ will be referred to as ‘stream health’). There are a variety of methods used for assessing stream health, but the majority look at the interplay of stressor variables and response variables; stressor variables being parameters that drive change in the ecosystem, and response variables being parameters that react to stressors (MacDonald et al. 1991; Haynes 1999b; Karr 1999; Stewart et al. 2001; Roy et al. 2003; Coulter et al. 2004; Lear et al. 2009; Young & Collier 2009; Death & Collier 2010). Stressors are generally abiotic factors whilst response variables are biotic factors (Purcell et al. 2009; Friberg 2010). In this study the response variables consist of abundance, diversity and composition of the freshwater macroinvertebrate communities.

Freshwater communities comprise both producers (e.g., phytoplankton, periphyton and macrophytes) and consumers (e.g., filter/gathers, grazers, shredders and predators) that form intricate food webs (Hauer & Lamberti 2007). Freshwater macroinvertebrates are extremely important in the functioning of freshwater ecosystems as they contribute towards crucial functions such as nutrient cycling, regulating rates of primary production, assisting in litter decomposition, water clarity, and thermal stratification as well as being food for higher-level organisms (Goodnight 1973; MacDonald et al. 1991; Strayer 2006; Böhm et al. 2012).

The freshwater communities are dependent upon the characteristics of their aquatic habitats which support all their biological functions meaning their presence is dependent upon a habitat with availability of suitable characteristics. Therefore it is necessary to have a combined approach of understanding the physical and biological parameters of streams and rivers. Biophysical monitoring is needed to provide a descriptive summary of the health of such aquatic ecosystems (DeFries & Eshleman 2004).

The most widely used response variables for assessing stream health by researchers are those involving macroinvertebrate communities (Goodnight 1973) (in correlation with water quality parameters as stressor variables). This is because the habitat type, long- term hydrologic cycle and water quality variation within the riverine systems influence

22 the adaptive strategies of benthic macroinvertebrates leading to corresponding differences in macroinvertebrate assemblages (Arkle & Pilliod 2010). Additionally, macroinvertebrates in their aquatic phases have limited mobility and are restricted to their immediate habitat while vertebrates are more mobile and can move either upstream or downstream to escape degrading habitat conditions (MacDonald et al. 1991).

Freshwater plants and microbes on the other hand can also be good indicators of anthropogenic impacts on stream health. It is well known that a bloom in native freshwater plants indicates an increase in nutrient loads or light into streams while the presence of microbes such as fecal coliform indicates potential sewage seepages which are one of the most well recognized sources of nutrient input. Other common nutrient sources recognized are animal manure (Rosario et al. 2002), urban activities and agricultural fertilizers (Carpenter et al. 1998).

Various countries have designed matrices to assess stream/river health. For instance, the National River Health Program for Australia utilizes the SIGNAL 2 biotic index for river macroinvertebrates designed to be applicable to the Australian rivers. SIGNAL 2 is an improved version of the original SIGNAL (Stream Invertebrate Grade Number – Average Level) index. SIGNAL is a simple biotic index for Australian river macroinvertebrates developed by Dr. Bruce Chessman whereby a ‘grade number’ between 1 and 10 is assigned to the 110 common families. The ‘grade numbers’ reflect estimates of the sensitivities of these families to common types of stream pollution where grade 1 is awarded to the most resilient families and 10 for the most sensitive ones (Chessman 2001). In New Zealand waterways, the biotic indices such as Macroinvertebrate Community Index (MCI) and its derivatives (Quantitative Macroinvertebrate Community Index –QMCI and Semi- quantitative Macroinvertebrate Community Index -QMCI) are commonly used to interpret the state of streams. These indices rely on the allocation of pollution tolerance scores (1-10) to taxa usually at genera level where high scores indicate pristine conditions and low scores indicate degraded conditions (Wright-Stow 2001). Unfortunately, the stream health assessment matrices are not yet developed for Fiji streams. The current study therefore uses the

23 classical stream health assessment technique i.e. the macroinvertebrate community structure and associated biophysical parameters.

1.6 Rationale for this study Freshwater macroinvertebrate communities integrate the effects of the multiple stressors that accompany human-induced changes in land use and therefore the presence, abundance and the composition of the macroinvertebrates indicate stream health (Sickle et al. 2004). Changing landuse patterns in Fiji pose an increasing need for joint management of land and water resources to ensure stream health. Stream health has been shown to be reliably assessed on the basis of the presence and abundance of freshwater macroinvertebrate communities (Roth et al. 1996; Timm et al. 2001; Gray 2004; Sickle et al. 2004). This project therefore will focus on investigating how landuse affects stream health in two study areas in Fiji (Nakasaleka district of Kadavu and Nakorotubu district of Ra). Macroinvertebrate diversity, in-stream habitat features and water quality indicators will be assessed across a range of landuse patterns. The presence and corresponding abundances and community composition of macroinvertebrate taxa (biotic parameter) and the associated environmental parameters (abiotic parameter) will be assessed across streams draining catchments with varying percentage forest cover.

Figure 1.0: Generic conceptual model for effects of forest cover on the macroinvertebrate community structure in Fiji rural streams. The major landuse types that affect forest cover which directly or indirectly affect habitat features and physiochemical parameters that support macroinvertebrate community structure are shown inside the boxes.

The generic conceptual model (Figure 1.0) given above summarizes the introduction chapter as well as highlights the study design. This study is designed to investigate how the catchment forest cover affects the abundance, diversity and community composition of macroinvertebrates in rural streams of Fiji. In this study the only human impacts on the catchment forest cover are modification via agriculture and human settlements (Figure 1.0). The generic conceptual model (Figure 1.0) is based on the theory that the freshwater macroinvertebrates are dependent upon the characteristics of their aquatic habitats which support all their biological functions meaning their presence is dependent upon a habitat with availability of suitable characteristics (Allan 2004; DeFries & Eshleman 2004). The model (Figure 1.0) shows the most common anthropogenic factors that affect forest cover (in Fiji) which tends to affect the stream habitat features and associated physiochemical parameters of the stream water (directly or indirectly) and these eventually affect the macroinvertebrate community composition, abundance and diversity. 25

The methods used in this investigation are carried out by visual estimation of the physical nature of each stream (yellow box) feature and measuring the physiochemical parameters (purple boxes) of the streams water at specific sites where the freshwater macroinvertebrates were sampled (light blue oval).

1.7 Research aims and objectives The aim of the project was to investigate how landuse affects stream health in the provinces of Kadavu and Ra, Fiji.

The following objectives were designed to achieve this aim:

1. To assess biotic parameters of stream health using ‘response’ (freshwater macroinvertebrate community structure) indicators.

2. To assess abiotic parameters of the stream health using ‘stressor’ (water quality and in-stream habitat features) indicators.

3. To examine relationships between abiotic, biotic parameters and catchment forest cover (which abiotic factor explains the greatest amount of variation in biotic community structure).

1.8 Structure of this thesis This thesis is presented in five chapters. Chapter One presents the general introduction; defines the scope and the extent of the problem as well as reviews the literature and related work done globally, within the region and locally. It also presents rationale for this study, study objectives, hypothesis, layout of the thesis and the significance of this study.

Chapter Two presents the general methodology and describes the particular methods used per stated objectives to help achieve the research aim.

Chapter Three presents the results of the biotic (macroinvertebrate community structure) data gathered across a continuum of streams in Kadavu and Ra and the associated discussions.

Chapter Four presents the results of the abiotic data gathered during the current study. It also presents the results of the linking abiotic variables and catchment forest cover (assessed through GIS mapping) to the macroinvertebrate community structure with associated discussions.

The final Chapter Five summarizes major findings of this research and contains a general conclusion. It also makes recommendations for development of appropriate monitoring indicators for long-term management of rivers and associated catchments in Fiji.

1.9 Significance of the study This project is part of two larger projects being implemented by the Institute of Applied Science (IAS) of the University of the South Pacific (USP) and will be used as an awareness tool to inform these projects as well as assisting with the development of appropriate monitoring indicators. IAS-USP has been funded to undertake two projects using CBAM (Community Based Adaptive Management) approaches to Watershed Management. The first, COWRIE (Coastal and Watershed Restoration Integrity for Island Environments) funded under the CRISP (Coral Reef Initiative for the South Pacific) project is concentrating efforts in Ra, north-east Viti Levu.

The second, WANI (Water and Nature Initiative), funded under IUCN (International Union for the Conservation of Nature) Regional Office for Oceania is concentrating efforts in Nakasaleka, Kadavu. This research feeds scientific information and knowledge directly into both these projects as its study sites are areas in which these projects are operating.

Chapter 2. METHODOLOGY

2.1 Overview In this study, freshwater macroinvertebrate and physicochemical surveys were conducted in lowland rural streams originating from six catchments; three in Viti Levu and three in Kadavu (Figure 2.1). These freshwater surveys were carried out at three stations per stream within a catchment (Figure 2.2 & Figure 2.3). Thus there was a total of 18 stations surveyed (Table 2.1).

Figure 2.1: Map showing the study sites in Kadavu and Ra province as part of Fiji archipelago 2.2 Catchment selection criteria A range of catchments in Nakasaleka district of Kadavu and Nakorotubu district of Ra, Viti Levu were selected to represent a continuum of landuse, from highly degraded to intact. Effort was made to ensure that certain variables remained as constant as possible. These variables included:

I. Catchment area: a total of six catchments; three from Nakasaleka district and three from Nakorotubu district were chosen to be of same area; approximately 4km2. II. Stream length: each of the six streams was approximately two kilometers long and originating from an elevation of approximately 200m above sea level. III. Only year-round permanently flowing streams were used in the study.

2.3 Study site description

2.3.1 Nakasaleka, Kadavu

2.3.1.1 Location Kadavu is a mountainous island and of volcanic origin; lying to the south of the main Fiji archipelago. The island is located at roughly 19⁰S and 178⁰ East (Terry 1999). The study was carried out within Nakasaleka district, which is located on the north-eastern coast of Kadavu, and which contains a total of 49 catchments. Stream sites for this study were selected in three catchments: Lomanikoro (Naqewa stream), Namajiu (Korolevu stream) and Naivarube (Savumiri stream) (Figure 2.2).

Figure 2.2: Sampling stations in Kadavu province

2.3.1.2 Climate The climate of Kadavu is tropical (Terry 1999). The Vunisea climate station, located approximately 28km south-west of Nakasaleka, records an annual rainfall of approximately 2340mm. The monthly rainfall ranges from 105.1 mm in August to 361.6 mm in March. The station recorded an average monthly rainfall of 221 mm from February 2011-February 2012. The monthly rainfall ranged from 102 mm in November 2011 to 568 mm in January 2012. The average maximum temperature from February 2011-February 2012 was 28.7 C and the average minimum temperature of the same period was 22.9 C (Fiji Meteorological Service 2012).

It is important to note that the Vunisea climate station is the only weather station located in Kadavu and the recorded rainfall readings may not be similar to that of Nakasaleka. According to the Worldclim (2014), Nakasaleka receives an annual rainfall of approximately 2435mm.

30 2.3.1.3 Geology The geology of Kadavu Island is described by Terry (1999) and is igneous rocks (andesitic rocks), followed by sedimentary rocks, which included breccias and conglomerates. The soil type is comprised of ultisols (red-yellow podsols and humic latosols), alluvial and gley soils.

2.3.1.4 Vegetation The catchments (Namajiu, Kavala and Nakoro) surveyed in this study comprised lowland rainforest, secondary forest and, farmland (agriculture) (Mueller-Dombois 1998; Tuiwawa 2010).

2.3.2 Ra

2.3.2.1 Location Nakorotubu is one of the seven sub-districts in Fiji's Ra province with 27 villages. It is one of the four coastal districts but with an extensive inland undulating and rugged terrain located in the northeast of the island of Viti Levu. The study was carried out in three main catchments of the district: Nabukadra (Vucinivola stream), Naocobau (Naikawaqa stream) and Saioko (Taveu stream) (Figure 2.3).

Figure 2.3: Sampling stations in Ra province

2.3.2.2 Climate The Penang weather station is the closest weather station to Nakorotubu. This weather station lies approximately 23km north-west of my study sites. The Penang climate station records an average monthly rainfall of 2400mm. The monthly rainfall ranges from 46.5 mm in July to 510.4 mm in January. The station recorded an average monthly rainfall of 311 mm from February 2011-February 2012. The monthly rainfall ranged from 44 mm in September 2011 to 990 mm in January 2012. The average maximum temperature from February 2011-February 2012 was 30.1 C and the average minimum temperature of the same period was 22.7 C (Fiji Meteorological Service 2012). It should be noted that Penang is located on the drier side of Ra province within the rain shadow zone and therefore receives lesser rainfall than Nakorotubu which lies in the windward zone. According to the rainfall map in Ryan (2000), Nakorotubu receives an annual rainfall of 3400mm.

2.3.2.3 Geology The geology of Ra is described as dominated by volcanic rock such as basalt, with andesitic (andesitic breccias) and sedimentary rocks also present. Sedimentary rocks comprise conglomerate, mudstone and limestone (Rodda 1967).

2.3.2.4 Vegetation The general vegetation of Nakorotubu district catchments consists of lowland forest on high and moderate elevations , talasiga grassland on lower catchment areas, cash crops (taro, cassava) and yaqona cultivation on lower valley slopes, coastal strand, beach forest and mangroves on coastal flatlands (Mueller-Dombois 1998)

2.4 General description of streams Table 2.1: Survey site description and coordinates

Stream Catchment Reach Site Elevation Landuse Dominant macro-habitat Longitude Latitude Area (Km2) code (m) type (⁰E) (⁰S)

Kadavu Naqewa 4.12 Lower KNL 94 Native forest Run, riffle, pool & chute 178.36982 18.94804 Mid KNM 116 Agriculture & Secondary forest Run, riffle, pool & chute 178.36642 18.95591 Upper KNU 244 Native forest and Agriculture Run, riffle, pool & chute 178.36234 18.95937 Korolevu 4.59 Lower KKL 47 Native forest & Secondary forest Run, riffle, pool & chute 178.39647 19.01401 Mid KKM 64 Native forest Run, riffle, pool & chute 178.38930 19.01100 Upper KKU 122 Native forest Run, riffle, pool & chute 178.38931 19.01100 Savumiri 3.55 Lower KSL 46 Agriculture, Fisheries office, PWD camp Run, riffle, pool & chute 178.44645 18.98528 & road construction in progress.

Mid KSM 72 Agriculture & settlement Run, riffle, pool & chute 178.43716 18.98424

Upper KSU 142 Agriculture, Secondary forest & Run, riffle, pool & chute 178.42977 18.97751 Settlement Ra Taveu 3.34 Lower RTL 19 Agriculture, Secondary forest & School Run, riffle & pool 178.36680 17.54127 Mid RTM 115 Secondary forest, Agriculture & Run, riffle, pool & chute 178.36011 17.54380 Settlement Upper RTU 165 Native forest Run, riffle, pool & chute 178.35950 17.54515 Vucinivola 4.61 Lower RVL 19 Agriculture & Settlement Run, riffle & pool 178.38381 17.55129 Mid RVM 41 Native forest & Settlement Run, riffle, pool & chute 178.38066 17.55447 Upper RVU 56 Agriculture & Secondary forest Run, riffle, pool & chute 178.37699 17.55507 Naikawaqa 3.31 Lower RNL 18 Native & Secondary forest Run, riffle & pool 178.33747 17.51008 Mid RNM 246 Native forest Run, riffle & pool 178.33535 17.51933 Upper RNU 286 Secondary forest & Native forest Run, riffle & pool 178.33979 17.52457

2.5 Sampling period and weather

A total of 18 sites were sampled for invertebrates and water chemistry measurements in the six catchments over two seasons. The survey periods comprised February-March 2011, August-September 2011 and February-March 2012. February-March surveys were conducted over a period of higher rainfall (wet season) and fluctuating river levels resulting in potential in-stream disturbance. August-September surveys were carried out over a period of relatively dry weather (dry season) with sampling not affected by significant rainfall and high flows. Sampling locations are shown in Figure 2.2 and Figure 2.3. The three-letter site codes are composed from the initials of the province, the stream and the reach, for example KNL indicates a station sampled in Kadavu, in the stream of Naqewa, in the lower reach (Table 2.1).

2.6 Data collection On each sampling trip, the streams were sampled along a 20m reach (station) at three main reaches; upper reach, mid reach and lower reach. These reaches were selected according to riffle habitat availability at stations in the upper, middle and lower section of the stream. The lower reach was well above the tidal zone to maintain a freshwater habitat feature. The 20m reach (station) remained constant throughout dry and wet season sampling. At each station three types of data were collected; habitat data, physicochemical data and macroinvertebrate data.

Habitat characteristic data were recorded on a ‘habitat field sheet’ (presented in Appendix 7.1) adapted from Golder Associates Limited (New Zealand office) and modified to suit the current research objectives.

Physiochemical sampling was conducted in conjunction with each macroinvertebrate sampling session, immediately prior to the collection of macroinvertebrates.

2.7 Habitat characteristics At each station habitat characteristics were assessed in conjunction with physicochemical and macroinvertebrate sampling to assist with interpretation of macroinvertebrate assemblage data. Habitat data were recorded on standard habitat assessment form. Habitat characteristic features assessed were: I. Channel Description

a) Channel width: channel width (m) was measured using a 30 m measuring tape. b) Water depth (m) at wadeable sites (depth=0.5m) was measured using a measuring staff or estimated at sites that were deep/unsuitable for macroinvertebrate community assemblage (i.e. >0.5 m). c) Water velocity: velocity (V) was calculated by timing (T) a bottle cap over 3 m lengths (D) (i.e. V = D/T) in riffle habitats.

II. Habitat type: the relative proportion (%) of each habitat type (e.g. run, riffle, pool and chute) present at each station was visually determined.

III. Substrate composition

Streambed substrate composition was assessed along riffle habitats at each station. While walking across the riffle habitat from the left to the right channel margin, 100 streambed particles were selected and a calibrated stick was used to measure the second longest axis of each particle. The size measurements were recorded following the Wolman scale (Wolman 1954), which groups particles into size categories given in Table 2.2. The substrate composition was measured as the proportion of each size class of particles following the Wolman scale (Wolman 1954).

Table 2.2: General streambed substrate classification with respective sizes

Substrate class Code Substrate size (mm) Bedrock Bedrock Boulder B >256 Large cobble LC 128-256 Small cobble SC 64-128 Large gravel LG 32-64 Medium-large gravel MLG 16-32 Small-medium gravel SMG 8-16 Small gravel SG 2-8 Sand-silt SS <2 Wood Wood

IV. Streambank stability and cover a) Streambank stability: visual characterisation of streambank stability (true left bank (left bank when facing downstream) and true right bank (right bank when facing downstream) at each station as (i.e. stable, partially stable or unstable). Evidence of erosion in the form or natural undercutting or accelerated erosion was noted (e.g. trampling by animals). b) Bank cover: visual characterisation of bank cover (vegetation, rootmass, boulders and stony walls) was noted at each station on both true left bank and true right bank.

V. Presence of organic matter: observation of woody debris, leaf litter and detritus at sampling stations were noted. These provide an indication of potential food availability for certain macroinvertebrate functional feeding groups or additional stable habitat.

VI. Riparian character and channel shade

37 a) A general assessment of riparian vegetation characteristics was carried out at each sampling station. A list of plant species according to height class (0-1m, 1-3m, >3m) was noted at each station on the true left bank and true right bank. This provided an indication of modification or intactness of riparian zone. b) The percentage channel shade at each site was visually estimated (% canopy cover).

2.8 Physicochemical sampling and analysis Water physicochemical parameters were measured at each station using a calibrated multi-water quality meter (Horiba 50) or (Aquaread AP 1000). The parameters measured included temperature, dissolved oxygen, conductivity, pH, TDS (total dissolved solids), turbidity and salinity. For nutrient and faecal coliform analysis, water samples were collected from each station, kept on ice and transported to the Institute of Applied Science (IAS) analytical laboratory for analysis. The nutrients analysed were ammonia, total phosphorus, nitrates and nitrites. Nutrient analyses were conducted in accordance with APHA 21st Edition 2005 (American Public Health Association 2005e, a, d, b). Faecal coliform analyses were conducted in accordance with APHA 21st Edition 2005 (American Public Health Association 2005c). The laboratory analyses were performed within 24 hours of the water sample collection.

2.9 Macroinvertebrate sampling and processing

Twp different sampling techniques (Surber sampler and kick-net) were used to collect macroinvertebrates during the current study. The sampling techniques were not used simultaneously as the Surber sampler was introduced later in Fiji. The researcher sampled the first two streams (Vucinivola and Taveu) in Ra during the wet season using a kick-net. The area sampled by a kick-net was not clear and therefore the data gathered was qualitative which was not suitable for the current study which required quantitative data. Then a Surber sampler was introduced to Fijian streams by a freshwater ecologist from New Zealand (Nick Carter) and the data collected via this technique was suitable

38 for quantitative analysis. A Surber sampler was then chosen to sample the remaining streams as it provided a defined area of collection (0.1m2) although it was applicable to water depths less than 0.5m and only to the riffle habitat.

2.9.1 Sampling technique one: kick-net sampler

This technique was used to sample Vucinivola and Taveu stream stations in Ra (Figure 2.3) during the wet season sampling. A kick-net sampler (approximately 1mm mesh, sampling area=0.05m2) (Figure 2.4) was used in riffle habitats to collect macroinvertebrates. The kick-net was placed downstream of rocks which were then disturbed using kick-motion; dislodging macroinvertebrates into the net. The samples were concentrated and placed as one sample into a 1000ml specimen jar. Following collection, samples were preserved in 70% ethanol.

Figure 2.4: kick-net sampling technique

2.9.2 Sampling technique two: Surber sampler

This sampling technique was used for sampling Naikawaqa stream stations in Ra during wet season as well as dry season sampling for all three streams in Ra and both season sampling for streams in Kadavu.

At each station six replicate Surber samples (0.5 mm mesh, sampling area=0.1m2) were collected from riffle habitats following Protocol C3 (Stark et al. 2001). Samples were collected by placing the Surber sampler over an area of streambed in a riffle habitat and then disturbing the habitat by washing the substrates within the Surber sampler frame with the water flowing through the net collecting dislodged macroinvertebrates (Figure 2.5). Contents of the net were then emptied into a bucket full of water to remove large leaves or gravel. Upon removal of the latter, the bucket water was sieved and sieve content was placed into a labeled 250ml specimen jar; samples were preserved in 70% ethanol.

Figure 2.5: Surber Sampling technique

2.9.3 Specimen identification

In the laboratory, the preserved samples were drained of ethanol and placed in a 34 x 27 cm sorting tray. Water was added to immerse the sample. All invertebrates were picked and sorted using a blunt tip thumb forceps. A compound microscope and a dissecting microscope were used for the identification of invertebrates. Preserved specimens were identified to genus or species level and in some cases higher taxonomic group such as family or order or class. The sorted and identified samples were then counted, recorded and eventually placed in a labeled vial; preserved in 70% ethanol.

Advice and assistance with identification was provided by Dr. Alison Haynes (Freshwater Molluscan Biologist-IAS/USP honorary fellow) and Mr. Johnson Seeto (Lecturer–School of Marine studies/USP). Specimen taxonomic identification was carried out following the guides of Haynes and Rashni (in prep.), Haynes (2009), Hasse et al. (2006), Jeng et al. (2003), Nandlal (unpublished), Williams (1980) and (Winterbourn et al. 2006).

