A Study Into the Anthropogenic Impacts Affecting the Elands River, Mpumalanga

A study into the anthropogenic impacts affecting the Elands River, Mpumalanga.

MARTIN FERREIRA

DISSERTATION

SUBMITTED IN FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE

MAGISTER SCIENTIAE

AQUATIC HEALTH

IN THE

FACULTY OF SCIENCE

AT THE

UNIVERSITY OF JOHANNESBURG

SUPERVISOR: PROF. JHJ VAN VUREN

CO – SUPERVISOR: PROF. V WEPENE

ACKNOWLEDGEMENTS

Met dank en lof aan my Almagtige Skepper en Hemelse Vader, want “ ’n mens kan sy voornemens hê, maar die laaste word daaroor kom van die Here af ” (Spreuke 16:1)

I would like to thank and acknowledge with appreciation the following people and institutions:

My parents and my sisters, for their continuous love and support, especially my farther who has given me every opportunity to better myself.

My supervisors Professor Van Vuuren and Professor Wepener for their leadership and support. A special thank you to prof Vic whose door is always open for all students, no matter how small the problem.

Gordon O’Brien for his leadership and giving me this opportunity.

The WRC and the University of Johannesburg for providing funding and equipment to complete the study.

I would also like to thank my Paper Mill study team: Wynand, Cameron, Irene and Riaan. You started out as my colleges and ended up being great friends.

Oom Giel and Tannie Louisa for everything they have done for me, and making it possible for me to complete my studies.

A final thank you to Maryke Coetzee for loving me, supporting me and putting up with me despite myself.

- 2 -

TABLE OF CONTENT

List of Tables 7 List of Figures 10 Summary 17 Opsomming 21

CHAPTER 1: GENERAL INTRODUCTION 1.1 Introduction 25 1.1.1 The River Health Program 26 1.1.2 The Reserve 28 1.1.3 EcoClassification and EcoStatus 29 1.2 Aim and Objective 30 1.3 References 31

CHAPTER 2: LITERATURE SURVEY AND SITE SELECTION 2.1 Introduction 33 2.1.1 The River Health Program 34 2.1.2 The Reserve 34 2.1.3 A New Approach for the management of rivers 35 2.2 The Study Area 36 2.3 SAPPI and the Elands River 38 2.4 Biomonitoring 40 2.4.1 Water Quality 41 2.4.2 Sediment 44 2.4.3 Habitat 45 2.4.4 Riparian Vegetation 47 2.4.5 Ichthyofauna 51 2.4.6 Macro Invertebrates 54 2.5 Site Selection 57 - 3 -

2.6 References 67

CHAPTER 3: WATER QUALITY 3.1 Introduction 76 3.2 Materials and Methods 77 3.2.1 Field Surveys 77 3.2.2 Laboratory Analysis 78 3.2.3 Statistical Analysis 78 3.3 Results and Discussion 79 3.4 Conclusion and Recommendations 99 3.5 References 100

CHAPTER 4: SEDIMENT, HABITAT AND RIPARIAN VEGETATION 4.1 Introduction 104 4.2 Materials and Methods 106 4.2.1 Physical and Chemical Characteristics of Sediment 106 4.2.2 Habitat Quality Indices 107 4.2.3 Riparian Vegetation 108 4.2.3 Statistical Analysis 108 4.3 Results and Discussion 108 4.3.1 Sediment 108 4.3.2 Habitat Quality 114 4.3.3 Riparian Vegetation 118 4.4 Conclusion and Recommendations 126 4.5 References 127

CHAPTER 5: ICHTHYOFAUNA 5.1 Introduction 132 5.2 Materials and Methods 133

- 4 -

5.2.1 Field Surveys 133 5.2.2 Spatial and Temporal Analysis 134 5.2.3 The FAII 135 5.3 Results and Discussion 135 5.3.1 Spatial and Temporal Analysis 135 5.3.2 Biotic Indices 151 5.4 Conclusion and Recommendations 155 5.5 References 157

CHAPTER 6: INVERTEBRATES 6.1 Introduction 161 6.2 Materials and Methods 162 6.2.2 SASS 162 6.2.1 Spatial and Temporal Analysis 163 6.2.3 Laboratory Analysis 165 6.3 Results and Discussion 165 6.3.1 Spatial and Temporal Analysis 165 6.3.2 Biotic Indices 178 6.4 Conclusion and Recommendations 191 6.5 References 193

CHAPTER 7: ECOCLASSIFICATION 7.1 Introduction 197 7.2 Materials and Methods 199 7.2.1 Macro Invertebrate Response Assessment Index 199 7.2.2 Fish Response Assessment Index 199 7.2.3 Rating, Ranking and Weighting 200 7.3 Results and Discussion 201 7.3.1 Macro Invertebrate Response Assessment Index 201

- 5 -

7.3.2 Fish Response Assessment Index 203 7.4 Conclusion and Recommendations 204 7.5 References 205

CHAPTER 8: GENERAL CONCLUSION AND RECOMMENDATIONS 8.1 Conclusion 207 8.2 Recommendations 209

APPENDIX A 211

- 6 -

LIST OF TABLES

Table 2.1 Typical analysis of acid bleach effluent generated during the activities of the Ngodwana pulp and paper mill.

Table 2.2 Examples of industry practices that fall within a Category I industry.

Table 2.3 Criteria used in the assessment of habitat integrity.

Table 2.4 Vegetation species found in the Riparian zone of Resource Unit 2 (Roodewal, Doornhoek, Waterval Boven, Blouboshkraal) T = tree; S = shrub; R = reed; F = herbaceous; G = grass.

Table 2.5 Vegetation species found in the Riparian zone of Resource Unit 4 (Roodewal BB2, Houtboschoek, Elandshoek & Ngodwana) T = tree; S = shrub; R = reed; F = herbaceous; G = grass.

Table 2.6 The present ecological state if the fish in the Elands River within resource unit 2.

Table 2.7 Present ecological state of the invertebrate communities of the Elands River within resource unit 2.

Table 3.1 Nutrient and system variables results for the high flow survey (March 2005). Data (in mg/l unless otherwise stated) represents results of once of sampling.

Table 3.2 Nutrient and system variables results for the low flow survey (June 2005). Data (in mg/l unless otherwise stated) represents results of once of sampling.

Table 3.3 Results (mean ± standard deviation) obtained by Labware laboratories for system and nutrient variables during high flow conditions (3 Jan 2005 – 9 May 2005).

Table 3.4 Results (mean ± standard deviation) obtained by Labware laboratories for system and nutrient variables during low flow conditions (27 June 2005 – 24 Oct 2005).

Table 3.5 Examples of industry practices that fall within a category I industry. - 7 -

Table 3.6 Reference conditions for average monthly water temperature (°C) and dissolved oxygen (mg/l).

Table 3.7 Reference conditions for average monthly electrical conductivity (µs/cm).

Table 3.8 Reference conditions for selected water quality variables (pH units and mg/l).

Table 3.9 Reserve conditions for average monthly electrical conductivity (mS/m).

Table 4.1 Physical and chemical characteristics of the sediment collected during high flow conditions.

Table 4.2 Physical and chemical characteristics of the sediment collected during low flow conditions.

Table 4.3 The method applied in assigning ecological classes based on the results of habitat indices applied in the study.

Table 4.4 Results (IHAS and HQI scores and ecological classes) of the habitat quality indices during high flow conditions.

Table 4.5 Results (IHAS and HQI scores and ecological classes) of the habitat quality indices during low flow conditions.

Table 4.6 Vegetation species found in the Riparian zone at sites on the Elands and Crocodile Rivers (species name in bold indicate non – endemic species).

Table 5.1 Results (FAII score, ecological classes and abundances) of ichthyofaunal species collected during high flow conditions.

Table 5.2 Results (FAII score, ecological classes and abundances) of ichthyofaunal species collected during low flow conditions.

- 8 -

Table 5.3 The contribution of the various taxa to the similarity within groups (determined by using SIMPER) for both flow regimes.

Table 5.4 The FAII scoring system applied in this study for assigning ecological classes.

Table 6.1 Biotopes and sampling duration of the biotopes that were included in this study.

Table 6.2 The contribution of the various taxa to the similarity within groups (determined by using SIMPER). The groups were determined using Bray-Curtis cluster analysis and NMDS.

Table 6.3 The invertebrate scoring system (SASS 5) results and the method applied in assigning ecological classes in this study.

Table 6.4 Result for SASS 5 (SASS score, number of taxa found and the average score per taxa) as recorded during high flow conditions. IHAS scores are calculated with 100 as a maximum score.

Table 6.5 Result for SASS 5 (SASS score, number of taxa found and the average score per taxa) as recorded during low flow conditions.

Table 6.6 Abundances of invertebrate families collected during high flow conditions (S – Stones; VG – vegetation). Families indicated in grey represent indicator families.

Table 6.7 Abundances of invertebrate families collected during low flow conditions (S – Stones; VG – vegetation). Families indicated in grey represent indicator families.

Table 7.1 MIRAI results (EC and ecological class) for the site on the Elands and Crocodile Rivers for both flow regimes.

Table 7.2 FRAI results (FRAI % and ecological class) for the site on the Elands and Crocodile Rivers for both flow regimes.

- 9 -

LIST OF FIGURES

Figure 2.1 Elands River Comprehensive Reserve Determination Study Surface Water Resource Units. The main study area falls within resource unit 2.

Figure 2.2 Besides the Sappi Ngodwana mill there are various other impacts on the Elands River, often occurring on the tributaries associated with the Elands River.

Figure 2.3 Photograph of ER 1 (Hemlock) the characteristic cobble beds (A) and interspersing pools (B).

Figure 2.4 Photograph of ER 2 (Ryton Estates) and the site too shows the typical cobble beds and pools.

Figure 2.5 The photograph shows site ER 3 (Bambi bridge) down stream of the Ngodwana mill and springs.

Figure 2.6 Photograph of site ER 4, before the confluence with the Lupelule River (A) upstream and (B) downstream.

Figure 2.7 Photographs of site ER 5 at Lindenau, above the Lindenau weir. Photograph (A) facing downstream in the direction of the weir and (B) upstream.

Figure 2.8 Photographs of site HR 1, positioned on the Lupelule River before the confluence with the Elands River, showing (A) upstream and (B) downstream areas.

Figure 2.9 Photographs of site NR 1, above the Ngodwana Dam.

Figure 2.10 Photograph of site NR 2 below the Ngodwana Dam wall.

Figure 2.11 Photographs of CR 1, the upper Crocodile River site, situated above the confluence with the Elands River, facing (A) downstream and (B) upstream.

- 10 -

Figure 2.12 Photographs of CR 2, the second Crocodile Site, situated after the confluence with the Elands River.

Figure 2.13 Positioning of the sites that has been selected on the Elands and Crocodile Rivers and various tributaries for this study.

Figure 3.1 The DWAF gauging stations at Lindenau (X2H015Q01) on the Elands River.

Figure 3.2 Statistical results of historical water quality data available from the Geluk gauging station on the Elands River.

Figure 3.3 Statistical results of historical water quality data available from the Lindenau gauging station on the Elands River.

Figure 3.4 Statistical results of historical water quality data available from the Montrose gauging station on the Crocodile River.

Figure 3.5 PCA plot showing the dissimilarity among sites on the Elands and Crocodile rivers during both high (sites without suffix) and low flow (sites with the suffix – L) regimes based on water quality. This bi-plot describes 95.9 % of the variation in the data, where 69.7 % is displayed on the first axis, while 35.5 % is displayed on the second axis.

Figure 3.6 PCA plot showing the dissimilarity among sites on the Elands and Crocodile rivers during high flow conditions based on water quality. This bi-plot describes 95 % of the variation in the data, where 68.9 % is displayed on the first axis, while 26.1 % is displayed on the second axis.

Figure 3.7 PCA plot showing the dissimilarity among sites on the Elands and Crocodile rivers during low flow conditions based on water quality. This bi-plot describes 98.4 % of the variation in the data, where 75.4 % is displayed on the first axis, while 23 % is displayed on the second axis.

- 11 -

Figure 4.1 PCA plot showing the dissimilarity among sites on the Elands and Crocodile rivers during both high (sites without suffix) and low flow (sites with the suffix – L) regimes based on sediment characteristics. This bi-plot describes 77.3 % of the variation in the data, where 53.1 % is displayed on the first axis, while 24.2 % is displayed on the second axis.

Figure 4.2 PCA plot showing the dissimilarity among sites on the Elands and Crocodile rivers during high flow conditions based on sediment characteristics. This bi-plot describes 81.5 % of the variation in the data, where 36.1 % is displayed on the first axis, while 49.9 % is displayed on the second axis.

Figure 4.3 PCA plot showing the dissimilarity among sites on the Elands and Crocodile rivers during low flow conditions based on sediment characteristics. This bi-plot describes 95.8 % of the variation in the data, where 77.8 % is displayed on the first axis, while 18 % is displayed on the second axis.

Figure 4.4 Graphical descriptions of the habitat assessment indices including (A) IHAS and (B) HQI completed for the sites on the Elands and Crocodile rivers with blue bars representing high flow and red bars representing low flow conditions (Polynomial trend lines superimposed with solid line representing high flow and broken line representing low flow conditions).

Figure 4.5 Bray-Curtis similarity matrix-based cluster analysis (A) and two dimensional representation of the NMDS ordination (B) of the riparian vegetation at the sites on the Elands and Crocodile Rivers during low flow conditions. The NMDS ordination was completed with 30 iterations and showed a stress of 0.07.

Figure 5.1 Univariate diversity indices indicating the number of species (A) and Margalef’s species richness (B). Red bars indicating high flow while blue bars indicate low flow. A polynomial trend line as been overlain on the graph (broken line indicates high flow and solid line low flow.

Figure 5.2 Univariate diversity indices indicating Pielou’s evenness index (A) and Shannon- Wiener diversity index (B). Red bars indicating high flow while blue bars indicate low flow. A polynomial trend line as been overlain on the graph (broken line indicates high flow and solid line low flow.

- 12 -

Figure 5.3 Bray-Curtis similarity matrix-based cluster analysis (A) and two dimensional representation of the NMDS ordination (B) of the ichthyofauna collected at the sites on the Elands and Crocodile Rivers during both high (sites without suffix) and low flow (sites with the suffix – L) regimes. The NMDS ordination was completed with 30 iterations and showed a stress of 0.13.

Figure 5.4 Bray-Curtis similarity matrix-based cluster analysis (A) and two dimensional representation of the NMDS ordination (B) of the ichthyofauna collected at the sites on the Elands and Crocodile Rivers during high flow conditions. The NMDS ordination was completed with 30 iterations and showed a stress of 0.08.

Figure 5.5 Bray-Curtis similarity matrix-based cluster analysis (A) and two dimensional representation of the NMDS ordination (B) of the ichthyofauna collected at the sites on the Elands and Crocodile Rivers during low flow conditions The NMDS ordination was completed with 30 iterations and showed a stress of 0.04.

Figure 5.6 Ranked species K-dominance curves for the ichthyofauna communities collected at the sites on the Elands and Crocodile Rivers during high flow (A) and low flow conditions (B).

Figure 5.7 RDA plot showing the dissimilarity among sites on the Elands and Crocodile Rivers during both high (sites without suffix) and low flow (sites with the suffix – L) regimes based on ichthyofauna communities with variables superimposed. This bi-plot describes 60.5 % of the variation in the data, where 33.2 % is displayed on the first axis, while 27.3 % is displayed on the second axis.

Figure 5.8 RDA plot showing the dissimilarity among sites on the Elands and Crocodile Rivers regimes based on invertebrate communities with variables superimposed and the effect of flow regime removed. This bi-plot describes 59.5 % of the variation in the data, where 34.2 % is displayed on the first axis, while 25.3 % is displayed on the second axis.

Figure 5.9 RDA plot showing the dissimilarity among sites on the Elands River during high flow based on invertebrate communities with environmental superimposed. This bi-plot describes 97.8 % of the variation in the data, where 69.9 % is displayed on the first axis, while 27.8 % is displayed on the second axis.

- 13 -

Figure 5.10 RDA plot showing the dissimilarity among sites on the Elands River during low flow based on invertebrate communities with environmental superimposed. This bi-plot describes 86.6 % of the variation in the data, where 64.1 % is displayed on the first axis, while 22.5 % is displayed on the second axis.

Figure 5.11 Graphical description of FAII scores obtained for the sites on the Elands and Crocodile rivers. Red bars indicate high flow conditions with blue bars representing low flow conditions. A polynomial trend line was overlain onto the graph (broken line represent high flow and solid line represent low flow conditions).

Figure 6.1 Univariate diversity indices indicating the number of species (A) and Margalef’s species richness (B). Red bars indicating high flow while blue bars indicate low flow. A polynomial trend line as been overlain on the graph (broken line indicates high flow and solid line low flow.

Figure 6.2 Univariate diversity indices indicating Pielou’s evenness index (A) and Shannon- Wiener diversity index (B). Red bars indicating high flow while blue bars indicate low flow. A polynomial trend line as been overlain on the graph (broken line indicates high flow and solid line low flow.

Figure 6.3 Bray-Curtis similarity matrix-based cluster analysis (A) and two dimensional representation of the NMDS ordination (B) of the aquatic macro invertebrates collected at the sites on the Elands and Crocodile Rivers during both high (sites without suffix) and low flow (sites with the suffix – L) regimes. The NMDS ordination was completed with 30 iterations and showed a stress of 0.09.

Figure 6.4 Bray-Curtis similarity matrix-based cluster analysis (A) and two dimensional representation of the NMDS ordination (B) of the aquatic macro invertebrates collected at the sites on the Elands and Crocodile Rivers during high flow conditions. The NMDS ordination was completed with 30 iterations and showed a stress of 0.01.

- 14 -

Figure 6.5 Bray-Curtis similarity matrix-based cluster analysis (A) and two dimensional representation of the NMDS ordination (B) of the aquatic macro invertebrates collected at the sites on the Elands and Crocodile Rivers during low conditions. The NMDS ordination was completed with 30 iterations and showed a stress of 0.03.

Figure 6.6 Ranked species K-dominance curves for the invertebrate communities collected at the sites on the Elands and Crocodile Rivers during high flow (A) and low flow conditions (B).

Figure 6.7 RDA plot showing the dissimilarity among sites on the Elands and Crocodile Rivers during both high (sites without suffix) and low flow (sites with the suffix – L) regimes based on invertebrate communities with variables superimposed. This bi-plot describes 57.1 % of the variation in the data, where 39.2 % is displayed on the first axis, while 17.9 % is displayed on the second axis.

Figure 6.8 RDA plot showing the dissimilarity among sites on the Elands and Crocodile Rivers regimes based on invertebrate communities with variables superimposed and the effect of flow regime removed. This bi-plot describes 49.2 % of the variation in the data, where 26.2 % is displayed on the first axis, while 23 % is displayed on the second axis.

Figure 6.9 RDA plot showing the dissimilarity among sites on the Elands River during high flow based on invertebrate communities with variables superimposed. This bi-plot describes 75.9 % of the variation in the data, where 55.3 % is displayed on the first axis, while 20.6 % is displayed on the second axis.

Figure 6.10 RDA plot showing the dissimilarity among sites on the Elands River during low flow based on invertebrate communities with variables superimposed. This bi-plot describes 79.9 % of the variation in the data, where 42.9 % is displayed on the first axis, while 37 % is displayed on the second axis.

Figure 6.11 Graphical description of SASS 5 scores obtained for the sites on the Elands and Crocodile rivers. Red bars indicate high flow conditions with blue bars representing low flow conditions. A polynomial trend line was overlain onto the graph (broken line represent high flow and solid line represent low flow conditions).

- 15 -

Figure 6.12 Graphical description of the number of taxa recorded at the sites on the Elands and Crocodile rivers. Red bars indicate high flow conditions with blue bars representing low flow conditions. A polynomial trend line was overlain onto the graph (broken line represent high flow and solid line represent low flow conditions).

Figure 6.13 Graphical description of the ASPT recorded at the sites on the Elands and Crocodile rivers. Red bars indicate high flow conditions with blue bars representing low flow conditions. A polynomial trend line was overlain onto the graph (broken line represent high flow and solid line represent low flow conditions).

- 16 -

SUMMARY

Water is one of our key and indispensable natural resources. It plays a fundamental part in life (and the quality thereof), the environment, food production, hygiene, industry and power generation. Water is one of the major limiting factors in South Africa when it comes to economical growth and social development. In our country water is a scarce resource which is unevenly distributed both geographically and through time. As the demand for water increases, with increasing human populations and economic development, so to does the pollution of our river ecosystems.The Elands River is one of these natural resources that is under constant threat. It falls within the Incomati Water Management Area and is further sub divided into the Crocodile River sub area. This sub area is highly stressed, as it provides water for several human activities. The Elands River is a major tributary of the Crocodile River. The Crocodile River is a source of fresh water for several towns and is used by industry, rural and the agricultural communities (including tobacco farms). The Elands River in turn, is used for irrigation of vegetables. Both these rivers support a rich diversity of aquatic life. Along with its social and economical importance, the Elands River has immense ecological importance, as it holds great biodiversity including critically endangered biota.

The main anthropogenic impacts on the Elands and Crocodile rivers include: • The Sappi Ngodwana Mill and the associated pulp and paper activities • The influence of the Ngodwana dam wall on the flow and water quality within the lower Ngodwana River • Nutrient loading taking place due to the treated sewage that is released into the river in the upper reaches and in the vicinity of the Mill • Sedimentation and flow regulation that is taking place in the Crocodile River, upstream of the confluence with the Elands River • And the agricultural activities within the Elands River system.

The activities related to the Mill are the major concern in the study. The Sappi Ngodwana Mill is situated at the confluence of the Elands and Ngodwana rivers. The mill does not discharge effluent directly into the river. The effluent is however, irrigated onto the 514 hectares of farmlands adjacent to the Mill. The irrigated effluent has contaminated the groundwater in this area and the primary influence of this groundwater contamination is the deterioration of the

- 17 -

surface water quality as well as, negatively impacts the quantity of water in the Elands River. The groundwater enters the Elands River through three springs near Ngodwana namely Fraser’s eye, Northern eye and Eye X. The groundwater from both Fraser’s eye and Eye X has been contaminated with calcium, potassium, magnesium, sulphates and most importantly chlorides. All these substances contribute to the increase in conductivity in the Elands River, which in turn may have a possible impact on the ecological integrity of the system. The pulp and paper industry is a large consumer of water and few regrettable incidents over the years have given the industry a reputation as a major water polluter. The industry’s management of water is, however, of world class and every attempted is made to manage the environment in a sustainable manner. This study aims to assess the impact of these anthropogenic activities on the associated aquatic ecosystems.

Assessing the impact of anthropogenic activities on the aquatic environments, like the Elands River, has in the past been based mainly on the assessment of water quality. Earlier management of water resources has thus been based on the potability of water. Over the last decade management initiatives have expanded to include domestic, agricultural, recreational and most importantly instream (fish, invertebrates etc.) users. It has become common practise to use aquatic biota to assess the impacts of human activities of freshwater resources. The reason for this is that animals and plants can provide a long–term integrated reflection of water quality, quantity, habitat quality and other environmental conditions. Water and sediment quality was assessed by applying standard techniques and protocols. Additionally historical water quality data was obtained from the Sappi Ngodwana Mill and the Department of Water Affairs and Forestry. Habitat quality was assessed by implementation of habitat quality indices. This included that Integrated Habitat Assessment Index and the Habitat Quality Index. The vegetation at each site was identified in the field with the assistance provided by the members of the Elands River Valley Conservancy and using various field guides and the riparian zone was then demarked. The integrity of the fish community was assessed by implementing the Fish Assemblage Integrity Index and the Fish Response Assessment Index. The integrity of the aquatic macro invertebrate communities was also assessed. This was achieved through use of the South African Scoring System and the Macro Invertebrate Response Assessment Index. Finally, spatial and/or temporal trends were assessed by implementation of various multi variate statistical procedures.

- 18 -

The water quality of the Elands and Crocodile rivers appears to be in a good state. There is some concern, however, over the increase in nutrients and salts over the last decade. An increase in nutrients is key indicators of pollution related to agricultural activities or sewage. The increase in chlorides, sulphates and conductivity is associated with the contaminated groundwater reaching the Elands River through the three springs. Effluent from the pulp and paper mill industry has been known to cause such an increase in these variables. To ensure the effective management of our resources one can not focus on water quality alone. Sediment quality and habitat diversity and availability are also major abiotic determinants in ecological integrity of a river. The availability of sediment is restricted within the Elands River. This is a good sign the sediments are constantly being transported in the river or that the activities are not causing sedimentation. The instream habitat in the Elands and Crocodile rivers appear to be in a largely natural state. There are, however, some habitat modifications within the study area. These modifications are largely in the form of loss of marginal vegetation as habitat and flow modifications caused by the construction of dams.

The structure and taxonomic composition of biological communities integrate both the physical and chemical aspects of the environment. Water, sediment and habitat quality can be seen as environmental variables that affect the integrity of the biological communities as the abiotic and biotic components of a river ecosystem are ultimately linked. The biotic components assessed in the study include the riparian vegetation, fish and macroinvertebrate communities. Riparian vegetation forms a vital part of any river ecosystem. It has been well documented that riparian vegetation plays a number of important geomorphological, ecological and social roles which may have an influence on the condition and sustainability of the riverine ecosystem. The riparian vegetation within the study area appears to be in a modified state, largely due to encroachment of exotic and terrestrial species. This occurrence appears to be related to the agricultural activities within the area. Fish communities and individuals themselves pose various qualities that make them useful in biological monitoring. Fish in general have a long life span and would thus be able to reflect changes in the integrity over long periods of time. Results indicate that the fish community within the Elands and Crocodile rivers appear to be in a largely natural state. A change in community characteristics may have taken place at the sites downstream of the Ngodwana mill but species richness and presence of sensitive species indicate little modification. One of the biggest concerns regarding the fish communities sampled during the study was the low numbers of Chiloglanis bifurcus present in the community. This highly endangered species occurs in the Elands River and has a high preference for clean, fast

- 19 -

flowing water through all of its live stages. Resident aquatic macroinvertebrates are good, short- term indicators of ecological integrity because they integrate the effects of physical and chemical changes. They are adapted to live within certain environmental conditions and changes within this environment may adversely affect community composition and abundance. It has become clear that the invertebrate communities within the Elands and Crocodile rivers are generally in a good state with minimum modification. There is a loss in diversity and sensitive families at some study sites that may attributed to various anthropogenic impacts including the activities of the Ngodwana Mill and sewage works.

It is evident that the Elands and Crocodile River, along with their tributaries, are of great importance to the environment and the public alike. These systems support a rich variety of terrestrial and aquatic fauna and are a source of fresh water for several towns, cities and important industries. The study has shown that the anthropogenic activities in the study area do not have a major impact on the ecological integrity of the Crocodile River and the Elands River in particular. From an ecological perspective it is of the utmost importance that the water, habitat and sediment quality within the Elands River be maintained. This will ensure that invertebrate and fish communities alike are not negatively impacted and that no species are lost.

- 20 -

OPSOMMING

Water is een van ons belangrikste en mees onmisbare natuurlike hulpbronne. Dit speel ‘n fundamentele rol in lewe (en die kwaliteit daarvan), die omgewing, voedselproduksie, hiegiëne, industrië en kragopwekking. In Suid-Afrika is water een van die vernaamste faktore wat ekonomiesegroei en sosiale ontwikkeling beperk. Water word as ‘n skaars hulpbron, wat beide geografies en ruimtelik oneveredig versprei is, beskou. Soos die vraag na water met bevlokingsaanwas en ekonomies groei toeneem, neem die besoedeling van ons riviersisteme ook toe. Die Elandsrivier is een van hierdie natuurlike hulpbronne wat gedurig onder bedreiging verkeer. Die Elandsrivier val binne die Incomati Waterbestuursarea en word verder onderverdeel in die Krokodilrivier sub-area. Die sub-area is onder hoë druk aangesien dit water aan verskeie aktiwiteite verskaf. Die Elandsrivier is ‘n belangrike sytak van die Krokodilrivier. Die Krokodilrivier is ‘n bron van varswater vir verskeie dorpe en word deur verskeie industrië, informelev nedersettings en landbousektore (insluitend tabakplase) gebruik. Water vanuit die Elandsrivier word meestal vir besproeiїng van gewasse benut. Beide die riviere ondersteun ‘n diverse verskeidenheid van akwatiese lewe. Saam met die sosiale en ekonomiese belang, is die Elandsrivier van onskatbare ekologiese belang vanweë die unieke biodiversiteit wat ‘n aantal krities bedreigde spesies insluit.

Die vernaamste antropogeniese impakte op die Elands- en Krokodilriviere sluit in: die Sappi Ngodwana papiermeule en die geassosieërde pulp en papier aktiwiteite; die invloed van die Ngodwana dam op die vloei en water kwaliteit in die laer streke van die Ngodwanarivier; die toename in nutriënt konsentrasies a.g.v die verwerkte riool wat naby die meule vrygestel word; sedimentasie en vloei regulasie wat in die Krokodilrivier, voor die samevloei met die Elandsrivier plaasvind en die landbou aktiwiteit in die nabyheid van die Elandsrivier. Die kommer rakende die invloed van die meule se aktiwiteite op die akwatiese ekosisteem van die Elandsrivier het tot die uitvoer van die huidige studie gelei. Die Sappi Ngodwana meule is geleë naby die samevloei van die Elands- en Ngodwanariviere. Die meule stort nie uitvloeisel direk in die rivier nie, maar dit word op nabygeleë wyvelde besproei. Die besproeide uitvloeisel dring egter die grondwater van die omliggende opvangebied binne. Die primêre effek van díe besoedelde grondwater is die verlaging in die oppervlakwaterkwaliteit sowel as ‘n verandering in die watervlak (d.i. waterkwantiteit) in die Elandsrivier. Die grondwater bereik die sisteem d.m.v drie bronne (fonteine of te wel “eyes”) wat naby Ngodwana geleë is naamlik: Fraser’s

- 21 -

eye, Northern eye en Eye X. Die grondwater van beide Fraser’s eye en Eye X is gekontamineer met kalsium, kalium, magnesium, sulfate en van die meeste belang, chloriede. Al hierdie stowwe dra by tot die toename in konduktiwiteit in die Elandsrivier, wat die ekologiese integriteit van die sisteem beïnvloed. Die pulp en papier industrie is die grootste waterverbruiker in die omgewing en as gevlog van ‘n paar omgewingsinsidente het die industrie ‘n reputasie as ‘n groot waterbesoedelaar ontwikkel. Sappi se waterbestuurspraktyke is van wêreldgehalte en elke moontlike poging word aangewend om die omgewing op ‘n volhoubare manier te bestuur.