2.10 GIS

2.10.1 Forest Cover analysis Watershed boundaries and size were determined by using Berkeley Image Segmentation and ArcGIS 10.x program. High resolution satellite (Quickbird 1.5m pan-sharpened) imagery was segmented using Berkeley Image Segmentation (http://berkenviro.com/berkeleyimgseg/) to produce a vector file of areas of similar spatial and spectral properties. These shapes were then overlaid onto the base imagery and manually assigned into a binary class structure of forested or non-forested. ArcGIS 10.x was then used to clip this vector file to the watershed areas that had been manually digitised used 1:50,000 topographic maps supplied by the Ministry of Lands and spatial analysis was performed to determine the area and therefore percent of each catchment that was forested and non-forested.

2.11 Statistical analysis

2.11.1 Macroinvertebrate Metrics The metrics used in the data analysis include:

2.11.1.1 Taxonomic metrics Macroinvertebrate samples collected were taxonomically identified to possible lowest level. Identified species were counted and placed under respective taxonomical groups (phylum, class, order, family and species). Due to the extensive data, the species were grouped into higher groups and tabulated for analysis. Raw data are presented as Appendices 7.5-7.7.

2.11.1.2 Functional Metrics- Functional Feeding group (FFG) Functional feeding group categories are based on morphological-behavioral mechanism of food acquisition by various macroinvertebrate families/taxa. Each taxa recorded from the survey was given a FFG category based on the classical system of FFG classification adopted from Merritt and Cummins (2006) given in Table 2.2.

Table 2.1.2: General classification of Functional Feeding Group categories for aquatic macroinvertebrates

FFG Categories Mechanism of food acquisition General particle size range of food (mm) Shredders Chew conditioned coarse particulate >1 organic matter (CPOM) Gathering-collectors Acquire CPOM from the streambed <1 Filtering-collectors Suspension or filter feeders <1 Scrapers Feed on surface algae <1 Predators Feed on living prey >1 Piercer-herbivore Suck contents of algal cells <1 There has not been an in-depth FFG study done on all freshwater macroinvertebrates in Fiji, aquatic insects in particular. However, the invertebrate families recorded to date have been found to be similar to those in Australia and New Zealand streams. Hence the FFG guides used in the current study include Winterbourn (2000), Merritt and

Cummins (2006) and The Murray-Darling Freshwater Research Center (2014). Each macroinvertebrate species recorded was assigned a FFG category. Relative abundance of each FFG per stream was calculated by averaging total abundance of FFGs of three reaches (lower, mid and upper). 2.11.1.3 Community variables

Diversity measures of macroinvertebrate samples included S (total number of unique species), N (total individuals) and (Loge) Shannon-Weiner Diversity index. These were calculated for individual samples in PRIMER v. 6.1.15 with the PERMANOVA add-on v. 1.0.5. A Euclidean distance similarity matrix was made between samples for each of these diversity measures individually. Using a single-term PERMANOVA model (term: Stream), PERMANOVA was performed under an unrestricted permutation of raw data model to determine whether differences in individual diversity measures between streams were significant. To determine which streams differed from which, pairwise PERMANOVA was again performed under an unrestricted permutation of raw data model. Graphs were drawn in R version 2.15.2.

2.11.2 Multivariate relationships All multivariate analyses were conducted in the package Plymouth Routines in Multivariate Ecological Analysis (PRIMER) v. 6.1.15 with the PERMANOVA add-on v. 1.0.5. Biotic data were firstly transformed to a similarity matrix using the Log (10) Modified Gower similarity measure at the individual sample level. Examination of the effect of sampling method was done visually using a Multi-Dimensional Scaling plot (MDS). Principle Coordinate Analysis was performed on biotic data for all data from both streams, and then separately for Ra and Kadavu streams. Correlations with Principle Coordinate plot axis are based on Pearson correlation coefficients; both for biotic and subsequently abiotic variables.

PERMANOVA tests were undertaken with the two-parameter model Season nested within stream on the Modified Gower similarity matrix using a Permutation of residuals under a reduced model system with the number of permutations set to 9999. Analysis

43 comparing biotic and abiotic variables was performed on standardised environmental variables to account for different units of measurement. Some of the environmental variables were dropped upon viewing correlation of paired environmental variables in a draftsman plot prior to computation. In the draftsman plot, the environmental variable nitrate and nitrite showed a high degree of correlation with each other. Retaining both of these variables would have shown no useful information and therefore only nitrate was chosen to be included in multivariate analyses. Nitrate was selected due to its high importance in stream ecosystems. Nitrate is a source of food for primary producers (periphyton) which are consumed by many benthic macroinvertebrates. Fluctuations in nitrate levels in streams due to transformation of land cover affect macroinvertebrate diversity (Scrimgeour & Kendal 2003; Herringshaw et al. 2011).

A few other environmental variables were not included in the analysis due to certain assumptions although they were recorded during sampling. These include turbidity, velocity, riffle depth, salinity and TDS. Turbidity impacts on stream biota are usually effectively observed over continuous monitoring over suitable monthly intervals. In the current study, turbidity was taken at one point in time which would not clearly reflect turbidity impacts on stream biota and therefore not tested. Velocity determines streambed substrate composition (habitat for biotic community) and therefore substrate composition was chosen over velocity for testing. The substrate composition was measured as the proportion of each size class of particles following the Wolman scale (Wolman 1954). Riffle depths at all stations were less than 0.5m and therefore assumed to be not a significant parameter to cause any change in biotic community composition. Salinity values were zero or slightly above zeros at all sites sampled since they were well above tidal influence (fully freshwater zones) and therefore assumed not to cause any significant change in biotic community. Total dissolved solids (TDS) was dropped for the same reason as turbidity.

It is important to note that for Naikawaqa stream, biotic and physiochemical data from mid and upper reach were excluded from all multivariate (except MDS analysis) and community variable analysis. Naikawaqa stream unlike the remaining five streams had

44 very high elevations at upper (286m) and mid reach (246m). To maintain consistency of elevation range at the three reaches per stream for the purpose of direct comparison of data, upper and mid reach data for Naikawaqa was excluded.

The Distance Based Linear Model (DistLM) procedure within the PERMANOVA add- on was used to examine which of the environmental parameters (either singularly or in combination) explained most of the variation in the ecological dataset.

Chapter 3. RESULTS & DISCUSSION OF BIOTIC (MACROINVERTEBRATE COMMUNITY STRUCTURE) VARIABLES

3.1 Sampling Method: Surber sampler and kick-net macroinvertebrate

community composition comparison

Across the 156 stream-reach-season samples, kick-net was used to cover 12 stream- reach season samples. A multi-dimensional scaling plot (MDS plot) was constructed from the Modified Gower similarity matrix based on log (x+1) transformed data (Figure 3.0). The MDS ordination plot showed that the macroinvertebrate assemblages collected by the two sampling techniques are not clustered together and thus not similar. The macroinvertebrate assemblages collected by the Surber sampler are distinct from the kick-net sampler and thus not comparable. To avoid the effect of sampling method overwhelming any patterns of macroinvertebrate community structure, only data from Surber sampler is analysed further in this chapter.

Figure 3.1: Multi-Dimensional Scaling ordination plot showing macroinvertebrate assemblage similarity using the kick-net and Surber sampler.

Various sampling techniques for freshwater invertebrates have been developed and widely employed by freshwater ecologists around the world. Freshwater macroinvertebrate studies are based on either quantitative or rapid-assessment (non- quantitative or semi quantitative) objectives and therefore require specific sampling techniques. Carter and Resh (2001) discussed several sampling methods used for macroinvertebrate surveys with the common ones including kick-type devices (e.g. D- frame sampler, slack sampler and kick-net), fixed-quadrat samplers (e.g. Surber and Hess sampler) and artificial substrates (e.g. multiplates and rock baskets).

In the present study of riffle dwelling macroinvertebrates, kick-net and Surber sampler were used. The kick-net sampling technique was used only during wet season in two streams (Taveu and Vucinivola) in Ra. The area sampled by kick-net was not clear and therefore the data gathered was qualitative which was not suitable for the current study which required quantitative data. A Surber sampler was then chosen to sample the remaining streams as it provided a defined area of collection (0.1m2) although it was applicable to water depths less than 0.5m and only to the riffle habitat. The riffle habitat supports a majority of invertebrates due to the following reasons, (a) heterogeneous

47 nature of substrates which provide shelter and or interstitial spaces for instars/larvae of various sizes; (b) swift water flow in riffle habitats provide plentiful oxygen and varying size of food particles for the invertebrate community (Ourso 2001; Tiemann et al. 2004; Birmingham et al. 2005). The riffle habitat may not reflect invertebrate species representative of a site as some species prefer slow flow runs and overhanging streamside vegetation (Friberg 2006). However, riffle-dwelling invertebrates are known to be the most sensitive to change in environmental conditions as they prefer high dissolved oxygen levels and clean water (Ourso 2001; Tiemann et al. 2004; Birmingham et al. 2005).

Despite its applicability to a single aquatic habitat, a Surber sampler was used to collect macroinvertebrates that are known to be the most sensitive to change in environmental conditions and therefore preferred for the current study. A direct comparison cannot be made between Surber and kick-net collections as both methods were not employed simultaneously at all sites. However an effort was made to test if the biotic community collected by kick-net was comparable to Surber sampler to assist with data analysis. As evidenced by the MDS ordination (Figure 3.1), macroinvertebrate assemblage collected by Surber sampler and kick-net were distinct.

Corroborating findings made in this study, many previous studies have recorded differences in macroinvertebrate taxa collected by kick-net and Surber sampler. For example, a study across eleven sites in south-western Australia over three seasons showed that the community assemblage of macroinvertebrates collected using Surber sampler versus kick-net was different in each season (Percentage similarity - Sorensen's and Czekanowski's coefficients, with mean values of 66% and 60% for Winter, 61% and 49% for Spring and 66% and 49% for summer respectively) (Storey et al. 1991). They also found that whilst the total number of individual organisms was lower, the Surber sampler collected more rare and cryptic taxa than the kick-net method. Therefore when a study on diversity is required, such as for an Environmental Impact Assessment, it is recommended that a Surber sampler be used. The current study contrasts with Storey et al. (1991) in that the Surber sampler collected higher total number of

48 individuals than kick-net when employed in the similar habitat. This could be due to difference in types of habitats sampled. The current study focused on riffle habitat only while Storey et al. (1991) sampled in multiple habitats.

Hornig and Pollard (1978) also agree that collections by kick-net and Surber sampler provided distinct estimates of species compositions and density. However, unlike the study by Storey et al. (1991), Hornig and Pollard (1978) found that the kick-net was able to capture a higher number of taxa with greater abundance per sample with similar or lower variability than the Surber sampler as well as being the most cost-efficient sampling technique. The current study in contrast with Hornig and Pollard (1978) found that a kick-net captured a lower number of taxa with lower abundance than a Surber sampler.

It is clear from two contrasting studies (Hornig & Pollard 1978; Storey et al. 1991) that the utility of either method is dependent on a number of factors possibly including the micro-habitat type and the overall type of macroinvertebrate community being sampled.

Both Storey et al. (1991) and Hornig and Pollard (1978) found that the Surber sampler collects more adherent macroinvertebrates (e.g. Hydroptila sp.) because it directly disturbs the substrate through scrubbing stones and stirring up sediments within an enclosed square frame while the kick-net collected loosely attached species such as the swimming mayfly (Baetis sp.) because it required disturbing substrates via foot (kick- motion) while holding onto the shaft and thus largely collected dislodged free-living species. The current study was in agreement to some extent with both Storey et al. (1991) and Hornig and Pollard (1978) whereby the Surber sampler was able to capture adherent species such as Oxyethira spp. while the kick-net collected mostly free-living species (e.g. Abacaria fijiana) or those adhering to upper rock surface (Pseudocloeon sp.) which could also be captured with the Surber sampler.

In the current study, the difference in macroinvertebrate assemblage collected by the two samplers is not only due to difference in method of collection but also attributed to

49 difference in mesh sizes and area of collection of the two samplers. In this study a square-framed Surber sampler with a mesh size of 0.5mm (sampling area=0.1m2/sample) and a kick-net of triangular-shaped net frame with a mesh size of approximately 1mm (sampling area=0.05 m2/sample) were used. According to Wisconsin Department of Natural Resources (2000), size of openings in a net affects the size and proportions of organisms collected. Finer mesh sizes allows collection of a high proportion of early instars of aquatic insects (Slack et al. 1991) while larger mesh sizes miss many of these and in particular a triangular-shaped net framed kick-net with relatively small sampling surface area is likely to lose many dislodged aquatic macroinvertebrates (Wisconsin Department of Natural Resources 2000). The presence or absence of early instars of aquatic insects greatly influences the results of quantitative descriptions in a survey (Huryn & Wallace 2000; Letovsky et al. 2012) and therefore explains the variation in macroinvertebrate community collected by two different samplers with varying mesh size and sampling area.

3.2 Biological communities: macroinvertebrate assemblages

3.2.1 Taxonomic Metrics- proportion of species at Group level A total of 35,334 individuals collected from 18 sites via both Surber and kick-net were identified to the lowest possible taxonomic level (Appendices 7.5-7.7). For the purpose of multivariate analysis only a total of 30,916 individuals were used as kicknet data was excluded. Macroinvertebrates collected were distributed across the taxonomic groups shown in Table 3.1. A total of 140 distinct taxa (Table 3.1) in 53 families (Appendices 7.7-7.7) were collected. The most diverse group was class Insecta with 87 taxa representing 62% of the total number of taxa recorded. Of the 87 taxa from the class Insecta, 26 were from the order Trichoptera (caddisflies), 18 were Diptera (true-flies) and 12 were Coleoptera (water beetles). The next most diverse taxonomic group was from the phylum Crustaceans (27 taxa) followed by the phylum Mollusca (21 taxa), phylum Annelida (4 taxa) and phylum Chelicerata (1 taxon).

Table 3.1: Number of macroinvertebrate taxa recorded in each of the taxonomic group across 18 sites

Phylum Class Order Sub- Common Number order name of species Arthropoda Insecta Trichoptera - Caddisfly 26 Ephemeroptera - Mayfly 3 Lepidoptera - Moth 6 Odonata Zygoptera Damselfly 9 Odonata Anisoptera Dragonfly 1 Hemiptera - Water Bug 9 Hymenoptera - Ants 2 Orthoptera - Water Cricket 1 Diptera - True-Fly 18 Coleoptera - Water Beetle 12 Chelicerata Arachnida Araneae - Water Spider 1 Crustacea Malacostraca Decapoda - Shrimp 20 Decapoda - Prawn 4 Kingdom : Animalia Decapoda - Crab 3 Mollusca Gastropoda - - Snail 21 Annelida Clitellata Oligochaeta - Worm 2 Hirudinea - Leech 1 Polychaeta Phyllodocida - Bristleworm 1

3.2.2 New additions to freshwater macroinvertebrate taxonomy of Fiji

Freshwater macroinvertebrates have received very little attention in Fiji. An exception is molluscan diversity, which has been well studied by Dr. Alison Haynes at both national and regional level (Haynes 1985, 1987, 1988a, b, 1989, 1990; Haynes & Kenchington 1991; Haynes 1994, 2001b, a, 2009). There are existing lists of terrestrial insect diversity based on entomology studies on adults (Tillyard 1924; Mosely 1941; Ross 1953; Kelley 1989; Donnelly 1990; Evenhuis 2007b, c, a; Johanson & Olah 2012). However, a majority of these adults are yet to be matched with their appropriate nymphs or larvae. The least studied groups include aquatic bugs, beetles, true-flies and aquatic worms. Decapod crustaceans (prawns and shrimps) have been studied to species level (Choy 1984; Shokita et al. 1985; Choy 1991; Nandlal 2010; Lal 2011) but a majority of the inland and coastal streams in Fiji isles are yet to be surveyed.

In the current study, overall taxonomic diversity of the six streams was higher than the current existing data for freshwater macroinvertebrate families of Fiji. Literature to date states that there are a total of 45 known freshwater macroinvertebrate families comprising of phylum Insecta (25 families), phylum Crustacea (4 families), phylum Mollusca (8 families), phylum Nematoda (2 families), phylum Annelida (3 families), phylum Platyhelminthes (1 family) and phylum Porifera (2 families) (Ryan 1980;Choy 1984; Shokita et al. 1985; Haynes 1988a, b, 1990; Choy 1991; Haynes & Kenchington 1991; Haynes 2001a; Jeng et al. 2003; Hasse et al. 2006; Haynes 2009; Nandlal 2010; Lal 2011). Results of this study present a total of 53 aquatic macroinvertebrate families including an additional 14 new records of families to phylum Insecta for Fiji. These 14 families (highlighted in grey in appendices 7.5 & 7.6) are: Crambidae (aquatic moth- Order Lepidoptera) Dixidae, Empididae and Stratiomyidae (aquatic true-flies-Order Diptera) Elmidae, Gyrinidae, Helminthidae, Hydraenidae, Hydrophilidae, Psephenidae and Spercheidae (aquatic beetles-Order Coleoptera) Mesoveliidae, Veliidae and Saldidae (water bugs-Order Hemiptera)

These additional scientific data on freshwater macroinvertebrate diversity of Fiji were collected due to the area of study being surveyed for the first time and also because of the Surber sampler technique being used for the first time. There are high chances of these families being present in many streams yet to be surveyed. At this stage, it is not possible to declare if aquatic species belonging to these families are new to science or endemic to Fiji. This would require further molecular biology work.

Presence of the endemic shell-less gastropod, Acochlidium fijiense (picture given on the title page of the thesis), in the Vucinivola stream of Ra province was another significant finding as it has contributed to the existing biogeographical distribution of this species in Fiji. Prior to the current study, A. fijiense has only been recorded from two sites in Fiji. In the year 1983, A. fijiense was recorded seven kilometers upstream from the mouth of Nasekawa River (Vanua Levu) and four kilometers upstream from the mouth of Lami River (Viti Levu) in 1988.

3.2.3 Functional Metrics- proportion of macroinvertebrate in Functional Feeding Groups Functional Feeding Group (FFG) classifications were adapted from Merritt and Cummins (2006) and corresponding data per stream were graphed using Microsoft Excel. An overview of macroinvertebrates and their FFG categories is presented in Appendix 7.3 with the relative proportions of each group at each stream shown in Figure 3.2. Macroinvertebrates within the Predator FFG were comprised of the largest number of taxa across the streams sampled. Predators recorded included caddisflies, damselflies, dragonflies, waterbugs, water beetles, bristle worms, leeches, crabs and prawns (Appendix 7.3). Predators were represented by 40 taxa (Figure 3.2) but with relatively low abundance; comprising between 2% (Vucinivola, Savumiri and Naqewa streams) and 17% (Korolevu stream) (Figure 3.2 and 3.3) of the total community abundance at sites sampled. Scrapers recorded included caddisflies, moths, beetles and snails (Appendix 7.3). They were the second highest number of taxa for a functional feeding group recorded being represented by a total of 31 species dominated by the Neritid snails. Shredders and piercers-herbivores had the fewest number of taxa in a FFG represented by 10 and 3 species respectively (Figure 3.2).

45 40 35 30 25 20 15

Number of Taxa of Number 10 5 0 Scraper Predator Shredder Piercers-herbivore Filtering-collectors Gathering-collectors Fuctional Feeding Group Categories (FFG)

Figure 3.2: Total taxa number per Functional Feeding Group across all study sites 54 In Kadavu streams filtering-collectors represented between 39% (Korolevu stream) and 64% (Naqewa stream) (mean: 50%) of total community abundance. In Ra streams, filtering-collectors made up between 29% (Naikawaqa stream) and 40% (Taveu stream, dry-season only) of the total community abundance (mean: 35%). Gathering-collectors were the second most abundant FFG. They comprised between 26% (Naqewa stream) and 41% (Savumiri stream) (mean: 31%) of total abundance in Kadavu streams and between 19% (Naikawaqa stream) and 51% (Vucinivola stream, dry-season only) (mean: 31%) in Ra streams (Figure 3.3).

Scrapers comprised between 10% (Vucinivola stream) and 39% (Naikawaqa stream) (mean: 27%) in Ra streams and between 7% (Naqewa and Savumiri stream) and 12% (Korolevu stream) (mean: 9%) of the total abundance in Kadavu streams. Piercers- herbivores were only present in Kadavu streams and only in two streams; Savumiri (4% of total abundance) and Korolevu (5% of total abundance)) (mean: 3%). Shredders were virtually absent from both Ra and Kadavu streams, recorded only in Naikawaqa stream of Ra (2% of total community abundance) and Korolevu stream of Kadavu (1% of total community abundance) (Figure 3.3).

Kadavu Ra 100% 90% Filtering-collectors 80% 70% Gathering-collectors 60% Scraper 50% 40% Predator 30% Shredder 20% Community composition composition Community 10% Piercers-herbivores 0% Naqewa Korolevu Savumiri Naikawaqa Taveu Vucinivola Stream

Figure 3.3: Proportion of total abundance that each functional feeding group made at streams in Kadavu and Ra

In the current study Functional Feeding Groups (FFG) were only used for descriptive purposes; relative abundance of macroinvertebrates in specific FFGs per stream. This is because no detailed study on Fiji freshwater invertebrates (except gastropods) has been conducted in relation to either morpho-behavioral mode of food acquisition by specific family/genus/species or gut content analyses. The FFGs assigned to genus or families in the current study were adapted from Merritt and Cummins (2006) and The Murray- Darling Freshwater Research Center (2014).

The study streams in both Kadavu and Ra had more filtering-collectors (mean: 50% and 35% respectively) than any other type of consumer followed by gathering-collectors. This result is counter to findings in Haynes (1999b) where scrapers were the most abundant FFG in both Nabukavesi and Wainikovu stream representing 68% and 65% of the total community abundance respectively. Scrapers in the current study were the third most abundant FFG in both Kadavu and Ra (mean: 9% and 27% of the total community abundance respectively).

The contrast in the findings of the current study and Haynes (1999b) is likely due to difference in sampling methodologies employed, habitats sampled as well as period of sampling. The current study employed sampling with a Surber sampler (mesh size- 0.5mm) in riffle habitat only while Haynes (1999b) used a square net (mesh size-1mm) held downstream while 15 stones each (120-300mm across) were sampled in a riffle and run habitat and six net (mesh size-1mm) sweeps of overhanging vegetation. Haynes (1999b) sampled riffles, runs and streamside habitats which collaboratively presented invertebrate diversity representative of the site. In the current study only riffle habitats were sampled due to applicability of Surber sampler in water depth below 0.5m and this prevented collection of invertebrates from run pool and edge habitats. Also the smaller mesh size (0.5mm) of Surber sampler used in the current study possibly captured more insect instars compared to the square net of 1mm mesh size used by Haynes (1999b). The current study involved sampling once at the peak of the wet and dry seasons while Haynes (1999b) sampled at intervals of two month for three years. Due to time and fund limitations of the current study, data presented were collected at two points in time

56 while Haynes (1999b) collected regular data for three years and could better assess population change occurring due to natural processes. For instance Haynes (1987) found that gastropod (scrapers) population and biomass was high during the warmer season (November to March) which encompasses the wet season for our study in Fiji. During the warmer season, biofilms on rock surfaces are plentiful. The gastropods (scrapers) graze on these biofilms and thus an increase in their abundance without significant decrease in other benthic invertebrates (mostly confined to underside of rocks) (Haynes 1987) and therefore a higher scraper representatives in Haynes (1999b).

In the current study, shredders were only present in Naikawaqa (94% forested) and Korolevu (64% forested) although in very low proportions (2% and 1% of the total community abundance respectively) and this is explained by availability of forest cover and well vegetated riparian zone in these forested catchment streams. Shredders naturally contribute only a minor component of macroinvertebrate community biomass in Fiji and Pacific Island streams (Bright 1982; Resh et al. 1990; Haynes 1999b). The low proportions of shredders is explained by the nature of leaves (food) entering streams from surrounding native forests, which tend to be tough with thick cuticles that are broken down slowly (Haynes 1999b). It should be noted that the Korolevu stream although less forested than Naqewa (91% forested) had Shredders and this is because the area of the catchment that Korolevu drains has been left unfarmed for 20 years which allowed the riparian zone community to recover and possibly eventuating in colonization of shredders.

3.2.4 Macroinvertebrate community variables-diversity measures

Diversity measures S (total number of species), N (total individuals) and Loge (Shannon-Weiner Diversity index) were calculated for individual samples in PRIMER v. 6.1.15 with the PERMANOVA add-on v. 1.0.5. PERMANOVA results showed that there were significant differences between the streams for all three diversity parameters (S: p-value (under permutation) = 0.001 (Appendix 7.4: Table 7.2); N: p-value (under permutation) = 0.0072 (Appendix 7.4: Table 7.4) and Shannon-Weiner: p-value (under permutation) = 0.001 (Appendix 7.4: Table 7.6). To determine which streams differed from which, pairwise PERMANOVA, performed under an unrestricted permutation of the raw data model, were undertaken. The results of these are presented in Appendix 7.4: Tables 7.3, 7.5 and 7.7.

When interpreting these differences in the diversity measures, some of the sampling deficiencies in Naikawaqa, Taveu and Vucinivola streams must be borne in mind. For Naikawaqa stream, only data from the lower-reach (n=12, whereby n=no. of samples) and for Taveu and Vucinivola streams, data from all three reaches during the dry season (n=18) was used for the analysis while the remaining streams had data included from all three reaches (n=36). As explained previously in the methodology chapter, the upper and mid reach sites of the Naikawaqa stream were outliers and therefore data collected from these reaches were not included in the community variable analysis. Taveu and Vucinivola streams were sampled using a Surber sampler during the dry season only and therefore only dry season data was used for community variable analysis.

3.2.4.1 Macroinvertebrate community variables association with catchment forest cover

Many studies have shown the effects of forests on macroinvertebrate communities whereby different forest systems are associated with different water quality and macroinvertebrate assemblages (Cummins et al. 1989; Eggert & Burton 1994; Friberg et al. 1997; Whiles & Wallace 1997). Riparian forest is closely connected to stream biotic communities through two major phenomena: (1) its contribution to stream food webs through the input of leaf detritus which acts as primary energy supply (Sabater et al. 2000; Kominoski et al. 2012) and (2) its support of the terrestrial adult (reproductive

58 phase) which is essential for determining the abundance and distribution of the larval/nymphs stages of aquatic insects in streams (Huryn & Wallace 2000). Therefore, any alteration of riparian plant diversity or abundance may influence stream macroinvertebrate community structure.

3.2.4.2 Macroinvertebrate species richness across streams The mean species richness per sample ranged from 10 taxa in the highly forested Naikawaqa stream (n=12) to 18 taxa in the medium-forested Korolevu stream (n=36). Macroinvertebrate species richness (S) in Korolevu stream was not significantly different from that in Naqewa stream (Kadavu province) and Taveu stream (Ra province). Naikawaqa stream of Ra province had the lowest number of species and was significantly different from the rest of the streams (Figure 3.4).

Figure 3.4: Mean number of species collected across six streams. Error bars represent standard error. Shared letters for each stream denote no significant difference (p>0.05) based on pairwise PERMANOVA tests. The first three bars represent streams from Kadavu province (Korolevu, Naqewa and Savumiri) and the last three bars represent streams from Ra province (Naikawaqa, Taveu and Vucinivola). The level of catchment forest cover (HF= highly forested, MF= moderately forested and LF= least forested) is indicated on each column.

3.2.4.2.1 Macroinvertebrate species richness association with forest cover

Recent studies (Sponseller et al. 2001; Roy et al. 2003) have shown a clear cut relationship between catchment land cover and macroinvertebrate species richness. Sponseller et al. (2001), in a study of catchment land use and macroinvertebrate communities, found that highly forested catchments (percent forested: 92.5, 91.6 and 86.1) had high species richness (species number: 51, 47 and 54, respectively) while less forested catchments (percent forested: 80.7, 79.7 and 56.6) had low species richness (species number: 30, 45 and 24). Roy et al. (2003) found a clear cut relationship between catchment land cover and macroinvertebrate taxon richness. In a study of 30 streams in the Etowah River basin, Georgia, U.S.A, examining relationships between land cover in catchments and macroinvertebrate assemblage attributes, Roy et al. (2003) found that taxon richness was positively related to forested catchment (forest land cover) and negatively related to deforested catchments (urban land cover).

It was expected, based on these published studies, that higher species richness would be associated with a higher percentage forest cover in the catchment. This relationship was not borne out by the results in the current study whereby a total of only 10 species were recorded in the catchment with the highest forest cover (percent forested: 94%). The remaining highly and moderately forested catchments (percent forested: 91%, 79% and 64%) had similar species richness (species number: 17, 16 and 18, respectively) compared to the least forested catchments (percent forested: 50% and 29%, species number: 16 and 16).

This was possibly due to the following reasons: (a) the current study data being just a ‘snapshot’ (data collected only once per wet and dry season) of the stream and therefore the taxa richness for the dry and wet season may not have been well-captured. A better representation of taxa richness would have been possible through monthly data collection whereby macroinvertebrate species richness would have indicated possible changes in catchment forest cover or any disturbance to riparian forest zone, (b) the exclusion of mid and upper reach data from Naikawaqa stream and lack of wet season Surber sample data from Taveu and Vucinivola streams may have led to mis-

60 representation of actual macroinvertebrate community structure for the study as a whole.

3.2.4.3 Abundance of macroinvertebrates across streams

The total number of macroinvertebrates was highest in Savumiri stream (233 individuals) of Kadavu province and it was not significantly different from Naqewa (213 individuals) (Kadavu province) and Vucinivola (197 individuals) (Ra province) streams. Naikawaqa (124 individuals) and Taveu (125 individuals) streams of Ra province had the lowest number of total macroinvertebrates and they were not significantly different from Vucinivola (Ra province) and Korolevu (Kadavu province) (Figure 3.5).

Figure 3.5: Mean number of individuals in the macroinvertebrate community collected across six streams. Error bars represent standard error. Shared letters for each stream denote no significant difference (p>0.05) based on pairwise PERMANOVA tests. The first three bars represent streams from Kadavu province (Korolevu, Naqewa and Savumiri) and the last three bars represent streams from Ra province (Naikawaqa, Taveu and Vucinivola). The level of catchment forest cover (HF= highly forested, MF= moderately forested, LF= least forested) is indicated on each column.

3.2.4.3.1 Macroinvertebrate abundance association with forest cover

The macroinvertebrate diversity graph (Figure 3.5) showed that there was no relationship between macroinvertebrate taxa abundance and catchment forest cover. A possible reason for this could be differences at micro-habitat level factors such as (a) food availability (Gray & Ward 1979) at sampling sites which is dependent on riparian forest composition, canopy cover and light availability (Merritt and Cummins 2006) and (b) rock type preferences by macroinvertebrate taxa (Mary 2002). This hypothesis cannot be further discussed as in-stream food availability and rock type data were not collected during the current study. However recent studies (Sponseller et al. 2001; Roy et al. 2003) have shown strong correlation between catchment forest cover and macroinvertebrate taxa abundance.

3.2.4.4 Shannon Wiener diversity index across streams

Shannon Wiener diversity analyses showed that Korolevu stream (Shannon-Wiener Index of 2.1) of Kadavu province had the highest macroinvertebrate diversity and was significantly different from the rest of the streams. Naikawaqa stream (Shannon-Wiener Index of 1.1) of Ra province had the lowest macroinvertebrate diversity and was significantly different from the rest of the streams. Naqewa (Shannon-Wiener Index of 1.6), Savumiri (Shannon-Wiener Index of 1.7) (Kadavu province), Taveu (Shannon- Wiener Index of 1.8) and Vucinivola (Shannon-Wiener Index of 1.8) (Ra province) had intermediate macroinvertebrate diversity and were not significantly different from each other (Figure 3.6).

Figure 3.6: Mean Shannon Weiner diversity Index measured across six streams. Error bars represent standard error. Shared letters for each stream denote no significant difference (p>0.05) based on pairwise PERMANOVA tests. The first three bars represent streams from Kadavu province (Korolevu, Naqewa and Savumiri) and the last three bars represent streams from Ra province (Naikawaqa, Taveu and Vucinivola). The level of catchment forest cover (HF= highly forested, MF= moderately forested, LF= least forested) is indicated on each column.

64 3.2.4.4.1 Macroinvertebrate diversity association with forest cover

The macroinvertebrate diversity graph (Figure 3.6) showed that there was no relationship between species diversity and catchment forest cover. This is possibly due to the diversity of macroinvertebrates being more associated with riparian forest composition (Yoshimura & Maeto 2006; Compson et al. 2013; Lester et al. 1994; Piccolo & Wipfli 2002; Volk et al. 2003) than forest cover. This hypothesis agrees with Snyder et al. (2002) who found a strong correlation in macroinvertebrate community structure and forest composition during a study in North America investigating the macroinvertebrate community structure in streams draining hemlock and mixed hardwood forests. They found that more than eight percent of the taxa were strongly associated with hemlock forest. They also found that the streams draining hemlock forests supported significantly lower number of rare species and total densities compared to mixed hardwood forests.

3.2.5 Principal Coordinate Analysis (PCO) of macroinvertebrate assemblages: both Ra and Kadavu In the biotic community (macroinvertebrate assemblage) PCO for all six streams, PCO axes 1 (PCO1) and 2 (PCO2) explained the most variation among sites (14.5 and 8.8% of the total variation, respectively). Examination of the PCO plot shows that Naikawaqa and Taveu separate from Korolevu and Savumiri along PCO axis 1 in the Modified Gower similarity matrix. This is largely explained by the freshwater gastropod Neritilia rubida (Pearson correlation coefficient: PCO1=-0.68, PCO2=0.06) where N. rubida is more common in Naikawaqa and Taveu. Separation along PCO2 separates Korolevu and Savumiri from Naqewa and Taveu. The correlation with the abundance of individual species along PCO2 is less clear, with a mineral case-maker caddisfly Odontoceridae sp. B (Pearson correlation coefficient: PCO1=-0.30, PCO2=0.40), two net-spinner caddisfly larvae Abacaria ruficeps (Pearson correlation coefficient: PCO1=- 0.48, PCO2=-0.39), and Abacaria fijiana (Pearson correlation coefficient: PCO1=-0.39, PCO2=-0.45), and a freshwater gastropod Septaria sanguisuga (Pearson correlation coefficient: PCO1=-0.03, PCO2=-0.52) (Figure 3.7), all appearing to ordinate along the PCO2 axis. The mineral case-maker caddisfly Odontoceridae sp. B is more abundant in Korolevu and Savumiri than the two net-spinner caddisfly larvae A. fijiana, A. ruficeps and the freshwater gastropod, Septaria sanguisuga. The freshwater gastropod Septaria sanguisuga is more common in Naqewa than the other streams while the two net- spinner caddisfly larvae A. fijiana and A. ruficeps are more common in Vucinivola stream (Figure 3.7).

Note: Kadavu streams include Korolevu, Naqewa and Savumiri and Ra streams include Naikawaqa, Taveu and Vucinivola streams.

67 3.2.6 Principal Coordinate Analysis (PCO) of macroinvertebrate assemblages- Kadavu streams In the biotic community PCO for the Kadavu streams, the first two axes explained 23.2 percent (13.4% and 9.4% of the total variation, respectively) of the variation. Examination of the PCO plot shows that Korolevu stream separates from Naqewa along PCO axis 1 in the Modified Gower similarity matrix. This is largely explained by the mineral case-maker caddisfly Odontoceridae sp. B (Pearson correlation coefficient: PCO1=-0.63, PCO2=0.32) where Odontoceridae sp. B is more common in Korolevu stream. Separation along PCO2 separates Naqewa from Savumiri. This separation is explained by three species; a freshwater gastropod, Septaria sanguisuga (Pearson correlation coefficient: PCO1=0.50, PCO2=0.39), a finger-net caddisfly, Chimarra sp. A (Pearson correlation coefficient: PCO1=0.12, PCO2=0.50) and a riffle shrimp, Atyopsis spinipes (Pearson correlation coefficient: PCO1=0.49, PCO2=0.25) (Figure 3.8). The freshwater gastropod, Septaria sanguisuga, the finger-net caddisfly, Chimarra sp. A and a riffle shrimp, Atyopsis spinipes, are more abundant in Naqewa than Savumiri.

Figure 3.8: PCO ordination plot showing macroinvertebrate assemblage similarities among three streams in Kadavu. The vectors in the plot represent the species most responsible for the dissimilarities between sampling sites.

3.2.7 Principal Coordinate Analysis (PCO) of macroinvertebrate assemblages- Ra streams In the case of streams in Ra province; the first two axes of PCO explained 35.1 percent (26.4% and 8.7% of the total variation, respectively) of the variation in macroinvertebrate assemblage (Figure 3.9). Examination of the PCO plot shows that Vucinivola stream separates from Naikawaqa and Taveu along PCO axis 1 in the Modified Gower similarity matrix. This is largely explained by three species, a mayfly Pseudocloeon sp. (Pearson correlation coefficient: PCO1=-0.71, PCO2=0.15) and two net spinner caddisflies, Abacaria fijiana (Pearson correlation coefficient: PCO1=-0.62, PCO2=0.03) and Abacaria ruficeps (Pearson correlation coefficient: PCO1=-0.67, PCO2=0.05) whereby these three species are more common in Vucinivola stream. Separation along PCO2 separates Naikawaqa stream from Taveu. The correlation with species abundance along PCO2 is less clear, with a riffle shrimp Atyoida pilipes (Pearson correlation coefficient: PCO1=0.33, PCO2=0.65), two neritid snails, Neritina squamaepicta (Pearson correlation coefficient: PCO1=0.43, PCO2=0.63) and Neritilia rubida (Pearson correlation coefficient: PCO1=0.72, PCO2=0.47) and a shrimp, Caridina typus (Pearson correlation coefficient: PCO1=0.52, PCO2=-0.38), all appearing to ordinate with PCO2 axis. The riffle shrimp, Atyoida pilipes, and two neritid snails, Neritina squamaepicta and Neritilia rubida, are more abundant in Naikawaqa while the shrimp, Caridina typus is more abundant in Taveu (Figure 3.9).

Figure 3.9: PCO ordination plot showing macroinvertebrate assemblage similarities among streams in Ra. The vectors in the plot represent the species most responsible for the dissimilarities between sampling sites.

3.2.8 General discussion on PCO analysis: macroinvertebrate association with

catchment forest cover

Several studies have shown that catchment forest cover strongly influences biotic diversity in streams (Sandin & Johnson 2004; Stephenson & Morin 2009; Weijters et al. 2009). It is widely held that reduction in forest cover results in a decrease in biodiversity (Allan 1995). However at the species level, reduction in forest cover is more complicated as indirect effects of forest cover reduction such as increase in light availability, increased nutrient runoff and sedimentation tend to favor some species over others (Newbold et al. 1980 ; Stone & Wallace 1998; Sonoda et al. 2001; Quinn et al. 2004). In the current study, certain macroinvertebrate species were more abundant than the others across streams draining catchments with varying percent forest cover.

3.2.8.1 Macroinvertebrate taxa associated with least forested catchments

In the current study, the clinging mayfly nymph Pseudocloeon sp. and the net spinning caddisfly larvae Abacaria fijiana were most associated with streams (Vucinivola and Savumiri) draining the least forested catchments.

Table 3.2: Number of individual macroinvertebrate taxa per stream associated with the PCO analysis.

Taxa Streams Naikawaqa Naqewa Taveu Korolevu Vucinivola Savumiri Pseudocloeon 377 1274 417 713 1357 1613 sp. Abacaria 313 3315 535 1441 891 4164 fijiana

The clinging mayfly nymph Pseudocloeon sp. was more abundant in Savumiri and Vucinivola than Naikawaqa, Naqewa, Korolevu and Taveu (Table 3.2). The net spinning caddisfly larvae Abacaria fijiana was more abundant in Savumiri and Naqewa than Korolevu, Vucinivola, Taveu and Naikawaqa (Table 3.2).

Savumiri and Vucinivola streams had highly modified riparian zones (weedy vegetation and eroded banks), possibly suggesting that A. fijiana and Pseudocloeon sp. are more resilient to major effect of forest cover removal such as siltation in streams. This finding agrees with Haynes (1999b) who found that A. fijiana and Pseudocloeon sp. were more abundant in a stream (Nabukavesi) draining a logged catchment than one (Wainikovu) draining an unlogged catchment. Logging occurred 1-2 km upstream of the sampling site at Nabukavesi stream, A. fijiana and Pseudocloeon sp. accounted for 61% of the total invertebrate community abundance and only 42% of the total invertebrate community abundance at Wainikovu stream. An interesting finding by Haynes (1999b) was the persistence of A. fijiana and Pseudocloeon sp. during the logging period (1989- 1990) when sedimentation was high causing stream discoloration and a thick layer of mud covering much of the stony streambed. Moreover, A. fijiana and Pseudocloeon sp. were also the first species to appear immediately after logging was stopped (Haynes 1994, 1999b).

3.2.8.2 Macroinvertebrate taxa associated with highly forested catchments

Species most associated with the streams (Naikawaqa and Naqewa) draining the highly forested catchments included the finger-net Caddisfly Chimarra sp. A, the atyid shrimps Atyoida pilipes and Atyopsis spinipes and the small neritid gastropods Neritilia rubida, Neritina squamaepicta and the limpet gastropod Septaria sanguisuga, suggesting the species preference for well vegetated catchment streams and sensitivity to sediments.

Table 3.3: Number of individual macroinvertebrate taxa per stream associated with the PCO analysis.

Taxa Streams Naikawaqa Naqewa Taveu Korolevu Vucinivola Savumiri Chimarra sp. 154 194 77 170 30 27 A Atyoida 141 87 12 6 1 5 pilipes Atyopsis 26 148 24 6 1 4 spinipes Neritilia 1002 64 494 5 5 10 rubida Neritina 30 0 9 0 1 0 squamaepicta Septaria 0 65 4 0 2 0 sanguisuga

The finger-net Caddisfly Chimarra sp. A was more abundant in Naqewa, Korolevu and Naikawaqa than Taveu, Vucinivola and Savumiri (Table 3.3). The atyid shrimp Atyoida pilipes was more abundant in Naikawaqa and Naqewa compared to Taveu, Korolevu, Savumiri and Vucinivola (Table 3.3). The atyid shrimp Atyopsis spinipes was more abundant in Naqewa, Naikawaqa and Taveu compared to Korolevu, Savumiri and Vucinivola (Table 3.3).

The small neritid gastropod Neritilia rubida was more abundant in Naikawaqa, Taveu and Naqewa compared to Savumiri, Korolevu and Vucinivola (Table 3.3). The neritid gastropod Neritina squamaepicta was more abundant in Naikawaqa than Taveu and Vucinivola and absent in Kadavu streams (Table 3.3). It should be noted that Neritina squamaepicta was also not found in Kadavu streams previously (Haynes 1988a). The limpet gastropod Septaria sanguisuga was more abundant in Naqewa than Taveu and Vucinivola (Table 3.3) and absent from Korolevu, Savumiri and Naikawaqa. It should

74 be noted that only lower reach (mid-reach and upper reach excluded due to them being outliers for the elevation factor) data from Naikawaqa stream was used for PCO analysis where S. sanguisuga was not present. However, at this site less than 5mm juvenile (Septaria spp. - 40 individuals) limpet snails were found which were not identifiable to species level. The lower reaches of streams in Fiji are known for the presence of highly abundant young gastropods because the veligers swept by sea currents tend to settle at the stream/river mouth (Haynes 1988a, b, 1989, 1999b). In the mid and upper reaches of the Naikawaqa stream, S. Sanguisuga were present on submerged boulders but not in riffle habitat where the researcher conducted sampling.

The finger-net caddisfly Chimarra sp. A, belongs to the family Philopotamidae. This family is rated highly sensitive in Australian streams with a SIGNAL2 sensitivity grade of 8 (SIGNAL2 grade range= 1-10 where 1=resilient taxa and 10=highly sensitive taxa) (McCulloh 2009). Studies of streams in isolated intact primary forest catchments of Fiji such as the Emalu (Rashni 2013) area which consists of three (Tovatova, Mavuvu and Waikarakarawa) catchments have shown high abundance of Chimarra sp. as compared to a study in the Sabeto river of Nadi, Fiji. The Sabeto river watershed has been exposed to multiple anthropogenic disturbances such as gravel extraction, farming, burning of forest, grazing, urban sprawl and exploratory mining (Naiova 2013). The average density of Chimarra sp. A was 168 individuals/m2 in Emalu streams and 1 individual/m2 in the Sabeto river.

Haynes (1999b) found that the Atyid shrimp Atyoida pilipes were threefold more abundant (mean abundance: 1.13) in Wainikovu stream draining the unlogged catchment than Nabukavesi stream (mean abundance: 0.40) draining the logged catchment. Atyopsis spinipes was one of the three species in Nabukavesi stream which did not return after five years since the cessation of logging although the shrimp was present in Wainikovu stream draining the unlogged catchment (Haynes 1994, 1999b). Atyoida pilipes and Atyopsis spinipes were not found in Sabeto River (Maleli 2013), presumably because the continuous gravel extraction activity had modified the substrate

75 structure and overhanging vegetation coupled with alteration in flow regime and thus preventing suitable habitat conditions for the shrimp population to thrive.

Gastropods and decapod crustaceans (shrimps and prawns) represent the largest proportion of invertebrate biomass in tropical Pacific streams including those in Fiji (Haynes 1987, 1988a, 1990, 1994). The gastropods are the main scrapers in the fast flowing streams in Fiji and they graze with their radula teeth on biofilms covering the rocks in streams (Haynes 1988a, b, 1989, 1999b). The biofilms can be smothered by sediment loads caused by human induced activities such as forest removal and faulty farming practices (Haynes 1989, 1999b) which not only tend to deprive the gastropods of food, but also clog their respiratory structures leading to death and eventually decline in their population (Haynes 1989). A study in the Naisogo stream of Ovalau island of Fiji over a year by Haynes (1989) showed that conversion of hillside forest cover to village garden resulted in rapid flooding and siltation which negatively affected the neritid limpet gastropod (Septaria sanguisuga, Septaria porcellana, Septaria suffreni and Septaria macrocephala) population. The density of Septaria spp. decreased from a total of 118 snails/m2 in March 1987 to 51 snails/m2 in March 1989; more than a 50% decrease in population. The 20mm thick mud generated as a result of stripping of hillside vegetation covered the substrate and smothered the periphyton which the Septaria spp. were feeding on. The smothering of periphyton led to a scarcity of food for the Septaria snails and as resulting decreases in their population. This study (Haynes 1989) on Septaria snails verifies gastropod sensitivity to stream sedimentation.