Assessering van die impak wat antropogeniese aktiwiteite op die akwatiese omgewing, soos die Elandsrivier, was in die verlede op waterkwaliteit alleen gebasseer. Vroeë bestuur van waterbronne was dus op die drinkbaarheid van water gebasseer. Gedurende die laaste dekade het bestuursinisiatiewe uitgebrei om nou die huishoudelike, landbou, onspannings en die akwatiese (vis, invertebrate, ens.) gebruikers in te sluit. Dit het algemene praktyk geword om akwatiesebiota vir die assessering van die impak van antropogeniese aktiwiteite op varswater bronne te gebruik. Die rede hiervoor is dat diere en plante ‘n geїntegreerde, lang-termyn aanduiding van waterkwaliteit en waterkwantiteit, habitatkwaliteit en ander omgewings toestande verskaf. Waterkwaliteit en die eienskappe van sediment is in die huidige studie ondersoek deur standaard tegnieke en metodes toe te pas. Historiese waterkwaliteitsinligting is vanaf die Sappi Ngodwana meule en die Departement Waterwese en Bosbou verkry. Habitatkwaliteit was bepaal deur middel van habitatskwaliteit indekse bepaal. Dit het die “Integrated Habitat Assessment Index” sowel as die “Habitat Quality Index” ingesluit. Die plantegroei by elke lokaliteit is m.b.v taksonomiese sleutels en deur kundiges van die “Elands River Valley Conservancy” geїdentifiseer. Die oewerplantegroeisone was sodoende afgebaken. Die visgemeenskap integriteit was m.b.v. die “Fish Assemblage Integrity Index” en die “Fish Response Assessment Index” bepaal. Die integriteit van die akwatiese makro-invertebraat gemeenskap was m.b.v. die “South African Scoring System” en die “Macro Invertebrate Response Assessment Index” bepaal. Die tyd / ruimtelike tendense in die gemeenskapstruktuur is d.m.v. ’n kombinasie van enkel- en multivariante statistiese metodes ondersoek.

Die waterkwaliteit van die Elands- en Krokodilriviere blyk in ‘n goeie toestand te wees. Daar is egter probleme t.o.v. ‘n toename in nutriënt en soutkonsentrasies gedurende die laaste dekade ondervind. Die toename in nutriënte is gewoonlik ‘n goeie aanduiding van organiese (riool) en

- 22 -

landbou (kunsmus) besoedeling terwyl die toename in chloriede, sulfate en konduktiwiteit met die besoedelde grondwater afkomstig vanaf die fonteine geassosieër word. Uitvloeisel van die pulp en papier industrie is juis bekend daarvoor om ‘n toename in hierdie veranderlikes teweeg te bring. Om die effektiewe bestuur van ons varswaterbronne te bewerkstellig kan mens egter nie alleen op waterkwaliteit fokus nie. Sedimentkwaliteit, habitatsdiversiteit en beskikbaarheid daarvan is ook belangrike abiotiese komponente van ekologiese integriteit van ‘n rivier. Die beskikbaarheid van sediment is egter in die Elandsrivier beperk. Die afwesigheid van hoë persentasies sediment is ‘n goeie aanduiding dat sediment konstant vervoer word in die rivier, of dat die aktiwiteite in die gebied nie tot sedimentasie bydra nie. Die habitat in die Elands- en Krokodilriviere blyk om hoofsaaklike in ‘n natuurlike toestand te wees. Daar is egter afwykings in die studie area waargeneem. Hierdie afwykings sluit onder andere die verlies van rivieroewer-plantegroei as habiat, sowel as die vloeiveranderinge meegebring deur die konstruksie van sommige damme, in. Die doel van die huidige studie is om die moontlike impakte, wat hierdie antropogeniese aktiwiteit op die geassosiërde akwatiese ekosisteem het, te assesseer.

Die struktuur en taksonomiese samestelling van die biologiese gemeenskappe integreer beide die fisiese en chemise aspekte van die omgewing. Water, sediment en habitatkwaliteit kan gesien word as omgewings veranderlikes wat die integriteit van die biologiese gemeenskappe beïnvloed, aangesien hierdie abiotiese en biotiese komponente almal met mekaar verbind is. Die biotiese komponente wat tydens die studie ondersoek is, sluit die oewerplantegroei-, invertebraat- en visgemeenskappe in. Oewerplantegroei vorm ‘n belangrike deel van enige rivierekosisteem. Dit word algemeen aanvaar dat hierdie plantegroei talle geomorfologiese, ekologiese en sosiale rolle vervul wat ‘n effek het op die toestand en volhoubaarheid van ‘n rivierekosisteem het. Daar is vasgestel dat die oewerplantegroei in die studie area in ‘n gewysigde toestand is. Dit word hoofsaaklik aan die voorkoms van eksotiese- en terrestriële indringerspesies toegeskryf. Hierdie verskynsel is klaarbleiklik a.g.v verskei landbou aktiwiteite wat in die studie area voorkom. Visgemeenskappe en individuele spesies het verskeie eienskappe wat hulle ideale biomoniteringsorganismes maak. Vis in die algemeen is lank lewend en gee dus ‘n aanduiding van die integriteit van die sisteem oor ‘n lang tydperk. Resultate dui daarop dat die visgemeenskap in die Elands- en Krokodilriviere meestal in ‘n natuurlike toestand blyk te wees. ‘n Verandering in die gemeenskapseienskappe het by die lokaliteite stroom-af van die Ngodwana meule plaasgevind, maar die feit dat daar ’n hoë

- 23 -

spesiesrykdom en sensitiewe spesies teenwoordig is, dui op min verandering. Een van die grootste probleme rakend die visgemeenskappe in die studie area is die lae getalle Chiloglanis bifurcus wat aangetref is. Hierdie hoogs bedreigde spesie kom voor in die Elandsrivier en benodig helder, goeie kwaliteit water gedurende al die lewenstadia. Akwatiese makro- invertebrate, is op hul beurt, goeie kort-termyn aanduiers van ekologiese integriteit aangesien hulle die effek van fisiese en chemise verandering oor tyd integreer. Hulle is aangepas om in sekere omgewings toestande te oorleef en verandering in hierdie toestande kan die gemeenskap samestelling en rykdom beïnvloed. Tydens die studie het dit duidelik geword dat die akwatiese invertebraat gemeenskap in die studie area oor die algemeen in ‘n goeie toestand is met minimale wysigings. Daar is wel ‘n verlies in diversiteit en sensitiewe families by sommige lokaliteite. Hierdie veranderinge kan aan die aktiwiteite van die pulp en papier meule, asook die naby geleë riool werke gekoppel word.

Dit was opvallend dat die Elands- en Krokodilriviere, saam met hul sytakke, groot publieke belangstelling uitlok. Hierdie stelsels onderhou a verskeidenheid terrestriële en akwatiese fauna en is ‘n bron van varswater vir verskeie dorpe, stede en belangrike industrië. Die huidige studie het getoon dat die antropogeniese aktiwiteite in die studie area nie ‘n grootskaalse negatiewe impak op die ekologiese integriteit van die twee sisteme het nie. Vanuit ‘n ekologiese perspektief is dit van uiterste belang dat die water, habitat en sediment kwaliteit behoue bly. Dit sal verseker dat beide vis- en invertebraatgemeenskappe nie negatief beïnvloed word en dat geen spesies verlore gaan nie.

- 24 -

CHAPTER 1

GENERAL INTRODUCTION

1.1 INTRODUCTION

The Elands River is one of these natural resources that are under constant threat. It falls within the Incomati Water Management Area and is further sub divided into the Crocodile River sub area (DWAF, 2004a). This sub area is highly stressed, as it provides water for several human activities. The Elands River is a major tributary of the Crocodile River, which in turn is a source of fresh water for several towns and is used by industry, rural and agricultural communities (including tobacco farms). Both these rivers support a rich diversity of aquatic life (Weddepohl et al., 1991). Along with its social and economical importance, the Elands River has immense ecological importance, as it holds great biodiversity including critically endangered biota. Chloride contamination in the Elands River has been identified in a study completed by O’Brien (2003). This has been linked to the activities of the Ngodwana Pulp and Paper Mill. Some of the recommendations made in the study were to further investigate the effect of the not only chloride contamination, but the eutrophication taking place due to sewage works in the study area. Because of the ecological, social and economical importance of the Elands River, along with the possible impacts that the Elands River has on the Crocodile River this study was necessary to determine the effects of anthropogenic impacts on the Elands and Crocodile rivers Extensive water allocation studies have been undertaken for the Elands River and this includes the intermediate and comprehensive reserves which are available for the Elands River (Godfey and Roux, 2000; Godfrey et al.,2000; Hill, 2005). - 25 - Chapter 1

The Sappi Ngodwana Mill is situated on the banks of the Elands River, near the confluence with the Ngodwana Mill. The pulp and paper industry has in the past had a significant impact on the environment in three major areas: waste water, air pollution from the combustion of fuels and from chemical pulping processes and solid waste disposal (DWAF, 2004b). The production of waste water and the chemical pulping process is of main concern in this study. In chemical pulp mills using the kraft process, most of the water can be recovered by evaporation and combustion of the solid material. If bleaching with chlorine or chlorine dioxide occurs, water from the bleaching process cannot be recovered and is usually discharged. The reason for this is that chlorine chemicals interfere with the combustion process. The discharge of this water is usually of great concern from an environmental point of view and the salt content of these pulp mill effluents presents the biggest disposal challenge. Although organic materials can be treated in conventional biological processes, the salts remain in the effluent. The effluent from mills in the interior regions of South Africa can not be diluted sufficiently due to scarce water resources and new procedures are always under development to deal with this problem. A few regrettable incidents over the years have given the industry a reputation as a major water polluter. One such an example is an accidental effluent spill that caused major deaths among fish and invertebrates in the Elands River (Kleynhans 1992; Weddepohl et al., 1991). In recent times it has become apparent that the industry’s management of water is of a world class standard and therefore better than in most other countries (DWAF, 2004b).

1.1.1 RIVER HEALTH PROGRAM

In the past South African citizens where denied equal access to water resources. Previous acts made no allowance for sustainability or equitability. It upheld the private ownership of water which was closely linked to riparian rights. However, changes were brought about, fueled by our countries first democratic elections in 1994. The adoption of the Constitution of the Republic of South Africa (Act no. 108 of 1996) laid the foundation for most of these changes, especially chapter 2: The Bill of Rights and more specifically section 24 (a) and (b) and 27 (b). The Bill of Rights set the stage for the development of our National Water Act (NWA). In this act the government sees itself as the trustee of South African waters and commits itself to do the following: • Guarantee access to water for basic human needs. • To make sure that environmental requirements are met. • To take into account the interconnected nature of water resources. 26 Chapter 1

• To make provision for the transfer of water between catchments. • To respect South Africa’s obligation towards neighboring countries. • And to fulfill its commitment as the custodian of our nations’ water (DWAF, 2003).

To reach these goals and to fulfill these promises several programs and procedures were developed. These programs help to protect our resources and improve management practices. One of the most important programs is the National River Health Program (RHP). This program was initiated by DWAF in 1994. The purpose for the implementation of this program was to gather information concerning the ecological state of our rivers. It is based on the concept that aquatic communities (like fish, invertebrates and riparian vegetation) combine and reflect the effect of anthropogenic disturbances that occur in the river over an extended period of time (Roux, 2001). The rational is that the integrity of the aquatic biota provides a direct, holistic and integrated approach to measure the integrity of a river as a whole (Roux, 1999). It is however important to note that a distinction can be made between integrity and health of an ecosystem. Integrity refers to how close an ecosystem is to natural conditions, while health represent a condition that is desired by humans using the resource. The RHP was the first program to contribute to setting and monitoring resource objectives. The main objectives of the RHP are to:

• Measure, assess and report on the ecological state of aquatic ecosystems. • Detect and report spatial and temporal trends in ecological state of aquatic ecosystems identify and report on emerging problems. • Ensure that reports provide scientific and managerial relevant information (Roux, 1999).

For the RHP to be truly successful as a management and information tool, a step – wise procedure must be in place for linking data and information with management actions (Roux, 2001). During this study, basic procedures as implemented in the RHP were followed. The most important of these was the site selection procedure. The first step in assessing impacts in the Elands and Crocodile Rivers was the selection of reference conditions in the form of a reference site. The reference site represents the natural, unimpacted characteristics of the resource that is being studied. All sites forming part of this study was compared to these reference conditions to determine whether any impacts are occurring.

27 Chapter 1

1.1.2 THE RESERVE

It is clear that the RHP has been developed to determine the ecological state of our rivers. Although this is important it is not enough. One not only needs to determine the state of our rivers, but also manage and protect this resource to ensure sustainability. With this in mind the South African government has developed several measures to ensure the protection of our water resources. These include source and resource directed measures. Resource directed measures focuses on the water resource as an ecosystem and specify objectives to reflect the required level of protection of the resource (Pegram and Görgens, 2001). The resource directed measures (RDM) where designed to be applied to water resources as a system i.e. a catchment. According to DWAF (2003) these RDM are:

A. The establishment of the reserve: The reserve refers to the quantity and the quality of water necessary to satisfy our basic human needs, and to protect aquatic ecosystems, in order to ensure ecological sustainability and use of relevant water resources. The aim of setting the reserve is thus to prevent the exploitation of resources and to ensure the sustainable use of aquatic ecosystems.

B. The classification of water resources: The classification of water resources refers to the determination of the present ecological state of water resources. Resource protection is given effect through this classification, as this procedure leads to the specification of an acceptable degree of change from natural conditions for a particular natural resource. Classification of resources is helpful in the setting of management scenarios.

C. Setting of resource quality objectives (RQOs): These are numerical and narrative descriptors of conditions that need to be met in order to achieve the required management scenarios. The main purpose of the RQOs is to establish clear goals. It becomes apparent that the RHP will only provide information on item 2 (B) mentioned above and forms only part of a greater management plan.

28 Chapter 1 link habitat, water and the aquatic biota. This definition stresses the fact that water resources are linked and with other features and process of nature. It is thus important to manage them in such a way that takes into account their dynamic character. RDM as a strategy reflects awareness of this problem.

1.1.3 ECOCLASSIFICATION AND ECOSTATUS

A new procedure has been under development to further support management and protection of rivers. This procedure is known as EcoClassification (the term used for ecological classification). It refers to the determination and categorisation of the present ecological state of rivers compared to their natural state. Ecological evaluation in terms of expected reference conditions, followed with the integration with an assigned category, represent the ecological status (Kleynhans and Louw, 2005). EcoClassification allows for gathering of useful information in deriving future ecological objectives for the river. EcoClassification should not be confused with the classification system as indicated in the National Water Act.

It may however, have a role in the reserve determination process and in monitoring. The EcoClassification process plays an important part in the reserve due to the fact that flow (water quantity) cannot be recommended without prior knowledge of the present ecological state. The ecological categories determined in the EcoClassification form an important part of the reserve process. Monitoring as used in the RHP further makes use of EcoClassification to assess the severity of change from reference conditions.

The main steps followed during EcoClassification include: • Determination of reference conditions. • Determination of the present ecological state as well as for the EcoStatus. • Determine trends for each component and the EcoStatus. • Determine reasons for the present ecological state and determine whether it is flow related. • Determine ecological importance and sensitivity of the resource. • Consider the above and the present ecological state and suggest recommended ecological categories.

29 Chapter 1

• Determine alternative ecological categories for each component and for EcoStatus (Kleynhans and Louw, 2005).

The two main components assessed in the new approach are the abiotic components (drivers) and the biotic components (responders). The basis of all the indices used with the EcoClassification procedure is the determination of how various factors influence biotic communities and abiotic components. These factors are generally known as metrics. Metrics are systems of parameters or ways of quantitative assessment of a process that is to be measured, along with the processes to carry out such measurement. Metrics define what is to be measured. They are usually specialised by the subject area, in which case they are valid only within a certain domain and cannot be directly benchmarked or interpreted outside it (Kleynhans and Louw, 2005). For the determination of EcoStatus and completion of EcoClassification a ranking and weighting approach has been adopted. The principle of following a ranking-weighting approach is that not all driver or biological response metrics have the same relative ecological significance in all types of rivers. That is, a particular metric may be seriously modified but it may be of relatively low significance in terms of the functioning and integrity of the river. In another river (or a different section of the same river) this metric may, however, be of very high ecological importance.

1.2 AIMS AND OBJECTIOBJECTIVESVES

The aim of this study was to use methods and procedure provided by the RHP and RDM along with multivariate statistical procedures to determine the impact of the anthropogenic activities in the study area on the ecological integrity of the Elands and Crocodile rivers, Mpumalanga. To reach this aim the following objectives needs to be met:

The assessment of the spatial and temporal changes in water quality. The assessment of the spatial and temporal sediment quality. The assessment of the current habitat quality at selected sites on these rivers. The assessment of the riparian vegetation community structure and integrity at selected sites on these rivers. The assessment of the ichthyofaunal community structure and integrity at selected sites on these rivers.

30 Chapter 1

The assessment of the aquatic macro-invertebrate community structure and integrity at selected sites on these rivers.

1.3 REFERENCES

DWAF (Department of Water Affairs and Forestry) (2003). Resource Directed Measures. Module 1: Introductory module, Department of Water Affairs and Forestry, Pretoria.

DWAF (Department of Water Affairs and Forestry) (2004a). DWAF Report No. P WMA 05/000/00/0303: Internal Strategic Perspectives: Inkomati Water Management Area – Version 1 (Department of Water Affairs and Forestry, Pretoria.

DWAF (Department of Water Affairs and Forestry) (2004b). Draft Background Information: South African Forestry, Pulp and Paper Industries: Draft Version 1. Department of Water Affairs and Forestry, Pretoria.

Godfrey L and Roux D (2000). Intermediate reserve determination for the Elands River catchment, Incomati system, Mpumalanga. Technical Report for the Department of Water Affairs and Forestry, by the Division of Water Environment and Forestry Technology, CSIR, Pretoria. Report No. ENV-P-C 2000-090 pp 1 -134.

Godfrey L, Jackson TB and Roux DJ (2000). Rapid ecological reserve determination for the Elands River Catchment, Mpumalanga. Technical Report for the Department of Water Affairs and Forestry, by the Division of Water Environment and Forestry Technology, CSIR, Pretoria. Report No. ENV-P-C 2000-043.

Hill L (2005). Elands Catchment Comprehensive Reserve Determination Study, Mpumalanga Province, Ecological Classification and Ecological Water Requirements (quantity) Workshop Report, Contract Report for Sappi-Ngodwana, Submitted to the Department: Water Affairs and Forestry, by the Division of Water Environment and Forestry Technology, CSIR, Pretoria. Report No. ENV-P-C 2004-019 pp 1 - 98.

31 Chapter 1

Kleynhans CJ and Louw MD (2005). River EcoClassification. Manual for EcoStatus Determination (Version 2). MODULE A: EcoClassification process and EcoStatus determination, KV 168/05. Department of Water Affairs and Forestry, Pretoria.

Kleynhans CJ, Schulz GW, Engelbrecht JS and Roussaeu FJ (1992). The impact a of paper mill effluent spill on the fish population of the Elands and Crocodile Rivers (Incomati System, Transvaal).Water SA. Vol. 18 (2) pp 73 – 80.

O’Brien GC (2003). An ecotoxicological investigation into the ecological integrity of a segment of the Elands River, Mpumalanga, South Africa. M.Sc. dissertation.

Pegram GC and Görgens AHM (2001). A guide to non – point source assessment to support water quality management of surface water resources in South Africa. WRC report no. TT 142/01. Water Research Commission, Pretoria.

Roux DJ (2001). Development of the procedures for the implementation of the national River Health Programme in the province of Mpumalanga. WRC report No 850/1/01. Water Research Commission, Pretoria.

Roux DJ, Kleynhans CJ, Thirion C, Hill L, Engelbrecht JS, Deacon AR and Kemper NP (1999). Adaptive assessment and management of riverine ecosystem: The crocodile/Elands River case study. Water SA. Vol. 25 (4) pp 500 – 511.

Weddepohl JP, Pauer JJ, Du Plessis HM, Harris J, Heath RGM, Archibald REM and Chutter FM (1991). Sappi Ngodwana Mill water quality in the Elands River. Technical Report for Sappi, by the Division of the Environment and Forestry Technology, Report No. DWT 000862, CSIR, Pretoria.

CHAPTER 2

LITERATURE SURVEY AND SITE SELECTION

2.1 INTRODUCTION

As the demand for water increases, with increasing human populations and economic development, so to does the pollution of river ecosystems. Users such as domestic, agricultural, recreational and industrial sectors all depend on fresh flowing water (Roux et al., 1996). River systems on a global level are heavily degraded by these human activities and impacts (Jungwirth et al., 2000; Muhar et al., 2000). It is thus of the utmost importance for both river conservation and management to determine which basic processes, functions and structures make up the ecological integrity of these running waters. Several programs and procedures have been developed to assist. Although the conservation of biological diversity has been the main aim of conservation biology, the phrase “biological integrity” has formed the cornerstone of all these programs. Biological or biotic integrity refers to a biological system’s ability to function and maintain itself in the face of changes in environmental conditions (Angemeier and Karr, 1994).

The structure and taxonomic composition of biological communities integrate both the physical and chemical aspects of the environment (Griffith et al., 2005). Biotic communities are thus often used as indicators of biological integrity. According to Dale and Beyeler (2001) ecological indicators should be easily measured, be sensitive to stresses on the system, respond to stress in a predictable manner, be anticipatory, predict changes that can be averted by management actions, be integrative, have a known response to disturbances, anthropogenic stresses, and changes over time, and have low variability in response. Nearly all modern management programs that are useful in the protection of our resources and improvement of management practices make use of some form of ecological indicators. Some of these programs that are widely used in South Africa include the River Health Program (RHP) and the Reserve determination studies.

33 Chapter 2

2.1.1 RIVER HEALTH PROGRAM

One of the most important programs is the National RHP. This program was initiated by DWAF in 1994. The purpose for the implementation of this program was to gather information concerning the ecological state of our rivers. It is based on the concept that aquatic communities (like fish, invertebrates and riparian vegetation) combine and reflect the effect of anthropogenic disturbances that occur in the river over an extended period of time (Ballance et al., 2001). The rational is that the integrity of the aquatic biota provides a direct, holistic and integrated approach to measure the integrity of a river as a whole (Roux et al., 1999). It is however, important to note that a distinction can be made between integrity and health of an ecosystem. Integrity refers to how close an ecosystem is to natural conditions, while health represent a condition that is desired by humans using the resource. The RHP was the first program to contribute to setting and monitoring resource objectives. The main objectives of the RHP are to:

• Measure, assess and report on the ecological state of aquatic ecosystems. • Detect and report spatial and temporal trends in ecological state of aquatic ecosystems, identify and report on emerging problems. • Ensure that reports provide scientific and managerial relevant information (Roux, 2001).

For the RHP to be truly successful as a management and information tool, a step – wise procedure must be in place for linking data and information with management actions (Roux, 2001). It is clear that the RHP has been developed to determine the ecological state of rivers in South Africa. Although this is important it is not enough. One not only needs to classify our rivers, but manage and protect this resource to ensure sustainability.

2.1.2 THE RESERVE

With this in mind the South African government developed resource directed measures (RDM). The RDM where designed to be applied to water resources as a system i.e. a catchment. These RDM are (DWAF, 2003):

34 Chapter 2

A. The establishment of the reserve: The reserve refers to the quantity and the quality of water to satisfy basic human needs, and to protect aquatic ecosystems, in order to ensure ecological sustainability and use of relevant water resources. The aim of setting the reserve is thus to prevent the exploitation of resources and to ensure the sustainable development of aquatic ecosystems.

B. The classification of water resources: The classification of water resources refers to the determination of the Present Ecological State (PES) of water resources. Resource protection is given effect through this classification, as this procedure leads to the specification of an acceptable degree of change from natural conditions for a particular natural resource. Classification of resources is helpful with setting of management scenarios.

C. Setting of resource quality objectives (RQOs): These are numerical and narrative descriptors of conditions that need to be met in order to achieve the required management scenarios. The main purpose of the RQOs is to establish clear goals. It becomes apparent that the RHP will only provide information on the classification of resources and forms only part of a greater management plan.

In order to better understand the RDM one needs to define a water resource. According to DWAF (2003) a water resource is an ecosystem that includes the physical and structural habitats (both instream and riparian), the water and aquatic biota and all the processes which link habitat, water and the aquatic biota. This definition stresses the fact that water resources are linked with other features and processes of nature. It is thus important to manage them in a way that takes into account their dynamic character. RDM as a strategy reflects awareness of this problem.

2.1.3 A NEW APPROACH FOR THE MANAGEMENT OF RIVERS

A new procedure was developed to further support management and protection of rivers. This procedure is known as EcoClassification (the term used for ecological classification). It refers to the determination and categorisation of the present ecological state of rivers compared to their natural state. Ecological evaluation in terms of expected reference conditions, followed with the integration with an assigned category, represent the ecological status (EcoStatus).

35 Chapter 2

EcoClassification allows for gathering information useful in deriving future ecological objectives for the river. EcoClassification should not be confused with the classification system as indicated in the National Water Act (Kleynhans and Louw, 2005). It may however, have a role in the reserve determination process and in monitoring. The EcoClassification process plays an important part in the reserve due to the fact that flow (water quantity) can not be recommended without prior knowledge of the present ecological state. The ecological categories determined in the EcoClassification form an important part of the reserve process. Monitoring as used in the RHP also makes use of EcoClassification to assess the severity of change from reference conditions.

2.2 THE STUDY AREA

The Elands River, a major tributary of the Crocodile River, rises in a gently sloping Highveld zone, near the town of Machadodorp (Roux et al., 1999). It is swiftly flowing in its upper reaches, with long sections of shallow rapids with interspersing deeper rocky pools. The upper half of the Elands River widens to about 5 m and has a moderate to steep slope. The lower half of the Elands River is a fast flowing section. The river is 15-20 m wide with large rocky pools and abundant riffle and rapid areas (Ballance et al., 2001). Below the confluence with the Ngodwana River the rapids become shorter and the pools deeper. At Lindenau the valley becomes narrower and the river becomes more inaccessible (Weddepohl et al., 1991). Land use in the upper reaches of the Eland River is largely dominated by forestry and citrus farming, while water from the Crocodile River supports a large area of irrigated agriculture. The Elands River supports a rich variety of aquatic life and riparian vegetation. The Elands River falls within the Incomati Water Management Area in is further sub divided in to the Crocodile River sub area (Figure 2.1). The first significant dam to be constructed in the catchment was the Longmere Dam on the White River completed in 1940. This was followed by more medium sized dams such as the Witklip, Ngodwana and Klipkoppie. This coincided with the steady process of small dam construction. The Crocodile Catchment covers about 10 450 km2, with the estimated water use in 1997 reaching 580 million cubic metres per year, of which irrigation accounted for 49%, forestry 43% and industry, commerce and mining the remaining 8%. There is an estimated 42 300 ha of irrigation lands in the catchment and an estimated 1 775 km2 of plantations. These two activities are the major users of water in the catchment. Industrial water use in the catchment is limited and consists mostly of the Sappi paper mill at Ngodwana and the sugar mills at Malelane (DWAF, 2004a).

36 Chapter 2

Figure 2.1 Elands River Comprehensive Reserve Determination Study Surface Water Resource Units (Hill, 2005).

37 Chapter 2

The main impacts on the Elands River in the study area, other than that of the Mill includes: (1) activities within the lower Ngodwana River which may cause modifications to water quality and flow, (2) the possibility that exotic fish species might be entering the system from the Ngodwana Dam and the upper reaches of the Elands River, (3) nutrient loading taking place due to the inadequate treatment of the sewage being released into the river in the upper reaches and in the vicinity of the Mill, (4) sedimentation and flow regulation that is taking place in the Crocodile River, upstream of the confluence with the Elands River and (5) the agricultural activities within the Elands River system. The Elands River and the associated systems has been the centre of various research programs (James and Barber, 1991a, 1991b; Weddepohl, 1991; Roux et al., 1999; Kleynhans et al., 1992, 1999; Godfrey and Roux, 2000; Ballance et al, 2001; O’Brien, 2003; Hill, 2005).

2.3 SAPPI AND THE ELELANDSANDS RIVER

There are various water users within the study area. This includes the domestic and agricultural users, but most importantly the industrial sector. All though the amount of water used by the industrial sector is minimal (8%) the impact of this sector cause the biggest concern. The major industrial activity in the study area is Sappi’s pulp and paper activities. The Sappi Ngodwana Mill is situated at the confluence of the Elands and Ngodwana rivers. Water is supplied by the Ngodwana Dam which was constructed in 1983 (Weddepohl et al., 1991). The Mill uses the Kraft pulping process to produce bleached and unbleached pulp. A major percentage of pulp produced is consumed in the production of newsprint, Kraft linerboard and white top linerboard. The rest is sold as market pulp. Of the Mill's total output, 70% is sold in the local market and the balance is exported to countries throughout the world. Resource recovery is very important and water use is kept to a minimum. The best available technology has been used, in each successive expansion phase, to ensure state-of-the-art manufacturing capability whilst minimising environmental impact. In 1995, for example, the Mill commissioned one of the world's first ozone bleaching plants. This eliminated the use of chlorine and thus reduced the amount of chlorine present in effluent (O’Brien, 2003). Historically the activities of the Mill have caused some environmental concerns. On 23 September 1989 effluent generally known as black liquor (soap skimming) was accidentally released into the Ngodwana River. The effluent flowed into the Elands River and eventually into the Crocodile River. The spill lasted about 2 hours and following the spill a large quantity of water was released from the Ngodwana Dam to

38 Chapter 2 dilute this effluent. This helped to some degree but it still caused massive fish mortalities. According to Kleynhans et al. (1992) the mortalities where mainly caused by the high biological oxygen demand and the sulphur containing substances of the black liquor. Kleynhans et al. (1992) further stated that the fish populations recovered to some degree within two years of the spill.

A few regrettable incidents over the years have given the industry a reputation as a major water polluter. However, it is a sound generalization that the industry’s management of water is of a world class standard and therefore better than in most other countries (DWAF, 2004b). The Ngodwana Mill has been designed to use little water during operation. This has largely been attained due to the internal reuse of water and the recycling of process steam. However, due to this tight water budget, the acid bleached streams that are produced in the process are highly concentrated with high chlorine content. This waste water typically contains 5 g/l of total solids with the major constituents being 1 – 2 g/l of chloride and 0.3 – 0.8 g/l of total organic carbon. It has a low pH (2 – 4) and a temperature of 50 to 60 ºC (Davies et al., 1992). The effects of pulp and paper Mill effluents have received attention over the past decade. Furthermore, the pulp and paper industry has several problems associated with the treatment and discharge of effluents. The two major concerns are the high concentrations of dissolved salts and the poor biodegradability of the effluent (Table 2.1).