In the current study, the high abundance of the neritid gastropod Septaria sanguisuga in Naqewa stream could possibly be due to suitable habitat availability. Septaria sanguisuga generally thrives in streams with clear water and high flow (approximately 1m/sec) (Haynes 1988a) and was abundant in Naqewa stream which had a gorge present upstream and mostly sloping stable aquatic habitats. Additionally, the well- forested catchment coupled with well-vegetated riparian zone may have prevented sediment loads in Naqewa and thus providing clear water for S. sanguisuga to thrive.

Neritilia rubida is one of the widely distributed Indo-Pacific gastropods and prefers similar habitat conditions across the Pacific island streams (Haynes 1999b). This gastropod is known to colonize short (1-3km) coastal streams (Haynes 2000, 2001a) and in this study they were found to be most abundant in the lower reach of Naikawaqa stream. A study conducted on the island of Moorea in French Polynesia by Haynes (1999a) showed that Neritilia rubida was more abundant in Afareaitu River (53 snails/m2) than Opunohu river (0 snails/m2) and Maatea river (3 snails/m2). The Afareaitu river catchment was unfarmed while pineapple plantations were observed in catchments drained by Opunohu River and Maatea River (wetted channel width above tidal range = 2-5m (pers. comm. Haynes 2014)). It should be noted that the channel width of Opunohu River and Maatea River in French Polynesia is approximately similar to the streams (wetted channel width above tidal range = 2-6m) in Ra and Kadavu. Neritina squamaepicta is also known to occur in short coastal streams and like any other gastropod it would prefer clear and well oxygenated water and hence its abundance in forested catchment stream.

3.2.9 Patterns in macroinvertebrate assemblage composition and variability A PERMANOVA analysis was done with the model season nested within stream on the Modified Gower log10 transformed similarity matrix of macroinvertebrate abundance data. The results (Table 3.2) indicate that both the terms season (stream) and stream are significant, indicating that macroinvertebrate community structures between seasons within streams and across streams are significantly different from each other. However, perhaps a more important result is the very high residual sum of squares left after the two model parameters have been fitted (70.9 out of the total sum of squares of 104.8). This indicates that for samples within the same streams in the same seasons there is very high variation in the macroinvertebrate community structures.

Table 3.4: PERMANOVA results on the macroinvertebrate data examining the effects of season within streams.

Modified Gower similarity Source of variation df SS MS Pseudo-F P(perm) perms Stream 5 25.834 5.1668 10.637 0.0001 9801 Season (stream) 4 8.091 2.0227 4.1642 0.0001 9766 Residual 146 70.919 0.48575 - - - Total 155 104.84 - - - -

3.2.9.1 Inter sample variation within streams

In the current study, macroinvertebrate taxa showed great inter-sample variation within streams during wet and dry season. The inter-sample variation is most likely due to the reach and stream-scale patchy nature of macroinvertebrates not accurately captured by a fixed quadrat frame of a Surber sampler. Although fixed-quadrat replicates are generally considered to have higher precision and mostly represented by lower coefficients of variation, not all studies have agreed on this. For example, Hornig and Pollard (1978) found that Surber sample collections had high variability; average coefficients of variation ranging between 50 percent (U.S. Environmental Protection Agency 1973) and over 80 percent (Hornig & Pollard 1978). Downes et al. (1993) found that significant variation in macroinvertebrate densities occurred between riffles and groups of stones within same riffle. Pollard and Kinney (1979) concluded that

78 relatively high variability in collections of small-area methods such as Surber sampler is inevitable due to the inherent patchiness of the stream dwelling benthic macroinvertebrate fauna.

Another likely reason for the variation could be diverse life histories of macroinvertebrates at genus or species level. Life histories of macroinvertebrates are complicated; species in a specific invertebrate group tend to have a distinct life history period (Moss 1980). The life history of each distinct freshwater macroinvertebrate group is not well studied in Fiji, except for gastropods (Haynes 1985, 1988a, b; Haynes & Kenchington 1991; Haynes 1998, 2001a, 2009) and crustaceans (Nandlal 2010; Lal 2011). A study in Nabukavesi and Wainikovu streams of Fiji over three years showed almost similar fluctuations in the abundance of the Mayfly nymph Pseudocloeon sp. and the Caddisfly larvae Abacaria fijiana (Haynes 1999b) within a year. Abundance was highest in months May-June and the least in September-November. Haynes (1999b) inferred that these Mayfly nymphs and Caddisfly larvae hatched into adults in the late months of the year prior to the hot cyclone season. During the hot cyclone season with high water temperature and heavy rain condition, these organisms were possibly in egg or adult stage. A study over two years on the endemic thiarid gastropod of Fiji, Fijidoma maculata, in the Wainibuka river showed that the young produced through parthenogenesis took 550 days to grow from 1.0-1.6mm high to 14mm high. The density of F. maculata population varied from 800-2438m2 for one year but decreased to 250m2 after two tropical cyclones (Haynes 1988b). These studies (Haynes 1988b, 1999b) suggest that presence and absence and fluctuations in abundance of specific invertebrate species in Fiji streams can be dictated by natural ecological processes.

In the current study macroinvertebrates were sampled only once in the wet and once in the dry season with a gap of four months between the sampling periods. Presumably, the four month gap involved changes in life cycle stages (eggs-instars-nymphs-adults) of distinct invertebrates causing fluctuations in species presence/absence and

79 corresponding abundance which may have changed macroinvertebrate patchiness across substrates.

Displacement downstream or ‘drift’ is another possible factor that possibly caused intra- sample variation in this study. The drift effect is referred to as downstream transport of stream organisms via inevitable directional flow in streams. This phenomenon is yet to be studied at the specific family/genus/species level in Fiji. Freshwater macroinvertebrates have specific adaptations to avoid turbulence in streams though they may be displaced from their habitats by water currents to form ‘drift’. Most invertebrates are motile and this places greater risk of displacement on them. However they have evolved to have special features that allow them to adapt to moving water. For example, the suction pad feet of Septaria snails allow them to cling to rocks. These snails are adapted to chute habitats with high water velocity (Haynes 2001b). The flattened body of motile damselflies (Nesobasis spp.) allows the animal to crawl over and under stones in riffles. Aquatic moths (Nymphula spp.) lower the risk of displacement by spinning flat sheets of silk and attaching them to rocks under which they may form cocoons. The blackfly larvae (Simulium spp.) spins a pad of silk which adheres to the stones and it attaches to it with hooks at its hind end (Haynes & Rashni 2015).

Despite these adaptations, macroinvertebrate displacement and downstream drift commonly occur in streams and it is a complex process. This is because distinct species and even different size classes of the same species drift to a different degree coupled with the process partially being determined by external factors (Moss 1980; Svendsen et al. 2004; Barbero et al. 2013). For example, Leptoceridae species displayed a daytime drift pattern (Thornton 2008) while for species such as the Amphipod Gammarus sp. and the Mayfly Baetis sp. drift increases just after sunset, occasionally with peaks late in the night (Peterka 1969). There has been no observed distinct diurnal drift pattern for Chironomids (Collier & Wakelin 1992; Thornton 2008). Drift is also related to food availability. Food availability or epithelium cover on submerged substrates and leaf– litter caught between substrates determines community composition of

80 macroinvertebrates per sample. Studies have shown that when food supplies are scarce drift is relatively high. In an artificial stream experiments carried out by Hildebrand (1974), showed that drift of animals feeding on algae attached to rocks was high during scarcity of algae possibly allowing dispersal of these organisms to sites richer in food.

The above mentioned studies (Peterka 1969; Hildebrand 1974; Collier & Wakelin 1992; Thornton 2008) on the drift phenomenon suggest that invertebrate composition in habitats sampled can differ depending on time of the day sampled and food availability at any particular time of the day and/or week and therefore invertebrate drift is being a possible additional factor causing inter-sample variation in the current study.

Chapter 4. RESULTS & DISCUSSION OF ABIOTIC (ENVIRONMENTAL) VARIABLES

Physio-chemical (Dissolved oxygen, Temperature, pH, Conductivity, Total dissolved solids, Turbidity and Salinity), aquatic habitat data (Flow and Riffle depth) including canopy cover, microbe (faecal coliform) and nutrient (ammonia, nitrate, nitrite and total phosphorus) data were graphed using the SPSS chart Builder (IBM SPSS Statistics 19) (Appendix 7.2, Table 7.1 (a-o)). The streambed substrate composition data were graphed using Microsoft Excel and presented in Appendix 7.2, Figure 7.1 (a-f). The physiochemical results for Taveu and Vucinivola stream are not shown in the wet season graphs as kick-net samples and the corresponding water quality results were not included in the analysis. A single measure of each physico-chemical parameter was recorded per site preventing statistical analysis being done.

Table 4.1: Water quality results for the six streams in comparison to the Fiji freshwater and ANZECC Guidelines

Parameter (range) Streams Guidelines

Naikawaqa Taveu Vucinivola Naqewa Korolevu Savumiri Fiji*** ANZECC

Temperature (⁰C) 21.4-24.9 23.6-24.4 22.95-25.38 20.06-24.5 21.28-23.8 21.62-25.2 No data No data pH 5.7-6.8 6.8-7.0 6.9-8.0 7.0-7.9 6.6-7.2 6.8-7.7 6.5-8.5 6.5-8.0

Dissolved oxygen (mg/L) 6.01-8.59 8.57-9.94 7.7-9.07 8.36-9.84 8.54-8.86 7.88-10.73 No data 5 or greater

Conductivity (mS/cm) 0.09-0.17 0.14-0.14 0.13-0.15 0.1-0.11 0.02-0.07 0.07-0.12 No data 0.2-2.5

Nitrate (mg/L) 0.30-1.55 0.10-0.39 0.11-0.42 0.16-0.42 0.12-0.56 0.08-0.37 No data No data Ammonia (mg/L) 0-0.13 0-0.03 0 0.01-0.14 0-0.03 0-0.11 No data No data Total Phosphorus (mg/L) 0-0.29 0.05-0.11 0.05-0.10 0.07-0.14 0-0.13 0.05-0.17 No data 0.35-0.37 Faecal coliform 0-1600 0-198 140-600 0-28 5-40 0-320 <150 * <150 * (c/100mL) <1000* * <1000**

Note: only measurement range of tested environmental variables are presented for both wet and dry seasons (except Taveu and Vucinivola which have only dry season data) No data= no data is available for the variable in the guidelines used. *= for primary contact. **= for secondary contact. ***specific criteria for class 2 (freshwater) (Ministry of Urban Development Housing and Environment 1996).

Chemical parameters indicated that none of streams were polluted. All readings recorded were within the safe range for biotic community survival (Table 4.1) in accordance with Australian and New Zealand Environment and Conservation Council and Agriculture and Resource Management Council of Australia and New Zealand (2000) guidelines. The exception was the feacal coliform counts which were high (>500 c/100ml) in mid and upper reaches of Naikawaqa stream during the wet season and the mid reach of Vucinivola. This was possibly due to cattle grazing in these areas. The water samples were collected after previous night rain which possibly led to leaching from cattle dung into the streams resulting in high FC counts, a temporary impact on streams. According to the Australian and New Zealand Environment and Conservation Council and Agriculture and Resource Management Council of Australia and New Zealand (2000) FC counts of <150 c/100ml is acceptable for primary contact (recreational activities that involve whole body contact with the body of water e.g. swimming or rafting) and <1000 c/100ml is acceptable for secondary contact (recreational activities that involve only the limbs in contact with the body of water e.g. rowing or fishing). The streams sampled were coastal streams impacted by human activities which pose a concern for human use but perhaps not for biotic community survival.

4.1 Forest Cover analysis

The percentage forest cover per catchment drained by each of the six streams is summarized in Table 4.2 and presented in Figure 4.2 and Figure 4.3. There was no significant difference (t (d.f. =4) =0.588, p=0.59) in the mean forestation for the two sets of catchments (Catchments in Ra (mean percent forested: 74%) and Kadavu (mean percent forested: 61%)).

Table 4.2: Streams and associated percent forested catchments at study sites

Island/province/district Streams Percent forested catchment (%) Kadavu island streams Naqewa 91 (Nakasaleka district) Korolevu 64 Savumiri 29 Viti Levu island Naikawaqa 94 streams Taveu 79 (Nakorotubu district, Vucinivola 50 Ra province)

Kadavu Ra

100 90 80 70 60 50 40 30 forested catchment forested catchment (%) 20 10 0 Naqewa Korolevu Savumiri Naikawaqa Taveu Vucinivola Stream

Figure 4.1: Percentage forest cover of catchments studied in Kadavu and Ra province

Lomanikoro

Kavala

Namajiu

Legend Forested 00.25 0.5 1 1.5 2 2.5 3 3.5 4 Kilometers Non forested

Figure 4.2: Nakasaleka study site catchment boundaries with proportion of forested and non-forested land cover.

Nacobau

Saioko

Nabukadra

Legend Non-forested 00.511.520.25 Kilometers Forested

Figure 4.3: Nakorotubu study site catchment boundaries with proportion of forested and non-forested land cover.

4.2 Linking abiotic variables and forest cover to biological communities The results of PCO analyses associating environmental variables with macroinvertebrate community assemblage at (i) combined province stream level are shown in Figure 4.4 and (ii) distinct province stream levels are shown in Figures 4.5 and 4.6.

4.2.1 Abiotic correlation with macroinvertebrate assemblage: both Kadavu and Ra streams Examination of the abiotic correlation vector based on Pearson correlation coefficient of standardized environmental variables overlaid on the macroinvertebrate assemblage PCO plot (Figure 4.4) shows that Naikawaqa and Taveu separate from the rest of the streams along PCO axis 1. Two environmental variables, percent forested (Pearson correlation coefficient: PCO1=-0.60, PCO2=-0.33) and conductivity (Pearson correlation coefficient: PCO1=-0.56, PCO2=-0.15) in turn correlate with the PCO1 axis. The proportion of the catchment that is forested in Naikawaqa and Taveu is greater than that in all other catchments aside from Naqewa which is in Kadavu. Along PCO axis 2, no environmental factors (absence of vectors) were shown on the plot. This is most probably due to the environmental factors having weaker correlations (<0.5) to PCO2 and therefore are not shown on the plot.

Note: Kadavu streams include Korolevu, Naqewa and Savumiri and Ra streams include Naikawaqa Taveu and Vucinivola.

Figure 4.4: PCO ordination plot showing the environmental variable associated with clusters among streams of Ra and Kadavu province. The vectors in the plot represent the abiotic variables most responsible for differences between all sampling sites.

4.2.2 Abiotic variable correlation with macroinvertebrate assemblage: Kadavu streams Examination of the abiotic correlation vectors based on the Pearson correlation coefficients of standardized environmental variables overlaid on the macroinvertebrate assemblage PCO plot of Kadavu streams (Figure 4.5), shows that the factor ‘percent forested’ aligns with the separation of Naqewa from Korolevu and Savumiri. Percent forested ordinates with PCO1 (Pearson correlation coefficient: PCO1=0.51, PCO2=0.50) which explains 13.4 % of total variation in the biotic dataset. Naqewa is more forested than other Kadavu streams. The abiotic factor, Faecal coliform (Pearson correlation coefficient: PCO1=-0.02, PCO2=-0.57), aligns with the separation of Savumiri from Korolevu and Naqewa and explains 9.8% of total variation in biotic dataset along PCO2.

Figure 4.5: PCO ordination plot showing the environmental variable associated with clusters among streams in Kadavu. The vectors in the plot represent the abiotic factors most responsible for distinguishing differences between sampling sites in the Kadavu streams.

4.2.3 Abiotic variable correlation with macroinvertebrate assemblage: Ra streams Vectors representing the proportional contribution of standardized abiotic factors to differences between sampling sites were overlaid on the macroinvertebrate assemblage PCO plot of Ra streams (Figure 4.6). The environmental factors of Percent forested (Pearson correlation coefficient: PCO1=0.80, PCO2=-0.15), Canopy cover (Pearson correlation coefficient: PCO1=0.67, PCO2=-0.23), Temperature (Pearson correlation coefficient: PCO1=0.53, PCO2=-0.35) and conductivity (Pearson correlation coefficient: PCO1=0.57, PCO2=-0.39) ordinate with PCO1 which explains 26.4% of total variation in biotic dataset. The streams Taveu and Vucinivola on the lower end of PCO1 and Naikawaqa on the upper end of PCO1 suggest that percent forested, canopy cover, temperature and conductivity are factors explaining the variation in the macroinvertebrate community along PCO1. Naikawaqa is more forested than other Ra streams. Environmental factors of Ammonia level (Pearson correlation coefficient: PCO1=0.49, PCO2=0.66) and Nitrate level (Pearson correlation coefficient: PCO1=0.63, PCO2=0.58) ordinate with PCO2 which explains only 8.7% of total variation in the biotic dataset. Nitrate and Ammonia levels are generally slightly above detection level.

Figure 4.6: PCO ordination plot showing the environmental variables associated with clusters among streams in Ra. The vectors in the plot represent the abiotic factors most responsible for distinguishing differences between sampling sites in the Ra streams.

4.2.4 Distance based linear models (DistLM) results

Distance based linear modeling was performed as part of PERMANOVA analyses using Modified Gower similarity matrix of the biotic community and the corresponding standardized abiotic variables in order to describe the predominant predictor variable(s) that influence the biotic community. The results obtained are presented in Table 4.3. The best solution single variable model highlights that the factor “percent forested” has the highest R2 correlation value (7.6197E-2) and is therefore the single factor that best explains the greatest proportion of variance in the macroinvertebrate assemblage. The best solution with two variables highlights that percent forested and conductivity together have an R2 value of 0.124 with the two (Distance-based redundancy analysis) dbRDA axes.

Table 4.3: PERMANOVA (DistLM) results on the macroinvertebrate data examining the correlation between biotic community and the corresponding environmental variables.

MARGINAL TESTS

Variable SS (trace) Pseudo-F P Rˆ2 % forested 7.9889 12.702 0.0001 7.62E-02 Conductivity 7.1752 11.313 0.0001 6.84E-02 Canopy cover 5.2191 8.0677 0.0001 4.98E-02 Nitrate 5.0273 7.7562 0.0001 4.80E-02 Temperature 4.6876 7.2076 0.0001 4.47E-02 Ammonia 3.2636 4.9477 0.0001 3.11E-02 pH 3.2175 4.8757 0.0001 3.07E-02 Total Phosphorus 3.0644 4.6367 0.0001 2.92E-02 Dissolved Oxygen 2.8577 4.3152 0.0001 2.73E-02 Substrate composition 2.5959 3.9098 0.0001 2.48E-02 Faecal coliform 2.4037 3.6135 0.0002 2.29E-02

4.2.5 Best predictory abiotic variable for variation in macroinvertebrate community structure

In the current study as evidenced by DistLM results, of all abiotic variables tested, “percent forested” (highlighted in Table 4.3) was the overall best predictor of macroinvertebrate community composition. This finding is in agreement with studies that have shown strong association between forest cover and macroinvertebrate assemblages (Roy et al. 2003; Zhang et al. 2009) and also where forest cover at catchment-scale is strongly correlated with ecological response (Bopp 2002; Hughes 2007; Stephenson & Morin 2009).

In a study investigating the affect of the spatial extent of forest cover on stream biotic community in National Capital Region of Canada, Stephenson and Morin (2009) found that catchment forest cover explained the most variation in macroinvertebrate assemblage composition (adjusted R2 = 0.51). Analyses of the partial effects of forest cover on stream invertebrate biomass and community structure metrics at three different scales (entire catchment, riparian and reach) identified catchment and reach scales as being most influential. They inferred that in order to protect stream function and structure from catchment-wide impacts, maintenance of the entire catchment is as equally important as protection of reach and riparian buffers.

In a study of 30 streams in the Etowah River basin, Georgia, U.S.A examining relationships between land cover in catchments and stream biotic community, Roy et al. (2003) found strong association between catchment cover and macroinvertebrate assemblages whereby land cover explained 29–38% of the variation in several macroinvertebrate indices. They inferred that catchment urbanization which reduces forest cover cause increased sediment transport and solutes and reduced streambed substrate size which results in less diverse and more resilient macroinvertebrate communities.

Electrical conductivity (highlighted in Table 4.3) was another explanatory factor for macroinvertebrate community composition in this study.

4.3 Methodological Limitations

Sampling methods affect the macroinvertebrate assemblage sampled. It is necessary to define the study purpose and therefore the most appropriate method and consistently use it. The macroinvertebrate assemblages collected by both kick-net and Surber sampler were distinct and as a result not comparable. Therefore, biotic data collected by the most consistently used sampling technique (Surber sampler) was used for analysis in this study.

Macroinvertebrates between individual samples within reaches within streams and within seasons were found to be very variable. This was due to differences in habitat at a micro-scale. Surber samples were collected randomly across riffle habitats containing different types of rocks per site instead of similar rock types. Fine scale habitat variation is considered very important and a recommendation for future work is that additional effort be made to control for this factor. In the current study, a single measurement of environmental variables was taken per site which prevented statistical analysis on environmental variable associations. Therefore it is recommended that environmental variable measurements should be taken per biotic community sample replicate to ensure a more comprehensive analysis.

Chapter 5. CONCLUSIONS This is a pioneering study on catchment-scale level effect and corresponding ecological response in Fijian streams. This study has provided a broad perspective on the freshwater macroinvertebrate communities across six catchments in two provinces of Fiji; Nakasaleka district of Kadavu province and Nakorotubu district of Ra province, and some of the important environmental factors explaining the variation in the macroinvertebrate community structure. These factors, summarised within a conceptual model (Figure 1.0) in the Introduction chapter, include the percentage catchment forest cover, stream habitat features and associated physiochemical parameters of the stream water that affect (directly or indirectly) the macroinvertebrate community composition, abundance and diversity. In both provinces, the catchment forest cover was affected by two major landuse types, i.e. agriculture and human settlements. Results indicate the following:

Catchment forest cover is an important environmental factor that influences the macroinvertebrate community composition in Fijian streams.

Despite high inter-sample variability, multivariate analysis linking environmental and macroinvertebrate community data suggests that in the studied rural streams percent forest cover as an environmental factor explained the greatest amount of the variation in the macroinvertebrate community structure. As evidenced by DistLM results, the best solution single variable model highlighted that percentage forested was the predictor variable with the highest R2 correlation value (7.6197E-2), explaining the most variation in biotic community structure of streams studied.