Table 2.1 Typical analysis of acid bleach effluent generated during the activities of the Ngodwana pulp and paper Mill (Davies et al., 1992).

Determinant Ngodwana Mill

pH 1.94

Conductivity 1769 mS/m

Total organic carbon 800 mg/l

Calcium 86 mg/l

Magnesium 511 mg/l

Sodium 510 mg/l

Chloride 2110 mg/l

39 Chapter 2

The Mill does not discharge effluent directly into the river. The effluent is however, irrigated onto the 514 hectares of farmlands known as “SAPPI” farm adjacent to the mill. Cut-off trenches have been created at the lower section of the farm to collect any runoff from the farm preventing run-off from entering the Elands River.

This water is redirected to the effluent storage dams bordering to the farm. The irrigated effluent has contaminated the local groundwater system (Claassen, 2005). Groundwater enters the Elands River through three springs near Ngodwana namely Fraser’s eye, Northern eye and Eye X. The primary influence of this groundwater contamination is the deteriation of the surface water quality as well as, influencing the quantity of water in the Elands River (Golder Associates, 2004). These eyes contribute approximately 34 000 m3/d of flow to the Elands River. Surveys completed during the determination of the comprehensive reserve indicated that the groundwater entering the Elands River from two of the springs was severely impacted. The groundwater from both Fraser’s eye and Eye X has been contaminated with calcium, potassium, magnesium, sulphates and most importantly chlorides. All these substances contribute to the increase in conductivity in the Elands River, which in turn has an impact on the ecological integrity of the system (Colvin and Engelbrecht, 2005)

2.4 BIOMONITORBIOMONITORINGING

More than a century ago, people recognised that human activities produced pollution harmful to the biota. They therefore made an effort to track the extent of biological degradation; biological degradation was even considered an indicator for the presence of human activities. So began biological monitoring (Karr and Chu, 2000). In the past water quality monitoring focused on physical and chemical measurements. It has however, become recognised that by using other indicators in addition to traditional water quality measurements can, to a great extent, enhance the assessment and management of water resources (Hohls, 1996). Biomonitoring can thus be defined as the use of living organisms as biological indicators of ecosystem or environmental health. It is further stated that animals and plants can provide a long – term integrated reflection of water quality, quantity, habitat quality and other environmental conditions. There are certain advantages and disadvantages associated with biomonitoring (Hohls, 1996).

40 Chapter 2

A. Advantages:

• Detects changes in water quality. • Detects changes caused to water flow. • Standardised method to compare water quality of different resources. • Easy and practical means of assessing water resources. • Inexpensive and quick. • Detects water quality changes that may be missed by chemical sampling.

B. Disadvantages:

• May require lots of training. • Usually open to subjective interpretation. • Makes no provision of exact figures of water quality. • Reflects change but can not point out the cause. • Has no legal standing.

The complexity of biological systems and the diverse influences of human activities require a multimetric approach that reflects all important aspects of stream biology ranging from individual to assemblage level and responds to anthropogenic activities in a detectable manner (Kerans and Karr, 1994). These indices are based on the natural attributes of organisms and can be defined as a tool that can be used to simplify and interpret data in such a way that non – specialists, managers and the general public can understand it. Various abiotic and biotic components, when compared to reference conditions, can be used to indicate the health or integrity of an aquatic ecosystem. These include the following:

2.4.1 WATER QUALITY

Water quality can be defined as the combined effects of the physical attributes and the chemical constituents of an aquatic ecosystem (Palmer et al., 1996). Water quality, with regards to the physical – chemical characteristics of the water resource and concentration of anthropogenic substances, may be one of the most important factors influencing aquatic ecosystem health. The reason for this is that all changes in the biotic components of these aquatic ecosystems are

41 Chapter 2

linked to changes water quality (Dallas and Day, 1993). It is however, important to note that water quality may not only be changed by pollution. According to Malan et al. (2003) alterations in flow may cause changes to water quality. These changes are nonetheless, complex and difficult to predict. Flow is thus a critical part of the reserve determination process. The task of the flow requirement component in the reserve determination is to provide both quantified and descriptive information about water flow, with information on frequency, magnitude and duration. Methods for quantifying water quality however, still focus on magnitude or concentration (Palmer et al., 2005). From the many possible water quality variables, only some are included in the determination of water quality. These include: inorganic salts, nutrients, physical variables and the toxic substances mentioned in the South African Water Quality Guidelines. The South African Water Quality Guidelines are used by the Department of Water Affairs and Forestry as its primary source of information and decision-support to judge the fitness of water for use and for other water quality management purposes (DWAF, 1996a). The users of water in South Africa include: domestic users, the industrial and agricultural sectors and aquatic ecosystems. For each of these users there are a specific set of water quality guidelines that has been determined. The concentration range at which a variable (chemical or physical) will have no effect is known as the Target Water Quality Range (TWQR) for that variable. Not all the users are of concern in the study area. Industries that fall within Category 1 are not important as no such industries are present in the study area (Table 2.2). Very few agricultural activities that involve livestock are found in the study area.

Waste waters associated with the pulp and paper Mill industry are high in organic and inorganic material leading to high chemical oxygen demand (COD), biological oxygen demand (BOD) and total dissolved solids (TDS). The high BOD is sustained because often there is insufficient nitrogen to support degradation. The pulp and paper mill effluent contains toxic substances such as soaps of resin acids and sodium salts of unsaturated fatty acids. Theses acids are relatively resistant to bio–degradation (Owens, 1991). At the Ngodwana Mill there is no debarking effluent, and black liquor can only reach wastewater through spills and leaks. To reach regulatory requirements and to facilitate biological treatment, effluent can be neutralized with

CaCO3. Bleaching liquors are high in chloride content, and this may cause problems for down stream users such as tobacco farmers with very strict chloride requirements (DWAF, 1996b). Organochlorines are released during the bleaching process when chlorines are used to remove

42 Chapter 2 residual lignin (Zokufa et al., 2001) and these components have the potential to induce long– term chronic toxicity at sub–lethal concentrations

Water quality in the Elands River has been closely monitored by the management of the Ngodwana Mill. A complete report was compiled by Weddepohl et al. in 1991 for the Ngodwana Mill and the effect that their activities have on the water quality of the Eland River. This report states that most chemical variables occur at low concentrations and the conductivity seldom exceeds 200 µs/cm and the dissolved oxygen levels are close to saturation. One of the major problems with regards to water quality in the Elands River has been salinity, especially due to high concentrations of chlorine. Chlorine, in the past, played a key part in the bleaching process and the replacement of chlorine with ozone has considerably reduced this problem (O’Brien, 2003).

Table 2.2 Examples of industry practices that fall within a Category I industry (adapted from DWAF, 1996c).

Heat Exchange Steam Generation Process Water Product Water

Evaporative cooling High pressure boilers Solvent agent Beverages Solution cooling Heat transfer medium Dairy Water heating Humidification Petrochemical Lubrication

Several other activities in the vicinity of the river may have an adverse effect on the water quality within the Elands River. These activities include: the sewage treatment facilities that are found in the upper reaches near Waterval Boven and at the Ngodwana village, the rural settlements near Waterval Boven and Lindenau and agricultural activities upstream of the Sappi Ngodwana pulp Mill. Sewage discharge is a major threat to the ecological integrity of any river system and may cause increased oxygen demand and nutrient loading within a river. This can lead to toxic algal blooms and degradation in the integrity and health of a system (Morrison et al., 2001). The potential effect of agricultural activities depends on the type and the extent of agriculture. Runoff, both surface and subsurface, often causes deteriation of water quality by increasing concentrations of plant nutrients (Dallas and Day, 1993) such as nitrogen and phosphate. A report regarding water quality released by Sappi indicate various water quality requirements for indicator species of fish (Weddepohl et al., 1991). These requirements take in to account that the Elands River in 1991 was classed as a cold water ecosystem. With this in

43 Chapter 2 mind it has been agreed that the dissolved oxygen levels should be greater then 6.5 mg/l for 99% of the time or 5 mg/l for 95% of the time. Temperature is just as important for fish, as high temperature means less dissolved oxygen. Should temperatures increase this could lead to increased metabolism, oxygen demand and respiration for the various fish species. Temperature requirements have been set at 28º C or less for 95% of the time or less the 30º C for 99% of the time. With chloride concentrations being a major problem in the past certain set values have been decided on that needs to be maintained. DWAF set this concentration at 20 mg/l. This value is considerably lower than the lowest harmful value for fish, which is 400 mg/l for trout (Weddepohl et al., 1991). These values, mentioned above, in large apply to the macro invertebrate fauna of the Elands River as well.

2.4.2 SEDIMENT

Aquatic sediments are formed from the deposition of particles and colloids and can act as both a source and a sink of pollutants. Although sediment may play a part in the provision of habitat for organisms, they may affect organisms in various ways. Long term toxicant input in sediment may often lead to the occurrence of contaminant levels far higher than that in the surrounding water. This is mainly due to the partitioning of substances onto sediment – based binding sites. Thus, sediment plays a key role in the bioavailability of substances. Two of the most important characteristics of sediment that needs to be taken into account are (1) the grain size and (2) the organic content within the sediment. Conflicting results have been reported on the influence of grain size on bioavailability. The basic consensus is that organisms usually ingest the smaller particles which are richer in organic material and this leads to the increased uptake of toxicants. Grain size is used to characterize the physical characteristics of sediments. It is most common to characterise grain size as percentages of gravel, sand slit and clay (USEPA, 1986). Grain size may influence both chemical and biological variables and plays an important role in the transport and availability of nutrients and contaminant (Walling and Moorehead, 1989). Furthermore, the diversity and abundance of invertebrate assemblages increases along with increasing sediment particle size, from mud to cobbles (Hill, 2005). The quantity and quality of organic matter in sediments are recognised as major factors affecting benthic fauna. Total organic matter (as determined by combustion) is generally an overestimate of food availability, mainly because various inorganic compounds may be oxidised at about 500 ºC (Pusceddu et al., 1999). Apart from the nutritional value organic matter plays an important role in the bioavailability of toxicants. Hydrophobic contaminants usually bind to the organic material in

44 Chapter 2 sediment thus, the amount and type of organic material (e.g. mud or plant material) needs to be taken into account. Sediment quality is of particular concern when focusing on invertebrate communities. Seemingly minor changes in substrate particle size, organic content, and even texture can influence the associated invertebrate community structure. Sediment can also influence the feeding or breeding patterns of fish communities (Berkman et al., 1986).

Chlorinated compounds have the tendency to reside in sediment. This is extremely important when taking in account that chlorine is one of the most common constituents of pulp and paper mill effluents. The effluent released by the pulp and paper mill industry may affect the sediment adjacent to the mill in two ways. Firstly, the effluent might contain large quantities of organic material that may cause increased organic carbon content within the sediment. This may lead to higher bioavailability of toxic compounds, while decaying material will lead to the decreased availability of oxygen. Secondly, the effluent may alter the grain size assembly of the sediment, inevitably obliterating that habitat for aquatic organisms (Sibley et al., 1997)

2.4.3 HABITAT Habitat integrity refers to the maintenance of a balanced, integrated composition of physio- chemical and habitat characteristics on a temporal and spatial scale that are comparable to the characteristics of natural habitats of the region (DWAF, 1999). Essentially, the habitat integrity status of a river will provide the template for a certain level of biotic integrity to be realised. In this sense the assessment of the habitat integrity of a river can be seen as a precursor of the assessment of biotic integrity. Habitat availability and diversity are major factors that influence aquatic community structures. Where the habitat is diverse and largely unimpacted, a healthy biological community is likely to occur. Changes in community structure may be attributed to either the deteriation in water quality or the alteration of habitat, or both.

The procedure presently used for the assessment of the habitat integrity for rivers is that developed by Kleynhans (1996). The methodology is based on the qualitative assessment of a number of pre-weighted criteria which indicate the integrity of the in-stream and riparian habitats available for use by riverine biota. Habitat integrity scores are then sequentially determined for five kilometre segments of the river which are derived from the application of a standardised formula. Any habitat integrity assessment which is site based and does not make use of aerial photography may be seen as insufficient or of low confidence (Table 2.3).

45 Chapter 2

Table 2.3 Criteria used in the assessment of habitat integrity. Adapted from DWAF (1999).

CRITERION RELEVANCE Water abstraction Direct impact on habitat type, abundance and size. Also implicated in flow, bed, channel and water quality characteristics. Riparian vegetation may be influenced by a decrease in the supply of water. Flow modification Consequence of abstraction or regulation by impoundments. Changes in temporal and spatial characteristics of flow can have an impact on habitat attributes such as an increase in duration of low flow season, resulting in low availability of certain habitat types or water at the start of the breeding, flowering or growing season. Bed modification Regarded as the result of increased input of sediment from the catchment or a decrease in the ability of the river to transport sediment. Indirect indications of sedimentation are stream bank and catchment erosion. Purposeful alteration of the stream bed, e.g. the removal of rapids for navigation is also included. Channel modification May be the result of a change in flow which may alter channel characteristics causing a change in marginal instream and riparian habitat. Purposeful channel modification to improve drainage is also included. Water quality modification Originates from point and diffuse point sources. Measured directly or agricultural activities, human settlements and industrial activities may indicate the likelihood of modification. Aggravated by a decrease in the volume of water during low or no flow conditions. Inundation Destruction of riffle, rapid and riparian zone habitat. Obstruction to the movement of aquatic fauna and influences water quality and the movement of sediments Exotic macrophytes Alteration of habitat by obstruction of flow and may influence water quality. Dependent upon the species involved and scale of infestation. Exotic aquatic fauna The disturbance of the stream bottom during feeding may influence the water quality and increase turbidity. Dependent upon the species involved and their abundance. Solid waste disposal A direct anthropogenic impact which may alter habitat structurally. Also a general indication of the misuse and mismanagement of the river. Vegetation removal Impairment of the buffer the vegetation forms to the movement of sediment and other catchment runoff products into the river. Refers to physical removal for farming, firewood and overgrazing. Includes both exotic and indigenous vegetation. Exotic vegetation Excludes natural vegetation due to vigorous growth, causing bank instability encroachment and decreasing the buffering function of the riparian zone. Allochtonous organic matter input will also be changed. Riparian zone habitat diversity is also reduced. Bank erosion Decrease in bank stability will cause sedimentation and possible collapse of the river bank resulting in a loss or modification of both instream and riparian habitats. Increased erosion can be the result of natural vegetation removal, overgrazing or exotic vegetation encroachment.

The two main indices that were applied during this study included the Integrated Habitat Assessment Index (IHAS) and the Habitat Quality Index (HQI). The ultimate aim of the IHAS is to summarise and numerically reflect the quantity, quality and diversity of biotopes available for habitation by macroinvertebrates at a sampling site (McMillan, 1998). The scoring system is based on a total of 100 points, split into two sections: Sampling Habitat (55 points) and Stream

46 Chapter 2

Condition/Characteristics (45 points). The Sampling Habitat section is further divided into three sub-sections: Stones-in-Current (20 points), Vegetation (15 points), and Other Habitat (20 points), including stones-out-of-current, gravel, sand and mud. The Stream Condition section provides an evaluation of a site in terms of its physical characteristics and the degree of disturbance present, including estimates of aspects such as stream width, depth and velocity. For the HQI the various metrics that represent the habitat quality of a site are rated from zero to 20, with 20 being a maximum score and indicating excellent habitat quality. The HQI not only takes in to account the quality of the instream habitat in terms of substrate and flow, but also assess the possible impacts of anthropogenic activities in the surrounding area (farming, construction, rural settlements, etc.).

2.4.4 RIPARIAN VEGETATION

Riparian vegetation forms a vital part of any river ecosystem. It has been well documented that riparian vegetation plays a number of important geomorphological, ecological and social roles which may have an influence on the condition and sustainability of the riverine ecosystem. These roles include: stabilisation of river channel, banks and flood plains, flood reduction, maintenance of water temperature and quality, provision of habitat, etc. It is clear from these roles that the assessment of the riparian vegetation is immiscible in any biomonitoring program or integrity assessment (Kemper, 2001). When naturally vegetated landscapes are converted to urban or agricultural uses, physical and biological relationships with adjacent streams are affected, usually resulting in habitat degradation and negative impacts on stream biota (Roth et al., 1996).

Riparian vegetation in the Elands River and surrounding systems is important for moderating water temperatures through shading, as well as for providing habitat to instream fauna, riverbank stability, and particulate organic material (Barling & Moore, 1991; Pinto et al., 2006). Furthermore, the organic inputs from riparian vegetation are major food sources for river organisms. This resource unit (Figure 2.1) occurs within the Acocks Veld type 9 (Lowveld Sour Bushveld) and type 10 (Lowveld). Table 2.4 and 2.5 clearly shows that one of the major concerns regarding the integrity of the riparian zone is the various exotic species occurring in both the upper and lower reaches of the Elands River.

47 Chapter 2

Table 2.4 Vegetation species found in the Riparian zone of Resource Unit 2 (Roodewal, Doornhoek, Waterval Boven, Blouboshkraal) T = tree; S = shrub; R = reed; F = herbaceous; G = grass (Hill, 2005). Species Riparian Type Morphological Units Indicator Active Seasonal Macro Species Channel Channel Channel bank Acacia karroo T * Acacia robusta T * Acacia sieberiana R T * Buddleja auriculata S * Buddleja salviifolia S * Choristylis rhamnoides S * Cliffortia sp. R S * Combretum erythrophyllum R T * Commelina africana var F * * barberae/lacispatha Diospyros lycioides R S * Dombeya pulchra T * Eriosema burkei F * Euclea undulata S * Greyia sutherlandii T * Gymnosporia buxifolia S * Leucosidea sericea R S * * Myrica lanceolata S * * Myrica serrata R S * * Persicaria serrulata F * Phragmites australis R R * * Plectranthus sp F * * Rhus chirindensis T * Rhus pentheri S * Salix mucronata R S * Themeda triandra G * Typha capensis R R * Ziziphus mucronata T * Exotic species Acacia dealbata T * * Acacia mearnsii T * * Casuarina cunninghamiana T * *

48 Chapter 2 Eucalyptus sp. T * * Lantana camara S * * Melia azedarach T * * Senna didymobotrya S * * Sesbania punicea S * * Solanum incanum F * Solanum mauritianum S * * Tithonia diversifolia F *

Table 2.5 Vegetation species found in the Riparian zone of Resource Unit 4 (Roodewal BB2, Houtboschoek, Elandshoek & Ngodwana) T = tree; S = shrub; R = reed; F = herbaceous; G = grass (Hill, 2005) Species Riparian Type Morphological Units Indicator Active Seasonal Macro Species Channel Channel Channel bank Acacia ataxacantha S *

Acacia karroo T *

Acacia sieberiana R T *

Breonadia salicina R T *

Buddleja salviifolia S * *

Cliffortia sp. R S *

Choristylis rhamnoides S *

Combretum erythrophyllum R T *

Crotalaria recta

Cyperus latifolia R G * *

Diospyros lycioides R S *

Dombeya burgessiae S *

Dombeya rotundifolia S *

Ficus sur R T *

Ischaemum faciculatum G * *

Myrica serrata R S *

Phragmites australis R R * *

Salix mucronata R S *

Typha latifolia R R *

Exotic species

Acacia dealbata T * *

49 Chapter 2 Eucalyptus sp. T * *

Lantana camara S * *

Morus alba T * *

Melia azedarach T * *

Arundo donax R * *

Psidium quajava T * *

Senna didymobotrya S * *

Sesbania bispinosa S * *

Sesbania punicea S * *

Solanum mauritianum S * *

Tagetes minuta F * *

The current index used in South Africa to assess the integrity of riparian vegetation is the Riparian Vegetation Index (RVI). This index was developed using that of Kleynhans (1996) for the assessment of habitat integrity. It uses the same scoring and weighting systems. For this index however, specific riparian vegetation criteria were selected with weightings applicable to each. The following five criteria are applied during the assessment of the riparian vegetation:

• The presence and degree of farming with the riparian vegetation zone which has sled to the displacement or removal of the natural riparian vegetation. • The presence of permanent or semipermanent forms of construction within the riparian vegetation zone. • The presence of weirs or impoundments which may lead to the inundation of the riparian vegetation zone. • The degree of erosion within the riparian vegetation zone. • The abundance of invasive plant species within the riparian vegetation zone.

At present the new index for use in EcoStatus is still under development. When completed, this index will be known as the Riparian Vegetation Response Assessment Index.

50 Chapter 2

2.4.5 ICHTHYOFAUNA

Many groups of organisms have been proposed as indicators of environmental integrity or health and till today no clear favorite has emerged. Fish communities and individuals themselves pose various qualities that make them useful in biological monitoring (Kotze et al., 2004). Fish in general have a long life span and would thus be able to reflect changes in the integrity over long periods of time (chronic exposure to adverse conditions). Fish are also mobile, and although they may move away from polluted areas, they may reflect conditions of a large area of the ecosystem of concern. Their presence, therefore, can also infer the presence of other aquatic organisms since they occupy the top of the food chain in most aquatic systems. Fish and indices using fish in general are critical to various steps in better managing and protecting water. It has already been mentioned that they indicate environmental health, but they may also be used as part of the reserve determination where flow plays a key part. Reason being, that flow is directly related to the processes (geomorphological) that provide the habitat that fish require during various stages of their life cycle, and habitat also involves the cover that fish species require (Hill, 2005). Various indices, using fish, has been developed to indicate biological integrity.

Fish Assemblages Integrity Index (FAII)

The current index, being used to assess the status of fish communities in South Africa, is the FAII. The FAII aims to measure the biological integrity of a river as based on the attributes of the fish assemblages native to the river. Alien species (introduced indigenous and exotic species) are not included as metrics in the FAII. Their presence and distribution are noted but interpreted as possible causes for a decline in the FAII score. This index takes into account three aspects of fish assemblages; (1) The relative intolerance of indigenous fish species expected to occur at a segment. Intolerance in this context refers to the degree to which a species is able to withstand changes in the environmental conditions under which it occurs. This includes modification of physical habitat characteristics as well as chemical characteristics of the water habitat. (2) The frequency of occurrence of a species (number of sites in which the species occur). Abundance is not included as a metric in the index due to the difficulty in obtaining quantitative information on this. (3) The rating of the general health (abnormalities or disease) of the individuals caught. Even under unimpaired conditions, a small percentage of individuals can be expected to exhibit some anomalies. A comparison is made between the

51 Chapter 2 expected fish assemblages and the observed assemblages to determine the integrity of the river (Kleynhans, 1999).

Fish Response Assessment Index (FRAI)

FRAI is an index that has been developed for use in EcoClassification and EcoStatus. For the FRAI the number of species expected for the reference condition should be compared with the observed (sampled) data to determine deviation from the reference situation. Where sampling is not representative (not all habitats were sampled, for instance) or effective (difficult conditions to employ a particular sampling method), some generally common species may be absent. In the Elands River, for example, Labeobarbus polylepis and Anguilla mossambica are species of fish that are difficult to sample. In such a situation the species likely to be present – based on habitat, presence of closely related species and other environmental conditions - may be used to supplement the list of “observed” species. If such an approach is adopted it is essential that this be indicated explicitly, as it will have an influence on the confidence of the fish ecological integrity determination. The metrics assessed in the FRAI include: velocity-depth preferences, cover preferences, flow requirements, migration, physico-chemical preferences and the effect of introduced species (Kleynhans and Louw, 2005). The ranking and weighting system adopted for the FRAI is the same as for other indices that for part of EcoClassification and EcoStatus. The rating, however, is based on an increase or decrease from reference conditions.

Species that are expected to occur in the Elands River include the mountain catlet (Amphilius uranoscopus), chubbyhead barb (Barbus anoplus), rosefin barb (Barbus argenteus), smallscale yellowfish (L. polylepis), banded tilapia (Tilapia sparrmanii), southern mouthbrooder (Pseudocrenilabrus philander), shortspine rock catlet (Chiloglanis pretoriae), Inkomati rock catlet (Chiloglanis. bifurcus) and longfin eel (A. mossambica). Rainbow trout may occur in the upper parts of this unit and largemouth bass (Micropterus salmoides) are increasing in abundance and proliferating in the lower sections of river. The indigenous sharptooth catfish (C. gariepinus; has also been introduced into the Elands River. These introduced species have the potential to have a significant impact on the indigenous species in the Elands River (Godfrey and Roux, 2000; Hill, 2005,). It has been mentioned earlier that a spill at the Ngodwana Mill in September of 1989 caused massive fish mortalities. Due to this unfortunate event several studies have been completed to assess the recovery of the Elands River with regards to fish population

52 Chapter 2 and assemblages (James and Barber, 1991a; Godfrey and Roux, 2000; O’Brien, 2003 and Hill, 2005). These recent studies have shown that the fish communities are returning to normal and are close to a natural state (Table 2.6).

One of these studies found that the recovery rate of the fish in this river was relatively slow. This is especially the case with regards to C. bifurcus. This species is highly sensitive to changes in the environment. This species is critically endangered and is increasingly threatened due to extraction and regulation of river water as well as pollution (Skelton, 2001). Virtually all species where destroyed just below the confluence of the Ngodwana and the Elands Rivers. It took between 6 and 11 months for the fish populations to recover to such a point as to which areas down stream of were the spill occurred were comparable with areas upstream (Kleynhans et al., 1992). Recent studies have shown that the situation has bettered and fish populations have returned to “normal.” The concern has been raised however, that resin acid which forms part of the effluent persists in both water and sediment. Continuous exposure to such acids may cause sub lethal toxic effects. Barber and James (1991) stated that the ecological status of the Elands River with regards to fish populations were of great concern. They found that fish numbers and species diversity remained low, even though invertebrate life was re – established in the water looked clean and clear. They did however; find large numbers of juvenile fish at the site of the spill, indicating that the river at that point in time should signs of recovery. The release of large quantities of water from various sources to dilute the effluent two hours after the spill may have altered the habitat and may have been the cause for the slow recovery rate.

Table 2.6 The present ecological state if the fish in the Elands River within resource unit 2 (see Figure 2.1) (Hill, 2005). FISH PES METRIC GROUP METRIC GROUP: WEIGHTED SCORE FOR CALCULATED GROUP SCORE FLOW-DEPTH METRICS 94.57 27.02 FLOW MODIFICATION METRICS 94.37 26.96 COVER METRICS 88.79 22.83 HEALTH/CONDITION METRICS 82.96 16.59 IMPACT OF INTRODUCED SPP (NEGATIVE) 37.14 -4.24

Fish - Present Ecological State (PES) 89.16 Fish PES Category B

53 Chapter 2

Pulp and paper mill effluent may cause a variety of biological effects. Studies have shown that they increase age to maturity, smaller gonads, increased hepatic mixed–function oxidase (MFO) activity, reduced fecundity with age in females and reduction in male secondary sex characteristics (Munkittrick et al., 1992). A delay in sexual maturity and a lower egg production has been observed by Robinson et al., (1994) during laboratory exposure tests. Other effects include elevated dioxin levels, changes in liver size, growth rate and gonadal size and decreased steroid hormone levels. Compared to the amount of literature focusing on how paper and kraft mill effluents affect fish physiology (Karels et al., 2001; Sepulveda et al., 2001; Khan and Payne, 2002), the lack of information and agreement on the consequences of fish community responses to the effluents represents a crucial gap in understanding how these effluents impact freshwater ecosystems (Greenfield and Bart, 2005). Karels and Niemi (2002), for example, did report changes to the community structures of fish, while Kovacs et al. (2002) found no substantial differences in fish community structure between sites upstream and downstream from mill outfalls.

2.4.6 MACROINVERTEBRATES

Aquatic macroinvertebrates are adapted to live within certain environmental conditions. Changes within this environment may adversely affect community composition and abundance. Resident aquatic macro-invertebrates are good short-term indicators of ecological integrity because they integrate the effects of physical and chemical changes. Thus, integration of biological- with chemical- and physical indicators will provide information on the ecological status of the river (Hill, 2005). The environments mentioned above, in which invertebrates are found, are also known as biotopes. There are mainly three biotopes. (1) Stones: which refer to stones in and out of the current and bedrock. These stones are usually found in areas where the movement of water prevents the settling of silt or sediment. (2) Vegetation: refers to the aquatic vegetation, whether it is marginal or submerged. (3) GSM (gravel, sand and mud) refers to fine stones, silt or sediment deposited over time as well as mud (Dickens and Graham, 2002). The importance of habitat quality and quantity and the role it plays in the community structure of macroinvertebrates has often been overlooked. Often one finds that macroinvertebrate community structures are seriously modified, yet these changes are not related to water quality (Dickens and Graham, 2002). The current habitat assessment index being used to assess the

54 Chapter 2 quality available to invertebrates is the Integrated Habitat Assessment System (IHAS). Some doubts has, however, been risen about the efficiency of this system (Ollis et al., 2006).

The current index, being used to assess the status of riverine macroinvertebrates in South

Africa, is the South African Scoring System (SASS). The index is based on the presence of aquatic invertebrate families and the perceived sensitivity to water quality changes of these families. Different families show different sensitivities to pollution. These sensitivities range from highly tolerant families (e.g. Muscidae and Psychodidae) to highly sensitive families (e.g. Oligoneuridae). The index has gone through several upgrades and version 5 is currently in use. SASS is an accredited protocol that has been tested and widely used in South Africa as a biological index of water quality. SASS results are expressed both as an index score (SASS score) and the average score per recorded taxon (ASPT value). From this data it is possible to establish the integrity or health of a river. The problem of seasonal and spatial variation may occur, negatively impacting the comparison of data. It has become evident, through research projects that certain invertebrate families are more common in certain areas and occur more frequently during certain times of the year (Dallas, 2004a, 2004b). A new index, the MIRAI (Macro Invertebrate Response Assessment Index) has been under development for use in the new EcoClassification and EcoStatus procedures. The MIRAI is used to determine the ecological integrity of macroinvertebrate communities. It integrates the ecological requirements of the invertebrate taxa in a community or assemblage and their response to modified habitat conditions (Thirion, 2005). Although the MIRAI can be determined using information collected during a standard SASS survey (Dickens and Graham 2002), it can also be determined using more detailed information. The metrics that make up the MIRAI includes: flow modification, habitat, water quality, connectivity and seasonality. The ranking, weighting and rating system adopted for the MIRAI, is the same as for other indices that for part of EcoClassification and EcoStatus.

Various studies have been conducted on the invertebrate assemblages in the Elands River (Weddepohl et al., 1991; James and Barber,1991; O’Brien, 2003). These studies have found that the present effluent irrigation practises of the Ngodwana Mill are not having a negative impact on the invertebrate assemblages within the Elands River (Table 2.7). It became clear, after the sampling in 1991, that some populations have not yet recovered after the spill and there were some unusual occurrences among populations. These occurrences at the Lupelule River and

55 Chapter 2

Lindenau could be attributed to local disturbances and not the spill. Taxa that characterise the Elands River include (Hill, 2005): Planaria; Oligochaeta; Potamonautes; Hydracarina; Plecoptera (stoneflies) – Perlidae; Ephemeroptera (mayflies) – Baetidae, Caenidae, Heptageniidae, Leptophlebiidae; Oligoneuridae and Trichorythidae; Odonata (dragonflies) – Aeshnidae, Gomphidae, Libellulidae, Coenagrionidae and Chlorocyphidae; Hemiptera (bugs) – Corixidae, Veliidae , Notonectidae, Pleidae, Gerridae and Naucoridae; Trichoptera (caddis Flies) Leptoceridae , Philopotamidae, Ecnomidae, Hydroptilidae and Hydropsychidae; Coleoptera (beetles) – Dytiscidae, Gyrinidae, Hydrophilidae, Psephenidae and Elmidae; Diptera (flies) – Chironomidae, Ceratopogindae, Simuliidae, Tanbanidae, Tipulidae, Culicidae, Dixidae and Athericidae and Gastropoda (snails and limpets) – Lymnaeidae, Planorbinae, Physidae and Ancylidae .

The main activity of concern in the study area remains the pulp and paper Mill. According to O’Brien (2003), the main impact the pulp and paper mill industry has on the environment is in the form of salinity. The effects of salts on aquatic macroinvertebrates have extensively been studied (Palmer et al., 1996; Zokufa et al.; 2001, Williams et al.; 2003, Kefford et al., 2004). It has become clear through exposure of aquatic invertebrates to paper mill effluent that these effluents are toxic (Van Wijk and Hutchinson, 1995; Zokufa et al., 2001). This toxicity may cause mortalities and eventually alter the community structure. The Mill manages effluent discharge and irrigation is the main route of exposure. It was demonstrated by Zokufa et al. (2001) that the contaminated groundwater is contributing to salinisation and may have an impact on the aquatic macroinvertebrate communities. Pulp and paper mill effluent may contain inorganic nutrients, which cause eutrophication and the stimulation of algal growth and food supplies to invertebrates. A small number of reports also sate that the problem mentioned above may cause changes in feeding and food – web interactions (Hall et al., 1991; Lowell et al., 1995). High concentrations of organic debris may disturb light penetration that in turn influences primary production and filter feeding.

56 Chapter 2

Table 2.7 Present ecological state of the macroinvertebrate communities of the Elands River within resource unit 2 (see Figure 2.1) (Hill, 2005). PES Site Date SASS ASPT ECOREGION CLASS Doornhoek Aug-00 142 6.17 B Great Escarpment Mountains Weltevreden Oct-96 161 6.71 B Great Escarpment Mountains Weltevreden Aug-00 172 6.62 B Great Escarpment Mountains Hemlock Jul-93 140 6.36 C Great Escarpment Mountains Hemlock Aug-93 133 5.78 C Great Escarpment Mountains Hemlock Oct-93 105 5.25 C Great Escarpment Mountains Hemlock Jun-94 115 6.05 C Great Escarpment Mountains Hemlock Sep-94 211 6.81 A Great Escarpment Mountains Hemlock Jun-95 165 7.50 A Great Escarpment Mountains Hemlock Oct-96 185 7.40 A Great Escarpment Mountains Hemlock May-99 179 6.88 B Great Escarpment Mountains Hemlock Jul-99 193 6.66 A Great Escarpment Mountains Hemlock Sep-99 200 7.14 A Great Escarpment Mountains Hemlock Aug-00 211 6.81 A Great Escarpment Mountains Malaga May-99 145 6.90 B Great Escarpment Mountains Malaga Jul-99 154 7.33 B Great Escarpment Mountains Malaga Sep-99 134 6.38 C Great Escarpment Mountains Eerste Geluk Oct-96 139 7.3 C Great Escarpment Mountains ER2a Oct-96 180 7.50 A Great Escarpment Mountains ER2 Aug-00 180 6.67 B Great Escarpment Mountains Goedgeluk Oct-96 137 8.6 C Great Escarpment Mountains Goedgeluk Aug-00 159 7.23 B Great Escarpment Mountains Elandshoek Jul-93 137 6.23 C Great Escarpment Mountains Elandshoek Aug-93 133 5.78 C Great Escarpment Mountains Elandshoek Oct-93 154 5.92 B Great Escarpment Mountains Elandshoek Jun-94 111 5.55 C Great Escarpment Mountains Elandshoek Sep-94 166 6.92 B Great Escarpment Mountains Elandshoek Jun-95 185 6.85 A Great Escarpment Mountains Elandshoek Oct-96 202 7.5 A Great Escarpment Mountains Elandshoek Aug-00 161 7.66 A Great Escarpment Mountains

2.5 SITE SELECTION

The site selection was based on the activities shown in Figure 2.2, and the main aim of the study. This was to assess the influence of multiple impacts on the Elands River and Crocodile River below the confluence. The main impacts on the Elands River in the study area, other than that of the Mill includes: (1) activities within the lower Ngodwana River which may cause modifications to water quality and flow, (2) the possibility that exotic fish species might be entering the system from the Ngodwana Dam and the upper reaches of the Elands River, (3) nutrient loading taking place due to the inadequate treatment of the sewage being released into the river in the upper reaches and in the vicinity of the Mill, (4) sedimentation and flow regulation that is taking place in the Crocodile River, upstream of the confluence with the Elands River and (5) the agricultural activities within the Elands River system The sites are as follows: 57 Chapter 2

Sedimentation Nutrients Metals Flow regulation

Nutrients Unknown

Mill and Aliens springs Nutrients

Aliens

Nutrients

Metals salinity

Figure 2.2 Besides the Sappi Ngodwana Mill there are various other impacts on the Elands River, often occurring on the tributaries associated with the Elands River.

ER 1 (Elands River site 1 - Hemlock) This site in the Elands River positioned at Hemlock Bridge (also see Figure 2.13), situated above any impacts from the activities of the Ngodwana Mill (Figure 2.3). This site has historically been utilised as the “reference-pristine” site on the Elands River and is included in this study due to the historical (flow and biomonitoring) data associated with this site (Godfrey and Roux, 2000; Balance et al.; O’Brien, 2003; Hill, 2005). It has become clear during this survey period that the site may be subjected to impacts from further upstream that may have an impact on the water quality. Along with this, extensive fishing by local communities has been taking place. This has led to the possible replacement of this site with the upper Ngodwana River site (NR 1) as a potential reference site. The site shows the characteristic riffle and rapid habitat with interspersing pools. Substrate is dominated by cobbles and boulders with very little sediment available. The water is clean and clear and flow is generally categorised as “fast shallow”. The canopy at this site is open and the vegetation is dominated by stems and shoots with very little leafy vegetation present.

58 Chapter 2

A B

Figure 2.3 Photograph of ER 1 (Hemlock) the characteristic cobble beds (A) and interspersing pools (B).

ER 2 (Elands River site 2 – Ryton Estate) The site is positioned above the Ngodwana Mill, above the runoff and potential leaching zone of Sappi Farm (Figure 2.4). This site is positioned just below the agricultural activities of Ryton Estates and will serve as the pulp and paper activities’ “reference” site on the Elands River. The pulp and paper activities reference sites are those sites which are strategically positioned upstream of any areas which are perceived to be impacted by activities related to the pulp and paper industry. These “reference” sites will be used to assess the nature, magnitude and relevance of contributing activities/impacts to the effects of the pulp and paper activities on the aquatic ecosystems being assessed in-order to omit these impacts from the other multiple impacts on the system. The site shows the characteristic riffle and rapid habitat with interspersing pools. Substrate is dominated by cobbles and boulders with very little sediment available. The water is clean and clear and flow is generally categorised as “fast shallow”. The canopy at this site is open and the vegetation is dominated by stems and shoots with very little leafy vegetation present.

59 Chapter 2

Figure 2.4 Photograph of ER 2 (Ryton Estates) and the site too show the typical cobble beds and pools.

ER 3 (Elands River site 3 - Bambie Bridge) This site is positioned below the Ngodwana Mill (Figure 2.5) below all aquifer discharge points in the vicinity of the Ngodwana Mill (discharge springs/eye’s). This site is strategically positioned to monitor the effects and response of the aquatic ecosystem to the activities relating to the pulp and paper industry. This is a current Sappi water quality monitoring site, and flow and historical biomonitoring data are available. The characteristic riffle and rapid habitat with interspersing pools is still present at this site, although the pools become larger and deeper and the cobble beds shorter. Substrate is still dominated by cobbles and boulders, but the availability of sediment increases. Flow changes from generally “fast shallow” to “fast deep”. The canopy at this site is open and the vegetation is dominated by stems and shoots with very little leafy vegetation present.

Figure 2.5 The photograph shows site ER 3 (Bambi bridge) down stream of the Ngodwana Mill and springs.

60 Chapter 2

ER 4 (Elands River site 4 before the Lupelule River confluence) This is the second monitoring site on the Elands River below the paper Mill, positioned above the confluence of the Elands and Lupelule rivers (Figure 2.13). This site will characterise the diluting and ecosystem recovery in the Elands River from site ER 3. This site characterises the contribution of discharge from the Lupelule River into the effected area (Ngodwana activities) of the Elands River. The characteristic riffle and rapid habitat with interspersing pools is still present at this site, although the pools become larger and deeper and the cobble beds shorter (Figure 2.6). Substrate is still dominated by cobbles and boulders with limited sediment availability. Flow at this site is predominantly “fast deep”. The canopy at this site is open and the vegetation is dominated by stems and shoots with very little leafy vegetation present.

A B

Figure 2.6 Photograph of site ER 4, before the confluence with the Lupelule River (A) upstream and (B) downstream.

ER 5 (Elands River site 5 - Lindenau Site) This site is positioned above the Lindenau gauging station on the Elands River (Figure 2.7). For similar reasons to site ER 4 this site is positioned above the confluence of the Elands and Crocodile rivers. This is a current water quality monitoring site (Sappi and DWAF) with historical flow and biomonitoring data available. The characteristic riffle and rapid habitat is present at this site, yet very little pool areas are present. Substrate is dominated by cobbles and boulders with limited sediment availability. Flow at this site is predominantly “fast deep” and the river widens at this site. The canopy at this site is open and the vegetation is dominated by stems and shoots with very little leafy vegetation present.

61 Chapter 2

A B

Figure 2.7 Photographs of site ER 5 at Lindenau, above the Lindenau weir. Photograph (A) facing downstream in the direction of the weir and (B) upstream.

HR (Lupelule River) This site is positioned on the Lupelule River, above the confluence of the Lupelule and Elands Rivers (Figure 2.8). This site will serve as a reference site to determine any possible influences that may enter the Elands River from this river. The area in which the site is situated is more commonly known as the Houtboschoek area. Therefore the sites will be referred to as HR in this study. The characteristic riffle and rapid habitat is present at this site with very shallow pools. Substrate is dominated by cobbles and boulders with high sediment availability. Flow at this site is predominantly “fast shallow”. A canopy covers the river and the vegetation is dominated trees as the river flows through a plantation. The vegetation consists largely of stems and shoots with very little leafy vegetation present.

A B

Figure 2.8 Photographs of site HR , positioned on the Lupelule River before the confluence with the Elands River, showing (A) upstream and (B) downstream areas.

62 Chapter 2

NR 1 (Ngodwana River site 1) This site is positioned well above the Ngodwana Dam and is proposed as an additional “pristine – reference” site for the Elands River (Figure 2.13). Other than forestry plantations the Ngodwana River is clear of any anthropogenic impacts. A reconnaissance survey carried out in 2004 indicated that this river is in a fairly natural state and can be regarded as pristine. As mentioned earlier this might become the reference site to which all other sites will be compared in this study. It currently serves as a fish sanctuary for the endangered Chiloglanis bifurcus species. The site shows the characteristic riffle and rapid habitat with interspersing pools. Substrate is dominated by cobbles and boulders with very little sediment available (Figure 2.9). The water is clean and clear and flow is generally “fast shallow”. The canopy at this site is open and the vegetation is dominated by stems and shoots with very little leafy vegetation present.

Figure 2.9 Photographs of site NR 1, above the Ngodwana Dam

NR 2 (Ngodwana River site 2) This site is positioned below the Ngodwana Dam just above the confluence of the Ngodwana and Elands rivers (Figure 2.13). This site in conjunction with Site NR 1 will be used to characterise the contribution of the Ngodwana River to the Elands River. The Ngodwana Dam has a significant impact on this site, potentially altering the biological template functioning at this site. The site shows none of the characteristic riffle and rapid habitat or interspersing pools. Substrate is dominated by cobbles and boulders with very little sediment available. The water is turbid and flow is generally “fast shallow”. The canopy at this site is open and the vegetation is dominated by stems and shoots with very little leafy vegetation present.

63 Chapter 2

Figure 2.10 Photograph of site NR 2 below the Ngodwana dam wall.

CR 1 (Crocodile River site 1 - Poplar Creek) This site is positioned on the upper Crocodile River. It is positioned above the confluence of the Crocodile and Elands rivers (Figure 2.13) and falls within a different bioregion. This is a current water quality monitoring site (Reserve Determination) and historical flow and biomonitoring data is available. This site is positioned to characterise the integrity state of the Crocodile River before the Elands River enters the Crocodile River. The site shows the characteristic riffle and rapid habitat with interspersing pools. Substrate is dominated by cobbles and boulders with very little sediment available. The water is clean and clear and flow is generally “fast shallow”. The canopy at this site is open and the vegetation is dominated by stems and shoots with very little leafy vegetation present.

A B

Figure 2.11 Photographs of CR 1, the upper Crocodile River site, situated above the confluence with the Elands River, facing (A) downstream and (B) upstream.

64 Chapter 2

CR 2 (Crocodile River site 2 - Rivulets) This site on the Crocodile River is situated below the confluences of the Elands and Houtbosloop rivers (Figure 2.12). This site acts as the final monitoring site for the activities relating to the Ngodwana Mill. This is a current water quality monitoring point (Sappi). Historical flow and biomonitoring data are available and the site shows the characteristic riffle and rapid habitat with interspersing pools. Substrate is dominated by cobbles and boulders with very little sediment available. The water is clean and clear and flow is generally “fast shallow”. The canopy at this site is open and the vegetation is dominated by stems and shoots with very little leafy vegetation present.

Figure 2.12 Photographs of CR 2, the second Crocodile Site, situated after the confluence with the Elands River.

65 Chapter 2

Figure 2.13 Location of the sampling sites on the Elands and Crocodile Rivers and various tributaries.

66 Chapter 2

2.6 REFERENCES

Angemeier PL and Karr JR (1994). Biological integrity versus biological diversity as policy directives, protecting biotic resources. Bioscience. Vol. 44 pp 690 – 697.

Ballance A, Hill L, Roux DJ, Silberbauer M and Strydom W. (2001). State of the Rivers report: Crocodile, Sabie – Sand and Olifants River systems. WRC Report No. TT 147/01. Water Research Commission, Pretoria.

Barling, RD and Moore ID (1991). Role of buffer strips in management of waterway pollution: a review. Environmental Management. Vol. 18 pp 543–558.

Berkman HE, Rabeni CF and Boyle TP (1986). Biomonitors of Stream Quality in Agricultural Areas: Fish versus Invertebrate. Environmental Management. Vol. 10 (3) pp. 413-419.

Claassen M (2005). Draft Integrated Report: The Ecological Reserve and the Economic Value of the Aquatic Ecosystem in the Elands River. Contract Report for Sappi-Ngodwana, Submitted to the Department: Water Affairs and Forestry, by the Division of Water Environment and Forestry Technology, CSIR, Pretoria. Report No. ENV-P-C 2005.

Colvin C and Engelbrecht P (2005). Elands River reserve determination: phase 2, Final report. CSIR Environmentek report number ENV-S-C 2004-032.

Dale VH and Beyeler SC (2001). Challenges in the development and use of ecological indicators. Ecological Indicators. Vol. 1 pp 3 – 10.

Dallas HF and Day JA (1993). The effect of water quality variables on riverine ecosystems: A review. WRC report No TT 61/93. Water Research Commission, Pretoria.

Dallas HF (2004a). Spatial variability in macroinvertebrate assemblages: comparing regional and multivariate approaches for classifying reference sites in South Africa. African Journal of Aquatic Science. Vol. 29 (2) pp 161 – 171.

67 Chapter 2

Dallas HF (2004b). Seasonal variability of macroinvertebrate assemblages in two regions of South Africa: implications for aquatic bioassessment. African Journal of Aquatic Science. Vol. 29 (2) pp 173 – 184.

Davies CJ, Hart OO and Buckley CA (1992). An investigation into water and effluent management in the pulp and paper industry: technical performance evaluation of the pilot plant treatment of pulp bleach effluents by cross – flow microfiltration and reverse osmosis. WRC report No. 188/1/92.Water Research Commission, Pretoria.

Dickens CWS and Graham PM (2002). The South African Scoring System (SASS), Version 5, Rapid bioassessment method for rivers. African Journal of Aquatic Science. Vol. 27 pp 1 – 10.

DWAF (Department of Water Affairs and Forestry) (1991). Water quality management policies and strategies in the Republic of South Africa. Department of Water Affairs and Forestry, Pretoria.

DWAF (Department of Water Affairs and Forestry) (1996a). South African water quality guidelines. Vol. 7: Aquatic ecosystems. Department of Water Affairs and Forestry, Pretoria.

DWAF (Department of Water Affairs and Forestry) (1996b). South African water quality guidelines. Vol. 4: Agricultural Use for Irrigation. Department of Water Affairs and Forestry, Pretoria.

DWAF (Department of Water Affairs and Forestry) (1996c). South African Water Quality Guidelines (second edition), Vol. 3: Industrial Use. Department of Water Affairs and Forestry, Pretoria.

DWAF (Department of Water Affairs and Forestry) (1999). Resource Directed Measures for Protection of Water Resources. Vol. 3: River Ecosystems Version 1.0. Department of Water Affairs and Forestry, Pretoria.

DWAF (Department of Water Affairs and Forestry) (2003). Resource Directed Measures. Module 1: Introductory module.Department of Water Affairs and Forestry, Pretoria.

68 Chapter 2

DWAF (Department of Water Affairs and Forestry) (2004b). Draft Background Information: South African Forestry Pulp and Paper Industries: Draft Version 1.0. Department of Water Affairs and Forestry, Pretoria.

Godfrey LK and Roux DJ (2000). Intermediate Reserve Determination for the Elands River Catchment, Incomati System, Mpumalanga. Technical Report for the Department Water Affairs and Forestry Technology, Report NO. ENV – P – C 2000, CSIR, Pretoria pp 1 – 134.

Golder Associates (2004). Draft environmental impact report: Proposed expansion of the Sappi Ngodwana mill, near Nelspruit, Mpumalanga, Vol. 1 of 2.

Greenfield DI and Bart HL (2005). Long-term fish community dynamics from a blackwater stream receiving kraft mill effluent between 1973 and 1988. Hydrobiologia. Vol. 534 pp 81–90.

Griffith MB, Hill BH, McCormick FH, Kaufmann PR, Herlihy AT, Selle AR (2005). Comparative application of indices of biotic integrity based on periphyton, macroinvertebrates, and fish to southern Rocky Mountain streams. Ecological Indicators. Vol. 5 pp 117–136.

Hall T, Haley RK, La Fleur LE (1991). Effects of biologically treated bleached kraft mill effluent on cold water stream productivity in experimental stream channels. Environmental Toxicology and Chemistry. Vol. 10 pp 1051 – 1060.

Hill L (2005). Elands Catchment Comprehensive Reserve Determination Study, Mpumalanga Province, Ecological Classification and Ecological Water Requirements (quantity) Workshop Report, Contract Report for Sappi-Ngodwana, Submitted to the Department Water Affairs and Forestry, by the Division of Water Environment and Forestry Technology, CSIR, Pretoria. Report No. ENV-P-C 2004-019 pp 1 -98.

Hohls DR (1996). National biomonitoring program for riverine ecosystems: Framework document for the program. NBP Report series No 1. Institute for water quality studies,

69 Chapter 2

Department of Water Affairs and Forestry, Pretoria, South Africa.

James NPE and Barber HM (1991a). A survey of the fishes of the Elands and Crocodile Rivers in the vicinity of the Sappi Kraft pulp and paper mill at Ngodwana, Eastern Transvaal. South African Institute of Aquatic Biodiversity. Investigational report No. 37.

James NPE and Barber HM (1991b). A survey of the fishes of the Elands and Crocodile Rivers in the vicinity of the Sappi Kraft pulp and paper mill at Ngodwana, Eastern Transvaal. South African Institute of Aquatic Biodiversity. Investigational report No. 37.

Jungwirth M, Muhar S and Schmutz S (2000). Fundamentals of fish ecological integrity and their relation to the extended serial discontinuity concept. Hydrobiologia. Vol. 422/423 pp 85– 97.

Karels AE and Niemi A (2002). Fish community responses to pulp and paper mill effluents at the southern Lake Saimaa, Finland. Environmental Pollution. Vol. 116 pp 309–317.

Karels AE, Markkula E and Oikari A (2001). Reproductive, biochemical, physiological, and population responses in perch (Perca fluvatilis L.) and roach (Rutilus rutilus L.) downstream of two elemental chlorine-free pulp and paper mills. Environmental Toxicology and Chemistry. Vol. 20 pp 1517–1527.

Karr JR and Chu EW (2000). Sustaining living rivers. Hydrobiologia. Vol. 422/423 pp 1 – 14.

Kefford BJ, Palmer CG, Pakhomova L and Nugegoda D (2004). Comparing test sytems to measure salinity tolerances of freshwater invertebrates. Water SA. Vol. 30 (4) pp 499 – 506.

Kemper NP (2001). RVI: Riparian Vegetation Index. WRC report NO 850/3/01 Water Research Commission, Pretoria, South Africa.

Kerans BL and Karr J (1994). A benthic index of biotic integrity (B-IBI) for rivers of the Tennessee Valley. Ecological Applications. Vol. 4 pp 768 – 785.

70 Chapter 2

Khan RA. and Payne JF (2002). Factors influencing EROD activity in feral winter flounder (Pleuronectes americanus) exposed to effluent from a pulp and paper mill in Newfoundland. Bulletin of Environmental Contamination and Toxicology. Vol. 68 pp 791–800.

Kleynhans CJ (1996). A qualitative procedure for the assessment of the habitat integrity status of the Luvuvhu river (Limpopo System, South Africa) Journal of aquatic ecosystem health. Vol. 5 pp 1 – 4.

Kleynhans CJ (1999). The development of a fish index to assess the biological integrity of South African rivers. Water SA. Vol. 25 (3) pp 265 – 278.

Kleynhans CJ (2005). River EcoClassification. Manual for EcoStatus Determination (Version 2). MODULE D: Fish Response Assessment Index (FRAI), KV 168/05. Department of Water Affairs and Forestry, Pretoria.

Kotze PJ, Steyn GJ, du Preez HH and Kleynhans CJ (2004). Development and application of a fish based Sensitivity – weighted Index of Biotic Integrity for use in the assessment of biotic integrity of the Klip River, Gauteng, South Africa. African Journal of Aquatic Science. Vol. 29(2) pp 129 – 43.

Kovacs TG, Martel PH and Voss RH (2000). Assessing the biological status of fish in a river receiving pulp and paper mill effluents. Environmental Pollution. Vol. 118 pp 123–140.

71 Chapter 2

Lowell RB, Culp JM, Wrona FJ (1995). Stimulation of increased short term growth and development of mayflies by pulp mill effluent. Environmental Toxicology and Chemistry. Vol. 14 pp 1529 – 1541.

Malan H, Bath A, Day J and Joubert A (2003). A simple flow – concentration modeling method for integrating water quality and quantity in rivers. Water SA. Vol. 29 (3) pp 305 – 311.

McMillan PH (1998). An integrated habitat assessment system (IHAS Version 2) for the rapid biological assessment of rivers and streams. CSIR Research Report No. ENV-P-I 98132, Water Resources Management Programme, Council for Scientific and Industrial Research, Pretoria.

Meyer JS (2002). The utility of the terms “bioavailability” “bioavailable fraction” for metals. Marine Environmental Research. Vol. 53 pp 417 – 423.

Morrison G, Fatoki OS, Persson L and Ekberg A (2001). Assessment of the impact of point source pollution from the Keiskammahoek Sewage Treatment Plant on the Keiskamma River – pH, electrical conductivity, oxygen – demanding substances (COD) and nutrients. Water SA. Vol. 27 (4) pp 475 – 480.

Muhar S, Schwarz M, Schmutz S and Jungwirth M (2000). Identification of rivers with high and good habitat quality: methodological approach and applications in Austria. Hydrobiologia. Vol. 422/423 pp 343–358.

Munkittrick KR, Van der Kraak GJ, McMaster ME, Portt CB (1992). Response of hepatic MFO activity and plasma sex steroids to secondary treatment of BKME and mill shut – down. Environmental Toxicology and Chemistry. Vol. 11 pp 1427 – 1439.

O’Brien GC (2003). An ecotoxicological investigation into the ecological integrity of a segment of the Elands River, Mpumalanga, South Africa. M.Sc. dissertation, Rand Afrikaans University, Johannesburg, South Africa.

Ollis DJ, Boucher C, Dallas HF and Esler KJ (2006). Preliminary testing of the Integrated Habitat Assessment System (IHAS) for aquatic macroinvertebrates. African Journal of Aquatic Science. Vol. 31(1) pp 1–14.

72 Chapter 2

Owens, JW (1991). The hazard assessment of pulp and paper effluent in the aquatic environment: a review. Environmental Toxicology and Chemistry. Vol. 10 pp 1511 – 1540.

Palmer CG, Goetsch PA, O’Keeffe JH (1996). Development of a recirculating artificial stream system to investigate the use of macro – invertebrates as water quality indicators. WRC report No 475/1/96. Water Research Commission, Pretoria, South Africa.

Palmer CG, Rossouw N, Muller WJ and Scherman PA (2005). The development of water quality methods within ecological reserve assessment, and links to environmental flow. Water SA. Vol. 31 (2) pp 161 – 170. Pinto BCT, Araujo FG and Hughes RM (2006). Effects of landscape and riparian condition on a fish index of biotic integrity in a large southeastern Brazil river. Hydrobiologia. Vol. 556 pp 69–83.

Pusceddu A, Sar`a G, Armeni1 M, Fabiano M and Mazzola A (1999). Seasonal and spatial changes in the sediment organic matter of a semi-enclosed marine system (W-Mediterranean Sea). Hydrobiologia. Vol. 397 pp 59–70.

Robinson RD, Carey JH, Solomon KR, Smith IR, Servos MR, Munkittrick KR (1994). Survey of receiving – water environmental impacts associated with discharge from pulp mills. 1. Mill characteristics, receiving – water chemical profiles and laboratory toxicity tests. Environmental Toxicology and Chemistry. Vol. 13 pp 1075 – 1088.

Roth NE, Allan DJ and Erickson DL (1996). Landscape influences on stream biotic integrity assessed at multiple spatial scales. Landscape Ecology. VOL. 11 (3) pp 141-156.

Roux DJ (2001). Development of the procedures for the implementation of the national River Health Programme in the province of Mpumalanga. WRC report No 850/1/01. Water Research Commission, Pretoria, South Africa.

Roux DJ, Jooste, SHJ and MacKay HM (1996). Substance – specific water quality criteria for the protection of South African freshwater ecosystems: methods for derivation and initial results for some inorganic toxic substances. South African Journal of Science. Vol. 92 pp198 – 205.

73 Chapter 2

Sepulveda MS, Ruessler DS, Denslow ND, Holm SE, Schoeb TR and Gross TS (2001). Assessment of reproductive effects in largemouth bass (Micropterus salmoides) exposed to bleached/unbleached kraft mill effluents. Archives of Environmental Contamination and Toxicology. Vol. 41 pp 475–482.

Sibley PK, Legler J, Dixon DG, Barton DR (1997). Environmental health assessment of the benthic habitat adjacent to a pulp mill discharge. I. Acute and chronic toxicity to benthic macroinvertebrates. Archives of Environmental Contamination and Toxicology. Vol. 32 pp 274 – 284.

Skelton PH (2001). A field guide to the freshwater fishes of Southern Africa. Southern publishers (Pty.) Ltd., Halfway House pp 1 – 395.

Thirion C (2005). River EcoClassification. Manual for EcoStatus Determination (Version 2). MODULE E: Macro Invertebrate Response Assessment Index (MIRAI), KV 168/05. Department of Water Affairs and Forestry, Pretoria.

USEPA (United States Environmental Protection Agency) (1986). Recommended protocols for measuring conventional sediment variables in Puget Sound. USEPA, Cincinnati, Ohio.

Walling DE and Moorehead PW (1989). The particle size characteristics of fluvial suspended sediment: an overview. Hydrobiologia. Vol. 176/177 pp 125-149.

Van Wijk DJ, Hutchinson TH (1995). The ecotoxicity of chlorate to aquatic organisms: a critical review. Ecotoxicology and environmental safety. Vol. 32 pp 244 – 253.

74 Chapter 2

CSIR, Pretoria.

Williams ML, Palmer CG, Gordon AK (2003). Riverine macroinvertebrate responses to chlorine and chlorinated sewage effluents – acute chlorine tolerance of Baetis harrisoni (Ephemeroptera) from two rivers in Kwazulu – Natal, South Africa. Water SA. Vol. 29 (4) pp 483 – 488.

Zokufa TS, Scherman PA, Palmer CG (2001). Tolerance of selected riverine indigenous macroinvertebrates from the Sabie River (Mpumalanga) and the Buffalo River (Eastern Cape), to complex saline kraft and textile effluents. WRC report No. 783/1/01. Water Research Commission, Pretoria

CHAPTER 3

WATER QUALITY

3.1 INTRODUCTION

As the demand for water increases, with increasing human populations and economic development, so to does the pollution of our river ecosystems. Users such as domestic, agricultural, recreational and industrial sectors all depend on fresh flowing water (Roux et al., 1996). To ensure the existence of water as a resource, and the biodiversity that relies on the resource, the quality of water becomes extremely important. Add to this the social importance of water and water quality can be seen as one of the most important factors to be considered when it comes to conservation and sustainability of this valuable resource. Water quality can be defined as the combined effects of the physical attributes and the chemical constituents of an aquatic ecosystem (Palmer et al., 1996). However, it is important to note that water quality may not only be changed by pollution and according to Malan et al. (2003), alterations in flow may cause changes to water quality. These changes are nonetheless, complex and difficult to predict.