This finding holds great importance for the agricultural sector and terrestrial Protected Areas (PAs) in Fiji. Agricultural expansion in catchments leads to forest cover reduction in Fiji. For example, at lower altitudes, land is more easily converted for agriculture or forestry. Forest cover reduction may lead to (a) increase in sediment loads, (b) alteration in nutrient cycling due to excess nutrient input (various fertilizer types used for agricultural practices), (c) reduction in canopy cover and (d) alteration in

96 channel hydrology of the riverine network and which may alter streambed conditions which will eventually influence the macroinvertebrate community structure. Understanding of the concept of ‘forest cover and its relationship with the associated stream invertebrate fauna’ is very much needed in the agricultural sector of the Fijian economy as agriculture plays an important role in the process of rural development and it is the rural streams that support some of Fijis’ endemic/area endemic and threatened freshwater macroinvertebrates. For example, as part of the current study, the endemic and patchily distributed shell-less gastropod, Achochlidium fijiense, was recorded for the third time since its discovery in the 1980s. Prior to this study, A. fijiense had only been recorded from the Nasekawa River (Vanua Levu island, in the year 1983) and Lami River (Viti Levu island, in the year 1988) despite surveys in other rivers in Fiji. Although Vucinivola stream drains the least forested catchment out of the three studied in the Nakorotubu district of Ra province, the upper reaches of the Vucinivola stream is well vegetated, unfarmed and protected as it is the source of drinking water for the Nabukadra village. It is possible that the protected forest supports the stream habitat conditions suitable for thriving of the A. fijiense population. Nabukadra village is one of the project sites for the Institute of Applied Sciences, whereby the Natural Resource Management Unit of IAS was funded to undertake the COWRIE (Coastal and Watershed Restoration Integrity for Island Environments) project using CBAM (Community Based Adaptive Management) approaches to Watershed Management. The COWRIE project encouraged native tree nursery set-up, reforestation and sustainable farming practices. Encouraged by this catchment forest restoration initiative, in the year 2011, the Nabukadra village farmers substituted ‘bird drops’ (collected from Vatu-i-Ra passage) for the processed fertilizers in their farms thus contributing to stream rehabilitation and natural nutrient cycling in stream network.

The concept of ‘forest cover and its relationship with the associated stream invertebrate fauna’ also holds relevance to the terrestrial Protected Areas (PAs) in Fiji. The Protected forest areas indicate biodiversity hotspots where most of the endemic or area endemic species of Fiji are found. Most of the PAs in Fiji concentrate conservation efforts in inland forest/highlands such as the Sovi basin in Namosi highlands, Savura,

Monasavu (hydro-dam), Nadarivatu highlands-Tomaniivi-Wabu Reserves and Ravilevu-Taveuni Reserves. The headwaters of major rivers and streams lie in the highlands and it is the headwaters that support the unique freshwater invertebrate diversity; for example, the rissooidean freshwater gastropods (spring snails- Fluviopupa spp.) that evolve within the headwaters of catchments and are washed down and distributed via the tributaries. These spring snails are unique in each catchment and thus contribute to endemism among the Fijian freshwater ecosystem biodiversity. The Fluviopupa spp. are crenobionts, only thrive in the purest water and are highly sensitive to any kind of anthropogenic development and therefore the presence of spring snails is an indicator for excellent water quality.

Another example is the endemic freshwater gastropod, Fijidoma maculata, which is endemic to Viti Levu island of Fiji. It has only been recorded from the headwaters of the Rewa (Wainibuka) River, Wailoa River (Laselevu), upper Ba and the Sigatoka River. Despite several riverine biotic surveys in Fiji over the last decade, F. maculata has not been recorded anywhere else. The protection of Fijian forests and their associated ecosystems means protection of the endemic and or area endemic freshwater macroinvertebrates. It is hoped that this information becomes an asset to the Forestry and Agriculture Department of Fiji, in that, that the information hopefully influences the legislation and forest protection and management initiatives such as the National Forest Monitoring System (NFMS) and the Sustainable Forest Management (SFM) program to ensure the integrity of freshwater ecosystems and the associated biotic community and the ecosystem services (e.g. water source, food security and cultural totems) in the face of rapid modification of forests due to agricultural expansion.

Several species were identified that appear to be possible indicators of the health of watersheds when considering the intactness of the forest within those watersheds.

Species associated with highly forested catchment streams include the finger-net caddisfly Chimarra sp. A, the atyid shrimps Atyoida pilipes and Atyopsis spinipes and the neritid gastropods Neritilia rubida, Septaria sanguisuga, and Neritina squamaepicta. These species were highly abundant in streams with percent forest cover of 94% and 91% respectively and low in abundance/absent in deforested catchments. Species associated with least forested catchment streams include the clinging mayfly nymph Pseudocloeon sp. and the net spinning caddisfly larvae Abacaria fijiana. These species were low in abundance in forested catchments.

This finding has shown the association of freshwater macroinvertebrate species with varying percentage forest cover in a watershed. It appears that the streams draining highly forested catchments have more ‘sediment sensitive’ species (the finger-net caddisfly Chimarra sp. A, the atyid shrimps Atyoida pilipes and Atyopsis spinipes and the neritid gastropods Neritilia rubida, Septaria sanguisuga, and Neritina squamaepicta) in comparison to streams draining less forested catchments. In contrast, the streams draining deforested catchments have more ‘sediment resilient’ species (the clinging mayfly nymph Pseudocloeon sp. and the net spinning caddisfly larvae Abacaria fijiana). It is possible that the forested catchment streams provide the suitable habitat conditions, food and breeding ground for the sensitive species to thrive and invertebrates in deforested catchment streams may be deprived of these suitable conditions.

In the rural areas of Fiji, modification of catchment cover is mostly concentrated at lower altitudes with agriculture and human settlement expansion being the most common anthropic impact on the forest cover reduction. This modification of land reduces catchment forest cover including the riparian forest and bank vegetation which may lead to an increase in sedimentation and, when coupled with rainfall, sediment loads alter the streambed conditions that support a diverse macroinvertebrate community structure in streams. The continuous input of sediment loads in streams may negatively affect the invertebrate community structure in several ways. These include (a) smothering of the biofilms (food for algal grazers) on submerged rocks; species die out of starvation, (b) clogging the gills of invertebrates; species die of suffocation, (c)

99 increasing turbidity and blocks sunlight which prevents photosynthesis and thus limits quick biofilm recovery to sustain the remaining invertebrate population (d) and eventually leading to ‘phase shift’ in community as ‘sediment resilient’ species outweigh the population of ‘sediment sensitive species’. Thus, it appears that the ‘sediment sensitive’ and ‘sediment resilient’ species in Fijian streams can be used as bioindicators of watershed health. It is hoped that the information gathered through the current study will be used to facilitate development of appropriate stream health and watershed health indicators and contribute to lentic system and watershed management in Fiji.

This study has revealed important additions to the freshwater macroinvertebrate fauna of Fiji.

A total of fourteen freshwater macroinvertebrate families not previously recorded in Fiji were documented during the study. These include: one family of aquatic moth (Crambidae), three families of aquatic true-flies (Dixidae, Empididae and Stratiomyidae, seven families of aquatic beetles (Elmidae, Gyrinidae, Helminthidae, Hydraenidae, Hydrophilidae, Psephenidae and Spercheidae) and three families of water bugs (Mesoveliidae, Veliidae and Saldidae). Presumably, these new records were due to the method of collection used, i.e. the Surber sampler. Prior to this study, Surber sampling was not trialed in Fiji. Interestingly, these new records were an outcome of sampling within just a single instream micro-habitat, i.e. the mid-stream riffle. The current study was centered on riffle-dwelling invertebrates because they are known to be the most sensitive to changes in environmental conditions as they prefer high dissolved oxygen levels and clean water. However, the freshwater macroinvertebrates occupy a range of micro-habitats in Fijian streams such as riffle, run, stream edge, submerged vegetation, pool and just recently discovered, the interwoven root mass in the mid-stream that mock riffle habitat, and therefore it is likely that with further study in other micro-habitats and in other freshwater system supportive islands of Fiji, this body of knowledge will continue to significantly grow. It is hoped that the documentation of 14 new records of the Fijian freshwater macroinvertebrate families

100 get included in the National Resource Inventory and Fiji's Biodiversity Strategy and Action Plan (NBSAP) and a collaborative effort is undertaken by the University of the South Pacific and the Department of Environment of Fiji to extend further research work in understanding the biogeographical distribution, ecology and biology of these newly recorded fauna as freshwater invertebrate diversity/ecology/biology is understudied in Fiji and is a ‘rare science’ in the Pacific.

To round up, the three major findings of this study highlighted the importance of (a) maintaining catchment forest cover which impacts the freshwater ecosystem macroinvertebrate diversity, (b) certain macroinvertebrates which appear to be bioindicators of watershed health and (c) the need for expanding freshwater infauna surveys across Fiji as a total of 14 new records of Fijian freshwater invertebrate families were documented from surveying only six streams. It is hoped that these vital information raises awareness amongst the government departments of Fiji (Department of Environment, Forestry and Agriculture) whose legislations and projects influence the Fijian freshwater ecosystems in general.

5.1 Recommendations The sampling methodology employed in the current study has been adapted from sampling protocol developed for New Zealand rivers and streams as there are no existing sampling protocols for tropical island streams. The sampling methodology employed might not be the most appropriate one for the Fiji streams. This is due to (1) differences in landmasses, geology and tropical climate that dictate stream hydrology and structure compared to those in temperate continents and (2) tropical island freshwater ecosystems have evolved differently in isolation with increasing speciation and endemism coupled with specialized habitats. It is therefore recommended that an appropriate freshwater macroinvertebrate sampling protocol and methodology applicable to Fiji river and stream ecosystems be developed. This study presented freshwater macroinvertebrate assemblages across varying percentage forest cover in watersheds. What still remains unclear is how much 101

forest cover is cleared before an actual ecological response in macroinvertebrate community is seen. Research on how big the buffer zone of forest cover is required to maintain stream ecosystem integrity will be beneficial for the environmental, economic and social sectors in Fiji as it will assist with planning for sustainable developments in watersheds.

The discovery of an additional 14 families of freshwater macroinvertebrates in the current study reflects a gap in knowledge of freshwater macroinvertebrate fauna for Fiji and therefore calls for surveys of different freshwater bodies across Fiji in order to enhance knowledge on freshwater macroinvertebrates. Data gathered from surveying freshwater bodies across Fiji will be used to facilitate development of freshwater invertebrate conservation and management for Fiji.

The heterogeneous nature of rocks in stream support diverse biotic fauna. It would be interesting to study the macroinvertebrate community supported by different rock types of similar size in Fijian streams.

102

Chapter 6. REFERENCES Abell, R., Thieme, M.L., Revenga, C., Bryer, M., Kottelat, M., Bogutskaya, N., Coad, B., Mandrak, N., Balderas, S.C., Bussing, W., Stiassny, M.J., Skelton, P., Allen, G.R., Unmack, P., Naseka, A., Ng, R., Sindorf, N., Robertson, J., Armijo, E., Higgins, J.V., Heibel, T.J., Wikramanayake, E., Olson, D., López, H.L., Reis, R.E., Lundberg, J.G., Pérez, M.H.S. & Peter, P. (2008). Freshwater ecoregions of the world: a new map of biogeographic units for freshwater biodiversity conservation. BioScience, 58, 403-414.

Alberti, M.D., Booth, K., Hill, B., Cobourn, C., Avolio, S., Coe & Spiandelli, D. (2007). The impact of urban patterns on aquatic ecosystems: an empirical analysis in Puget lowland sub-basins. Landscape and Urban Planning 80, 345- 361.

Allan, J.D. (1995). Modification of running waters by humankind In: Stream ecology: structure and functioning of running waters (ed. Allan, JD). Chapman & Hall London, pp. 320-330.

Allan, J.D. (2004). Landscapes and riverscapes: the influence of land use on stream ecosystems. Annual Review of Ecology, Evolution, and Systematics, 35, 257- 284.

American Public Health Association (2005a). 4500-P. B. & 4500-P E. In: Standard methods for the examination of water and wastewater (eds. Clesceri, LS, Eaton, AD & Greenberg, AE). American Public Health Association (APHA) Washington, D. C.

American Public Health Association (2005b). Method 9221A, B, C and E multiple-tube fermentation technique for members of the coliform group. In: Standard methods for the examination of water and wastewater (eds. Clesceri, LS, Eaton, AD & Greenberg, AE). American Public Health Association (APHA) Washington, D.C.

American Public Health Association (2005c). Method 9222D Faecal coliform membrane filter procedure. In: Standard methods for the examination of water and wastewater (eds. Clesceri, LS, Eaton, AD & Greenberg, AE). American Public Health Association (APHA) Washington, D.C.

- American Public Health Association (2005d). Method No. 4500-NO3 I. In: Standard methods for the examination of water and wastewater (eds. Clesceri, LS, Eaton, AD & Greenberg, AE). American Public Health Association (APHA) Washington, D. C.

103

American Public Health Association (2005e). Method No. 4500 NH3 - H. In: Standard methods for the examination of water and wastewater (eds. Clesceri, LS, Eaton, AD & Greenberg, AE). American Public Health Association (APHA) Washington, D. C.

Arkle, S.B. & Pilliod, S.D. (2010). Prescribed fires as ecological surrogates for wildfires: A stream and riparian perspective. Forest Ecology and Management, 259, 893–903.

Asia Development Bank (2003). Special evaluation study on Asian Development Bank capacity building assistance for managing water supply and sanitation to Republic of the Fiji islands, Kiribati, Papua New Guinea, and Republic of the Marshall Islands. Asian Development Bank, pp. 2-5.

Atherton, J., Olson, D., Farley, L. & Qauqau, I. (2006). Watershed assessment for healthy reefs and fisheries: Fiji watersheds at risk. Wildlife Conservation Society–South Pacific (WCS) Fiji.

Australian and New Zealand Environment and Conservation Council & Agriculture and Resource Management Council of Australia and New Zealand (2000). National water quality management strategy: Australian and New Zealand guidelines for fresh and marine water quality Australian Government Publishing, Canberra, Australia.

Bailey, C.R., Norris, H.R. & Reynoldosn, B.T. (2004). Bioassessment of Freshwater ecosystems: using the reference condition approach. Kluwer Academic Publishers, Boston.

Barbero, M.D., Oberto, A.M. & Gualdoni, C.M. (2013). Spatial and temporal patterns of macroinvertebrates in drift and on substrate of a mountain stream (Cordoba, Central Argentina). Acta Limnologica Brasiliensia, 25, 375-386.

Benstead, J.P., March, J.G., Pringle, C.M., Ewel, K.C. & Short, J.W. (2009). Biodiversity and ecosystem function in species-poor communities: community structure and leaf litter breakdown in a Pacific island stream. The North American Benthological Society, 28, 454-465.

Bernhadt, S.E. & Palmer, A.M. (2011). The environmental costs of mountaintop mining valley ﬁll operations for aquatic ecosystems of the Central Appalachian. Annals of the New York Academy of Sciences, 1223, 39–57.

Birmingham, M., Gautsch, J., Heimdal, H.D., Krier, K., Leopold, R., Luzier, J. & Wilton, T. (2005). Benthic macroinvertebrate indexing. Iowa Department of Natural Resources, Iowa.

104

Böhm, M., Kemp, R., Baillie, J.E.M. & Collen, B. (2012). Chapter 1: The unravelling underworld. In: Spineless: status and trends of the world’s invertebrates (eds. Collen, B, Böhm, M, Kemp, R & Baillie, JEM). Zoological Society of London United Kingdom.

Bopp, J.A. (2002). Combined effects of water chemistry, canopy cover, and stream size on benthic macroinvertebrates along a Central Appalachian stream continuum. Master of Science in Wildlife and Fisheries Resources: Davis College of Agriculture, Forestry, and Consumer Sciences. West Virginia University Morgantown, United States.

Boseto, D. (2006). Diversity, distribution and abundance of Fijian freshwater fishes. Masters of Science in Marine Science. University of the South Pacific Suva, Fiji, p. 256.

Boseto, D., Morrison, C., Pikacha, P. & Pitakia, T. (2007). Biodiversity and conservation of freshwater fishes in selected rivers on Choiseul Island, Solomon Islands. The South Pacific Journal of Natural Sciences, 3, 16-21.

Bright, G.R. (1982). Secondary benthic production in a tropical island stream. Limnology & Oceanography, 27, 472-480.

Bunn, E.S., Abal, E.G., Smith, M.J., Choy, S.C., Fellows, C.S., Harch, B.D., Kennard, M.J. & Sheldon, F. (2010). Integration of science and monitoring of river ecosystem health to guide investments in catchment protection and rehabilitation. Freshwater Biology, 55, 223-240.

Burcher, L.C., McTammany, E.M., Benfield, F.E. & Helfman, S.G. (2008). Fish assemblage responses to forest cover. Environmental Management, 41, 336– 346.

Burcher, L.C., Valett, M.H. & Benfield, F.E. (2007). The land-cover cascade: relationships coupling land and water. Ecology, 88, 228-242.

Burton, B. (1999). BHP admits Ok Tedi mine is environmental disaster. In: Asia Times. Asia Times Online Co., Ltd.

Carpenter, C. & Lawedrau, A. (2002). Effects of forestry activities on surface water quality in the Pacific region: a case study of the Rewa River catchment, Fiji Islands. International Forestry Review, 4, 307-309.

105

Carpenter, S.R., Caraco, N.F., Correll, D.L., Howarth, R.W., Sharpley, A.N. & Smith, V.H. (1998). Nonpoint pollution of surface waters with phosphorus and nitrogen. Ecological Applications, 8, 559-568.

Carter, J.L. & Resh, V.H. (2001). After site selection and before data analysis: sampling, sorting, and laboratory procedures used in stream benthic macroinvertebrate monitoring programs by USA state agencies. Journal of North American Benthological Society, 20, 658–682.

Chadwick, A.M., Dobberfuhl, R.D., Benke, C.A., Huryn, D.A., Suberkropp, K. & Thiele, E.J. (2006). Urbanization affects stream ecosystem function by altering hydrology, chemistry, and biotic richness. Ecological Applications, 16, 1796- 1807.

Chakrapani, G.J. (2005). Factors controlling variations in river sediment loads. Current Science, 4, 569-575.

Chessman, B. (2001). SIGNAL scoring system for river bio-assessment by community groups. Land & Water Australia. Australia.

Choy, S.C. (1984). On the freshwater palaemonid prawns from the Fiji islands (Decapoda, Caridea) Crustaceana, 47 269-277.

Choy, S.C. (1991). The Atyid shrimps of Fiji with description of a new species. Zoologische Mededelingen, 65 343-362.

Clarke, A., Nally, R.M., Bond, N. & Lake, P.S. (2008). Macroinvertebrate diversity in headwater streams: a review. Freshwater Biology, 53, 1707–1721.

Collier, K.J. & Wakelin, M.D. (1992). Drift of aquatic macroinvertebrate larvae in Manganuiateao River, Central North Island, New Zealand. New Zealand Natural Sciences, 19, 15-26.

Compson, Z.G., Adams, K.J., Edwards, J.A., Maestas, J.M., Whitham, T.G. & Marks, J.C. (2013). Leaf litter quality affects aquatic insect emergence: contrasting patterns from two foundation trees. Oecologia, 173, 507–519.

Coulter, B.C., Kolka, K.R. & Thompson, A.J. (2004). Water quality in agricultural, urban and mixed land use watersheds. Journal of the American Water Resources Association, 40, 1593-1601.

106

Covich, A. P. (2006). Dispersal-limited biodiversity of tropical insular streams. Polish Journal of Ecology, 54(4), 523-547.

Cowie, R.H. & Holland, B.S. (2006). Dispersal is fundamental to biogeography and the evolution of biodiversity on oceanic islands. Journal of Biogeography, 33, 193– 198.

Craig, D.A. (2003). Geomorphology, development of running water habitats, and evolution of black flies on Polynesian islands. BioScience, 53, 1079-1093.

Crowe, A. & Hay, J. (2004). Effects of fine sediment on river biota. Cawthron New Zealand.

Cummins, K.W., Wilzbach, M.A., Gates, D.M., Perry, J.B. & Taliaferro, W.B. (1989). Shredders and riparian vegetation. BioScience, 39, 24-30.

Cummins, W.K. & Klug, J.M. (1979). Feeding ecology of stream invertebrates. Annual Review of Ecology and Systematics, 10, 147-172.

Darwall, W., Seddon, M., Clausnitzer, V. & Cumberlidge, N. (2012). Chapter 2: Freshwater invertebrate life. In: Spineless: status and trends of the world’s invertebrates (eds. Collen, B, Böhm, M, Kemp, R & Baillie, JEM). Zoological Society of London United Kingdom.

Death, G.R. & Collier, J.K. (2010). Measuring stream macroinvertebrate responses to gradients of vegetation cover: when is enough enough? . Freshwater Biology, 55, 1447-1464.

DeFries, R. & Eshleman, K.N. (2004). Land-use change and hydrologic processes: a major focus for the future. Hydrological Processes, 18, 2183-2186.

Department of Environment (2010). Volume 1: freshwater resources inventory of the Republic of Fiji. In: Natural resource inventory report of the Republic of Fiji 2010. Department of Environment (DoE) Suva, pp. 12-16.

Dobson, M. & Frid, C. (2009). Ecology of aquatic systems Second edn. Oxford University Press, New York.

Donnelly, T.W. (1990). The Fijian genus Nesobasis Part 1: Species of Viti Levu, Ovalau, and Kadavu (Odonata: Coenagrionidae). New Zealand Journal of Zoology, 17, 87-117.

107

Downes, B.J., Lake, P.S. & Schreiber, E.S.G. (1993). Spatial variation in the distribution of stream invertebrates: implications of patchiness for models of community organization. Freshwater Biology, 30, 119-132.

Dudgeon, D., Arthington, H.A., Gessner, O.M., Kawabata, Z., Knowler, J.D., Le´veˆque, C., Naiman, J.R., Prieur-Richard, A., Soto, D., Stiassny, J.L.M. & Sullivan, A.C. (2006). Freshwater biodiversity: importance, threats, status and conservation challenges. Biological Reviews, 81, 163–182.

Duncan, D. (2012). Freshwater under threat; Pacific Islands: vulnerability assessment of freshwater resources to environmental change. United Nations Environmental Programme (UNEP) and Secretariat of the Pacific Community: Applied Geoscience and Technology Division Bangkok, pp. 3-24.

Eggert, S.L. & Burton, T.M. (1994). A comparison of Acroneuria lycorias (Plecoptera) production and growth in northern Michigan hard-soft-water streams. Freshwater Biology, 32, 21-31.

Ellison, J.C. (2009). Wetlands of the Pacific Island region. Wetlands Ecology and Management, 17, 169–206.

Engelman, R., Pauly, D., Zeller, D., Prinn, R.G., Pinnegar, J.K. & Polunin, N.V.C. (2008). Introduction: climate, people, fisheries and aquatic systems. In: Aquatic ecosystems: trends and global prospects (ed. Polunin, NVC). Cambridge University Press New York.

Evenhuis, N.L. (2007a). Checklist of Diptera of Fiji. In: Bishop Museum Technical Report (ed. Evenhuis, NL). Bishop Museum Hawaii.

Evenhuis, N.L. (2007b). Checklist of Ephemeroptera of Fiji. In: Bishop Museum Technical Report (ed. Evenhuis, NL). Bishop Museum Hawaii.

Evenhuis, N.L. (2007c). Checklist of Trichoptera of Fiji In: Bishop Museum Technical Report (ed. Evenhuis, NL). Bishop Museum Hawaii.

Falkland, T. (2003). Water resources managment. In: Pacific regional consultation on water in small island countries (eds. Carpenter, C, Stubbs, J & Overmars, M). Asian Development Bank & South Pacific Applied Geoscience Commission Sigatoka, Fiji.

Fiji Bureau of Statistics (2012). 2011 facts & figures. Fiji Bureau of Statistics (FBoS). Suva, Fiji.

108

Fiji Bureau of Statistics (2013). Census:housing characteristics. (ed. Statistics, FBo). Fiji Bureau of Statistics (FBoS). Suva, Fiji.

Fiji Department of Energy (c.2010). Fiji renewable energy report. Fiji Department of Energy (FDoE) Fiji.

Fiji Meteorological Service (2012). Daily rainfall data for Penang Mill from February 2011 to February 2012. Fiji Meteorological Service (FMS) Nadi, Fiji.

Food and Agriculture Organization of the United Nations (2009). National Fishery Sector Overview Fiji Food and Agriculture Organization of the United Nations (FAO) Rome, pp. 4-6.

Food and Agriculture Organization of the United Nations (2010). State of world fisheries and aquaculture: world review of fisheries and aquaculture. Food and Agriculture Organization of the United Nations (FAO) Rome, p. 89.

Friberg, N. (2006). Macroinvertebrates and biotic interactions. In: Running waters: historical development and restoration of lowland Danish streams (eds. Jensen, KS, Friberg, N & Murphy, J). National Environmental Research Institute Denmark, pp. 83-91.

Friberg, N. (2010). Pressure-response relationships in stream ecology: introduction and synthesis. Freshwater Biology, 55, 1367–1381.

Friberg, N., Winterbourn, M. J., Shearer, K. A. & Larsen, S. E. (1997). Benthic communities of forest streams in the South Island, New Zealand: effects of forest type and location. Archiv fur Hydrobiologie, 138, 289-306.

Fuchs, A.E., Hinch, G.S. & Mellina, E. (2003). Effects of streamside logging on stream macroinvertebrate communities and habitat in the sub-boreal forests of British Columbia, Canada. Canadian Journal of Forest Research, 33, 1408-1415.

Fuller, S.L.H. (1974). Clams and Mussels (Mollusca: Bivalvia). In: Pollution Ecology of Freshwater Invertebrates (eds. Hart, CW & Fuller, SLH). Academic Press New York.

Gillett, R. (2009). Fisheries in the economies of the Pacific island countries and territories. Asian Development Bank (ADB) Mandaluyong.

Goodnight, J.C. (1973). The use of aquatic macroinvertebrates as indicators of stream pollution. Transactions of the American Microscopical Society, 92, 1-13.

109

Gorman, O.T. & Karr, J. R. (1978). Habitat structure and stream fish communities. Ecology, 59, 507-515.

Gravelle, A.J., Ice, G., Link, E.T. & Cook, L.D. (2009). Nutrient concentration dynamics in an inland Pacific Northwest watershed before and after timber harvest. Forest Ecology and Management, 257, 1663–1675.

Gray, L. (2004). Changes in water quality and macroinvertebrate communities resulting from urban stormflows in the Provo River, Utah, U.S.A. Hydrobiologia, 518 33–46.

Gray, L.J. & Ward, J.V. (1979). Food Habits of Stream Benthos at Sites of Differing Food Availability. American Midland Naturalist, 102, 157-167.

Gutknecht, D., Jolankai, G. & Skinner, K. (2008). Patterns and processes in the catchment. In: Ecohydrology: processes, models and case studies: an approach to the sustainable management of water resources (eds. Harper, D, Zalewski, M & Pacini, N). CAB International (CABI) Oxforshire, UK, pp. 18-27.

Harding, J.S., Benfield, E.F., Bolstad, P.V., Helfman, G.S. & Jones, E.B.D., III (1998). Stream biodiversity: the ghost of landuse past. Proceedings of the National Academy of Sciences of the United States of America, 95, 14843-14847.

Harvey, C.B. & Hill, R.W. (1991). Effects of snails and fish on benthic invertebrate assemblages in a headwater stream. Journal of the North American Benthological Society, 10, 263-270.

Hasse, M., Ponder, W.F. & Bouchet, P. (2006). The Genus Fluviopupa Pilsbry, 1911 from Fiji (Caenogastropoda, Rissooidea). Journal of Molluscan studies, 72, 119- 136.

Hauer, F.R. & Lamberti, G. A. (2007). Methods in stream ecology. Elsevier Inc. Sydney, pp. 293-486.

Haynes, A. (1985). The ecology and local distribution of non-marine aquatic gastropods in Viti Levu, Fiji. The Veliger, 28, 204-210.

Haynes, A. (1987). Species richness, abundance and biomass of benthic invertebrates in a lowland tropical stream on the island of Viti Levu, Fiji. Archiv fuer Hydrobiologie, 110, 451–459.

Haynes, A. (1988a). The gastropods in the streams and rivers of five Fiji Islands: Vanua Levu, Ovalau, Gau, Kadavu, and Taveuni. The Veliger, 30, 377-383.

110

Haynes, A. (1988b). A population study of the Fijian freshwater thiarid gastropod Fijidoma maculata (Mousson). Archiv fuer Hydrobiologie, 113, 27-39.

Haynes, A. (1989). Effects of erosion caused by faulty farming practices on the biota of Fijian streams and rivers. In: Agricultural Development in the Pacific Islands in the 90s: Proceedings of the International Conference and Workshop, Suva, 31 March to 1 April 1989 (eds. Haynes, RJ & Naidu, R). Institute of Agricultural Science Suva, Fiji, pp. 300-304.

Haynes, A. (1990). The numbers of freshwater gastropods on Pacific islands and the theory of island biogeography. Malacologia, 31, 237-248.

Haynes, A. (1994). The effects of development on Fijian island freshwater invertebrates. Memoirs of the Queensland Museum, 36, 87-91.

Haynes, A. (1998). A population study of the Fijian freshwater thiarid gastropod Fijidoma maculata (Mousson). Archiv fuer Hydrobiologie, 113, 27-39.

Haynes, A. (1999a). A comparison of the distribution of the freshwater neritid gastropods on the islands of Moorea (French Polynesia) and Ovalau (Fiji). Malacological Review, 8, 13-18.

Haynes, A. (1999b). The long term effect of forest logging on the macroinvertebrates in a Fijian stream. Hydrobiologia, 405, 79-87.

Haynes, A. (2000). The distribution of freshwater gastropods on four Vanuatu islands: Espiritu Santo, Pentecost, Efate and Tanna (South Pacific). Annales de Limnologie, 36, 101-111.

Haynes, A. (2001a). Freshwater Snails of the Tropical Pacific Islands. Institute of Applied Sciences, Suva, Fiji.

Haynes, A. (2001b). A revision of the genus Septaria Ferussac, 1803 (Gastropoda: Neritimorpha). Annalen des Naturhistorischen Museums in Wien, 103 B, 177- 229.

Haynes, A. (2009). Snails in Fiji’s rivers and streams. Institute of Applied Sciences, Suva, Fiji.

Haynes, A. & Kenchington, W. (1991). Acochlidium fijiensis sp. nov. (Gastropoda: Opisthobranchia: Acochlidiacea) from Fiji. The Veliger, 34, 166-171.

Haynes, A. & Rashni, B. Guide to freshwater insects of Fiji (In prep.).

111

Helfrich, L.A. & Neves, R.J. (2009). Sustaining America’s aquatic biodiversity: freshwater fish biodiversity and conservation. Virginia Cooperative Extension Virginia.

Herringshaw, C.J., Stewart, T.W., Thompson, J.R. & Anderson, P.F. (2011). Land use, stream habitat and benthic invertebrate assemblages in a highly altered Iowa watershed. The American Midland Naturalist, 165, 274-293.

Hershler, R. (1998). A systematic review of the hydrobiid snails (Gastropoda: Rissooidea) of the Great Basin, western United States. Part I. Genus Pyrgulopsis. Veliger, 41, 1-132.

Hershler, R. (1999). A systematic review of the hydrobiid snails (Gastropoda: Rissooidea) of the Great Basin, western United States. Part II. Genera Colligyrus, Eremopyrgus, Fluminicola, Pristinicola, and Tryonia. Veliger, 42, 306-337.

Hettler, J., Irion, G. & Lehmann, B. (1997). Environmental impact of mining waste disposal on a tropical lowland river system: a case study on the Ok Tedi Mine, Papua New Guinea. Mineralium Deposita, 32, 280-291.

Hildebrand, S.G. (1974). The relation of drift to benthos density and food level in an artificial stream. Limnology and Oceanography, 19, 951-957.

Hindara, A., Wright, F.R., Nilsenc, P., Larssenb, T. & Hogberget, R. (2003). Effects on stream water chemistry and forest vitality after whole-catchment application of dolomite to a forest ecosystem in southern Norway. Forest Ecology and Management, 180, 509–525.

Hornig, C.E. & Pollard, J.E. (1978). Macroinvertebrate sampling techniques for streams in semi-arid regions: comparison of the surber method and a unit-effort traveling kick method. Office of Research and Development, Environmental Monitoring and Support Laboratory Nevada.

Hughes, D. (2007). Stream biotic health and land cover in the Soque river watershed. In: 2007 Georgia Water Resources Conference. The University of Georgia Georgia.

Hughes, H.R. (1977). Aquatic weed control in Fiji: studies of the ecology of Hydrilla verticillata in the Rewa River and possible methods of control; Research reports (1969-1976) Department of Agriculture. Fiji.

112

Huryn, A.D. & Wallace, J.B. (2000). Life history and production of stream insects. Annual Review of Entomology, 45, 83–110.

Japan International Cooperation Agency & Ministry of Agriculture Fisheries and Forests (1998). The study on watershed management and flood control for the four major Viti Levu rivers in the Republic of the Fiji islands. Japan International Cooperation Agency (JICA) & Ministry of Agriculture, Fisheries and Forests (MAFF) Fiji.

Jeng, M., Liu, H., Tzeng, C. & Ng, P.K.L. (2003). On the taxonomy and ecology of Labuanium trapezoideum (Decapoda,brachyura, Sesarmidae), a crab living on riverine cliffs in Taiwan. Crustaceana, 76, 227-240.

Jenkins, A.P., Jupiter, S.D., Qauqau, I. & Atherton, J. (2010). The importance of ecosystem-based management for conserving aquatic migratory pathways on tropical high islands: a case study from Fiji. Aquatic Conservation: Marine and Freshwater Ecosystems, 20, 224-238.

Johanson, K.A. & Olah, J. (2012). Revision of the Fijian Chimarra (Trichoptera, Philopotamidae) with description of 24 new species. Zootaxa, 3354, 1–58.

Jurado, B.G., Masterson, M., Harrington, R. & Kelly-Quinn, M. (2008). Evaluation of sampling methods for macroinvertebrate biodiversity estimation in heavily vegetated ponds. Hydrobiologia, 597, 97-107.

Kalkman, V. J & Orr, A. G. (2013). Field guide to the damselflies of New Guinea. Brachytron, 16, 3-120.

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

Kelley, R.W. (1989). New species of micro-caddisflies (Trichoptera: Hydroptilidae) from New Caledonia, Vanuatu and Fiji. Proceedings of the Entomological Society of Washington, 91, 190–202.

Keppel, G. & Tuiwawa, M. (2007). Dry zone forests of Fiji: species composition, life history traits, and conservation. New Zealand Journal of Botany, 45, 545-563.

Kling, G., Hayhoe, K., Johnson, L.B., Magnuson, J.J., Polasky, S., Robinson, S.K., Shuter, B.J., Wander, M.M., Wuebbles, D.J. & Zak, D.R. (2003). Chapter Three: ecological vulnerability to climate change: aquatic ecosystems; river and stream ecosystems. In: Confronting climate change in the Great Lakes region: Impacts

113

on our communities and ecosystems. Union of Concerned Scientists, Cambridge, Massachusetts, and Ecological Society of America Washington, D.C., pp. 25-29.

Kominoski, J.S., Larranaga, S. & Richardson, J.S. (2012). Invertebrate feeding and emergence timing vary among streams along a gradient of riparian forest composition. Freshwater Biology, 57, 759–772.

Kondolf, G.M., Smeltzer, M. & Lisa, K. (2002). Freshwater gravel mining and dredging issues. Washington Department of Fish and Wildlife Berkeley, CA, p. 130.

Kristensen, R.M. & Funch, P. (2000). A new class with complicated jaws like those of Rotifera and Gnathostomulida. Journal of Morphology, 246, 1-49.

Lal, M.M. (2011). Spawning and larval rearing of the Monkey river prawn Macrobrachium lar (Fabricius,1798) (Crustacea: Decapoda: Caridea: Palaemonidae) in the Fiji Islands. Master of Science in Marine Science: School of Marine Studies. University of the South Pacific Fiji.

Larinier, M. (2000). Dams and fish migration. World Commission on Dams Toulouse, France.

Lear, G., Boothroyd, G.K.I., Turner, J.S., Roberts, K. & Lewis, D.G. (2009). A comparison of bacteria and benthic invertebrates as indicators of ecological health in streams. Freshwater Biology, 54, 1532-1543.

Leroy, C.J. & Marks, J.C. (2006). Litter quality, stream characteristics and litter diversity influence decomposition rates and macroinvertebrates. Freshwater Biology, 51, 605–617.

Lester, P.J., Mitchell, S.F. & Scott, D. (1994). Effects of riparian willow trees (Salix fragilis) on macroinvertebrate densities in two small Central Otago, New Zealand, streams New Zealand Journal of Marine and Freshwwater Research, 28, 267–276.

Letovsky, E., Myers, I.E., Canepa, A. & McCabe, D.J. (2012). Differences between kick sampling techniques and short-term Hester-Dendy sampling for stream macroinvertebrates. BIOS, 83, 47–55.

Lovell, E.R. & Sykes, H.R. (2007). Fiji. In: Status of Coral Reefs in the Southwest Pacific: 2004 (ed. Sulu, R). IPS Publications Suva, Fiji, pp. 32-35.

114

Lucey, J. & Dorris, Y. (2001). Biodiversity in Ireland: a review of habitats and species. In: Ireland’s Environment: a millennium report (eds. Stapleton, L, Lehane, M & Toner, P). Environment Protection Agency Ireland.

MacDonald, S.B., Mullins, W.G. & Lewis, S. (1991). Macroinvertebrates as indicators of stream health. The American Biology Teacher 53, 462-466.

Mackay, R.J. & Kalff, J. (1973). Ecology of two related species of caddis fly larvae in the organic substrates of a woodland stream. Ecology, 54, 499-511.

Magierowski, R.H., Davies, P.E., Read, S.M. & Horrigan, N. (2012). Impacts of land use on the structure of river macroinvertebrate communities across Tasmania, Australia: spatial scales and thresholds. Marine and Freshwater Research, 63, 762-776.

Maleli, N. (2013). EIA of Tuvatu gold mine. Environment and Resource Management Consultant Fiji.

Marinov, M. & Pikacha, P. (2013). On a dragonfly collection from the Solomon Islands with overview of fauna from this Pacific archipelago (Insecta: Odonata). Faunistic Studies in South-East Asian and Pacific Island Odonata. Journal of the International Dragonfly Fund, 4, 1-48.

Marinov, M. & Waqa-Sakiti, H. (2013). An illustrated guide to dragonflies of Viti Levu, Fiji. USP Press, Suva, 144p.

Marshall, W. (2002). Australian mining giant leaves environmental disaster in Papua New Guinea. Available at: http://www.wsws.org/en/articles/2002/04/png- a09.html Last accessed 27th November 2012.

Mary, N. J. (2002). Spatio-temporal variations in macroinvertebrate assemblages of New Caledonian streams. Bulletin Français de la Pêche et de la Pisciculture, 364, 197-215.

McClain, E.M., Bilby, E.R. & Triska, J.F. (2001). Nutrient cycles and responses to disturbance. In: River ecology and management: lessons from the Pacific coastal ecoregion (eds. Naiman, RJ & Bilby, ER). Springer New, York, pp. 347- 359.

McCulloh, B.A. (2009). The spatiotemporal pattern of macroinvertebrate community composition in two subtropical streams in north-eastern NSW, Australia. Master of Science: School of Environmental Science and Management. Southern Cross University Australia.

115

McDowall, R.M. (1997). The evolution of diadromy in fishes (revisited) and its place in phylogenetic analysis. Reviews in Fish Biology and Fisheries 7, 443-462.

McDowall, R.M. (2008). Diadromy, history and ecology: a question of scale. Hydrobiologia, 602, 5–14.

Mckee, B.A. (2003). RiOMar: the Transport, transformation and fate of carbon in river- dominated ocean margins.Report of the RiOMar workshop, 1-3 November 2001. Tulane University. New Orleans, LA., pp. 5-21.

Merritt, R.W. & Cummins, K.W. (2006). Trophic relationships of macroinvertebrates. In: Methods in stream ecology (eds. Hauer, FR & Lamberti, GA). Elsevier Inc. Sydney.

Ministry of Urban Development Housing and Environment (1996). Fiji's draft sustainable development Bill: integrated and consolidated environmental and resource management legislation Ministry of Urban Development, Housing and Environment Fiji.

Mosely, M.E. (1941). Fijian Trichoptera in the British Museum. Annals and magazine of natural history, 11, 361–374.

Moss, B. (1980). Upland streams and rivers. In: Ecology of fresh waters: man and medium (ed. Moss, B). Blackwell Scientific Publications Oxford, pp. 62-86.

Mueller-Dombois, D. (1998). Chapter III Eastern Melanesia. In: Vegetation of the tropical Pacific islands (eds. Mueller-Dombois, D & Fosberg, FR). Springer- Verlag New York, Inc. New York.

Murty, D.S. & Cavuilati, S.T. (1977). A report on the status of the grass carp project, Fiji. New Caledonia.

Nandlal, S. Key to the species of the genus Macrobrachium from Fiji.

Nandlal, S. (2010). A new species for culture in the Pacific: evaluation of the potential of the indigenous Macrobrachium lar (Fabricius,1798). Doctor of Philosophy in Marine Science: School of Marine Studies. University of the South Pacific. Fiji.

National Geographic Society (2012). Hydropower: going with the flow. Available at: http://environment.nationalgeographic.com/environment/global- warming/hydropower-profile/ Last accessed 20th November 2012.

116

Nejadhashemi, A.P., Wardynski, B.J. & Munoz, J.D. (2011). Evaluating the impacts of land use changes on hydrologic responses in the agricultural regions of Michigan and Wisconsin. Hydrology and Earth System Sciences Discussions, 8, 3421–3468.

Nelson, S. G (1999). Chapter 9: Aquatic ecosystems. In: The Pacific Islands: Environment & society (eds. Rapaport, M). The Bess Press, Inc, p 109.

Newbold, J.D., Erman, D.C. & Roby, K.B. (1980 ). Effects of logging on macroinvertebrates in streams with and without buffer strips. Canadian Journal of Fisheries and Aquatic Sciences, 37, 1076-1085.

Nunn, P.D. (1990). Recent environmental changes on Pacific Islands. The Geographical Journal, 156, 125–140.

Office of Technology Assessment (1993). Harmful Non-Indigenous Species in the United States. Office of Technology Assessment (OTA) Washington, DC, pp. 6- 88.

Olson, D., Farley, L., Patrick, A., Watling, D., Tuiwawa, M., Masibalavu, V., Lenoa, L., Bogiva, A., Qauqau, I., Atherton, J., Caginitoba, A., Tokota’a, M., Prasad, S., Naisilisili, W., Raikabula, A., Mailautoka, K., Morley, C. & Allnutt, T. (2009). Priority forests for conservation in Fiji: landscapes, hotspots and ecological processes. Oryx, 44, 57–70.

Ourso, R.T. (2001). Effect of urbanization on benthic macroinvertebrate communities in streams. In: Anchorage, Alaska: U.S. Geological Survey Water-Resources Investigations Report Alaska, USA, p. 38

Pacini, N., Harper, M.D., Ittekkot, V., Humborg, C. & Rahm, L. (2008). Nutrient processes and consequences. In: Ecohydrology: processes, models and case studies: an approach to the sustainable management of water resources (eds. Harper, D, Zalewski, M & Pacini, N). CAB International (CABI) Oxfordshire, UK.

Paulay, G. (1994). Biodiversity on Oceanic Islands: its origin and extinction. American Zoologist 34, 134-144.

Peterka, J.J. (1969). Distribution, standing crops and drift of benthic invertebrates in a small Wisconsin stream. Wisconsin Academy of Sciences, Arts and Letters, 57, 155-161.

117

Peterson, B.J., Deegan, L., Helfrich, J., Hobbie, J.E. & Huller, M. (1993). Biological responses of a tundra river to fertilization. Ecology, 74, 653-672.

Petrie, J.E. (2008). The accuracy of river bed sediment samples. Masters of Science in Civil Engineering: Faculty of the Virginia Polytechnic Institute. State University Blacksburg, Virginia.

Piccolo, J.J. & Wipfli, M.S. (2002). Does red alder (Alnus rubra) in upland riparian forests elevate macroinvertebrates and detritus export from headwater streams to downstream habitats in Southeastern Alaska? . Canadian Journal of Fisheries and Aquatic Sciences, 59, 503–513.

Pippard, H. (2012). The current status and distribution of freshwater fishes in the Pacific Islands of Oceania. IUCN Oceania Regional office. Fiji., p. 14.

Pires, A.M., Cowx, I.G. & Coelho, M.M. (2000). Benthic macroinvertebrate communities of intermittent streams in the middle reaches of the Guadiana Basin, Portugal. Hydrobiologia, 435, 167-175.

Poff, L.N., Bledsoe, P.B. & Cuhaciyan, O.C. (2006). Hydrologic variation with land use across the contiguous United States: Geomorphic and ecological consequences for stream ecosystems. Geomorphology, 79, 264–285.

Polhemus, D.A., Englund, R.A., Allen, G.R., Boseto, D. & Polhemus, J.T. (2008). Freshwater biotas of the Solomon Islands: analysis of richness, endemism and threats. Bishop Museum Honolulu, Hawai’i.

Pollard, J.E. & Kinney, W.L. (1979). Assessment of macroinvertebrate techniques in an energy development area: a test of the efficiency of three macrobenthic sampling methods in the White River. United States Environmental Protection Agency, Office of Research and Development Nevada, USA.

Poole, G.C. & Berman, C.H. (2001). Pathways of human influence on water temperature dynamics in stream channels. Environmental Management, 27, 787- 802.

Postel, S.L., Daily, G.C. & Ehrlich, P.R. (1996). Human appropriation of renewable fresh water. Science, 271, 78-788.

Printemps, J. (2008). Mapping potential erosion risks in North Viti Levu (Fiji) using the USLE model and a GIS. Coral Reef InitiativeS for the Pacific (CRISP) Noumea.

118

Purcell, H.A., Bressler, W.D., Paul, J.M., Barbour, T.M., Rankin, T.E., J, C.L. & Resh, H.V. (2009). Assessment tools for urban catchments: Developing biological indicators based on benthic macroinvertebrates. Journal of the American Water Resources Association, 45, 306-319.