The Elands River as a resource has immense importance (Godfrey and Roux, 2000). Not only are there several users of water in the study area, but the Elands River carries with it a high ecological importance and sensitivity. Several species of fish occur in the river that is sensitive to changes in flow and water quality. One of these species is the endangered Chiloglanis bifurcus that depends on clean water through all stages of its life cycle (Godfrey and Roux, 2000). Several activities in the vicinity of the river may have an adverse effect on the water quality within the Elands River. These activities include: the sewage treatment facilities that are found in the upper reaches near Waterval Boven and at the Ngodwana village, the rural settlements near Waterval Boven and Lindenau, agricultural activities upstream of the Sappi Ngodwana kraft and paper mill and the mill and its activities itself. Sewage discharge is a major threat to the ecological integrity of any river system and it may cause increased oxygen demand and nutrient loading within a river. This can lead to toxic algal blooms and degradation in the integrity and health of a system (Morrison et al., 2001). The potential effect of agricultural activities depends on the type and the extent of agriculture. Runoff, both surface and subsurface in nature, often causes deterioration of water quality by increasing concentrations of plant nutrients (Dallas and Day, 1993). These nutrients include nitrogen and phosphate.

76 Chapter 3

Pulp and paper making requires large quantities of water and water is used as the transport medium in pulp and paper mills and, in chemical pulping, the pulping reactions occur in an aqueous medium. As a consequence, contaminated water is produced (DWAF, 2004). Effluent from this mill is not released in the river directly. It is irrigated on to pasture lands from where it drains into the aquifer. Irrigation of effluent at the Ngodwana Mill started in 1967, while the mill only started monitoring effluent quantity and quality by 1985. According to Zokufa et al. (2001) the groundwater that is affected by the irrigation project enters the Elands River via 3 eyes i.e. Fraser’s Eye, Eye X and Northern Eye and a diffuse seepage front along the length of the Elands River immediately adjacent to the irrigation area. One of the major problems this activity creates is an increase in the salinity, especially due to high concentrations of chlorine. Chlorine, in the past, played a key part in the bleaching process. Water quality in the Elands River has been closely monitored by the management of the Ngodwana Mill. A complete report was compiled by the CSIR in 1991 for the Ngodwana Mill and the effect that their activities have on the water quality of the Eland River (Weddepohl et al., 1991).

The objective of this chapter is to assess the current status of the water quality in the Elands River. This involves an assessment of physical and chemical variables at selected sites.

3.2 MATERIAL AND METMETHODSHODS

3.2.1 FIELD SURVEYS

Water samples were collected at the sites described in chapter 2 (Figure 2.13) during high (March 2005) and low flow conditions (June 2005). Water samples were collected in acid- washed polyethylene bottles from each site. The polyethylene bottles were rinsed with water from the site before a sample was taken and after collection, these samples were stored at –4 ºC until further analysis. Along with the collection of the samples, in situ measurement of the following variables was undertaken using a WTW 340i Multi meter:

• Oxygen saturation (%) • Dissolved oxygen concentration (mg/l) • Temperature (o C) • pH • Conductivity (µS / cm)

77 Chapter 3

3.2.2 LABORATORY ANALYSIS

Water samples were frozen and transported to the laboratory. After defrosting a Merck Spectroquant Spectrophotometer and relevant test kits were used to measure various chemical variables. These variables were: • chemical oxygen demand (COD) • ammonium, • chloride, • nitrates, • nitrites, • sulphates • ortho – phosphates • turbidity.

3.3.3 STATISTICAL ANALYSIS

Historical data was obtained from DWAF for the gauging stations at Geluk (X2H011Q01), Lindenau (Figure 3.1) (X2H015Q01) and Montrose (X2H013Q01) as well as from the Sappi, Ngodwana Mill. The gauging stations are situated above ER1 (Geluk) and below ER 5 (Lindenau). Labware laboratories (in house analysis by Sappi) complete weekly water quality tests on behalf of the Ngodwana Mill at several sites. Some of these sites correspond with sites in this study. These sites include sampling points at Hemlock (ER1), Lindenau (ER5) and Rivulets (CR2). These sites are key sites in this study as they represent water quality above the Ngodwana Mill, below the Ngodwana Mill and water quality of the Crocodile River after the confluence with the Elands River.

Canoco version 4.5 was used to complete ordination of the sampling sites. The Principle Component Analysis (PCA) is based on a linear response model relating species and environmental variables (Van den Brink et al., 2003). Results of the ordination is a map of the samples being analysed on a 2 dimensional bases, where the placements of the samples reflect the (dis)similarities between the samples; in this case the sampling sites.

78 Chapter 3

Figure 3.1 The DWAF gauging stations at Lindenau (X2H015Q01) on the Elands River.

3.3 RESULTS AND DISCDISCUSSIONUSSION

Water from the Elands River is important for several users (Kleynhans, 1999). These include:

• Agricultural use: citrus and vegetable farmers depend on the Elands River as a source of water for irrigation purposes. In the Crocodile River there are several agricultural users, including tobacco farmers that have specific water quality needs. • Domestic use: several rural settlements have originated in the Lindenau area. This local community is dependent on the Elands River to fulfill its basic human needs. The Crocodile River supplies water to several towns and cities. • Recreational use: The Elands River has become a favourite fishing destination amongst fly fishermen. Several holiday resorts and camp sites have been developed on the banks of the river. • Aquatic ecosystem: The most important user of water in the Elands River is ultimately the aquatic ecosystem. The Elands River has great ecological importance as it houses fish species that are critically endangered.

79 Chapter 3

ER 1 Table 3.1 and Table 3.2 show that there is no serious degradation in water quality at this site. The only concern is the higher COD (14 mg/l) during the high flow survey. Table 3.3 also shows this elevated COD 14.2 mg/l (± 6.66). It is possible that this elevation is brought about by the presence of sewage from the Waterval Boven area. It is a known fact that the COD can increase in a river due to sewage discharge (Morrison et al., 2001; Nhapi and Tirivarombo, 2004). O’Brien (2003) came to the conclusion that sewage spills in the upper reaches of the Elands River are occurring and that it may contribute to this higher COD and even elevated nutrient levels. Elevated nutrient levels are usually due to sewage are usually seen as phosphate, nitrites, nitrates and ammonium. None of these variables were found to be in an elevated state during the high or low flow period (Tables 3.1 and 3.2) at this site. Historical data (Figure 3.2) does show that in the last decade there has been an increase in the levels of nitrates, ammonium and phosphates above this site.

This historical data was obtained from the DWAF gauging station (X2H011Q01). This station is situated at Geluk, below the town of Waterval Boven. These values do however, fall within the target water quality range (TWQR) for all water users. The TWQR is the concentration of a substance at which it will have no effect on the particular user. It is used by DWAF to determine guidelines for water quality with regards to substances and variables. Although elevated, the observed COD value falls with the TWQR for all users, except industrial users that falls within category I (DWAF, 1996a). Seeming that there is no, and possibly never will be category I industry (see Table 3.5) within the study area this is of little importance. The presence of elevated level of faecal coliforms (76.42 cfu / 100 ml) strongly point to the presence of sewage. Furthermore, the faecal coliform count is above the recommended range (0 counts per 100 ml) for domestic use. At these concentrations, as observed during high flow conditions (Table 3.3), there is a significant risk of infectious disease transmission Water temperature at this site was higher (270 C) than reference conditions. The reference conditions for the Elands River were determined during the intermediate ecological reserve determination study (Godfrey and Roux, 2000).

This elevated temperature change may cause death in extreme cases, but with lower ranges it may ultimately influence the movement, migration and even behaviour of organisms. At a population level, temperature changes can influence population density, diversity and abundance (Palmer et al., 1996). The dissolved oxygen concentration at this site was 6.67 mg/l

80 Chapter 3

(Table 3.1) during high flow conditions while it was measured at 9.74 mg/l during low flow conditions (Table 3.2). These concentrations fall within the reference conditions for these time periods. These conditions are set at 6.3 – 9.5 mg/l for high flow and 7.3 – 10.9 mg/l for low flow conditions (Table 3.6). The increase in oxygen during low flow conditions is mainly due to the decrease in temperature. Lower temperatures and salinity increase the solubility of oxygen in water (Palmer et al., 1996).

The conductivity at this site during high flow was 156 µs/cm and at low flow 164 µs/m (Tables 3.1 and 3.2). This exceeded the reference conditions of 97 µs/cm and 127 µS/cm for high and low flow respectively (Table 3.7). These values are, however, not in excess of 230 µs/cm and 300 µs/cm for high and low flow conditions as set by reserve conditions (Table 3.9). Historical data shows that the average conductivity at this site is 189 µS/cm above reference conditions for this resource unit (Figure 3.2). An increase in the dissolved material (and thus the conductivity) in a river environment may be attributed to the sewage discharged, return flows from agricultural activities, and the effluent from various industries (Morrison et al., 2001). Data from the Ngodwana Mill indicates that the conductivity at this site during high flow conditions (Table 3.3) was 129.1 µs/cm (± 53.2). Although in excess of reference conditions (Table 3.7) these values are not in excess of reserve conditions (Table 3.9). During low flow conditions the data from the mill indicated that the average conductivity at this site was 208 µs/cm (± 25.8). This value falls within both reference and reserve conditions.

The pH at this site, as with temperature, did not fall within the reference conditions set for this resource unit. Reference conditions for pH are set at 7.07 (Table 3.8). In both high flow (8.78) and low flow (7.97) conditions these values were exceeded (Tables 3.1 and 3.2). It is a well known fact that an increase in pH may adversely affect biota (Morrison et al., 2001). This is due to the fact that it may influence the toxicity of certain substances. Free ammonia (NH3), for example, becomes far more toxic to aquatic biota at a pH of more than 8.5 (DWAF, 1996d). This increase in pH is not a common occurrence but may be attributed to anthropogenic eutrophication in the study area (Dallas and Day, 1993).

81 Chapter 3 Table 3.1 Nutrient and system variables results for the high flow survey (March 2005). Data (in mg/l unless otherwise stated) represents results of once of sampling. Ortho - Turbidity Ammonium Chloride COD Nitrate Nitrite Sulphates Oxygen Oxygen Temperatur Conductivity pH phosphates (NTU) (%) e (ºC) (µs/cm)

ER1 <0.01 1 14 <0.25 0.01 < 0.01 8 1 6.67 83.9 27 156 8.78 ER2 <0.01 0 4 <0.25 < 0.01 < 0.01 7 2 6.26 77.3 22.8 161 8.38 ER3 <0.01 75 4 <0.25 < 0.01 0.02 21 4 7.41 89.4 22.2 481 8.11 ER4 <0.01 52 0 <0.25 < 0.01 0.01 37 7 6.45 83.5 23.3 448 8.6 ER5 <0.01 61 0 <0.25 < 0.01 0.01 60 2 7 87.6 23.2 443 8.67 HR <0.01 0 3 0.72 < 0.01 < 0.01 3 3 7.24 80.6 19.8 200 7.79 CR1 <0.01 0 10 0.38 < 0.01 < 0.01 < 25 6 5.29 67 22.5 219 7.7 CR2 <0.01 35 2 <0.25 0.01 0.01 5 6 6.5 80.3 21.2 122 8.01 NR2 <0.01 0 0 0.51 < 0.01 0.02 < 25 5 7 82.8 21.3 242 8.03

Table 3.2 Nutrient and system variables results for the low flow survey (June 2005). Data (in mg/l unless otherwise stated) represents results of once of sampling. Ortho - Turbidity Ammonium Chloride COD Nitrate Nitrite Sulphates Oxygen Oxygen Temperature Conductivity pH phosphates (NTU) (%) (ºC) (µs/cm)

ER1 <0.01 3 2 <0.25 < 0.01 0.02 <25 0 9.74 104.9 13.1 164 7.97 ER2 <0.01 4 6 <0.25 0.01 1 <25 2 9.8 110 15.3 167 7.98 ER3 <0.01 60 5 0.39 0.12 0.03 61 5 10.16 115.5 16.7 640 8.14 ER4 <0.01 45 4 <0.25 0.21 0.06 125 5 10.4 114 15.4 588 8.48 ER5 <0.01 30 5 <0.25 0.05 0.05 34 2 9.6 102.5 14.1 597 8.22 HR <0.01 2 0 <0.25 0.02 0.02 <25 8 8.85 99.8 16.1 199 8.07 NR1 <0.01 1 2 <0.25 0.13 0.04 111 6 9.75 103.5 12.3 88 7.83 NR2 <0.01 4 5 <0.25 0.01 0.03 <25 18 8.44 96.5 16.4 152 7.55 CR1 <0.01 3 5 <0.25 0.01 0.03 <25 2 9.85 102.2 12.7 127 8.07 CR2 <0.01 18 5 <0.25 0.02 0.02 <25 11 9.77 102.8 14.2 278 8.02

82 Chapter 3

Table 3.3 Results (mean ± standard deviation) obtained by Labware Laboratories for system and nutrient variables during high flow conditions (3 Jan 2005 – 9 May 2005).

ER1 ER5 CR2

Conductivity (µS/cm) 128.1 (± 53.2) 296.7 (± 96.3) 169.8 (± 40.2)

pH 8.2 (± 0.22) 8.2 (± 0.23) 8.1 (± 0.12)

Chlorides (mg/l) 3.7 (± 2.38) 35.8 (± 9.33) 13.6 (± 2.63)

Faecal Coli forms (cfu/100ml) 74.47 (± 56.90) 76.58 (± 61.36) 172.4 (± 42.51)

COD (mg/l) 14.2 (± 6.66) 10.8 (± 4.66) NA Ammonia (mg/l) 0.4 (± 0.82) 0.17 (± 0.14) NA

Ortho - phosphates (mg/l) 0.0 (± 0.00) 0.0 (± 0.00) NA

Nitrates (mg/l) 0.0 (± 0.06) 0.1 (± 0.08) NA Sulphates (mg/l) 6.7 (± 2.48) 44.9 (± 13.57) 17.9 (± 4.17)

Table 3.4 Results (mean ± standard deviation) obtained by Labware Laboratories for system and nutrient variables during low flow conditions (27 June 2005 – 24 Oct 2005).

ER1 ER5 CR2 Conductivity (µs/cm) 208.8 (± 25.8) 792.1 (± 112.5) 239.5 (± 21.0)

pH 8.2 (± 0.19) 8.3 (± 0.16) 8.1 (± 0.08)

Chlorides (mg/l) 4.6 (± 0.54) 77.2 (± 11.18) 18.0 (± 3.48)

Sulphate (mg/l) 5.5 (± 0.44) 73.9 (± 18.76) 16.8 (± 1.79)

Faecal Coli forms (cfu/100ml) 17.29 (± 18.87) 47.85 (± 42.42) 149.33 (± 37.75)

Table 3.5 Examples of industry practices that fall within a category I industry.

Heat Exchange Steam Generation Process Water Product Water

Evaporative cooling High pressure boilers Solvent agent Beverages

Solution cooling Heat transfer medium Dairy

Water heating Humidification Petrochemical

Lubrication

83 Chapter 3

ER 2 The conductivity at this site was 161 µs/cm and 16.7 µs/cm during high and low flow conditions respectively (Tables 3.1 and 3.2). This was in excess of reference conditions, but not the reserve conditions (Tables 3.7 and 3.9). The pH during both high flow (8.38) and low flow (7.97) conditions exceed reference conditions (Tables 3.1 and 3.2). As mentioned reference conditions for this resource unit are set at a pH of 7.07 (Table 3.8). Despite this higher pH, these values are still within the TWQR for all users in the study area, including domestic, agricultural and industrial users and the aquatic ecosystem. Although there are no signs of eutrophication at this site, the increase in pH and electrical conductivity does point to some impacts from the agricultural activities.

ER 3 Tables 3.1 and 3.2 indicate that there is a definite change in water quality as the river flows past the mill. The elevation in chloride is of major concern. Chloride is an anion of chlorine. It occurs as chloride in nature and can not occur as chlorine. Although the values of chloride are high at site ER3, these concentrations are still within the water quality guidelines (0 – 100 mg/) for domestic use (DWAF, 1996b). The main source of chlorides within the river is from the springs; however, the Ngodwana sewage plant releases treated sewage near the mill. Chloride is used in the sewage treatment process as an algaecide, bactericide and for the removal of odours. However, this contribution is possibly small. For industrial use the water quality guidelines for industries that fall within category II is of concern. There are however no industries in the vicinity of the Elands River that fall within a category I industry. Activities that do fall within this category are summarised in Table 3.5). It is also unlikely that any industry that falls within category I will operate within the area. The TWQR for category II industries, with regard to chloride, ranges between 0 and 40 mg/l of chloride (DWAF, 1996b). This value is exceeded at site ER3. There is, however, no proof of any such industries directly down stream of the mill (in the vicinity of the affected sites). The TWQR for industries that fall within a category III and IV are set at much higher values (0 – 100 mg/l for category III and 0 – 500 for category IV). Should the water quality objectives be met for category II industries, the TWQR for these categories will automatically be met.

One of the major concerns, with regards to chloride concentrations, is related to agriculture. Farming activities with the vicinity of the Elands River mostly consists of citrus farming. The

84 Chapter 3 chloride TWQR for crop yield and quality is set at <140 mg/l (DWAF, 1996c). At no point in the Elands River is this value exceeded. A variation from this TWQR does occur in two cases: when the agricultural activities include avocados and tobacco. The only foreseeable problem is the tobacco industry as some tobacco farms occur downsteam of the Crocodile River. Chloride concentrations as low as 25 mg/l, may effect the burning properties and storage life of tobacco leaves (DWAF, 1996c). According to results this value was exceeded during high flow conditions. Constant monitoring by the Ngodwana Mill (Tables 3.3 and 3.4) during this period does show that chloride concentrations do not exceed 25 mg/l. The adsorption rate of avocado leaves is slower than that of other citrus plants and concentrations of 70 – 105 mg/l of chloride can cause leave foliage in these plants. These values are exceeded at site ER3 during high flow conditions (Table 3.1). As there is no avocado farming in this vicinity of the site, it is not of major concern. Other possible farming activities that might be affected are livestock and the TWQR for chlorides, with regards to livestock is 0 – 1500 mg/l of chloride. The concentration of chlorides falls within this excepted range at all times and at all sites.

Nitrate values during low flow conditions (Table 3.2) are higher in comparison to sites upstream (i.e. 0.39 mg/l). The most likely source of nitrates at this site is the release of treated sewage by the Ngodwana sewage works, although elevated nitrate concentration are often found in groundwater. These higher nitrate values may thus be linked to the groundwater entering the river above the site. Nitrite levels (0.12 mg/l) are also higher at this site, compared to upstream sites during low flow conditions (Table 3.2). Although elevated, the concentration of nitrates and nitrites during low flow conditions does fall within the TWQR for all users. The source of these nitrites is most likely sewage. All other variables fall within the TWQR aquatic ecosystems as well as domestic, agricultural, industrial and recreational users.

Most system variables fall within reference conditions except for pH, temperature and conductivity (Tables 3.1 and 3.2). The pH values of 8.11 during high flow and 8.14 for low flow (Tables 3.1 and 3.2) exceed the reference conditions of 7.07 (Table 3.8). Temperatures are higher then reference conditions for the low flow period (Table 3.7). The effect of drastic temperature changes has already been mentioned. The higher conductivity at this site is due to the elevated chloride concentrations. The conductivity at this site during high flow was 481 µs/cm, while this increased to 640 µs/cm during low flow (Tables 3.1 and 3.2). These values are well above the conductivity values for both reference and reserve conditions (Tables

85 Chapter 3

3.7 and 3.9). This elevated conductivity may have adverse ecological effects on the aquatic biota at this site.

ER 4 As with ER 3, there are elevated chloride concentrations at this site (Tables 3.1 and 3.2). The values range from 52 mg/l during high flow to 45 mg/l during low flow conditions. This does fall within the TWQR for all users in the vicinity of this site except industries that fall within a category II. There are no such industries near this site so this is of no concern. During low flow conditions nitrite levels were also elevated and the concentration of 0.21 mg/l is possibly due to the treated sewage released at ER 3. Although it may be toxic at certain concentrations (Palmer et al., 1996), the concentrations at this site are not of major concern. It may, nevertheless, be contributing to the eutrophication in the system. The elevated chloride levels at this site coincided with a higher conductivity. Tables 3.1 and 3.2 shows that during high flow (448 µs/cm) and low flow (588 µs/cm) the reference and reserve conditions (Tables 3.7 and 3.9) were exceeded.

ER 5 The water quality at this site is of immense importance. It is the last site on the Elands River where water quality is monitored, before the river joins the Crocodile River. Historical data is available from the DWAF gauging station (X2H015Q01) situated at this site. Result obtained during high flow conditions (Table 3.1) indicate that chloride concentrations (61 mg/l) at this site are higher compared to reference sites. During low flow conditions (Table 3.2) these concentrations decreased to 30 mg/l. Water quality monitoring conducted over a 19 week period by the Ngodwana Mill (Table 3.3) shows that during high flow conditions the average chloride concentration was 35.8 mg/l (± 9.33). Monitoring by the mill during low flow conditions over an 18 week period (Table 3.4) shows an increase in chloride concentration to 77.2 mg/l (± 11.18). Although the TWQR for industrial use (category I and II industries) are exceeded, there are no such industries within the vicinity of this site.

Furthermore, historical data indicate that the average chloride concentrations are not in excess of the TWQR for any users (Figure 3.2). Sulphate concentrations (60 mg/l) during high flow (Table 3.1) as with the concentrations measured by the Ngodwana Mill during both high flow and low flow conditions (Tables 3.3 and 3.4) are in excess of the TWQR for category I

86 Chapter 3

industrial users. The conductivity at this site was in excess of both reference and reserve conditions (Tables 3.7 and 3.9). During high flow the conductivity was 443 µs/cm (Table 3.1) and this increased to 597 µs/cm during low flow conditions (Table 3.2). Constant monitoring by the mill at this site shows that over the 19 week period the conductivity at this site averaged 296.7 µs/cm (Table 3.3). All these values are well in excess of reference and reserve conditions. The 18 week period for low flow conditions did, however, show that this value increased to 792.1 µs/cm (± 112.5) (Table 3.4). Historical water quality data shows that the average conductivity at this site over the past 35 years was 166 µs/cm (Figure 3.2). Even this value is above the reference conditions for this resource unit.

The pH values at this site (Tables 3.1, 3.2, 3.3 and 3.4) exceeded reference values for this resource unit during the entire study. Although the reference value of 7.07 (Table 3.8) is exceeded, historical data show that the pH at this site has always been in the vicinity of 7.9 (Figure 3.2). The water temperature at this site during high flow conditions (Table 3.1) exceeded the reference conditions of 21.10 C (Table 3.6) for this resource unit.

Table 3.6 Reference conditions for average monthly water temperature (°C) and dissolved oxygen (mg/l) (Claassen, 2005). The area in blue represents high flow, while green represents low flow.

Resource Unit 1 Resource Unit 2

(X2H011) (X2H015) Temperature (o C) Oxygen (mg/l) Temperature (o C) Oxygen (mg/l)

January 17.8 6.4 – 9.6 21.9 6.2 – 9.2 February 17.7 6.4 – 9.6 20.8 6.4 – 9.6 March 16.6 6.6 – 9.8 21.1 6.3 – 9.5 April 14.6 6.8 – 10.2 18.1 6.7 – 10.1 May 12.5 7.3 – 10.9 14.7 7.2 – 10.8 June 9.5 7.6 – 11.4 12.0 7.3 – 10.9 July 9.5 7.8 – 11.8 12.3 7.3 – 10.9 August 11.0 7.4 – 11.2 13.8 7.0 – 10.4 September 15.2 6.8 – 10.2 16.4 6.6 – 10.0 October 15.9 6.7 – 10.1 18.5 6.5 – 9.7 November 16.9 6.6 – 9.8 21.1 6.3 – 9.5 December 18.1 6.4 – 9.6 21.9 6.2 – 9.2

87 Chapter 3

Table 3.7 Reference conditions for average monthly electrical conductivity (µs/cm) (Claassen, 2005). The area in blue represents high flow, while green represents low flow.

Resource Unit 2

(X2H015) Electrical conductivity 5th %tile (factor of median) 95th %tile (factor of median)

January 113 7.5 150 February 97 8.0 150 March 97 8.5 140 April 110 9.0 130 May 119 .9.0 120 June 127 9.0 115 July 138 9.0 115 August 140 9.0 115 September 140 9.0 115 October 137 8.5 115 November 126 7.5 120 December 111 7.5 130

Table 3.8 Reference conditions for selected water quality variables (pH units and mg/l) (Claassen, 2005).

Resource Unit 1 Resource Unit 2 (X2H011) (X2H015)

5th %tile Median 95th %tile 5th %tile Median 95th %tile pH 6.22 7.115 7.6 6.26 7.07 7.65

-5 -5 NH3 2.74 x 10 0.0002 0.0014 3.09 x 10 0.0002 0.0013

PO4 0.0025 0.007 0.065 0.0025 0.007 0.041

TIN 0.04 0.1 0.31 0.04 0.07 0.345

At a population level, temperature changes can influence population density, diversity and abundance. It is important to note the biotope at which the temperature is measured (e.g. a pool, backwater or in the current) may show different temperatures. Faecal coliform counts for this site was measured at 76.58 cfu/100 ml (± 61.36) during high flow and 47.85 cfu/100 ml (± 42.42) during low flow conditions (Tables 3.3 and 3.4). Although not very high, these values are in excess of the TWQR for faecal coliform count with regards to domestic use. Rural

88 Chapter 3 settlements have developed at this site and the local community is depended on water at this site to fulfil basic human needs. Although these coliform counts are not high they do pose a potential risk for the local community at this site. The rural settlement may be contributing to this increase in faecal coliforms through there daily use of the river.

Table 3.9 Reserve conditions for average monthly electrical conductivity (mS/m) (Claassen, 2005).

Resource Unit 2

(X2H015) Electrical conductivity 5th %tile 95th %tile ms/m January 27 20 40 February 23 18 34 March 23 19 32 April 26 23 34 May 28 25 34 June 30 27 34 July 32 29 37 August 33 30 38 September 33 30 38 October 35 29 40 November 30 22 36 December 26 20 34

HR Conductivity at this site was higher than reference conditions (Table 3.7). Conductivity values were 200 µs/cm and 199 µs/cm during high and low flow periods respectively (Tables 3.1 and 3.2). There are no known impacts above this site so the elevated conductivity values can only be a natural occurrence. The higher concentration in dissolved material is probably due to weathering of the rocks over which the river flows. The only other water quality variable at this site that did no fall within reference conditions is pH during both high flow (7.97) and low flow (8.07) conditions (Table 3.1 and 3.2). Reference conditions are set at a pH of 7.07 (Table 3.8). Despite this higher pH, these values are still within the TWQR for all users in the study area, including domestic, agricultural and industrial users and the aquatic ecosystem. Elevated nitrate concentration also occurred at this site, with nitrate concentration reaching 0.72 mg/l during high flow conditions.

89 Chapter 3

This is higher then the TWQR of 0.6 mg/l for domestic use but falls within the TWQR for other users in the study area. These high nitrate concentrations may be attributed to a slow turnover rate by nitrogen fixating bacteria. Chapter 4 does, however, indicate that the sediments in the Lupelule River have a very high organic content. Where organic concentrations are high one often finds that turnover rates are high and thus does not explain these high concentrations (Howarth et al., 1988).

NR 1 As with the other sites in this study, pH values at this site did not fall within the reference conditions (Table 3.8). pH values at this site were 8.06 and 7.83 during high and low flow conditions respectively (Tables 3.1 and 3.2). Although not within reference conditions these values still are within the TWQR for all users in the study area. High nitrite (0.13 mg/l) and sulphate (111 mg/l) concentrations where found at this site during low flow conditions (Table 3.2). There are no anthropogenic activities above this site and these elevated conditions can only be seen as naturally occurring, possibly due to leaching from the geological formations, in this area. One of the lithostratographic units in the water management area includes the compact sedimentary strata. The residual soils are in general very shallow and significant mineral deposits include gold, arsenic, copper and sulphur (Botha et al., 2003). The latter is a possible source of high sulphate concentrations at this site.

NR 2 During high flow conditions (Table 3.1), nitrate concentrations (0.51 mg/l) where above the TWQR for domestic use. This concentrations still falls within the TWQR for other users in the area. Conductivity at this site was 219 µs/cm during high flow (Table 3.1) and 152 µs/cm during low flow conditions (Table 3.2). Although not in excess of reserve conditions (Table 3.9) these values are in excess of reference conditions (Table 3.7). pH values at this site did not fall within the reference conditions for the study area (Table 3.8). During high flow the pH was 7.7 and 7.55 during low flow conditions (Tables 3.1 and 3.2). Although not within reference conditions these values still are within the TWQR for all users in the study area. During high flow conditions the temperature at this site was 22.5 0 C (Table 3.1), which is higher then the temperatures during reference (Table 3.6) conditions. The same occurrence was observed during low flow conditions when the temperature of 16.4 0 C (Table 3.2) exceeded the reference

90 Chapter 3 conditions for this area during this time period (Table 3.6).The high temperatures during high flow corresponded with a decrease in the oxygen concentrations. This lower oxygen concentration of 5.24 mg/l (Table 3.1) is below the recommended range for the time of year (Table 3.6). The turbidity at this site reached 18 NTU during low flow (Table 3.1). The temperature changes and turbidity problems are most probably related to the reduced flow at this site. The construction of the dam wall has obscured natural flow and this possibly causes the degradation of water quality at this site. The obstruction in flow may be impacting the entire aquatic ecosystem at this site (Deksissa et al., 2003).

CR 1 During high flow conditions (Table 3.1) high nitrate concentrations where observed. The observed concentrations of 0.38 mg/l is higher then the TWQR for domestic use but falls within the TWQR for other users. Along with the high nitrate concentrations, higher COD values where observed (Table 3.1). Although not extremely high the COD value of 10 mg/l possibly indicates that the breakdown of certain chemicals is occurring. These chemicals are most probably from return flow from the numerous agricultural activities in the area. The elevated nitrate and COD levels are possible indicators of sewage pollution (Morrison et al., 2001). During high flow the pH was 8.01 and 8.07 during low flow conditions (Tables 3.1 and 3.2). Although not within reference conditions these values still are within the TWQR for all users in the study area.

91 Chapter 3

Figure 3.2 Statistical results of historical water quality data available from the Geluk gauging station on the Elands River (Website 1).

92 Chapter 3

CR 2 Water quality at this site is extremely important in this study. The site is situated below the confluence with the Elands River, thus any impacts that the Elands River has on the Crocodile River should be reflected by the water quality at this site. During low flow conditions, temperatures at this site were measured at 14 0 C (Table 3.2). This is higher then the reference conditions for temperatures in this resource unit (Table 3.6). Conductivity at this site exceeded reference conditions (Table 3.7). During high flow the conductivity was 242 µs/cm (Table 3.1) while this increased to 278 µs/cm during low flow conditions (Table 3.2). Data obtained from the Ngodwana Mill (Tables 3.3 and 3.4) shows that during high flow the conductivity was 169.8 µs/cm (± 4.02) and during low flow 239.5 µs/cm (± 21.0).