Quinn, J.M., Cooper, A.B., Davies-Colley, R.J., Rutherford, J.C. & Williamson, R.B. (1997). Land use effects on habitat, water quality, periphyton, and benthic invertebrates in Waikato, New Zealand, hill-country streams. New Zealand Journal of Marine and Freshwater Research, 31, 579-597.

Quinn, J.M., Ian, K.G., Boothroyd, I.K.G. & Smith, B.J. (2004). Riparian buffers mitigate effects of pine plantation logging on New Zealand streams 2. invertebrate communities. Forest Ecology and Management, 19, 129–146.

Ram, A.N., Gawander, J.S., Jokhan, A.D. & Garan, K. (2007). Soil loss and declining sugarcane yields on sloping land in Fiji. Proceedings of the Internantional Society of Sugar Cane Technologists, 26, 190-198.

Rashni, B. (2013). Chapter 7: Freshwater Macroinvertebrates. In: A Rapid Biodiversity Assessment & Archaeological Survey of the Fiji REDD+ Pilot Site: Emalu Forest, Viti Levu (eds. Tuiwawa, MV, Sarah, P & Tuiwawa, SH). Institute of Applied Sciences Fiji, pp. 70-81.

Resh, V.H. (2005). Stream crossings and the conservation of diadromous invertebrates in South Pacific island streams. Aquatic Conservation: Marine and Freshwater Ecosystems, 15, 313–317.

Resh, V.H., Barnes, J.R. & Craig, D.A. (1990). Distribution and ecology of benthic macroinvertebrates in the Opunohu river catchment, Moorea, French Polynesia. Annales de Limnologie, 26, 195–214.

Resh, V.H. & de Szalay, F.A. (1995). Chapter 23: Streams and rivers of Oceania. In: River and stream ecosystems of the world (eds. Cushing, CE, Cummins, KW & Minshall, GW). Elsevier Amsterdam, The Netherlands, pp. 717-737.

Ricciardi, A., Neves, R.J. & Rasmussen, J.B. (1998). Impending extinctions of North American freshwater mussels (Unionoida) following the zebra mussel (Dreissena polymorpha) invasion. Journal of Animal Ecology, 67, 613–619.

Rodda, P. (1967). Outline of the geology of Viti Levu. New Zealand Journal of Geology and Geophysics, 10, 1260-1273.

119

Rosario, D.B.R., Betts, A.E. & Resh, H.V. (2002). Cow manure in headwater streams: tracing aquatic insect responses to organic enrichment. Journal of the North American Benthological Society, 21, 278-289.

Ross, H.H. (1953). Relationship of the Fijian species of Apsilochorema (Trichoptera, Rhyacophilidae). Occasional papers of the Bernice Pauahi Bishop Museum of Polynesian Ethnology and Natural History, 20, 175–182.

Roth, E.N., Allan, D.J. & Erickson, L.D. (1996). Landscape influences on stream biotic integrity assessed at multiple spatial scales. Landscape Ecology, 11, 141-156.

Roy, A.H., Rosemond, A.D., Paul, M.L., Leigh, D.S. & Wallace, J.B. (2003). Stream macroinvertebrate response to catchment urbanization (Georgia, U.S.A). Freshwater Biology, 48, 329-346.

Ryan, P. A. (1991). The success of the Gobiidae in tropical Pacific insular streams. New Zealand Journal of Marine and Freshwater Research, 18, 25-30.

Ryan, P.A. (1980). Namalycastis Vuwaensis n.sp. (Polychaeta: Nereidae) from the Nadrau Plateau, Fiji. New Zealand Journal of Zoology, 7, 509-512.

Ryan, P.A. (2000). Introduction. New Zealand, Exisle Publishing Limited.

Sabater, F., Butturini, A., Marti´, E., Mun˜ Oz, I., Romani, A., Wray, J. & Sabater, S. (2000). Effects of riparian vegetation removal on nutrient retention in a Mediterranean stream. Journal of the North American Benthological Society, 19, 609-620.

Sandin, L. & Johnson, K.R. (2004). Local landscapes and regional factors structuring benthic macroinvertebrate assemblages in Swedish streams. Landscape Ecology, 19, 501-515.

Schoonover, J.E., Lockaby, B.G. & Pan, S. (2005). Changes in chemical and physical properties of stream water across an urban-rural gradient in western Georgia. Urban Ecosystems, 8, 107-124.

Scrimgeour, G.J. & Kendal, S. (2003). Effects of livestock grazing on benthic invertebrates from a native grassland ecosystem. Freshwater Biology, 48, 347- 362.

Secretariat of the Pacific Community (2007). National integrated water resource management (IWRM) diagnostic report Fiji Islands. (ed. Taylor, P). Applied

120

Geoscience and Technology division (SOPAC), Secretariat of the Pacific Community (SPC) Suva, Fiji.

Secretariat of the Pacific Community (2012). Catalogue of Rivers. Secretariat of the Pacific Community (SPC) Suva, Fiji.

Seeto, J. & Baldwin, W.J. (2010). A checklist of the fishes of Fiji and a bibliography of Fijian fishes. University of the South Pacific Suva, Fiji.

Shokita, S., Takano, M., Lal, S.N. & Vereivalu, T. (1985). Environmental Survey of rivers and biology of inland water prawns in Fiji. Ministry of Agriculture and Fisheries Suva, Fiji.

Sickle, V.J., Baker, J., Herlihy, A., Bayley, P., Gregory, S., Haggerty, P., Ashkenas, L. & Li, J. (2004). Projecting the biological condition of streams under alternative scenarios of human land use. Ecological Applications, 14, 368-380.

Slack, K.V., Tilley, L.J. & Kennelly, S.S. (1991). Mesh-size effects on drift sample composition as determined with a triple net sampler. Hydrobiologia, 209, 215– 226.

Smith, G. C., Covich, A. P., & Brasher, A. M. (2003). An ecological perspective on the biodiversity of tropical island streams. BioScience, 53(11), 1048-1051.

Snyder, C.D., Young, J.A., Lemarié, D.P. & Smith, D.R. (2002). Influence of eastern hemlock (Tsuga canadensis) forests on aquatic invertebrate assemblages in headwater streams. Canadian Journal of Fisheries and Aquatic Sciences, 59, 262–275

Sonoda, K., Yeakley, J.A. & Walker, C.E. (2001). Near-stream landuse effects on stream water nutrient distribution in an urbanizing watershed. Journal of the American Water Resources Association, 37, 1517-1532.

Sponseller, R.A., Benfield, E.F. & Valett, H.M. (2001). Relationships between land use, spatial scale and stream macroinvertebrate communities. Freshwater Biology, 46.

Stark, J.D., Boothroyd, I.K.G., Harding, J.S., Maxted, J.R. & Scarsbrook, M.R. (2001). Protocols for sampling macroinvertebrates in wadeable streams. New Zealand Macroinvertebrate Working Group Report No. 1. Prepared for the Ministry for the Environment.Sustainable Management Fund Project No. 5103 New Zealand, p. 57.

121

Stephenson, J.M. & Morin, A. (2009). Covariation of stream community structure and biomass of algae, invertebrates and fish with forest cover at multiple spatial scales. Freshwater Biology, 54, 2139-2154.

Stewart, S.J., Wang, L., Lyons, J., Horwatich, A.J. & Bannerman (2001). Influences of watershed, riparian-corridor, and reach-scale characteristics on aquatic biota in agricultural watersheds. Journal of the American Water Resources Association, 37, 1475-1487.

Stone, M.K. & Wallace, J.B. (1998). Long-term recovery of mountain stream from clear-cut logging: the effects of forest succession on benthic invertebrate community structure. Freshwater Biology, 39, 151-169.

Storey, A.W., Edward, D.H.D. & Gazey, P. (1991). Surber and kick sampling: a comparison for the assessment of macroinvertebrate community structure in streams of south-western Australia. Hydrobiologia, 211, 111-121.

Strayer, D.L. (2006). Challenges for freshwater invertebrate conservation. Journal of North American Benthological Society, 25, 271-287.

Strayer, D.L., Beighley, E.R., Thompson, C.L., Brooks, S., Nilsson, C., Pinay, G. & Naiman, J.R. (2003). Effects of land cover on stream ecosystems: roles of empirical models and scaling issues. Ecosystems, 6, 407-423.

Strayer, D.L., Caraco, N.F., Cole, J.J., Findlay, S. & Pace, M.L. (1999). Transformation of freshwater ecosystems by bivalves: a case study of zebra mussels in the Hudson River. BioScience, 49, 19-27.

Sundaresan, A. & Reddy, N. (1979). Salvinia molesta (Mitchell): a serious water weed in Fiji. Fiji Agricultural Journal, 40-41, 103-107.

Svendsen, C.R., Quinn, T. & Kolbe, D. (2004). Review of macroinvertebrate drift in lotic ecosystems. Seattle.

Taylor, J.N., Courtenay, W.R. & McCann, J.A. (1984). Known impacts of exotic fishes in the continental United States. In: Distribution, Biology, and Management of Exotic Fishes (eds. Courtenay, WR & Stauffer, JR). Johns Hopkins University Press Baltimore, pp. 322–373.

Terry, J.P. (1999). Kadavu Island, Fiji: Fluvial studies of a volcanic island in the humid tropical South Pacific. Singapore Journal of Tropical Geography, 20, 86-98.

122

The Murray-Darling Freshwater Research Center (2014). Major group identification. Available at: http://www.mdfrc.org.au/bugguide/display.asp?type=1&class=19 Last accessed 12th January 2014.

Thornton, D.P. (2008). Macroinvretebrate stream drift-an Australia example. Applied Ecology and Environmental Research, 6, 49-55.

Tiemann, J.S., Gillette, D.P., Wildhaber, M.L. & Edds, D.R. (2004). Effects of lowhead dams on riffle-dwelling fishes and macroinvertebrates in a Midwestern river. Transactions of the American Fisheries Society, 133, 705–717.

Tillyard, R.J. (1924). The dragonflies (Order Odonata) of Fiji with special reference to a collection made by Mr H.W. Simmonds, F.E.S. on the island of Viti Levu. Transactions of the Royal Entomological Society of London, 1923, 305–346.

Timm, H., Ivask, M. & Mols, T. (2001). Response of macroinvertebrates and water quality to long-term decrease in organic pollution in some Estonian streams during 1990-1998. Hydrobiologia, 464, 153-164.

Tuiwawa, M. (2010). Rapid assessment of the district Nakasaleka, Kadavu, Fiji Islands from 7th-12th August 2010. In: IAS Environmental Studies Report NO. 2010-05. Institute of Applied Science. University of the South Pacific. Suva, Fiji.

U.S. Environmental Protection Agency (1973). Biological field and laboratory methods for measuring the quality of surface waters and effluents. U. S. Environmental Protection Agency Cincinnati, Ohio, p. 176

United Nations Environment Programme (2010). Blue Harvest: inland fisheries as an ecosystem Service. United Nations Environment Programme (UNEP) Penang, Malaysia.

Volk, C.J., Kiffney, P., M & Edmonds, R.L. (2003). Role of riparian red alder (Alnus rubra) in the nutrient dynamics of coastal streams of the Olympic Peninsula, WA USA. In: Nutrient in salmon ecosystems: sustaining production and biodiversity (ed. Stockner, JG). American Fisheries Society Bethesda, pp. 213– 225.

Vörösmarty, J.C., McIntyre, B.P., Gessner, O.M., Dudgeon, D., Prusevich, A., Green, P., Glidden, S., Bunn, E.S., Sullivan, A.C., Liermann, R.C. & Davies, M.P. (2010). Global threats to human water security and river biodiversity. Nature, 467, 555-561.

123

Wade, M.L., Fanolua, S.F., Vargo, M.A., van Houte-Howes, K., Bardi, E. & Vargo, L.D. (2008). Exploiting macrofauna diadromy for assessing anthropogenic impact in American Samoa streams. Pacific Science, 62, 177-190.

Water Authority of Fiji (2012). History of Water Authority of Fiji. Available at: http://www.waterauthority.com.fj/en/history/ Last accessed 20th November 2012.

Watling, D. & Chape, S.P. (1992). Environment Fiji - the national state of the environment report. International Union for Conservation of Nature (IUCN) Gland, Switzerland.

Weijters, M.J., Janse, J.H., Alkemade, R. & Verhoeven, J.T. (2009). Quantifying the effect of catchment land use and water nutrient concentrations on freshwater river and stream biodiversity. Aquatic Conservation: Marine and Freshwater Ecosystems, 19, 104-112.

Welcomme, R.L., Cowx, I.G., Coates, D., Be´ne´, C., Funge-Smith, S., Halls, A. & Lorenzen, K. (2010). Inland capture fisheries. Philosophical Transactions of the Royal Society B: Biological Sciences, 365, 2881–2896.

Whiles, M.R. & Wallace, J.B. (1997). Leaf litter decomposition and macroinvertebrate communities in headwater streams draining pine and hardwood catchments. Hydrobiologia, 353, 107-119.

Williams, W.D. (1980). Australian Freshwater Life: the Invertebrates of Australian inland waters. Macmillan Company of Australia Pty Ltd, Melbourne.

Winterbourn, M.J. (2000). Feeding Ecology. In: New Zealand stream invertebrates: Ecology and implications for management (eds. Collier, KJ & Winterbourn, MJ). New Zealand Limnological Society Christchurch, pp. 100-124.

Winterbourn, M.J., Gregson, K.L.D. & Dolphin, C.H. (2006). Guide to the aquatic insects of New Zealand. 4th edn. Bulletin of the Entomological Society of New Zealand New Zealand.

Wisconsin Department of Natural Resources (2000). Guidelines for collecting macroinvertebrate samples from wadable streams. Bureau of Fisheries Management and Habitat Protection Monitoring and Data Assessment Section, Madison.

Wolman, M.G. (1954). A method of sampling coarse river-bed material. Transactions, American Geophysical Union, 35, 951-956.

124

World Watch Institute (2012). Use and capacity of global hydropower increases. Available at: http://www.worldwatch.org/node/9527 Last accessed 20th November 2012.

Worldclim (2014). Worldclim-global climate data. Available at: http://www.worldclim.org/ Last accessed 15th July 2014.

Wright-Stow, A.E. (2001). Biomonitoirng and the macroinvertebrate faunas of Canterbury stream. Masters of Science in Zoology. University of Canterbury Canterbury, New Zealand.

Yoshimura, M. & Maeto, K. (2006). Comparison of an aquatic invertebrate assemblage between an old-growth natural forest and planted coniferous forest basins in a Japanese temperate region: the Kuroson stream in the Shimanto River basin. Landscape and Ecological Engineering, 2, 81–89.

Young, G.R. & Collier, J.K. (2009). Contrasting responses to catchment modification among a range of functional and structural indicators of river ecosystem health. Freshwater Biology, 54, 2155-2170

Zhang, Y., Richardson, J.S. & Pinto, X. (2009). Catchment-scale effects of forestry practices on benthic invertebrate communities in Pacific coastal streams. Journal of Applied Ecology, 46, 1292–1303.

Zielske, S & Haase, M. (2014). New insights into tateid gastropods and their radiation on Fiji based on anatomical and molecular methods (Caenogastropoda: Truncatelloidea). London, Zoological Journal of the Linnean Society, 172, 71– 102.

125

Chapter 7. APPENDICES

7.1 Habitat Assessment form used during the study

126

127

7.2 Environmental variable graphs Table 7.1: (a-o): Bar graphs showing physical and chemical variables measured at wet and dry season across study sites

(a) Dissolved Oxygen Wet season Dry season

128 (b) Temperature Wet season Dry season

129 (c) pH Wet season Dry season

130

(d) Conductivity Wet season Dry season

131 (e) Total Dissolved Solids Wet season Dry season

132

(f) Turbidity Wet season Dry season

133

(g) Salinity Wet season Dry season

Zero ppt across all sites

134 (h) Ammonia Wet season Dry season

135

(i) Nitrate

Wet season Dry season

136 (k) Total Phosphorus Wet season Dry season

137 (l) Faecal coliform Wet season Dry season

138 (m) Canopy Cover Wet season Dry season

139 (n) Flow Wet season Dry season

140 (o) Riffle depth Wet season Dry season

141 (a) Streambed composition of Naqewa stream sampling stations

Lower Reach Mid Reach Upper Reach 100 90 Sand 80 Small gravel 70 Small medium gravel 60 Medium large gravel 50 Large gravel 40 Small cobble 30 Large cobble 20 Boulder Streambed composition (%) composition Streambed 10 Bedrock 0 Dry Wet Dry Wet Dry Wet Season (b) Streambed composition of Korolevu stream sampling stations

Lower Reach Mid Reach Upper Reach 100 Sand 90 Small gravel 80 70 Small medium gravel 60 Medium large gravel 50 Large gravel 40 Small cobble 30 Large cobble 20 Boulder 10 Bedrock Streambed composition (%)composition Streambed 0 Dry Wet Dry Wet Dry Wet Season (c) Streambed composition of Savumiri stream sampling stations

Lower Reach Mid Reach Upper Reach 100 90 Sand 80 Small gravel 70 Small medium gravel 60 Medium large gravel 50 Large gravel 40 Small cobble 30 Large cobble 20 10 Boulder Streambed composition (%) composition Streambed 0 Bedrock Dry Wet Dry Wet Dry Wet Season

142

(d) Streambed composition of Naikawaqa stream sampling stations

Lower Reach Mid Reach Upper Reach 100 90 Sand 80 Small gravel 70 Small medium gravel 60 Medium large gravel 50 Large gravel 40 Small cobble 30 Large cobble 20 Boulder 10 Streambed composition (%) (%) composition Streambed Bedrock 0 Dry Wet Dry Wet Dry Wet Season

(e) Streambed composition of Taveu stream sampling stations Lower Reach Mid Reach Upper Reach 100 90 Sand 80 Small gravel 70 Small Medium gravel 60 Medium large gravel 50 Large gravel 40 Small cobble 30 Large cobble 20 Boulder 10 Bedrock

Streambed composition (%) composition Streambed 0 Dry Dry Dry Season (f))p Streambed composition of Vucinivola stream samplingpg stations Lower Reach Mid Reach Upper Reach 100 90 Sand 80 Small gravel 70 Small medium gravel 60 Medium large gravel 50 Large gravel 40 Small cobble 30 Large cobble 20 Boulder

Streambed composition (%) composition Streambed 10 0 Bedrock Dry Dry Dry Season

Figure 7.1: (a-f) Streambed substrate composition across sites surveyed within Ra and Kadavu streams.

143

7.3 Functional Feeding groups for macroinvertebrates recorded during the study

Filtering collectors Gathering collectors Shredder Scraper Piercers-herbivores Predator Abacaria (caddisfly) Baetis (mayfly) Triaenodes Acochlidium (snail) Oxyethira Apsilochorema (caddisfly) (caddisfly) (micro-caddisfly) Hydropsychidae Pseudocloeon (mayfly) Triplectides Goera (caddisfly) Araneae (spider) (caddisfly) (caddisfly) Simulium Caenis (mayfly) Limonia (cranefly) Fluviopupa (snail) Gyrinidae (whirligig beetle) (black fly) Chimarra (caddisfly) Chironomus (midge) Tipula (cranefly) Septaria (snail) Corduliidae Atyopsis (shrimps) Chironomidae (midge) Paralimophila Neritina (snail) Oecetis (caddisfly) (cranefly) Atyoida (shrimps) Stratiomyidae Tipulidae Neritilia(snail) Coleoptera (Soldier fly) (cranefly) Caridina (shrimps) Psychoda (moth fly) Anisocentropus Clithon(snail) Dytiscidae (diving beetle) (caddisfly) Spercheidae (beetle) Oligochaeta (worm) Hymenoptera Thiara(snail) Hemigrapsus (crab) (ant) Empididae (dance fly) Nymphula (moth) Hemiptera (water bug) Culicidae (mosquitoes) Crambidae (moth) Hirudinea (leech) Dixidae (dixid midge) Hygraula (moth) Hydrobiosis (caddisfly) Nemobiinae Psephenidae Hydrophilidae (water-cricket) (water-penny beetle) Tanypodinae (midge) Elmidae (beetle) Labuanium (crab) Helminthidae (beetle) Limnogonus (water strider) Hydraenidae (beetle) Macrobrachium (Prawn) Odontoceridae (caddisfly) Mesovelia (water bug) Melanoides (snail) Microvelia (water bug) Scirtidae (marsh beetle) Namalycastis(bristle worm) Physastra (snail) Nesobasis (damselfly) Saldidae (shore bug) Varuna (crab) Rhyacophilidae (caddisfly) Corixidae (water boatman)

144

7.4 PERMANOVA analyses on macroinvertebrate species diversity measures

Table 7.2: PERMANOVA results on the species richness (S) data within stream. Significant value is in bold.

S- main test Source df SS MS Pseudo-F P(perm) perms St 5 494.4 98.879 6.3527 0.0001 4962 Res 150 2334.7 15.565 Total 155 2829.1

Table 7.3: PERMANOVA pair-wise test on the effect of stream on species richness (S). Significant values are in bold. P (perm) is p-value based on n where n = unique permutations.

S- pairwise test Groups t P(perm) Unique perms Korolevu, Naqewa 0.99067 0.3399 56 Korolevu, Savumiri 2.3181 0.0234 59 Korolevu, Naikawaqa 6.1749 0.0001 50 Korolevu, Taveu 1.9669 0.058 87 Korolevu, Vucinivola 2.0589 0.0488 42 Naqewa, Savumiri 1.1747 0.2538 65 Naqewa, Naikawaqa 4.4005 0.0001 53 Naqewa, Taveu 0.92395 0.3682 97 Naqewa, Vucinivola 0.8935 0.3972 94 Savumiri, Naikawaqa 3.5562 0.0007 97 Savumiri, Taveu 4.4991E-2 0.9721 100 Savumiri, Vucinivola 0.1446 0.9115 49 Naikawaqa, Taveu 3.2917 0.0038 47 Naikawaqa, Vucinivola 3.881 0.0008 80 Taveu, Vucinivola 8.9233E-2 0.9636 40

Table 7.4: PERMANOVA results on the total number of individuals (N) data within stream. Significant value is in bold.

N- main test Source df SS MS Pseudo-F P(perm) Unique perms St 5 2.8787E5 57574 3.3859 0.0072 9945 Res 150 2.5506E6 17004 Total 155 2.8385E6

145

Table 7.5: PERMANOVA pair-wise test on the effect of stream on total number of individuals (N). Significant values are in bold. P (perm) is p-value based on n where n = unique permutations.

N- pairwise test Groups t P(perm) Unique perms Korolevu, Naqewa 2.5393 0.0118 1302 Korolevu, Savumiri 2.5374 0.0111 1546 Korolevu, Naikawaqa 0.66766 0.5091 803 Korolevu, Taveu 0.73221 0.4707 898 Korolevu, Vucinivola 1.5039 0.1304 1882 Naqewa, Savumiri 0.44828 0.6621 1590 Naqewa, Naikawaqa 2.3758 0.0224 1754 Naqewa, Taveu 2.7739 0.008 1944 Naqewa, Vucinivola 0.52445 0.6134 1189 Savumiri, Naikawaqa 2.0833 0.0438 2190 Savumiri, Taveu 2.4754 0.0162 2393 Savumiri, Vucinivola 0.76323 0.4654 1412 Naikawaqa, Taveu 4.9754E-2 0.9642 544 Naikawaqa, Vucinivola 1.7056 0.0987 846 Taveu, Vucinivola 1.9437 0.0632 904

Table 7.6: PERMANOVA results on the Shannon-Weiner diversity index (Loge) data within stream. Significant value is in bold.