These values are within reserve values (Table 3.9), but shows that during high flow reference conditions (Table 3.7) are definitely exceeded. The values are relatively high when taking into account that further upstream on the Crocodile River (at Montrose) the conductivity is 129.9 µs/cm (Figure 3.2). The pH values where also higher then those of reference conditions (Table 3.8).The values observed during high and low flow conditions (Tables 3.1 and 3.2) were 8.03 and 8.02 respectively. The results obtained from constant monitoring by the Ngodwana Mill indicates that the pH values (8.1) during high and low flow conditions (Tables 3.3 and 3.4) are higher then reference conditions. One of the most important variables to take in consideration at this site is chloride concentrations. The reason for this is that this is the last site where the quality of the water that will be used by the agricultural sector is monitored by the mill. It has been mentioned how important chlorides concentrations are to tobacco farmers further downstream in the Crocodile River. Chloride concentrations were measured at 35 mg/l during high flow (Table 3.1) and 18 mg/l during low flow conditions (Table 3.2). Monitoring by the mill shows that these concentrations where 13.6 mg/l (± 2.63) during high flow and 18.0 (± 3.48) during low flow (Tables 3.3 and 3.4). These results indicated that chloride levels are within the TWQR for all users including the 25 mg/l as required by the tobacco industry. Faecal coliform concentrations were, however, alarmingly high. During high flow conditions these concentrations were 172.4 cfu/100 ml (± 42.51) while during low flow concentrations were 149.33 cfu/100 ml (± 37.75). The rural settlements further upstream of this site may be largely responsible for these elevated coliform concentrations, as there are no signs of any sewage related problems at Montrose (Figure 3.3). This high faecal coliform count makes the water unfit for domestic use and irrigation purposes.

93 Chapter 3

Figure 3.3 Statistical results of historical water quality data available from the Lindenau gauging station on the Elands River (Website 2).

94 Chapter 3

Figure 3.4 Statistical results of historical water quality data available from the Montrose gauging station on the Crocodile River (Website 3).

95 Chapter 3

The water quality relationships between sites are displayed using PCA plots. In a PCA plot each arrow points in the direction of steepest increase of values for the corresponding variable. The angles between arrows indicate the sign of the correlation between the variables: the approximated correlation is positive when the angle is acute and negative when the angle is larger than 90 degrees. The length of the arrow is a measure of fit for the variables. The distance between the sampling sites in the diagram approximates the dissimilarity of their water quality variables as measured by their Euclidean distance.

Statistical analyses of water quality variables were compared temporally (between flow regimes) and spatially (between sites). Figure 3.5 shows that during both flow regimes there is a distinct separation of some of the sites. The most important of these separations, was the separation of sites ER3, ER4 and ER5.

Figure 3.5 PCA plot showing the dissimilarity among sites on the Elands and Crocodile rivers during both high (sites without suffix) and low flow (sites with the suffix – L) regimes based on water quality. This biplot describes 95.9 % of the variation in the data, where 69.7 % is displayed on the first axis, while 35.5 % is displayed on the second axis.

96 Chapter 3

This is important as this indicates that the water quality at sites downstream of the Ngodwana Mill differ from other sites in the study. The variables that drive this distinct separation include chlorides and conductivity. Figure 3.5 further shows that as expected, there is a certain degree of correlation between chlorides and conductivity. These two variables are associated with the activities of the Ngodwana Mill. This is an indication that the mill has a definite impact on the water quality of the Elands River. The high similarity between sites ER 3, ER 4 and ER 5 during both flow regimes indicates that the dilution factor has little effect on the impacts of the mills’ activities. During both high flow (Figure 3.6) and low flow (Figure 3.7) conditions this separation of sites above and below the mill was also obvious. The same spatial trend was observed by O’Brien (2003) during a recent survey on the Elands River. It is important to note that the site CR 2 is grouped separate from the sites down stream of the mill. This indicates that the river shows definite recovery from the changes caused by chlorides and conductivity in the Elands River, although conductivity was still elevated after the confluence with the Elands River. During both high flow (Figure 3.6) and low flow (Figure 3.7) conditions the same variables were responsible for the trends observed, except for CR 2 during high flow and NR 1 during low flow. CR 2 during high flow conditions separated largely due to a high turbidity, while NR 1 during low flow separated due to a high sulphate concentrations. Statistical analysis further shows that the release of treated sewage in the vicinity of the mill had little impact on the water quality of the sites during both high and low flow conditions.

97 Chapter 3

Figure 3.6 PCA plot showing the dissimilarity among sites on the Elands and Crocodile rivers during high flow conditions based on water quality. This biplot describes 95 % of the variation in the data, where 68.9 % is displayed on the first axis, while 26.1 % is displayed on the second axis.

Figure 3.7 PCA plot showing the dissimilarity among sites on the Elands and Crocodile rivers during low flow conditions based on water quality. This biplot describes 98.4 % of the variation in the data, where 75.4 % is displayed on the first axis, while 23 % is displayed on the second axis.

98 Chapter 3

3.4 CONCLUSION AND RRECONMENDATIONSECONMENDATIONS

Unfortunately, the data available for most of the sites in this study are from once off samples during high and low flow conditions. Fortunately water quality data has been collected by the Sappi Ngodwana Mill (through Labware laboratories) at Hemlock (ER1), Lindenau (ER5) and Rivulets (CR2) for years. Along with this data, historical data are available from DWAF monitoring stations. These stations are located at Geluk (above ER1), Lindenau (ER5) and Montrose (above CR2). Water quality of the Elands River can thus be monitored by the data available from Hemlock (ER1) and Geluk. The impact of any activities between Hemlock and the confluence with the Crocodile River can be assessed by the data available from Lindenau (ER5). The possible effect of the Elands River on the Crocodile River is reflected by the water quality measured at Rivulets (ER2) on the Crocodile River. Finally, water quality of the Crocodile River upstream of its confluence with the Elands River can be reflected by the data available from the Montrose gauging station.

The water quality of the Elands and Crocodile rivers appears to be in a good state. There is some concern, however, over the increase in nutrients over the last decade. The higher COD levels measured at ER 1; along with the increased nutrients is an indication that sewage effluent is reaching the Elands River from the Waterval Boven area. An increase in nutrients and COD are key indicators of sewage pollution. Between the reference site (Hemlock) and the second site at Ryton Estates (ER2) there is little impact on the water quality in the Elands River. After the Elands River passes the site below Sappi Ngodwana an increase in chlorides, sulphates, conductivity, nitrates, nitrites, phosphates and faecal coliforms is observed. The increase in chlorides, sulphates and conductivity is associated with the contaminated groundwater reaching the Elands River through the three springs. Effluent from the pulp and paper mill industry has been known to cause such an increase in these variables. The increase in nutrients is, however, associated with the release of treated sewage by the Ngodwana sewage works below the Ngodwana Mill. Down stream of the mill there is a definite impact on the water quality of the Elands River. Although the sewage being released may cause some form of eutrophication, statistical analysis show that chloride and conductivity are the main contributors to the altered water quality down stream of the mill. There is also a contribution by the rural community situated at Lindenau to the changes in water quality. During the study it has become apparent that the river does show recovery after the confluence with the Crocodile River.

99 Chapter 3

Most of the nutrient and system variables fall within the recommended reference and reserve conditions. The major concern is the fact that pH values for all sites are higher than reference conditions for this resource unit. It is important to note that even historical data shows that the pH over the past 35 years has on average exceeded these reference conditions. It is recommended that the reference conditions for pH for this resource unit should be amended. The major concern regarding the Ngodwana Mill and its activities is the fact that effluent entering the system through the groundwater is causing an increase in the conductivity as well as the chloride and sulphate concentrations. The concentrations of these ions does fall within the TWQR for all users (including tobacco farmers) when the rivers reaches the last monitoring site at CR 2.

3.5 REFERENCES

Botha S, Palmer R, Nonthuys B (2003). Inkomati Water Management Area: Water Resources Situation Assessment – Main Report, DWAF Report No: P05000/00/0101.

Dallas HF and Day JA (1993). The effect of water quality variables on riverine ecosystems: A review. WRC report No TT 61/93. Water Research Commission, Pretoria.

Deksissa T, Ashton PJ, Vanrolleghem PA (2003). Control options for river water quality improvement: A case study of TDS and inorganic nitrogen in the Crocodile River (South Africa). Water SA. Vol. 29 (2) pp 209 -217.

DWAF (Department of Water Affairs and Forestry) (1996a). South African Water Quality Guidelines (second edition), Vol 3: Industrial Use. Department of Water Affairs and Forestry, Pretoria.

100 Chapter 3

DWAF (Department of Water Affairs and Forestry) (1996b). South African Water Quality Guidelines (second edition). Vol 1: Domestic Use. Department of Water Affairs and Forestry, Pretoria.

DWAF (Department of Water Affairs and Forestry) (1996c). South African Water Quality Guidelines (second edition). Vol 4: Agricultural Use: Irrigation. Department of Water Affairs and Forestry, Pretoria.

DWAF (Department of Water Affairs and Forestry) (1996d). South African Water Quality Guidelines (second edition). Vol 5: Agricultural Use: Livestock. Department of Water Affairs and Forestry, Pretoria.

DWAF (Department of Water Affairs and Forestry) (1996e). South African Water Quality Guidelines (second edition), Vol 7: Aquatic Ecosystem. Department of Water Affairs and Forestry, Pretoria.

DWAF (Department of Water Affairs and Forestry) (2004).Draft Background Information: South African Forestry, Pulp and Paper Industries: Draft Version 1. Department of Water Affairs and Forestry, Pretoria.

Howarth RW, Marino R, Lane J and Cole JJ (1988). Nitrogen fixation in freshwater, estuarine and marine ecosystems. 1. Rates and importance. Limnology and Oceanography. Vol. 33 (4) pp 669 – 687.

Kleynhans CJ (1999). The development of a fish index to assess the biological integrity of South African rivers. Water SA. Vol. 25 (3) pp 265 – 278.

Malan H, Bath A, Day J and Joubert A (2003). A simple flow – concentration modeling method for integrating water quality and quantity in rivers. Water SA. Vol. 29 (3) pp 305 – 311.

101 Chapter 3

Morrison G, Fatoki OS, Persson L and Ekberg A (2001). Assessment of the impact of point source pollution from the Keiskammahoek Sewage Treatmet Plant on the Keiskamma River – pH, electrical conductivity, oxygen – demanding substances (COD) and nutrients. Water SA. Vol. 27 (4) pp 475 – 480.

Nhapi I and Tirivarombo S (2004). Sewage discharges and nutrient levels in Marimba River, Zimbabwe. Water SA. Vol. 30 (1) pp 107 – 11.

O’Brien GC (2003). An ecotoxicological investigation into the ecological integrity of a segment of the Elands River, Mpumalanga, South Africa. M.Sc. dissertation.

Palmer CG, Goetsch PA, O’Keeffe JH (1996). Development of a recirculating artificial stream system to investigate the use of macro – invertebrates as water quality indicators. WRC Report No 475/1/96. Water Research Commission, Pretoria.

Van den Brink PJ, Van den Brink NW, Ter Braak CJF (2003). Multivariate analysis of ecotoxicological data using ordination: demonstration of utility on the basis of various examples. Australian journal of ecotoxicology. Vol. 9 pp 141 – 156.

Website 1: Historical water quality data for gauging station X2h011 (Geluk). Available at http://www.DWAF.gov.za/iwqs/wms/data/pdf/x2h011q01.pdf

Website 2: Historical water quality data for gauging station X2h015 (Lindenau). Available at http://www.DWAF.gov.za/iwqs/wms/data/pdf/x2h015q01.pdf

Website 3: Historical water quality data for gauging station X2h013 (Montrose). Available at http://www.DWAF.gov.za/iwqs/wms/data/pdf/x2h013q01.pdf

Weddepohl JP, Pauer JJ, Du Plessis HM, Harris J, Heath RGM, Archibald REM and Chutter FM (1991). Sappi Ngodwana Mill water quality in the Elands River. Technical Report for 102 Chapter 3 Sappi, by the Division of the Environment and Forestry Technology, Report No. DWT 000862, CSIR, Pretoria.

103

CHAPTER 4

SEDIMENT, HABITAT AND RIPARIAN VEGETATION

4.1 INTRODUCTION

Assessing the impact of anthropogenic activities on the aquatic environment has in the past been based mainly on the assessment of water quality (Botes and van Staden, 2005). Earlier management of water resources has thus been based on the potability of water. Over the last decade management initiatives have expanded to include domestic, agricultural, recreational and most importantly instre,am (fish, invertebrates etc.) users. To ensure the effective management of our resources one can not focus on water quality alone. Sediment quality and habitat diversity and availability are also major determinants in ecological integrity.

Aquatic sediments are formed from the deposition of particles and colloids and can act as both a source and a sink of pollutants. Sediment may play a part in the provision of habitat for organisms, and has an important role in ecotoxicology. Long term toxicant input in sediment may often lead to the occurrence of contaminant levels far higher than that in the surrounding water. This is mainly due to the partitioning of substances onto sediment–based binding sites. Sediments in the freshwater environment may thus pose potential hazards to sediment-dwelling organisms, aquatic organism and human health (USEPA, 2002). It is clear that sediment plays a key role in the bioavailability of substances. Two of the most important characteristics of sediment that needs to be taken into account are grain size and the organic content within the sediment. Grain size is used to characterize the physical characteristics of sediments and it is most common to characterise grain size as percentages of gravel, sand slit and clay (USEPA, 1986). Grain size may influence both chemical and biological variables and its plays an important role in the transport and availability of nutrients and contaminants (Walling and Moorehead, 1989). Furthermore, the diversity and abundance of invertebrate assemblages increases along with increasing sediment particle size, from mud to cobbles (Hill, 2005). The quantity and quality of organic matter in sediments are recognised as major factors affecting benthic fauna. The effluent released by the pulp and paper mill industry may affect the sediment adjacent to mill in two ways. Firstly, the effluent might contain large volumes of organic material that may cause increased organic content within the sediment.

104 Chapter 4

Secondly, the effluent may alter the grain size assembly of the sediment, inevitably obliterating that habitat for aquatic organisms (Sibley et al., 1997). Furthermore, the increase in the volume of water below the Ngodwana Mill may cause an increase in the wetted perimeter. This is important as it may lead to the loss of riffles and rapids and an increase in pooled areas.

The integrity of instream habitat is important as it reflects the quality of the physical habitat in which the aquatic ecosystem functions. Habitat assessments on a larger spatial scale, in particular, are being used to an increasingly greater extent (Muhar and Jungwirth 1998). The main reason for this is the fact that, without an assessment of the available physical habitat, it is not possible to determine whether any changes to the ecological integrity detected at a sampling site is attributable to water quality impacts as opposed to physical impacts (Ollis et al., 2006). Loss of habitat can be regarded as the most important factor that has contributed to the extinction of species (Ballance et al., 2001). Certain organisms have a preference for a specific habitat like fringing vegetation, cobbles, pools, etc. Loss of these habitats may lead to a loss of a particular species. The habitat available and the quality thereof, are thus major determinants of the aquatic community structure. Changes to the biological community in a river may be linked to changes to water quality, habitat or both. Habitat in the Elands and Crocodile rivers consists mainly of large cobble beds with an abundance of riffles and rapids interspersed with deep pools. It has been said that the habitat integrity in the study area has been impacted (Ballance et al., 2001). This is largely due to the irrigated agricultural and forestry plantations in the area.

One of the major determinants in habitat quality and sediment transport and deposition is a healthy riparian zone (Rios and Bailey, 2006). Riparian vegetation forms a vital part of any river ecosystem and these ecosystems often are adversely affected by human alteration. It has been well documented that riparian vegetation plays a number of important geomorphological, ecological and social roles which may have an influence on the condition and sustainability of the riverine ecosystem (Kemper, 2001). These roles include: stabilisation of river channel, banks and flood plains, flood reduction, maintenance of water temperature and quality, provision of habitat, etc. It is clear from these roles that the assessment of the riparian vegetation is immiscible in any biomonitoring program or integrity assessment.When naturally vegetated landscapes are transformed to urban or agricultural uses, physical and biological relationships with adjacent streams are affected, usually resulting in habitat degradation and negative impacts on stream biota (Roth et al., 1996).

105 Chapter 4

Riparian vegetation in the Elands River and surrounding systems is important for moderating water temperatures through shading, as well as for providing habitat to instream fauna, riverbank stability, and particulate organic material (Barling and Moore, 1991; Pinto et al., 2006). Furthermore, the organic inputs from riparian vegetation are major food sources for river organisms. The activities in the study area have been mentioned in chapter 2. The main activities that may effect the integrity of the riparian vegetation includes the pulp and paper mill, forestation, agricultural activities and fires. These activities can result in streambank erosion, increased sedimentation, alteration of geomorphology of riparian habitats, loss of species diversity and assemblage composition of macroinvertebrate and fish communities (Rios and Bailey, 2006).

The objective of this chapter was to assess the quality of the sediment, as well as the quality and integrity of the physical habitat and riparian vegetation within the study area.

4.2 MATERIAL AND METMETHODSHODS

4.2.1 PHYSICAL AND CHEMICAL CHARACTERISTICS OF SEDIMENT

All the techniques applied in this study are based on the standard protocols of the United States Environmental Protection Agency (2001). Surface sediment samples (top 5 cm) were collected at each site mentioned in chapter 2 during both high flow (March 2005) and low flow (June 2005). Sediment was collected in honey jars and was frozen to prevent the loss of organic material through the digestion by invertebrate fauna or organic decomposition. In the laboratory the sediment samples were allowed to defrost and then a known amount of sediment from each site was placed into pre-weighed glass beakers. The beakers were placed in a Gallenkamp hot box oven and the sediment was dried at 60 ºC for 96 h. The sediment was then reweighed to determine the moisture content of the sediment sample at each site. The organic content of each sample was also determined. This was achieved by taking a known amount of the dried sediment (accurate to 0.0001 g) and placing them in porcelain crucibles and then incinerating the sample at 600 ºC for 6 hours in a Labcon type RM4 Muffle Furnace. The samples were then reweighed to determine the percentage of organic matter in each sample. The percentage organic content was classified as follows (USEPA, 2001): • Very low - < 0.05% • Low - 0.05 – 1%

106 Chapter 4

• Moderately low - 1 – 2% • Medium - 2 – 4% • High - > 4%

The remaining sediment was used to determine the grain size of each sample by using an Endecott EFL 2000/1 sieve system with sieves ranging from > 4000 µm to 53 µm. The following grain size categories were applied in this study (Cyrus et al., 2000): • > 4000 µm - Gravel • 4000 µm – 2000 µm - Very coarse sand • 2000 µm – 500 µm - Coarse sand • 500 µm – 212µm - Medium sand • 212 µm – 53 µm - Very fine sand • < 53 µm - Mud

4.2.2 HABITAT QUALITY INDICES

Habitat quality and diversity was assessed during both high flow (March 2005) and low flow (June 2005) at all the sites indicated in chapter 2 by implementing two indices: the Habitat Quality Index (HQI) and the Invertebrate Habitat Assessment Index (IHAS). Both these indices consist of various score sheets containing various observations. For the HQI the various metrics that represent the habitat quality of a site are rated from zero to 20, with 20 being a maximum score and indicating excellent habitat quality. The HQI not only takes in to account the quality of the instream habitat in terms of substrate and flow, but further assess the possible impacts of anthropogenic activities in the surrounding area (farming, construction, rural settlements, etc.). The ultimate aim of the IHAS is to summarise and numerically reflect the quantity, quality and diversity of biotopes available for habitation by macroinvertebrates at a sampling site (McMillan 1998,). The scoring system is based on a total of 100 points, split into two sections: Sampling Habitat (55 points) and Stream Condition/Characteristics (45 points).

107 Chapter 4

4.2.3 RIPARIAN VEGETATION

During low flow conditions the vegetation at each site was identified in the field with the assistance provided by the members of the Elands River Valley Conservancy and using various field guides (Van Wyk, 1992; Van Wyk and Van Wyk, 1997; Pooley, 1998; Bromilow, 2001; Henderson, 2001). The vegetation at ER 5 and NR 1 was not assessed as these sites were not accessible during the vegetation survey. Herbs, shrubs and trees were comprehensively identified; grass elements were however generally omitted, unless a particular grass species dominated an area within the riparian zone of a site. Based on the species present the riparian zone was then demarked using the riparian indicator species identified in Godfrey and Roux (2000) and Hill (2005).

4.2.4 STATISTICAL ANALYSES

Canoco version 4.5 was used to complete ordination of the sampling sites based on sediment characteristics. PCA is based on a linear response model relating species and environmental variables (Van den Brink et al., 2003). Results of the ordination is a map of the samples being analysed on a 2-dimensional bases, where the placements of the samples reflect the (dis)similarities between the samples; in this case the sampling sites. Bray-Curtis similarity matrices, constructed from the abundances of the various riparian vegetation species recorded at each site during the low flow period, were subjected to group averaged clustering and two- dimensional non-metric Multidimensional Scaling (MDS) ordination (Clarke and Warwick, 1994).

4.3 RESULTS AND DISCDISCUSSIONUSSION

4.3.1 SEDIMENT

Analysis of sediment characteristics for both flow regimes indicates that a moisture content of more then 10 % was observed for all the sites except for ER 5 during low flow conditions (Tables 4.1 and 4.2). Water content is a measurement of sediment porosity and is usually expressed as a percentage of the whole sediment weight. It is known to influence toxicity and is used to aid in the interpretation of sediment quality assessments (USEPA, 2001). The organic

108 Chapter 4 content classification system of USEPA (2001) indicates that during high flow ER 3 and ER 1, ER 2 and ER 4 during low flow showed an organic content ranging from medium to high. All the other sites, however, show a high organic content. The high organic content present in the sediment sampled at all the sites may have an impact on the bioavailability of toxicants. Apart from the important role in the bioavailability of toxicants, organic matters also have some nutritional value and possibly explain the high productivity within the Elands River.

Table 4.1 Physical and chemical characteristics of the sediment collected during high flow conditions.

Variables (%) ER 1 ER 2 ER 3 ER 4 ER 5 HR NR 1 NR 2 CR 1 CR 2

Water content 15.2 11.9 14.4 11.0 23.5 25.1 - 22.2 18.7 17.3 Organic content 7.1 5.4 2.6 6.9 17.2 32.2 - 9.2 16.5 17.6 Gravel 0.3 36.4 1.1 0.6 34.4 4.7 - 12.8 9.2 1.1

Very coarse sand 1.2 9.5 4.1 2.8 16.9 5.9 - 5.4 7.3 1.6 Coarse sand 8.2 16.2 29.4 62.9 24.4 25.6 - 20.8 31.2 10.9 Medium sand 26.0 24.7 37.9 29.2 11.4 25.4 - 29.5 23.5 19.1 Very fine sand and mud 64.2 13.3 27.4 4.6 32.0 38.4 - 31.5 28.9 67.4

Total organic matter (as determined by combustion) is generally an overestimate of food available for consumers, mainly because various inorganic compounds may be oxidised at about 500 ºC (Pusceddu et al., 1999). Extremely high organic contents were observed during both flow regimes at HR (Tables 4.1 and 4.2). As mentioned earlier, the Lupelule River does flow through a plantation and the canopy completely covers the river at this particular site.

Leaves from the canopy are constantly entering the river and the decaying foliage is the main contributor to this high organic content at this site. This high organic content might explain the unusual macroinvertebrate assemblages observed at this site (chapter 6). Effluent generated by paper mill are high in organic content (Sibley et al., 1997), yet the organic content at ER 3 (were the irrigated effluent enters the system) was not extremely high. The organic content of the sediment at CR 2 does not differ dramatically from the organic content observed before the confluence with the Elands River, the activities of the Ngodwana mill appears to have little impact on the sediment quality in the study area.

109 Chapter 4

Table 4.2 Physical and chemical characteristics of the sediment collected during low flow conditions.

Variables (%) ER 1 ER 2 ER 3 ER 4 ER 5 HR NR 1 NR 2 CR 1 CR 2

Water content 10.6 13.5 18.6 10.1 7.9 23.3 13.0 17.1 19.8 8.5 Organic content 2.8 3.7 8.4 2.2 2.6 25.4 4.3 17.2 16.8 2.4 Gravel 22.3 23.7 5.2 6.4 52.6 53.5 3.0 0.0 2.2 60.3 Very coarse sand 12.9 9.8 4.2 7.5 14.6 5.2 3.7 1.3 8.8 13.6 Coarse sand 30.8 32.3 13.5 60.9 27.4 13.9 44.8 13.7 25.4 20.2 Medium sand 24.5 27.9 40.9 22.2 3.7 13.4 31.8 38.9 51.5 3.7 Very fine sand and mud 55.3 60.2 54.4 83.1 31.1 27.3 76.6 52.6 76.9 24.0

The percentage particle size distribution is of importance in this study. It is useful in determining the soil type of the area and the possible effect that the agricultural activities and degradation of the riparian vegetation have on the river system. The presence of a high percentage of finer sediments is usually an indication that siltation is taking place. As with organic content, particle size also plays a role in the bioavailability of toxicants (Venter and van Vuren, 1997). Tables 4.1 and 4.2 show that, throughout the study, sediments consisted mostly of very fine sand and mud. In general the percentages of these fine sediments and mud were higher during low flow conditions. The grain size characteristics of the sediment can be expected to vary temporally in response to variations in water discharge and other environmental variables (Walling and Moorehead, 1989). As flow decreases, so too does the ability of the river to transport suspended sediments and these sediments then settles. The comprehensive reserve study indicated that only 0.5 – 0.2 % of the sediment sampled consisted of very fine sand and clay.

It further indicated that sediments are dominated by gravel, and very coarse sand (Hill, 2005). The present study shows the opposite, with coarse sand occurring in low percentages. This is attributed to the fact that sediment samples were taken from backwaters and pools. With the limited flow in these habitats, it could be expected that mud would occur in such high percentages. Very little sediment was present in the main channel as the river has the ability to constantly transport and remove sediment. The lack of sediment present at sites was noted during the assessment of the habitat integrity of the sites.

110 Chapter 4

Sediment quality is of particular concern when focusing on invertebrate communities. Seemingly minor changes in substrate particle size, organic content, and even texture can influence the associated invertebrate community structure (Berkman et al., 1986). Sediment can also influence the feeding or breeding patterns of fish communities (Berkman et al., 1986; Naden et al., 2003). The quality of fish and invertebrate assemblages (chapters 6 and 7) indicates that the sediment quality had little influence on the biotic community structure in the study area. The main reason for this is that the diversity and abundance of invertebrate assemblages increases along a gradient of increasing sediment particle size, from mud to cobbles (Hill, 2005). There was a concern that siltation may be taking place due to the numerous agricultural activities in the study area, as well as the degradation in the riparian vegetation. Siltation can be defined as the deposition of fine sediment either on the surface of the stream bed or within a gravel substrate (Naden et al., 2003). Deposition usually takes place in backwater areas, near channel margins or in dead zones where the velocity is reduced, around vegetation, and as surface layers within pools (Naden et al., 2003). Although the high prevalence of very fine sand and mud (Tables 4.1 and 4.2) may indicate that siltation is taking place, it is probably due to the areas at the site where samples were taken. The agricultural activities and the erosion occurring as a result of this at ER 2 appear to have no effect on the sediment quality down stream of this site.

The occurrence of instream vegetation is associated with the distribution of deposited sediment, which is itself influenced by the vegetation through its effect on the local hydraulics and hence the movement, storage and stabilisation of sediment (Sharpe and James, 2006). The amount of sand and mud present at ER 3 does not differ much from those observed at other sites (Tables 4.1 and 4.2). The presence of the aquatic vegetation is thus not due to the sediment characteristic of this site and it does not explain why this site is the only site that has aquatic vegetation. Nutrient levels at this site should be high due to the sewage and the pulp and paper mill activities. The growth of vegetation is also determined by nutrient supply, which is associated at least in part with fine sediment (Barko and Smart, 1986; Sharpe and James, 2006), the combination of nutrient availability and sediment grain size could explain the presence of the aquatic macrophytes.

111 Chapter 4

PCA biplots were developed to assess the spatial and temporal trends of the study sites. In a PCA plot each arrow points in the direction of steepest increase of values for the corresponding variable. The angles between arrows indicate the sign of the correlation between the variables: the approximated correlation is positive when the angle is accute and negative when the angle is larger than 90 degrees. The length of the arrow is a measure of fit for the variables. For sampling sites the distance between the symbols in the diagram approximates the dissimilarity of their sediment characteristics, measured by their Euclidean distance.

Figure 4.1 indicates that based on sediment characteristics there are no clear temporal and spatial trends amongst flow regimes. There appears to be a strong correlation between very fine sand and organic content and a negative correlation between coarse sand and very fine sand. Both these correlations could be expected. Figures 4.2 and 4.3 indicate that the during both flow regimes CR 1, ER 3 and NR 2 show some degree of similarity. The driving variable appears to be medium sand. It is interesting to note that all these sites are situated below flow modifications. Figures 4.2 and 4.3 further indicate that there are very little spatial patterns. This is a good indication that the activities of the Ngodwana mill have little impact on the physical sediment composition of the Elands River.

Figure 4.1 PCA plot showing the dissimilarity among sites on the Elands and Crocodile rivers during both high (sites without suffix) and low flow (sites with the suffix – L) regimes based on sediment characteristics. This biplot describes 77.3 % of the variation in the data, where 53.1 % is displayed on the first axis, while 24.2 % is displayed on the second axis.

112 Chapter 4

Figure 4.2 PCA plot showing the dissimilarity among sites on the Elands and Crocodile rivers during high flow conditions based on sediment characteristics. This biplot describes 81.5 % of the variation in the data, where 36.1 % is displayed on the first axis, while 49.9 % is displayed on the second axis.

Figure 4.3 PCA plot showing the dissimilarity among sites on the Elands and Crocodile rivers during low flow conditions based on sediment characteristics. This biplot describes 95.8 % of the variation in the data, where 77.8 % is displayed on the first axis, while 18 % is displayed on the second axis.

113 Chapter 4

4.3.2 HABITAT QUALITY

Graphical representation of habitat integrity for the sites in the study (Figure 4.4) shows that the habitat in the study is in a good state. Temporal trends indicate that there is little change in habitat between flow regimes. O’Brien (2003) found that habitat integrity, and more particular IHAS, showed an improvement during low flow conditions. In this study high flow results constantly show better IHAS result when compared to low flow, whereas HQI results show very little temporal trends. IHAS results (Figure 4.4 A) indicate that there is degradation in habitat integrity at all the sites on the Elands River when compared to reference conditions at site ER 1. This is largely due to the numerous agricultural activities below ER 1 as well as the additional volume of water entering the system at ER 3. IHAS scores for NR 1 were excellent, while scores for CR 2 were lower than all sites except for NR 2. The IHAS scores at site NR 2 were very low when compared to other sites in the study. HQI results (Figure 4.4 B) shows almost the same spatial trends as IHAS results with a degradation in habitat at all sites on the Elands River when compared to habitat conditions at ER 1. During low flow conditions the HQI score for NR 2 was, much lower than the scores obtained for the other sites in the study.