(Loge) Shannon Weiner diversity index - main test Source df SS MS Pseudo-F P(perm) Unique perms St 5 10.487 2.0974 18.495 0.0001 9943 Res 150 17.01 0.1134 Total 155 27.498

146

Table 7.7: PERMANOVA pair-wise test on the effect of stream on Shannon-Weiner diversity index (Loge). Significant values are in bold. P (perm) is p-value based on n where n = unique permutations.

(Loge) Shannon Weiner diversity index - pairwise test Groups t P(perm) Unique perms Korolevu, Naqewa 7.0173 0.0001 9831 Korolevu, Savumiri 6.0351 0.0001 9824 Korolevu, Naikawaqa 8.8338 0.0001 9816 Korolevu, Taveu 3.5842 0.0011 9839 Korolevu, Vucinivola 4.9767 0.0001 9827 Naqewa, Savumiri 0.40931 0.6839 9816 Naqewa, Naikawaqa 4.1604 0.0001 9829 Naqewa, Taveu 1.5081 0.1359 9817 Naqewa, Vucinivola 1.6786 0.0967 9840 Savumiri, Naikawaqa 4.111 0.0002 9809 Savumiri, Taveu 1.0991 0.2789 9845 Savumiri, Vucinivola 1.1431 0.265 9828 Naikawaqa, Taveu 3.9688 0.0006 9820 Naikawaqa, Vucinivola 4.9433 0.0001 9825 Taveu, Vucinivola 0.14777 0.8858 9864

147

7.5 Macroinvertebrate species and corresponding abundances recorded

across Kadavu sampling stations (Surber sampler data)

148

Trichoptera - Odontoceridae Odontoceridae sp. A 0 0 0 0 2 0 0 0 0 000000 Trichoptera - Odontoceridae Odontoceridae sp. B 34 95 78 21 91 67 3 6 3 4 3 25 3 3 5 Trichoptera - Leptoceridae Oecetis sp. 3 50 458 14 95 121 0 0 0 0 0 0 0 0 1 Trichoptera - Hydroptilidae Oxyethira fijiensis 0000 000000 0 0 003 Trichoptera - Hydroptilidae Oxyethira sp. A 11 12 27 36 41 52 0 13 0 05111116 Trichoptera - Hydroptilidae Oxyethira sp. B 0 3 0 6 7 0 0 0 0 000010 Trichoptera - Leptoceridae Triaenodes fijianus 2130 150000 0 0 001 Trichoptera - Leptoceridae Triplectides sp. 0000 020000 0 0 000 Ephemeroptera - Baetidae Baetis sp. 15 1 0 0 0 0 39 52 7 0 89 0 182 389 12 Ephemeroptera - Caenidae Caenis sp. 0001200000780 32000 Ephemeroptera - Baetidae Pseudocloeon sp. 175 117 73 192 56 100 88 395 118 189 363 204 114 157 12 Lepidoptera - Crambidae Crambidae sp. B 0000 000000 0 0 002 Lepidoptera - Crambidae Hygraula sp. 0000400101100130 Lepidoptera - Pyralidae Nymphula sp. A 32 8 8 11 12 30 0 8 2 045105016 Lepidoptera - Pyralidae Nymphula sp. C 0 0 0 0 1 0 0 0 0 000001 Lepidoptera - Pyralidae Nymphula sp. D 0 0 0 0 0 0 0 0 0 010001 Lepidoptera - Pyralidae Nymphula sp. F 0 0 0 0 0 0 0 0 0 000040 Odonata Anispotera Corduliidae Corduliidae 2 13 2 2 5 2 0 0 1 001002 Odonata Zygoptera Coenagrionidae Nesobasis sp. A 0 0 0 0 0 0 0 1 0 000000 Odonata Zygoptera Coenagrionidae Nesobasis sp. B 0005 740021 8 1 110 Odonata Zygoptera Coenagrionidae Nesobasis sp. E 1 3 0 1 3 4 0 2 0 140004 Odonata Zygoptera Coenagrionidae Nesobasis sp. F 5 11 3 0 0 0 0 2 0 0000011 Odonata Zygoptera Coenagrionidae Nesobasis sp. G 0 0 1 0 0 0 0 2 0 000000 Hemiptera - Gerridae Limnogonus sp. 0000 000000 0 0 001 Hemiptera - Mesoveliidae Mesovelia sp. A 0000 000000 0 0 000 Hemiptera - Mesoveliidae Mesovelia sp. B 0000 000000 0 1 000 Hemiptera - Veliidae Microvelia sp. A 1010 100010 1 0 024 Hemiptera - Veliidae Microvelia sp. B 0000 000000 0 0 000 Hemiptera - Saldidae Saldidae 0000 000000 0 0 010 Hymenoptera - Hymenoptera Hymenoptera sp. A 9 0 1 1 2 2 0 1 3 311301 Hymenoptera - Hymenoptera Hymenoptera sp. B 0 0 1 0 0 0 0 1 3 001000 Orthoptera - Nemobiinae Nemobiinae 2 2 0 1 0 1 5 0 1 001100 Diptera - Chironomidae Chironomidae sp. B 0 0 0 0 0 2 0 0 0 000000 Diptera - Chironomidae Chironomidae sp. C 2 3 8 3 0 0 0 0 0 000000 Diptera - Chironomidae Chironomidae sp. D 0 0 0 0 0 0 0 0 0 000000 Diptera - Chironomidae Chironomidae sp. E 0 1 9 1 1 2 0 0 0 000000 Diptera - Chironomidae Chironomus sp. 60 196 89 36 20 19 83 114 8 14 12 17 315 7 10 Diptera - Culicidae Culicidae 2 38 28 8 14 14 8 12 1 552431033 Diptera - Dixidae Dixidae sp. A 51232346711003810 Diptera - Dixidae Dixidae sp. B 0000 000120 0 0 000 Diptera - Empididae Empididae sp. 213111392401222531 Diptera - Tipulidae Limonia sp. 3001 000000 0 1 202 Diptera - Tipulidae Paralimnophila sp. 0000 010000 0 0 000 Diptera - Psychodidae Psychoda sp. 0102 000100 0 0 000 Diptera - Simuliidae Simulium jolli 27152425342140076713 Diptera - Simuliidae Simulium sp. B 0 0 0 0 0 3 1 1 1 000001 Diptera - Stratiomyidae Stratiomyidae 0000 000100 0 0 000 Diptera - Chironomidae Tanypodinae sp. 3 3 0 8 1 1 4 27 0 26252532 Diptera - Tipulidae Tipula sp. 0100 000100 0 0 000 Coleoptera - Dytiscidae Dytiscidae sp. A 0 0 0 0 0 0 0 0 0 000100 Coleoptera - Dytiscidae Dytiscidae sp. B 0 0 0 1 0 1 0 0 0 000000 Coleoptera - Elmidae Elmidae 0001 000000 0 0 000 Coleoptera - Helminthidae Helminthidae 0000 010000 0 0 000 Coleoptera - Hydraenidae Hydraenidae 0000 100000 0 0 000 Coleoptera - Hydrophilidae Hydrophilidae 0000 000000 0 0 000 Coleoptera - Psephenidae Psephenidae 0000 010000 0 0 000 Coleoptera - Scirtidae Scirtidae 0 1 4 0 1 3 0 0 0 000000 Araneae - - Araneae 2 0 1 1 2 0 2 1 3 000321 ca Decapoda - Atyidae Atyoida pilipes 0006 001069351017010 ca Decapoda - Atyidae Atyopsis spinipes 5001 0028233299 47001 ca Decapoda - Atyidae Caridina multidentata 0000 000000 0 0 000 ca Decapoda - Atyidae Caridina sp. A 0 0 0 0 0 0 0 1 0 000000 ca Decapoda - Atyidae Caridina sp. C 0 0 0 3 0 0 0 0 0 620000 ca Decapoda - Atyidae Caridina sp. D 0 0 0 0 0 0 0 2 0 001000 ca Decapoda - Atyidae Caridina sp. F 0 0 0 0 0 0 0 0 0 000030 ca Decapoda - Atyidae Caridina sp. H 0 0 0 0 0 0 4 0 0 000000 ca Decapoda - Atyidae Caridina sp. K 0 0 0 0 0 0 0 16 0 0 0 0 000 ca Decapoda - Atyidae Caridina sp. L 0 0 0 0 0 0 0 0 0 000010 ca Decapoda - Atyidae Caridina sp. M 0 0 0 0 0 0 0 0 0 000010 ca Decapoda - Atyidae Caridina sp. X 20 0 0 0 0 0 0 3 2 010011 ca Decapoda - Sesarmidae Labuanium trapezoideum 0000 000000 0 1 000 ca Decapoda - Palaemonidae Macrobrachium caledonicum 0001 000000 0 0 000 ca Decapoda - Palaemonidae Macrobrachium holthuis 0000 000000 0 0 000 ca Decapoda - Palaemonidae Macrobrachium lar 0000 000000 0 0 225 ca Decapoda - Palaemonidae Macrobrachium lepidactyloides 1003 000200 0 0 020 ca Decapoda - Grapsidae Varuna litterata 0000 004001 1 0 070 a - - Neritidae Clithon castanea 0000 000001 0 0 000 a - - Neritidae Clithon olivaceus 2000 00118137 156 000 a - - Neritidae Clithon pritchardi 1000 004006 0 0 000 a - - Tateidae Fluviopupa sp. 1000 0024171 4 32030 a - - Thiaridae Melanoides aspirans 0000 000000 0 0 510 a - - Thiaridae Melanoides tuberculata 0000 002211 0 1 582 a - - Neritidae Neritilia rubida 5000 00101204 1 28900 a - - Neritidae Neritina canalis 0001 000010 1 1 021 a - - Neritidae Neritina petitii 0000 001100 0 0 000 a - - Neritidae Neritina porcata 0000 000000 1 0 000 a - - Neritidae Neritina pulligera 0000 000000 0 0 431 a - - Neritidae Septaria bougainvillei 0000 005106 2 1 030 a - - Neritidae Septaria macrocephala 13000 00542331126 000 a - - Neritidae Septaria sanguisuga 0000 00013153 1618000 a - - Neritidae Septaria suffreni 4002 000000 0 2 140 Hirudinea - - Hirudinea 0 0 1 0 0 0 0 0 0 000000 Oligochaeta - Naididae Oligochaeta sp. A 4 13 3 2 2 2 1 11 5 326630 7.6 Macroinvertebrate species and corresponding abundances recorded

across Ra sampling stations (Surber sampler data)

150

Arthropoda Insecta Trichoptera - Odontoceridae Odontoceridae sp. B 18 12 15 2 6 5 10 27 29 5 7 9 Arthropoda Insecta Trichoptera - Leptoceridae Oecetis sp. 0 0 2 0 26 2 0 0 0 0 1 1 Arthropoda Insecta Trichoptera - Rhyacophilidae Rhyacophilidae sp. X 0 0 0 0 2 0 0 0 0 0 0 0 Arthropoda Insecta Trichoptera - Rhyacophilidae Rhyacophilidae sp. Y 0 39 0 0 1 0 0 0 0 0 0 0 Arthropoda Insecta Trichoptera - Rhyacophilidae Rhyacophilidae sp. Z 0 0 0 0 0 0 0 38 0 0 0 0 Arthropoda Insecta Trichoptera - Leptoceridae Triaenodes sp. A 0100 000000 00 Arthropoda Insecta Trichoptera - Leptoceridae Triplectides sp. 0700 000000 00 Arthropoda Insecta Ephemeroptera - Baetidae Baetis sp. 111758071303 30 Arthropoda Insecta Ephemeroptera - Baetidae Pseudocloeon sp. 7 35 200 17 97 21 45 210 162 775 246 253 Arthropoda Insecta Lepidoptera - Pyralidae Nymphula sp. A 0010011157001408 Arthropoda Insecta Odonata Aniosptera Corduliidae Corduliidae 0 2 0 0 0 0 0 0 0 0 0 0 Arthropoda Insecta Odonata Zygoptera Coenagrionidae Nesobasis rufostigma 0000 4100000 00 Arthropoda Insecta Odonata Zygoptera Coenagrionidae Nesobasis sp. A 0000 100000 00 Arthropoda Insecta Odonata Zygoptera Coenagrionidae Nesobasis sp. B 1 10 25 3 12 0 1 6 9 1 6 5 Arthropoda Insecta Odonata Zygoptera Coenagrionidae Nesobasis sp. C 0000 0670000 00 Arthropoda Insecta Odonata Zygoptera Coenagrionidae Nesobasis sp. E 0300 000200 20 Arthropoda Insecta Odonata Zygoptera Coenagrionidae Nesobasis sp. H 1000 000000 00 Arthropoda Insecta Odonata Zygoptera Coenagrionidae Nesobasis sp. X 0000 400000 00 Arthropoda Insecta Hemiptera - Corixidae Corixidae sp. A 0 0 1 0 0 0 0 0 0 0 0 0 Arthropoda Insecta Hemiptera - Corixidae Corixidae sp. B 0 1 1 0 0 0 0 0 0 0 0 0 Arthropoda Insecta Hemiptera - Hemiptera Hemiptera sp. X 0 0 0 0 1 0 0 0 0 0 0 0 Arthropoda Insecta Hemiptera - Gerridae Limnogonus sp. 0000 310000 00 Arthropoda Insecta Hemiptera - Veliidae Microvelia sp. A 1 21320 6 2801 1 2 6 1 Arthropoda Insecta Hemiptera - Veliidae Microvelia sp. B 0440 100000 00 Arthropoda Insecta Hymenoptera - Hymenoptera sp. A 0 0 0 0 0 0 0 0 4 1 0 1 Arthropoda Insecta Orthoptera - Nemobiinae Nemobiinae 1 1 0 0 0 0 0 0 0 0 1 0 Arthropoda Insecta Diptera - Chironomidae Chironomidae sp. B 0 0 0 24 1 0 0 0 1 0 0 0 Arthropoda Insecta Diptera - Chironomidae Chironomidae sp. C 1 1 1 0 0 0 0 0 0 1 0 3 Arthropoda Insecta Diptera - Chironomidae Chironomidae sp. D 0 0 0 0 0 0 0 0 2 0 2 0 Arthropoda Insecta Diptera - Chironomidae Chironomus sp. 0 16 20 39 16 20 14 10 40 213 68 57 Arthropoda Insecta Diptera - Culicidae Culicidae 2 5 1 7 5 0 9 4 7 48 3 11 Arthropoda Insecta Diptera - Dixidae Dixidae sp. A 0011 003058 58 Arthropoda Insecta Diptera - Empididae Empididae sp. 0100 244361023 Arthropoda Insecta Diptera - Psychodidae Psychoda sp. 0000 200130 00 Arthropoda Insecta Diptera - Simuliidae Simulium jolli 0000 000707 861 Arthropoda Insecta Diptera - Chironomidae Tanypodinae sp. 0 3 0 2 2 0 2 2 2 41 4 5 Arthropoda Insecta Diptera - Tipulidae Tipula sp. 0121 290000 00 Arthropoda Insecta Diptera - Tipulidae Tipulidae 0 0 0 0 0 0 1 0 0 0 0 0 Arthropoda Insecta Coleoptera - - Coleoptera sp. X 0 2 0 0 0 0 0 0 0 0 0 0 Arthropoda Insecta Coleoptera - - Coleoptera sp. Z 0 0 0 0 1 0 0 0 0 0 0 0 Arthropoda Insecta Coleoptera - Dytiscidae Dytiscidae sp. A 0 1 1 0 0 0 0 0 0 0 0 0 Arthropoda Insecta Coleoptera - Gyrinidae Gyrinidae sp. 0 0 0 0 16 0 0 0 0 0 0 0 Arthropoda Insecta Coleoptera - Scirtidae Scirtidae 0 3 0 0 0 0 0 0 1 0 0 1 Arthropoda Insecta Coleoptera - Spercheidae Spercheidae 0000 000010 00 Chelicerata Arachnida Araneae - - Araneae 1 0 1 3 4 0 0 0 2 1 2 0 Crustacea Malacostraca Decapoda - Atyidae Atyoida pilipes 000140107411 00 Crustacea Malacostraca Decapoda - Atyidae Atyopsis spinipes 051342211760 01 Crustacea Malacostraca Decapoda - Atyidae Caridina leucosticta 0010 000000 00 Crustacea Malacostraca Decapoda - Atyidae Caridina multidentata 0000 100000 00 Crustacea Malacostraca Decapoda - Atyidae Caridina sp. A 0010 011000 00 Crustacea Malacostraca Decapoda - Atyidae Caridina sp. B 0000 013000 00 Crustacea Malacostraca Decapoda - Atyidae Caridina sp. C 0030 010300 03 Crustacea Malacostraca Decapoda - Atyidae Caridina sp. D 0100 707302 00 Crustacea Malacostraca Decapoda - Atyidae Caridina sp. F 0001 000000 00 Crustacea Malacostraca Decapoda - Atyidae Caridina sp. G 0004 007000 00 Crustacea Malacostraca Decapoda - Atyidae Caridina sp. H 0002 401000 00 Crustacea Malacostraca Decapoda - Atyidae Caridina sp. J 0000 0091940 03 Crustacea Malacostraca Decapoda - Atyidae Caridina sp. K 8000 006300 00 Crustacea Malacostraca Decapoda - Atyidae Caridina sp. L 0002 000000 00 Crustacea Malacostraca Decapoda - Atyidae Caridina sp. N 0001 000000 00 Crustacea Malacostraca Decapoda - Atyidae Caridina sp. Q 0200 010000 00 Crustacea Malacostraca Decapoda - Atyidae Caridina sp. X 0010 002100 00 Crustacea Malacostraca Decapoda - Atyidae Caridina sp. Z 0000 000500 00 Crustacea Malacostraca Decapoda - Atyidae Caridina typus 19 34 0 0 0 0 0 0 0 0 0 0 Crustacea Malacostraca Decapoda - Palaemonidae Macrobrachium caledonicum 0000 000001 00 Crustacea Malacostraca Decapoda - Palaemonidae Macrobrachium lar 0010 000000 00 Crustacea Malacostraca Decapoda - Palaemonidae Macrobrachium lepidactyloides 0230 001020 01 Crustacea Malacostraca Decapoda - Grapsidae Varuna litterata 0000 004000 00 Crustacea Malacostraca Decapoda - Grapsidae Hemigrapsus sp. 0002 000000 00 Mollusca Gastropoda - - Acochlidacea Acochlidium fijiense 0000 000000 02 Mollusca Gastropoda - - Neritidae Clithon diadema 14 0 0 13 0 0 0 0 0 0 0 0 Mollusca Gastropoda - - Neritidae Clithon olivaceus 0009 004202891281 Mollusca Gastropoda - - Neritidae Clithon pritchardi 4009 0021006 54 Mollusca Gastropoda - - Tateidae Fluviopupa sp. 0000 100951216 Mollusca Gastropoda - - Thiaridae Melanoides aspirans 0010 006115001 Mollusca Gastropoda - - Thiaridae Melanoides lutosa 0170 000000 00 Mollusca Gastropoda - - Thiaridae Melanoides tuberculata 0641041004000 Mollusca Gastropoda - - Neritidae Neritilia rubida 318 0 0 684 0 0 365 83 46 2 3 0 Mollusca Gastropoda - - Neritidae Neritina canalis 0430 240220 00 Mollusca Gastropoda - - Neritidae Neritina petitii 0120 100273 21 Mollusca Gastropoda - - Neritidae Neritina porcata 0000 000222 05 Mollusca Gastropoda Neritidae Neritina pulligera 0020 040200 11 7.7 Macroinvertebrate species and corresponding abundances recorded

across Ra sampling stations (kick-net data)

152

pp Trichoptera - Philopotamidae Chimarra sp. A 00 0 Trichoptera - Philopotamidae Chimarra sp. B 4161 Ephemeroptera - Baetidae Baetis sp. 10 0 Ephemeroptera - Baetidae Pseudocloeon sp. 48 82 206 Lepidoptera - Pyralidae Nymphula sp. A 01016 Odonata Zygoptera Coenagrionidae Agriocnemis exsudans 00 0 Odonata Zygoptera Coenagrionidae Nesobasis rufostigma 00 0 Odonata Zygoptera Coenagrionidae Nesobasis sp. A 00 2 Odonata Zygoptera Coenagrionidae Nesobasis sp. B 22 1 Odonata Zygoptera Coenagrionidae Nesobasis sp. C 00 0 Odonata Zygoptera Coenagrionidae Nesobasis sp. D 00 0 Diptera - Chironomidae Chironomus sp. 61 1 Diptera - Culicidae Culicidae 0 0 0 Diptera - Empididae Empididae sp. 0 0 0 Diptera - Simuliidae Simulium jolli 15392 Diptera - Stratiomyidae Stratiomyidae 0 0 1 Diptera - Chironomidae Tanypodinae sp. 0 0 0 Coleoptera - - Coleoptera sp. X 0 1 0 Coleoptera - - Coleoptera sp. Y 0 1 0 nida Araneae - - Araneae 1 0 0 ostraca Decapoda - Atyidae Atyoida pilipes 01 14 ostraca Decapoda - Atyidae Atyopsis spinipes 82 0 0 ostraca Decapoda - Atyidae Caridina leucosticta 01 0 ostraca Decapoda - Atyidae Caridina sp. M 00 1 ostraca Decapoda - Atyidae Caridina sp. X 0150 ostraca Decapoda - Atyidae Caridina sp. Y 01 0 ostraca Decapoda - Atyidae Caridina sp. Z 02 0 ostraca Decapoda - Sesarmidae Labuanium trapezoideum 00 0 ostraca Decapoda - Palaemonidae Palaemonidae sp. 0 1 0 poda - - Neritidae Clithon diadema 15 0 0 poda - - Neritidae Clithon Olivaceus 31 24 49 poda - - Neritidae Clithon pritchardi 48 0 2 poda - - Tateidae Fluviopupa sp. 01 34 poda - - Thiaridae Melanoides aspirans 12 13 11 poda - - Thiaridae Melanoides plicaria 30 0 poda - - Thiaridae Melanoides tuberculata 20 7 poda - - Neritidae Neritilia rubida 126 35 30 poda - - Neritidae Neritina canalis 13 2 poda - - Neritidae Neritina petitii 05 2 poda - - Neritidae Neritina porcata 04 12 poda - - Neritidae Neritina pulligera 22 12 6 poda - - Neritidae Neritina variegata 00 1 poda - - Neritidae Septaria bougainvillei 00 0 poda - - Neritidae Septaria livida 13 0 0 poda - - Neritidae Septaria macrocephala 45 6 poda - - Neritidae Septaria sanguisuga 00 0 poda - - Neritidae Septaria suffreni 51021 poda - - Thiaridae Thiara amarula 20 0 poda - - Thiaridae Thiara scabra 20 0 ata Hirudinea - - Hirudinea 0 1 0 ata Oligochaeta - Naididae Oligochaeta sp. A 1 3 0 Total Abundance 479 471 871 1