The method applied in the study to determine the integrity of the habitat at the various sites is illustrated in Table 4.3. The habitat at all sites in the study consisted largely of cobble beds that were interspersed with pools. The length and depth of the pools varied among sites. At all the sites in the study was limited to the marginal vegetation present and it consisted largely of stems and shoots with very little leafy vegetation. Very little sediment occurred at any of the sites in the study area, except for the sites HR, NR 1, CR 1 and CR 2. An abundance of submerged aquatic vegetation was present only at site ER 3 during both flow regimes. This vegetation has not been identified, yet it was clear that it was not filamentous algae.

114 Chapter 4

90 80 70 60 50 40 30 IHAS Score (%) Score IHAS 20 10 0 ER1 ER2 ER3 ER4 ER5 HR NR1 NR2 CR1 CR2 Sites

100 90 80 70 60 50 40

HQI Score (%) Score HQI 30 20 10 0 ER1 ER2 ER3 ER4 ER5 HR NR1 NR2 CR1 CR2 Sites

115 Chapter 4

Table 4.3 The method applied in assigning ecological classes based on the results of habitat indices applied in the study. IHAS/HQI score (%) Condition Class Colour

70 - 90 Unmodified/Natural A

60 - 70 Few modifications/Largely natural B 35 - 60 Modified state C

0 - 35 Largely modified/Unnatural state D

The habitat at nearly all the sites in the study appeared to be in a natural state (class A) (Tables 4.4 and 4.5). The IHAS result for ER 2 during low flow (Table 4.5) showed that some modification of the habitat available to invertebrate communities occurred. This is largely due to the fact that the total area of cobbles covered with water decreased with the decrease in flow. There are also signs of erosion at this site with the right bank being near vertical. Despite this there is little evidence of sedimentation at this site. Sedimentation will have a negative impact on the habitat availability, as it covers cobbles and rocks that provide excellent habitat to aquatic organisms (Nedeau et al., 2003). There is a citrus farm situated on the right bank of the river at this site. The farming activities at this site may pose a potential risk to the integrity of the habitat in this section of the river.

O’Brien (2003) found that there was habitat degradation below the Ngodwana mill. He stated that this was largely due to the increase in the volume of water at site ER 3. Results from this study (Tables 4.4 and 4.5) indicate that this increase in the volume of water has little impact on the habitat integrity at ER 3, with even a slight improvement compared to the results obtained by O’Brien (2003). The reduction in habitat integrity at site ER 4 during high flow was brought about largely due to the presence of algae on the rocks at this site. The presence of algae is an indication of eutrophication. In chapter 3 it became clear that the activities at ER 3 (Ngodwana mill and sewage works is contributing to the nutrient loading in the system, and is responsible for the algal blooms at ER 4. The presence of algae on the rocks at this site reduces the availability of habitat to organisms, especially macroinvertebrates. Very little cobbles occur at this site and the rocks at this site are quite large. These larger rocks can not provide the same quality of habitat as cobble beds do. Flow at this site is also relatively homogenous. This is important as research has shown that different organisms prefer different flow velocities (Kleynhans, 1999; Dickens and Graham, 2002). This change in flow is probably brought about

116 Chapter 4 by the increased volume of water entering the river above ER 3. Tables 4.4 and 4.5 indicate that during both flow regimes the rural settlement at ER 5 had little impact on the habitat integrity at this site. The decrease in habitat integrity at Site HR during high flow was largely due to the lack of marginal vegetation, while the decrease in the wetted perimeter contributed to the changes observed at CR 2 during low flow conditions (Tables 4.4 and 4.5). The habitat at HR is unique compared to other sites in this study. Although cobble beds and pools are present, a canopy covers the entire river at this site. This is partly because the Lupelule River flows through a plantation.

Table 4.4 Results (IHAS and HQI scores and ecological classes) of the habitat quality indices during high flow conditions.

ER1 ER2 ER3 ER4 ER5 HR NR1 NR2 CR1 CR2

IHAS (%) 75 81 75 66 73 69 - 55 79 66

HQI (%) 80 72 79 70 75 73 - 63 85 85

IHAS Class A A A B A B - C A A

HQI Class A A A A A A - B A A

Table 4.5 Results (IHAS and HQI scores and ecological classes) of the habitat quality indices during low flow conditions. ER1 ER2 ER3 ER4 ER5 HR NR1 NR2 CR1 CR2

IHAS (%) 75 60 71 68 73 78 83 57 82 65

HQI (%) 82 73 73 70 75 79 91 42 83 87

IHAS Class A B A B A A A C A B

HQI Class A A A A A A A C A A

117 Chapter 4

The greatest concern regarding habitat integrity in the study area is the changes that have occurred at NR 2. The construction of the dam wall has caused serious changes to the natural flow regime at this site. There are no riffles or rapids present at this site, flow is homogenous, there are algae present on the rocks, and the substrate consist mostly of large boulders and bedrock. Although changes to water quality is a potential stressor it appears that reduced flow and the subsequent loss of habitat may pose a greater risk to the biotic community structure at this site. The biological template at this site may be seriously altered. The IHAS has proved to be an effective index to assess habitat quality for aquatic invertebrates in the study area. The fact remains that this is largely due to the large quantity of cobble beds present at all the sites. Where these cobble beds were not present (NR 2, for example) the IHAS results change quite dramatically. Ollis et al. (2006) has recently found that the IHAS was an effective method for assessing the quality and quantity of habitat available to invertebrates in this province (Mpumalanga). Doubt still remains about the effectiveness of this index in other ecoregions/geomorphological zones. Furthermore, it has become clear that the IHAS and the HQI are not sensitive enough to respond to subtle changes in habitat integrity.

4.3.3 RIPARIAN VEGETATION

The upper reaches of the Elands River occur within the Acocks Veld type 9 (Lowveld Sour Bushveld) (Hill, 2005). The potential riparian zone width is between 2 and 10m and vegetation cover is in some cases less than 75% as a result of disturbances such as burning, flooding and exotic plant species invasion. The present ecological status is estimated to be a class B (largely natural with few modifications). The lower sections, which is part of the main study area occurs within the Acocks Veld type 10 (Lowveld). The potential riparian zone width is between 2 and 22 m and the active channel width is between 6 and 20m. Vegetation cover is seldom less than 75 %. Disturbances that occur in this section of the study area includes: exotic vegetation invasion, forestry, erosion, flooding and burning. The present ecological status is estimated to be a class D (largely modified). Although there is evidence of alien removal, alien recruitment is prevalent (Hill, 2005). During the study conducted by O’Brien (2003) it was found that the riparian vegetation within the Elands River area was in a class D with some sites (ER 2) deemed to be in a class E.

118 Chapter 4

Riparian vegetation can often be used as indicators of certain morphological zones. These include the active channel, seasonal channel and the macro channel bank. A few vegetation species found in the study area are indicators of the active channel (Hill, 2005) (Also see appendix A). This includes Cliffortia sp, Cyperus sp. Phragmites australis, and Salix mucronata. Table 4.6 shows that at least one of the species mentioned above are present at the sites in the study area. The occurrence of S. mucronata and Cliffortia sp. on both banks will provide good clues for base flow requirements.

These species were not present at HR or the sites on the Crocodile River (CR 1 and CR 2). Combretum erythrophyllum, Diospyros lycioides, Ficus sur and P. australis are a few indicators species of the seasonal channel. C. erythrophyllum and can be good indicators of elevated flow requirements. Although C. erythrophyllum was found at most of the study sites, D. lycioides was not as wide spread (Table 4.6). Phragmites and Cyperus latifolius can provide a good indication of summer base flow requirements. Finally, many species are good indicators of the macro channel. These include Acacia karroo, A. sieberiana and Dombeya rotundifolia. Except for ER 4, Acacia sieberiana were found at all the sites in the study area. It is the most wide spread of the macro channel indicators, with A. karroo only present at NR 2 and CR 1 and D. rotundifolia only present at CR 1 (Table 4.6). The presence of A. karroo may indicate terrestrialisation due to possible reduced flows (Hill, 2005). The fact that both NR 2 and CR 1 are situated below dams that have caused flow modifications further supports this assumption.

The invasibility of natural plant communities by exotic species continues to be major ecological concern (Planty-Tabacchi et al., 1996). The greatest concern, regarding the riparian vegetation in the study area, remains the large number of exotic species present. About 14 % of the estimated total water use, by volume, is related to alien vegetation in Mpumalanga. Although the Trans African Concession (TRAC), together with Working for Water has made recent efforts to remove these exotic species, some of the endemic vegetation is ultimately removed with the invading plants. Although the control of alien vegetation, through such methods as active removal, it is still cost effective compared to other methods (Le Maitre, 2000). Table 4.6 clearly indicates that the riparian vegetation at all the sites had some elements of invading plants. Some of these exotic species may have an impact on the river. Both Eucalyptus spp. and Grevillea robusta is native to Australia. Eucalyptus spp. is cultivated in commercial plantations and often invades watercourses. Grevillea robusta, Jacaranda mimosifolia (native to S

119 Chapter 4 Table 4.6 Vegetation species found in the Riparian zone at sites on the Elands and Crocodile Rivers (species name in bold indicate non – endemic species). Woody/ Habitat Non-woody Exotic ER1 ER2 ER3 ER4 HR NR2 CR1 CR2 Acacia ataxacantha Forest margins x x x Acacia caffra Scrub/grassland x Acacia karroo Grassland x x Acacia mearnsii Woody Non-endemic x x x Acacia sieberiana var woodii Grassland x x x x x x x Adenopodia spicata Forest margins x Allophylus africanus Forest margins x Aloe spp. x x x Arundo donax Non-woody Non-endemic x x x x x Asclepias spp. x Asparagus spp. x x Berchemia zeyheri x Bidens pilosa x x Breonadia salicina Riverine forest x Bridelia micrantha Riverine forest x Caesalpinia decapetala Woody Non-endemic x Cardiospermum grandiflora Non-woody Non-endemic x x x Celtis africana Forest x x x x x Cestrum elegans x Choristylis rhamnoides Forest x Citrus Lemone x Clausena anisata Forest margins x Cliffortia strobilifera x x x x x Combretum erythrophyllum Streams x x x x x x x Combretum zeyheri Bushvelds often along rivers x

120 Chapter 4 Cussonia spicata Forest margins x Cyperus sp. x x x x x x Diospyros lyciodes Grassland x x Diospyros whyteana Forest x x x Dombeya pulchra Rocky grassland x x x x Dombeya rotundifolia Scrub forest x Ekebergia capensis Forest/bushveld x x Englerophytum mogalismontanum Rocky grassland x Erythroxylum delgagoense x Eucalyptus spp. Woody Non-endemic x x x x x Euclea crispa crispa Forest margins x x x x x x Euphorbia ingens Dry forest x Ficus sur Forest x x x Ficus sycomorus Widespread along lowveld rivers x Gelditisia tricanthos Woody Non-endemic x Grevillea robusta Woody Non-endemic x Grewia occidentalis Forest margins x x Gymnosporia buxifolia Forest margins x x Gymnosporia harveyana Forest x Gymnosporia senegalensis Bushveld riverbeds and dry riverbeds x Halleria lucida Forest margins x x Hippobromus pauciflorus Forest margins x Hyperacanthus amoenus Forest x Jacaranda mimosifolia Woody Non-endemic x x x Lantana camara Non-woody Non-endemic x x x x x x Leucosidea sericea x Lippia javanica Scrub grassland x x lpomoea alba x

121 Chapter 4 Macfadyena unguis-cati Non-woody Non-endemic x Maytenus undata Forest x x x x Melia azederach Woody Non-endemic x x x x Morella serrata Along drainage lines and streams x Passiflora subpeltata Non-woody Non-endemic x x Pavetta gardenifolia Rocky scrub x Phragmites australis x x x x x x x Phragmites mauritianus x Plectranthus sp. x x Populus deltoldes x Populus x canescens Woody Non-endemic x Psidium guajava Woody Non-endemic x x Pterocelastrus rotundifolius Open bushveld x Rhamnus prinoides x x Rhus chirindensis Forest margins x Rhus pentheri Forest margins x Rhus pyroides Streams x Rhus rehmanniana x Salix mucronata x x x x x Sansevieria spp. x Sedges x Senna septemtrionalis Woody Non-endemic x x x x x Sesbania bispinosa Non-woody Non-endemic x Sesbania punicea Non-woody Non-endemic x x Smilax anceps x Solanum mauritianum Non-woody Non-endemic x x x x x x x Syzigium cordatum Grassland x Tagetes minuata Non-woody Non-endemic x Tecoma stans Woody Non-endemic x

122 Chapter 4 Tithonia rotundifolia Non-woody Non-endemic x x x x Trema orientalis Forest x Typha capensis x x x Verbena bonariensis x Zanthoxylum capense Scrub forest x Ziziphus mucronata Scrub forest x x

123 Chapter 4

America), Macfadyena unguis-cati (native to Asia) and Populus x canescens (native to Europe and Asia) invades bushveld and river banks (Van Wyk et al., 2000). A. mearnsii is one of the species which presence is of great concern. It invades grasslands and watercourses and can consume large quantities of water (Van Wyk and Van Wyk, 1997).

A Bray-Curtis similarity analysis and resulting cluster and NMDS plot (Figures 4.5 A and B) of the vegetation surveyed at the sampling sites indicated that apart from ER 4, all the sites on the Elands River are grouped together. According to O’Brien (2003) the vegetation in this section of the river is mostly leafy with few grasses and reeds occurring. This may explain why NR 2, with the large number of leafy vegetation, has been grouped with these sites. This occurrence may be due to the increased volume of water in this section of the river (O’Brien, 2003). The reasons as to why ER 4 has been grouped separately from the other sites are not clear, yet it may be an indication of the changes in flow occurring at this site. HR has been grouped separately due to the large numbers of forest and forest margin elements that are present at this site. This was expected as the river does flow through a plantation and forested areas. CR 1 and CR 2 have been grouped separate from each other and the sites on the Elands River. These sites are situated in a different biome.

124 Chapter 4 A

125 Chapter 4

4.4 CONCLUSION AND RRECOMMENDATIONSECOMMENDATIONS

Habitat quality and quantity is of immense importance to river conservation and management as it provides the template on which biota function. Changes to habitat are often caused by anthropogenic impacts such as agriculture, construction of reservoirs, sedimentation, water abstraction, etc. These changes may cause a loss of diversity or the serious alteration of community structures of resident aquatic biota. Sediments often serve as habitat to a variety of aquatic organisms and become even more important, when one looks at its role in ecotoxicology. One of the major determinants in habitat quality and sediment transport and deposition is a healthy riparian zone.

The availability of sediment is restricted within the Elands River. This is a good sign that sediments are constantly being transported in the river. There are little signs of the sedimentation first thought to be occurring due to the agricultural activities at ER 2 and the loss of riparian vegetation in the study area. In very low flow conditions the possibility exists that these activities may contribute to the deposition of sediments. Sediment availability was much higher in the Crocodile River, largely due to the flow modifications caused by the Kwena Dam. Analysis of sediment characteristics indicated that sediments within the study area, unlike the results of the comprehensive reserve determination, consist mostly of very fine sand and mud. This is largely due to the fact that sediment samples were taken from backwater areas where deposition naturally occurs. This should not be seen as a sign of sedimentation. The organic content of the sediments in the Lupelule River is extremely high. Although this may be a natural occurrence, the effect it has on invertebrate community structures should be further investigated.

The instream habitat in the Elands and Crocodile rivers appear to be in a largely natural state. In general the habitat integrity fell within the recommended ecological category of a B class, as determined by the reserve. There are, however, some habitat modifications within the study area with the virtual absence of marginal vegetation as a habitat is one of the greatest concerns. The changes observed at NR 2 are linked to the construction of the dam wall and the resultant flow modifications. The Ngodwana River above the dam appears to be unimpacted by the dam. The biggest concern regarding habitat within the Elands River is the activities related to the Ngodwana mill. Although groundwater naturally enters the system through the springs situated

126 Chapter 4 near the mill, it has become apparent that there is a definite increase in the volume of water entering the system between sites ER2 and ER 3. The possible effects of this modification became clear at ER 4 with this site showing limited variety with regards to flow velocities. The possible effect that the increase in water has on community structures of fish and invertebrates should be further investigated.

The riparian vegetation within the study area appears to be in a modified state. This is largely due to encroachment of exotic and terrestrial species. This occurrence appears to be related to the agricultural activities within the area. As expected, the riparian vegetation communities within the Crocodile River differ from those found in the Elands River. Analysis has revealed that the riparian vegetation community of the Lupelule River (HR) is unique in terms of speciation. This is due to the large number of forest species occurring at the site. The efforts made by TRAC and the Working for Water group in controlling exotic encroachment appear to be successful and should be continued.

From an ecological perspective it is of the utmost importance that the habitat and sediment quality within the Elands River be maintained. This will ensure that macroinvertebrate and fish communities alike are not negatively impacted and that no species are lost.

4.5 REFERENCES

Ballance A, Hill L, Roux DJ, Silberbauer M and Strydom W (2001). State of the Rivers report: Crocodile, Sabie – Sand and Olifants River systems. WRC Report No. TT 147/01. Water Research Commission, Pretoria.

Barko JW and Smart RM (1986). Effects of organic matter additions to sediment on the growth of aquatic plants. Journal of Ecology. Vol. 71 pp 161-175.

Barling, RD and Moore ID (1991). Role of buffer strips in management of waterway pollution: a review. Environmental Management. Vol. 18 pp 543-558.

Berkman HE, Rabeni CF and Boyle TP (1986). Biomonitors of Stream Quality in Agricultural Areas: Fish versus Invertebrate. Environmental Management. Vol. 10 (3) pp. 413-419. 127 Chapter 4

Botes PJ and van Staden (2005). Investigation of trace element mobility in river sediments using ICP-OES. Water SA. Vol. 31 (2) pp 183 – 192.

Bromilow C (2001). Problem plants of Soth Africa. Briza Publications, Pretoria.

Clark KR and Warwick RM (1994). Change in marine communities: an approach to statistical analysis and interpretation. Manuel for the PRIMER statistical programme. Natural environment research council.

Cyrus DP, Wepener V, Mackay CF, Cilliers PM, Weerts SP and Viljoen A (2000). The effect of interbasin Transfer on the Hydrochemistry, Benthic Invertebrates and Ichthyofauna of the Mhlathuze Estuary and lake Nsenzi. WRC report No. 722/1/00. Water Research Commission, Pretoria.

Dickens CWS and Graham PM (2002). The South African Scoring System (SASS), Version 5, Rapid bioassessment method for rivers. African Journal of Aquatic Science. Vol. 27 pp 1 – 10.

Henderson L (2001). Alien weeds and invasive plants. A complete guide to declared weeds and invaders in South Africa. Agricultural Research Council, Pretoria.

Hill L (2005). Elands Catchment Comprehensive Reserve Determination Study, Mpumalanga Province, Ecological Classification and Ecological Water Requirements (quantity) Workshop Report, Contract Report for Sappi-Ngodwana, Submitted to the Department Water Affairs and Forestry, by the Division of Water Environment and Forestry Technology, CSIR, Pretoria. Report No. ENV-P-C 2004-019 pp 1 -98.

Kemper NP (2001). RVI Riparian Vegetation Index. WRC report NO 850/3/01. Water Research Commission, Pretoria.

128 Chapter 4

Kleynhans CJ (1999). The development of a fish index to assess the biological integrity of South African rivers. Water SA. Vol. 25 (3) pp 265 – 278.

Le Maitre DC, Versfeld DB and Chapman RA (2000). The impact of invading alien plants on surface water resources in South Africa: A preliminary assessment Water SA. Vol. 26 (3) pp 397 – 408.

Muhar S and Jungwirth M (1998). Habitat integrity of running waters - assessment criteria and their biological relevance. Hydrobiologia. Vol. 386 pp 195–202.

Naden P, Smith B, Jarvie H, Llewellyn N, Matthiessen P, Dawson H,Scarlett S and Hornby D (2003). Siltation in Rivers. A Review of Monitoring Techniques. Conserving Natura 2000 Rivers Conservation Techniques Series No. 6. English Nature, Peterborough.

Nedeau EJ, Merrit RW, Kaufman, MG (2003). The effect of an industrial effluent on an urban stream benthic community: water quality vs. habitat quality. Environmental Pollution. Vol. 123 pp 1-13.

Pinto BCT, Araujo FG and Hughes RM (2006).Effects of landscape and riparian condition on a fish index of biotic integrity in a large southeastern Brazil river. Hydrobiologia. Vol. 556 pp69– 83.

129 Chapter 4

Planty – Tabacchi A, Tabacchi E, Naiman RJ, Deferrari C and Decamps H (1996). Invasibility of species-rich communities in riparian zones. Conservation Biology. Vol. 10 (2) pp 598 – 607.

Pusceddu A, Sar`a G, Armeni1 M, Fabiano M and Mazzola A (1999).Seasonal and spatial changes in the sediment organic matter of a semi-enclosed marine system (W-Mediterranean Sea). Hydrobiologia. Vol. 397 pp 59–70.

Rios SL and Bailey RC (2006). Relationship between riparian vegetation and stream benthic communities at three spatial scales. Hydrobiologia. Vol. 553 pp153–160.

Roth NE, Allan DJ and Erickson DL (1996). Landscape influences on stream biotic integrity assessed at multiple spatial scales. Landscape Ecology. Vol. 11 (3) pp 141-156.

Sharpe RG and James CS (2006). Deposition of sediment from suspension in emergent vegetation. Water SA. Vol. 32 (2) pp 211 – 218.

USEPA (United States Environmental Protection Agency) (1986). Recommended protocols for measuring conventional sediment variables in Puget Sound. USEPA, Cincinnati, Ohio.

USEPA (United States Environmental Protection Agency) (2001). Methods for Collection, Storage and Manipulation of Sediments for Chemical and Toxicological Analyses: Technical Manual. EPA 823-B-01-002. U.S. Environmental Protection Agency, Office of Water, Washington, DC.

USEPA (United States Environmental Protection Agency) (2002). A Guidance Manual to Support the Assessment of Contaminated Sediments in Freshwater Ecosystems. Volume I - An

Ecosystem-Based Framework for Assessing and managing contaminated sediments. EPA-905- B02-001-A. USEPA, Cincinnati, Ohio.

130 Chapter 4

Van Wyk (1992). Field guide to the trees of the Kruger National Park. Struik Publishers (Pty) Ltd, Cape Town.

Van Wyk B and Van Wyk P (1997). Field guide to trees of Southern Africa. Struik Publishers (Pty) Ltd, Cape Town pp 1 – 506.

Van Wyk B, Van Wyk P and Van Wyk BE (2000). Photographic guide to trees of Southern Africa. Briza Publications, Pretoria pp 7 – 356.

Venter AJ and Van Vuuren JHJ (1997). The effects of gold mine related operations on the physical and chemical characteristics of sediment texture. Water SA. Vol. 23 pp 249 – 256.

Walling DE and Moorehead PW(1989). The particle size characteristics of fluvial suspended sediment: an overview. Hydrobiologia. Vol. 176/177 pp 125-149.

131 Chapter 5 CHAPTER 5

ICHTHYOFAUNA

5.1 INTRODUCTION

River systems on a global level are heavily degraded by various human activities and impacts (Jungwirth et al., 2000; Muhar et al., 2000). It is thus of the utmost importance for both river conservation and management to determine which basic processes, functions and structures make up the ecological integrity of these running waters. To achieve this, scientist and managers alike use living organisms as indicators. Many groups of organisms have been proposed as indicators of environmental integrity or health and untill today no clear favourite has emerged. According to Dale and Beyeler (2001) ecological indicators should be easily measured, be sensitive to stresses on the system, respond to stress in a predictable manner, be anticipatory, predict changes that can be averted by management actions, be integrative, have a known response to disturbances, anthropogenic stresses, and changes over time, and have low variability in response. Fish communities and individuals themselves show a variety of these qualities that make them useful in biological monitoring (Kotze et al., 2004). This includes the fact that fish are relatively long-lived and reflect changes in the condition of a river system. This makes them excellent long-term indicators of environmental integrity (River Health Program, 2005).

There are various environmental factors that may have an impact on the natural structure and integrity of fish communities. Habitat, for example, is a major environmental factor that influences fish communities (Schmutz et al., 2000). It plays a fundamental role in feeding, reproduction, and survival by effecting physiology, behavior, and genetics (Morgan, 2002). Relationships between physical habitat structure and stream-fish assemblage structure are well- documented. Along with habitat degradation, a decline in water quality and quantity and the introduction of exotic species into river systems have caused the extinction of several species of fish and altered populations in several countries (Lyons et al., 1995). These changes mentioned above are all linked to human activities. Human activities within the Elands River have in the past threatened the ecological integrity of the fish communities within the river. On the 23 September 1989, effluent generally known as black liquor (soap skimming) was accidentally released by the Ngodwana pulp and paper mill into the Ngodwana River

132 Chapter 5

The spill lasted for 2 hours and following the spill a large quantity of water was released from the Ngodwana Dam to dilute this effluent. This helped to some degree but it still caused massive fish mortalities. The mortalities where mainly caused by the high biological oxygen demand and the sulphur containing substances of the black liquor (Kleynhans et al., 1992). Virtually all species where destroyed just below the confluence of the Ngodwana and the Elands Rivers. It took between 6 and 11 months for the fish populations to recover to such a point as to which areas downstream of where the spill occurred were comparable with areas upstream (Kleynhans et al., 1992). James and Barber (1991a) stated that the ecological status of the Elands River with regards to fish populations were of great concern. They found that fish numbers and species diversity remained low, even though macroinvertebrate life was re – established in the water appeared clean and clear. They did however; find large numbers of juvenile fish at the site of the spill, indicating that the river at that point in time should signs of recovery. The release of large quantities of water from various sources to dilute the effluent two hours after the spill may have altered the habitat and may have been the cause for the slow recovery rate. Recent studies have, however, shown that the situation has bettered and fish populations have returned to “normal” (Godfrey and Roux, 2000; O’Brien, 2003; Hill, 2005).

The objective of this chapter was to assess the effect of anthropogenic activities on the fish communities within the study area on a temporal and spatial level.

5.2 MATEMATERIALRIAL AND METHODS

5.2.1 FIELD SURVEY

Fish were comprehensively sampled during both the high (March and 2005) and low (June and 2005) flow regimes. All fish were sampled in the habitats determined by Kleynhans (1999) at the sites described in chapter 2 (Figure 2.13), except for ER 5 and NR 1 and during high flow and NR 2 during both flow regimes. The habitats sampled (when present) include shallow and deep section of the river in fast and slow flowing water. The habitat available to the fish community was also noted by completing the Habitat Quality Index (HQI). Sampling techniques include the use of seine nets, gill nets and electroshocking. A medium size seine net was hauled through the pooled areas at some sites. The net is 30 m in length and has a 22 mm mesh size. Additionally gill nets were placed in these pools overnight. The technique most

133 Chapter 5 commonly applied in the study was electroshocking as it proved to be the most effective. This technique was applied at all the sites in the study.

5.2.2 SPATIAL AND TEMPORAL ANALYSIS

The multivariate statistical assessment approach in this study was based on the approach and rational adopted by O’Brien (2003) during his study on the Elands River. Bray-Curtis similarity matrices, constructed from the abundances of the various fish species recorded at each site on the different sampling occasions were subjected to group averaged clustering and two- dimensional non-metric Multidimensional Scaling (MDS) ordination (Clarke and Warwick, 1994). One-way Analysis of Similarities (ANOSIM) was used to determine the extent of the overall differences in ichthyofauna composition among the various sites and flow conditions. For each ANOSIM test, the null hypothesis assumes that there were no significant differences among groups was rejected when the significance level (P) was <5% (P<0.05). The extent of any significant differences produced by this test were determined using the R-statistic value (Clarke and Green, 1988), which can range from +1, i.e. all samples within each of the groups are more similar to each other than to any of the samples from other groups, down to approximately zero, when average similarities within and between groups are the same (i.e. the null hypothesis). Small negative values of R are possible by chance under the null hypothesis, but are not generally interpretable since they correspond to similarities between groups being smaller than within groups and the ANOSIM test for R is thus one-sided. When the pair wise comparisons in the ANOSIM test detected a significant difference in ichthyofauna compositions between sites and flow regimes, Similarity Percentages (SIMPER) was used to identify which species typified each of those habitat types.

To determine which environmental variables were possibly responsible for the various groupings a Redundancy Analysis (RDA) was completed for both using Canoco version 4.5. RDA is derivative of PCA with one additional feature. The values entered into the analysis are not the original data but the best-fit values estimated from a multiple linear regression between each variable in turn and a second matrix of environmental data. Thus the PCA is constrained to optimize a fit to the environment data so that this technique is the canonical version of PCA. Interpretation of RDA is through biplots (Shaw, 2003). These biplots is a map of the samples being analysed on a 2 dimensional bases, where the placements of the samples reflect the (dis)similarities between the samples; in this case the sampling sites.

134 Chapter 5

Indices of diversity and evenness were applied to describe the species abundance relationships among the ichthyofauna communities. The Shannon-Wiener diversity index was applied. This index incorporates both the species richness and equitability components (O’Brien, 2003). The species richness was expressed as Margalef’s index, while the evenness was expressed using Pielou’s evenness index. Margalef’s index measures the number of individuals present for a given number of species, incorporating both the total number of species and the total number of individuals. Dominance curves have been included in the study to indicate which sites reflect relative increased dominance of species.

5.2.3 THE FAII

The current index, being used to assess the status of fish communities in South Africa, is the Fish Assemblage Integrity Index (FAII). The FAII aims to measure the biological integrity of a river as based on the attributes of the fish assemblages native to the river. Alien species (introduced indigenous and exotic species) are not included as metrics in the FAII. Their presence and distribution are noted but interpreted as possible causes for a decline in the FAII score. This index takes into account three aspects of fish assemblages; (1) The relative intolerance of indigenous fish species expected to occur at a segment. Intolerance in this context refers to the degree to which a species is able to withstand changes in the environmental conditions under which it occurs. This includes modification of physical habitat characteristics as well as chemical characteristics of the water habitat. (2) The frequency of occurrence of a species (number of sites in which the species occur). Abundance is not included as a metric in the index due to the difficulty in obtaining quantitative information on this. (3) The rating of the general health (abnormalities or disease) of the individuals caught. Even under unimpaired conditions, a small percentage of individuals can be expected to exhibit some anomalies. A comparison is made between the expected fish assemblages and the observed assemblages to determine the integrity of the river (Kleynhans, 1999).

5.3 RESULTS AND DISCDISCUSSIONUSSION

5.3.1 SPATIAL AND TEMPORAL ANALYSIS

Tables 5.1 and 5.2 shows that, except for ER 1, a higher number of species was sampled during low flow conditions. Taking into account that C. gariepinus was sampled at ER 1 during high 135 Chapter 5 flow, but not low flow conditions (Table 5.1 and 5.2), ER 1 follows more or less the same trend. This is mainly due to the difficulty of sampling during high flow conditions. Polynomial trend lines indicate that a large number of species was sampled at ER 3 and CR 2. CR 2 falls within the Crocodile River, and this river does have larger species diversity. The higher number of species at ER 3 is possibly due to the change in habitat, with more pooled sections occurring at this site and an abundance of aquatic vegetation present compared to other sites in the study. This provides excellent habitat for some species such as P. philander, T. sparrmanii and the introduced M. acutidens that were all sampled at this site. ER 2 and ER4 showed a very low numbers of species occurring at these sites. Although some of the expected species were not sampled it is possible that they do occur at these sites. ER 4, however, does have little habitat diversity. The limited habitat available may have contributed to the low number of species sampled.

Table 5.1 Results (FAII score, ecological classes and abundances) of ichthyofaunal species collected during high flow conditions.

Expected species ER1 ER2 ER3 ER4 ER5* HR NR1* CR1 CR2 Amphilius uranoscopus 11 11 2 6 2 3 - Anguilla mossambica ------Barbus anoplus - - - - 4 1 - Barbus argenteus 16 19 25 38 54 - 31 Clarias gariepinus 2 - 1 - - - 1 Chiloglanis bifurcus ------Chiloglanis pretoriae 36 14 7 2 26 21 7 Labeobarbus marequensis ------4 Labeobarbus polylepis ------Micralestes acutidens 2 - 15 - - - 13 Pseudocrenilabrus philander - - - - - 56 41 Tilapia sparmanii 4 - 25 2 - - -

Total number of species 6 3 6 4 - 4 - 4 6 FAII score 79.5 79.5 76.9 76.9 - 79.5 - 77.6 75.5

Class B B C C - C - C C *Not Sampled

136 Chapter 5

Fish populations in the Elands River have been extensively studied over the past two decades. The main focus has always been the accidental spill in 1989 and several surveys to determine whether the fish communities have recovered. Shortly after the spill, a survey by Kleynhans in 1989 found no living fish. Surveys performed in the following years indicated some degree of recovery (James and Barber, 1991a), while the recent reserve study shows that fish population returned to normal (although in some cases abundances are still low). The section of the Elands River that forms part of this study carries with it great ecological importance and sensitivity. The endangered Incomati rock catlet (C. bifurcus), occurs in the stretch of river and has a high preference for clean, fast flowing water through all of its live stages (Godfrey and Roux, 2000). According to Hill (2005), species that are expected to occur in this resource unit include the mountain catlet (A. uranoscopus), chubbyhead barb (B. anoplus), rosefin barb (B. argenteus), smallscale yellowfish (L. polylepis), banded tilapia (T. sparrmanii), southern mouthbrooder (P. philander), shortspine rock catlet (C. pretoriae), Inkomati rock catlet (C. bifurcus) and longfin eel (A. mossambica). Rainbow trout (O. mykiss) and largemouth bass (M. salmoides) are exotic

Table 5.2 Results (FAII score, ecological classes and abundances) of ichthyofaunal species collected during low flow conditions.

Expected species ER1 ER2 ER3 ER4 ER5 HR NR1 CR1 CR2

Amphilius uranoscopus 8 4 3 2 14 1 12 8 39

Anguilla mossambica - - - - 1 - - 1 - Barbus anoplus - - - - - 15 6 - -

Barbus argenteus 21 26 38 38 14 38 19 - 17 Clarias gariepinus ------Chiloglanis bifurcus 1 - 1 - - - 4 - 1

Chiloglanis pretoriae 8 21 7 5 17 23 51 3 4

Labeobarbus marequensis ------1 Labeobarbus polylepis - - 1 - - - 4 - 1 Micralestes acutidens - - 13 ------

Pseudocrenilabrus philander - - 1 - - - - 2 2

Tilapia sparmanii 6 2 5 4 8 3 - - -

Total number of species 5 4 8 4 5 5 6 4 7

FAII score 87.2 87.2 82.1 82.1 82.1 82.1 94.9 79.6 91.8

Class B B B B B B A B A

137 Chapter 5 species that are increasing in abundance and proliferating in the resource unit. The indigenous sharptooth catfish (C. gariepinus) and the silver robber (M. acutidens) have also been introduced into the Elands River. These introduced species have the potential to have a significant impact on the indigenous species in this resource unit.

Primer version 5 was implemented to complete a range of univariate diversity tests. Margalef’s species richness index is an indication of species diversity and abundance. The polynomial trend lines superimposed on to Figure 5.1 B shows the same trend as Figure 5.1 A, with an increase in the abundance and diversity at ER 3. An increase in abundance and diversity is not always advantages to a river system, especially if the increase includes introduced species. At ER 3, for example, the increase includes the occurrence of introduced species and species which are moderately tolerant to changes in water quality and habitat. These exotic species are probably entering the system through the Ngodwana River. The dam has been stocked by anglers with a variety of exotic fish species. Figure 5.1 B further indicates that there was a low species abundance and diversity at ER 4. Chapter 4 does indicate that the habitat at this site is affected in terms of different flow velocities. This limited habitat availability is the main cause for this reduction in abundance and diversity. As with Figure 5.1 A the polynomial trend lines in Figure 5.1 B indicate a higher species diversity and abundance at CR 2. In comparison to CR 1 the diversity and abundance increases after the confluence of the Elands River with the Crocodile River, indicating a minimal impact by the Elands River. Pielou’s evenness index is an indication of how the individuals in a sample are distributed over the various species that make up a community. Figure 5.2 A indicates that there is a slight change in evenness during low flow conditions at all the sites below the reference site. The polynomial trend lines in Figure 5.2 A shows that the most obvious change occurred in evenness occurred at ER 4 during both flow regimes and at CR 2 during low flow conditions. This lack of evenness is a good indication of the dominance of a family or families at this site. During both flow regimes, B. argenteus dominated the community structure at ER 4. The fact that this species of fish is sensitive to water quality changes (based on its perceived tolerance to water quality changes) is a good indication that this species of fish does not dominate the community structure due to pollution. This further supports the assumption that changes to flow at this site may be driving the community structure. Figure 5.2 A indicates that as with ER 4 there is domination of the community by one species of fish at CR 2. A. uranoscopus is the dominant species at this site and this fish is particularly sensitive to pollution. This is a good indicator that the activities in

138 Chapter 5 the Elands River are not affecting the community structure of the fish at CR 2. The ecological integrity of the community might be influenced by flow modification further upstream in the Crocodile River. The Shannon-Wiener diversity index followed the same trend as Pielou’s evenness index. The polynomial trend lines in Figure 5.2 B clearly show that there is a loss in diversity at ER 4. A

9 8 7 6 5 4 3 2 Toatl number of species of number Toatl 1 0 ER1 ER2 ER3 ER4 ER5 HR NR1 CR1 CR2 Sites

1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 Margalef species richness species Margalef 0.2 0 ER1 ER2 ER3 ER4 ER5 HR NR1 CR1 CR2 Sites

139 Chapter 5

1.2

0.8

0.6

0.4

Pielou's eveness index eveness Pielou's 0.2

0 ER1 ER2 ER3 ER4 ER5 HR NR1 CR1 CR2 Sites

1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 Shannon - Wiener diversity index diversity Wiener - Shannon 0 ER1 ER2 ER3 ER4 ER5 HR NR1 CR1 CR2 Sites

140 Chapter 5

In order to obtain an indication of the temporal and spatial trends among the fish communities, Bray-Curtis similarity-based cluster analysis and NMDS were performed. The data used for this analysis was not transformed in any way. The Bray-Curtis cluster analysis completed for the fish sampled at all the sites on the Elands and Crocodile rivers showed no clear temporal trends (Figure 5.3 A). Some spatial trends were obvious, with the sites on the Elands River separating from those on the Crocodile River. Although the site on the Crocodile River upstream and downstream of the confluence with the Elands River has been grouped together during low flow conditions, it separates during high flow. This is a indication that the modification to natural flow caused by Kwena Dam might be impacting the ecological integrity of the fish communities within the Crocodile River. Even sampling at CR 1 during high flow was complicated by extremely strong flow conditions not experience at any other sites. The main reason why the sites on the Crocodile River are grouped separately from those on the Elands River is that different species of fish are expected to occur in the Crocodile River. L. marequensis, for example, was sampled in the Crocodile River but not at any sites in the Elands River. The population of L. polylepis that occurs in the Elands River has grown in the absence of L. marequensis and shows various alterations to mouth parts (Godfrey and Roux, 2000). The NMDS obtained for all the sites in the study for both flow regimes shows the same groupings as the Bray-Curtis cluster analysis (Figure 5.3 B). The NMDS plot was produced following 30 iterations and showed a stress of 0.13. A stress value close to 0.1 is an indication that this is a good ordination with no real prospect of misleading interpretation.

The Bray-Curtis cluster analysis and NMDS plots for high flow conditions indicates that ER 3, CR 1 and CR 2 separated from the other sites in the study (Figures 5.4 A and 5.4 B). While ER 3 probably separated to the high abundance of non-indigenous fish, the Crocodile River sites separated due to domination of the community by some species. Figure 5.5 A and 5.5 B indicate that during low flow conditions the Crocodile River sites separated from the site on the Elands River, but ER 3 did not separate from the other sites. The fact that ER 3 did not separate from the other sites is a further indication that the fish communities at ER 3 are not responding to water quality changes brought about by human activities at this site.

141 Chapter 5 A

3 4

142 Chapter 5 A B

A B

143 Chapter 5

There are, however, indications that the communities might be responding to habitat changes and more particularly changes to flow. Compared to the amount of literature focusing on how paper and kraft mill effluents affect fish physiology (Munkittrick et al., 1992; Robinson et al., 1994; Karels et al., 2001; Sepulveda et al., 2001; Khan and Payne, 2002), the lack of information and agreement on the consequences of fish community responses to the effluents represents a crucial gap in understanding how these effluents impact freshwater ecosystems (Greenfield and Bart, 2005). Karels and Niemi (2002), for example, did report changes to the community structures of fish, while Kovacs et al. (2002) found no substantial differences in fish community structure between sites upstream and downstream from mill outfalls. Due to the fact that the effluent reaches the Elands River through groundwater after it has been irrigated on adjacent pastures, it is not entirely surprising that the fish communities do not appear to be affected. As there is no direct release of effluent into the river the effluent should not create a physico - chemical barrier that prevents any species of fish to migrate upstream of the mill.

The dominance curves in Figure 5.6 are in the form of ranked species abundance curves. Often one finds that where pollution is taking place the community structure is usually dominated by opportunistic species. These opportunistic species then dominate the biomass and abundance of the community (O’Brien, 2003). A B

100 100

ER1 ER1 80 80 ER2 ER2 ER3 60 ER3 60 ER4

ER4 ER5 40 40 HR HR

Cumulative Dominance% Cumulative Dominance% Cumulative NR1 20 20 CR1 CR1

0 CR2 0 CR2 1 10 1 10 Species rank Species rank

Figure 5.6 Ranked species K-dominance curves for the ichthyofauna communities collected at the sites on the Elands and Crocodile Rivers during high flow (A) and low flow conditions (B).

144 Chapter 5

For the plots in Figure 5.6 the relative abundance of species at each site was expressed as a percentage of the total abundance at the particular sites. This is plotted against the cumulative abundance at each site and this cumulative plot is known as the k-dominance plot. Figure 5.6 A indicates that during high flow conditions nearly 80 % of the community was dominated by a single species at ER 4, while more than 60 % of the community was dominated at CR 1 and HR. Figure 5.6 B shows that only ER 4 showed signs of domination during low flow conditions. During both flow regimes the community at ER 4 was dominated by B. argenteus, while this species dominated at HR during low flow. B. anoplus was also sampled in large numbers at HR in comparison to other sites in the study. P. philander dominated the community at CR 1. Fish communities in rivers are structured typically by complexity of habitat, environmental variables and periodic phenomena, such as low-flows and floods, and associated shifts in water quality (Pires et al., 1999). The environmental factors driving fish assemblages in the study area appear to include changes to habitat and flow and not primarily water quality effects as first expected.

Result of a SIMPER analysis for both flow regimes indicates the intergroup relationships between fish species (Table 5.3). The four groups identified by the Bray-Curtis analysis and NMDS ordination are (Figure 5.3): Group 1: All the sites on the Elands River during both flow regimes Group 2: CR 1 and CR 2 during low flow conditions Group 3: CR 1 during high flow Group 4: CR 2 during high flow

Table 5.3 The contribution of the various taxa to the similarity within groups (determined by using SIMPER) for both flow regimes. Average Cumulative Average Contribution Taxon Abundance % Similarity % (per site) contribution Barbus argenteus 28.83 35.03 58.53 58.53 Group 1 Chiloglanis pretoriae 18.08 15.07 25.17 83.7 Amphilius uranoscopus 6.33 6.03 10.07 93.77 Pseudocrenilabrus philander 48.5 46.07 85.42 85.42 Group 2 Chiloglanis pretoriae 14 7.87 14.58 100 Group 3 Less than two sites in the group Group 4 Less than two sites in the group

145 Chapter 5

The average similarity of the sites in group 1 was 59.86 %, with B. argenteus, C. pretoriae and A. uranoscopus contributing to more than 93 % of the abundance within the group, and B. argenteus being the most dominant. The species that have a 50 % or higher contribution to the formulation of the group may be good indicators of any disturbance taking place at the sites. Within this group B. argenteus is quit sensitive to pollution (based on perceived tolerance to water quality changes) and this is a sign that water quality has not been degraded to such a level that species are being lost. The fact that the sites upstream and downstream of the Ngodwana Mill are grouped together is a further indication that the fish communities within the Elands River are not responding to the activities at ER3. There was a high percentage of dissimilarity between group 1 and the other groups (group 1 and 2 = 67.76, group 1 and 3 = 77.81 and group 1 and 4 = 56.87). This possibly is an indication of different variables driving these groupings, although it is probably to differences in expected diversity. The sites within group 2 showed a similarity of 53.93 % with P. philander and C. pretoriae, making a large contribution to the abundances within this group. As with group 1 there was a high percentage of dissimilarity between this group (group 2 and 1 = 67.7, group 2 and 3 = 87.07, group 2 and 4 =79.02) and the rest of the groups in the study. Both groups 3 and 4 consisted of only one site, so the similarity of the sites within the group could not be determined. Both these groups show a high dissimilarity compared to the other groupings.

To provide possible insight to the basis for the spatial groups of sites identified in the Bray- Curtis analysis and NMDS ordination the ANOISM procedure was applied to the original similarity matrix for the ichthyofauna collected from the sites on the Elands and Crocodile Rivers during both flow regimes. The ANOISM procedure is short for analysis of similarities. The procedure compares all the sites in the study over both flow regimes to yield a test statistic and level of significance (O’Brien, 2003). For the results of this test the R value is taken as the degree of similarity between sites. This value ranges between 1 and -1. The R value is a constant and has been chosen so that: • R can never lay outside the range 1 to -1 • R = 1 only if all replicates within group are more similar to each other then any replicates from different groups • R = approximately 0 if the similarities between and within groups will be the same on average.

146 Chapter 5

The results of the ANOISM test performed on the groups determined by the Bray-Curtis cluster analysis and NMDS ordination indicate that there was a significant difference (p < 0.05) between the sites on the Elands and Crocodile rivers (group 1 and 2 = 0.864, group 1 and 3 = 0.977 and group 1 and 4 = 0.715). The groupings of the sites on the Crocodile River did not show a significant difference with the R value for group 2 and 3 and group 2 and 4 being one.

The RDA biplot for both flow regimes indicate (Figure 5.7), as with the Bray-Curtis similarity matrices and related NMDS plots (Figure 5.30), that there is a separation of the sites on the Elands and Crocodile rivers. In a biplot each arrow points in the direction of steepest increase of values for the corresponding variable. The angles between arrows indicate the sign of the correlation between the variables: the approximated correlation is positive when the angle is accute and negative when the angle is larger than 90 degrees. The length of the arrow is a measure of fit for the variables. The distance between the sampling sites in the diagram approximates the dissimilarity of their water quality variables as measured by their Euclidean distance. Unlike in Figure 5.3, Figure 5.7 clearly shows a separation of ER 3 from the other sites in the study (p = 0.3). The separation of the Crocodile River sites from the rest of the sites in the study was expected as different species of fish occur in this river compared to the other rivers in the study area. ER 3 separates from the other sites largely due to a higher conductivity at this site.

To determine the effect that flow has on the fish communities the flow regime as a variable was removed. The resultant RDA biplot (Figure 5.8) also shows the dissimilarity of ER 3 compared to the other study sites. The driving variables still appears to be conductivity. With the removal of flow regime as a variable all the sites are still grouped together, although these groupings are not significant (p = 0.31). This may be an indication that flow regime has very little impact on the fish community structure in the study area. The driving variable in the separation of the Crocodile River sites from those on the Elands River appear to be a combination of variables, including habitat (based on HQI scores).

147 Chapter 5

148 Chapter 5

A RDA was completed for the sites on the Elands River during both high flow and low flow with discharge added as a variable (Figures 5.9 and 5.10). It is evident from the results that during both flow regimes there is a separation of the sites above the Ngodwana mill from those below the mill. ER 4 and ER 5 in turn show a high dissimilarity in comparison to the sites above the mill and ER 3. It appears that discharge and the increase in conductivity has little effect on the community structure of the fish within the Elands River as statistical analysis show no significant differences between the sites above and below the mill (p = 1).

149 Chapter 5

150 Chapter 5

It became evident that there is a strong relationship between discharge and conductivity. This occurrence was expected as the increase in discharge is directly related to the groundwater entering the system above ER 3. Figures 5.9 and 5.10 show that there is no real correlation between FAII scores and habitat quality (based on HQI scores). For both flow periods, it appears that a combination of conductivity and nutrients as well as discharge (these variables always fall between the ER3 and ER4/ER5 grouping) are responsible for this separation between the sites below the mill and those above the mill. The increased discharge is not only responsible for the increase in conductivity observed in the system, but also seems to be responsible for altering of available habitat due to sustained inundation of geomorphological structures.

5.3.2 BIOTIC INDICES

In general it is normal practise to combine sampling results for both surveys when assessing the ecological integrity of fish communities. The fact remains that spatial and temporal variability are not independed and spatial and temporal differences may be significant (Meador and Matthews, 1992). For this study the results for both flow regimes were kept separate. A graphical description of FAII scores (Figure 5.11) indicates that during low flow conditions FAII scores were higher than scores obtained during high flow conditions. This occurrence is often encountered due to the fact that sampling of fish is made difficult by strong flows.

151 Chapter 5

100.0 90.0 80.0 70.0 60.0 50.0 40.0 FAII score FAII 30.0 20.0 10.0 0.0 ER1 ER2 ER3 ER4 ER5 HR NR1 CR1 CR2 Sites

High flow result (Figure 5.11) shows that there is a slight reduction in the ecological integrity of the fish communities as the river passes the Ngodwana Mill. The polynomial trend line that has been superimposed indicates that this change is barely noticeable. The fish community does show recovery at CR 2. Low flow results show the same trend with a slight reduction in the integrity of the fish community from ER 3 when compared to the site above the mill. As with high flow conditions, the river shows recovery and the ecological integrity at CR 2 shows improvement. Although fish communities may be responding to water quality changes, the increase in the volume of water at ER 3 may have an impact on the habitat of the sites below the Ngodwana Mill (O’Brien, 2003). The increase in the volume of water potentially causes a loss of some riffle habitat which is of vital importance in the Elands River. The loss of habitat may be the largest contributing factor to changes in the fish communities in the study area.

The scoring system applied in this study to assign ecological classes is represented in Table 5.4. Based on a high and low flow assessment, only two sites were within the recommended ecological category during high flow (Table 5.1). Both these sites were situated above the activities of the Ngodwana Mill and the sewage treatment works at ER 3. The results further indicate that the agricultural activities at ER 2 have a minimal impact on the ichthyofauna communities. The sites below the mill and in the Crocodile River show a deviation from the 152 Chapter 5 recommended conditions. Studies have shown that the fish diversity is lower in rivers that receive paper mill effluent and the fish community of the affected system tends to be numerically dominated by stress-tolerant species (Greenfield and Bart, 2005). However, Kovacs et al. (2002) found no substantial differences in fish community structure between sites upstream and downstream from pulp and paper mill outfalls. In the study area, where effluent is irrigated it is highly unlikely that the pulp and paper mill effluent is responsible for changes in the ecological integrity of fish downstream of the mill. There is an increase in the volume of water downstream of the Ngodwana Mill that may be linked to the mill’s activities. This increase in the volume of water and its possible impact on habitat has been linked to changes in the community structures of fish (Godfrey and Roux, 2000; O’Brien, 2003). During spatial and temporal analyses the effect this has on the community structure becomes clear (especially at ER 4). The flow modifications caused by the Kwena Dam are probably largely responsible for the change in ecological integrity observed in the Crocodile River.

Table 5.4 The FAII scoring system applied in this study for assigning ecological classes. FAII score Class Condition Colour 90 – 100 A Natural/unmodified 80 – 89 B Minimally modified 60 – 79 C Moderately modified 40 – 59 D Largely modified 20 – 39 E Seriously modified 0 - 19 F Critically modified

During low flow conditions the ecological integrity of the fish communities at all the sites in the study fell within the recommended ecological category (Table 5.2). The results do indicate that there was a decrease in the FAII score at the sites below the mill, but the ecological class did not change. In general one expect to find an improvement in community structures during low flow, as sampling is usually complicated during high flow.

The studies conducted by both O’Brien (2003) and Kleynhans (1999) determined that the ecological category for the section of the Elands River below the confluence with the Ngodwana River was in a natural state (class A). The present study shows this section to be in a class C during high flow and class B during low flow. The techniques applied in sampling were, nevertheless, not efficient enough to sample all the species that is expected to occur in this 153 Chapter 5 section of river. L. polylepis, for example, is difficult to sample by means of electroshocking. This is especially true if the river is in high flow. A. mossambica on the other hand are most often present during the summer and occur in low abundances. The fact that these species of fish were not sampled is not necessary an indication that they are not present. However, should only results of the fish sampled in riffle and rapid habitats be taken into account, the relative FAII score would reflect this section of river to be in a natural state (class A). The study conducted by O’Brien (2003) found that the lower section of the Crocodile River was in a D class and was reaching largely modified conditions. Kleynhans (1999) determined that this section of river was in a moderately modified state (class C). The current study showed that CR 2 was in a class C during high flow conditions, but improved to a class A during low flow conditions. O’Brien (2003) found the site on the upper section of the Crocodile River to be in a seriously modified state (class E); while Kleynhans (1999) stated that this section of river was in a largely natural state. The current study indicated that during high flow CR 1 was in a class C, but improved to a class B during low flow conditions. Kleynhans (1999) found that the flow modification in the Crocodile River brought about by the Kwena Dam hampered effective sampling. It was stated that these flow modifications may be altering the ichthyofaunal community structure. The constant change in ecological integrity observed in this study, including the difference between the results of different studies further supports the latter statement.

During the study ER 5 and NR 1 was sampled only during low flow conditions. The reason for this is that during high flow conditions ER 5 was inaccessible due to dense vegetation and NR 1 was added as a study site during the latter part of the study. During the study all the species sampled during the reserve determination was found, although not at all the site. L. polylepis and A. mossambica was not constantly sampled. As mentioned earlier L. polylepis has been identified as an indicator species for this resource unit. Another indicator species (A. uranoscopus) was sampled at all the sites in the study, except at CR 2 during low flow conditions. Two introduced species were sampled in the study. C. gariepinus was sampled at ER 1 and ER 3, while M. acutidens was sampled in large numbers at ER 3, possibly due to habitat availability. M. acutidens is usually found in shoals in clear, flowing or standing, open waters (Skelton, 2001). The effect that these introduced species has on the indigenous fish communities is not known, although at first glance this impact appears to be minimal (Godfrey and Roux, 2000).

154 Chapter 5

One of the biggest concerns regarding the fish communities sampled during the study was the low numbers of C. bifurcus present in the community. James and Barber (1991b) estimated that about three percent of a total catch should be made up by this species. A study they completed on the conservation of this critically endangered fish showed that this species was not flourishing. The current study, about 14 years later, indicates that this fish species is till under threat. Even though O’Brien (2003) sampled quite a few individuals of this species during high flow conditions, they were sampled in very low abundances in the present study. Although the upper Ngodwana River has been identified as a sanctuary for C. bifurcus, the survey at NR 1 indicated low population numbers. It is possible that the several exotic species of fish in the dam may have an impact on the C. bifurcus population, yet C. pretoriae that colonises the same habitats, does not appear to be affected. The same situation is encountered in the Elands River, with C. pretoriae occurring in much larger numbers than C. bifurcus, despite the fact that the river does not seem to be highly polluted (chapters 3 and 6). The reasons behind these low numbers of C. bifurcus are thus not known. Despite this it is still alarming that so few individuals of this species have been sampled in recent surveys.

5.4 CONCLUSION AND RRECOMMENDATIONSECOMMENDATIONS

The main aims of this chapter were to the describe and quantify the ichthyofaunal communities at various sites on the Elands and Crocodile rivers, as well as to determine how these communities respond to changes to their environment and more specifically activities related to the Ngodwana Paper mill. Linking community structure and environmental variables are possible when enough data is available. For this particular study, enough site specific data was available to make such an assessment.

FAII results indicate that the fish community within the Elands and Crocodile rivers appear to be in a largely natural state (class B) and falls within the recommended ecological integrity of a B class. A change in community characteristics may have taken place but species richness and presence of sensitive species indicate little modification. Despite the fact that FAII results indicated that at most sites the fish community structure are in a B class, the fish communities are not approaching an A class (natural) as stipulated in the RQOs set out by the comprehensive

155 Chapter 5 reserve determination. This occurrence is due to sampling techniques and is most likely not related to anthropogenic impacts.

FAII results, along with the community structure analyses indicate that the fish communities within the Elands River are not responding primarily to changes in water quality. There is little difference between sites upstream and downstream of the main activities of concern. These activities include the sewage treatment works at ER 3, the agricultural activities at ER 2, the informal settlements at ER 5 and the pulp and paper mill activities at the confluence with the Ngodwana River. There is, however, some domination of the community at ER 4. Results indicate that this may be due largely to changes in habitat and not water quality changes. The changes in fish habitat within the Elands and Crocodile rivers appear to be due to flow modifications. The flow modifications in the Crocodile River are due to the construction of the Kwena Dam and return flow from the various agricultural activities further upstream. The flow modifications in the Elands River are, however, due to an increase in the volume of water entering the system at ER 3 and may be linked to the irrigation practises of the Ngodwana Mill. The increase in water causes changes to flow velocities that ultimately drive the community structures at ER 4.

There are some unusual features among macroinvertebrate communities in the Lupelule River (see chapter 6) and assessment of the fish community at this site showed the same trend. Unusual large numbers of B. argenteus and B. anoplus was sampled at this site in comparison to other sites in the study. B. anoplus is quit common in this river compared to the other rivers in the study area and this occurrence should be further investigated.

As with several other studies conducted on the Elands and Crocodile rivers, C. bifurcus is still found in alarmingly low numbers. The focus should thus shift from assessing the conservation status of this fish to active breeding and repopulation programs if the continuous existence of this species in the Elands River is to be assured. Along with such a program habitat rehabilitation is of the utmost importance.

The families identified in the reserve as indicator species should be kept as indicators of flow and water quality. The contribution that habitat changes and water quality in the Elands River

156 Chapter 5 makes to community structures within the Crocodile River must be further investigated, should the reserve determination of the Crocodile River materialise. Along with this it is recommended a study should be conducted that focuses on the community structure of the fish within the Crocodile River below the Kwena Dam. This will give more insight into the effect of flow modifications on the fish communities caused by the dam, as well as return flow from the various agricultural activities within the vicinity of this section of river.

5.5 REFERENCES

Clark KR and Green RH (1988). Statistical design and analysis for a ‘biological effects’study. Marine Ecology progress Series. Vol. 46 pp 213 – 226.

Dale VH and Beyeler SC (2001). Challenges in the development and use of ecological indicators. Ecological Indicators. Vol. 1 pp 3 – 10.

Greenfield DI and Bart HL (2005). Long-term fish community dynamics from a blackwater stream receiving kraft mill effluent between 1973 and 1988. Hydrobiologia Vol. 534 pp 81–90.

Hill L (2005). Elands Catchment Comprehensive Reserve Determination Study, Mpumalanga Province, Ecological Classification and Ecological Water Requirements (quantity) Workshop Report, Contract Report for Sappi-Ngodwana, Submitted to the Department Water Affairs and Forestry, by the Division of Water Environment and Forestry Technology, CSIR, Pretoria. Report No. ENV-P-C 2004-019 pp 1 -98.

157 Chapter 5

Jungwirth M, Muhar S and Schmutz S (2000). Fundamentals of fish ecological integrity and their relation to the extended serial discontinuity concept. Hydrobiologia. Vol. 422/423 pp 85– 97.

Karels AE and Niemi A (2002). Fish community responses to pulp and paper mill effluents at the southern Lake Saimaa, Finland. Environmental Pollution. Vol. 116 pp 309–317.

Karels, A., E. Markkula & A. Oikari (2001). Reproductive, biochemical, physiological, and population responses in perch (Perca fluvatilis L.) and roach (Rutilus rutilus L.) downstream of two elemental chlorine-free pulp and paper mills. Environmental Toxicology and Chemistry. Vol. 20 pp 1517–1527.

Kleynhans CJ (1999). The development of a fish index to assess the biological integrity of South African rivers. Water SA. Vol. 25 (3) pp 265 – 278.

Kovacs TG, Martel PH and Voss RH (2002). Assessing the biological status of fish in a river receiving pulp and paper mill effluents. Environmental Pollution. Vol. 118 pp 123–140.

Lyons J, Navaro – Perez S, Cochran PA, Santana EC and Guzman – Arroyo M (1995). Index of biotic integrity based on fish assemblages for the conservation of streams and rivers in West – Central Mexico. Conservation Biology. Vol. 9 (3) pp 569 – 584.

Meador MR and Matthews WJ (1992). Spatial and temporal patterns in fish assemblage structure of an intermittent Texas stream. American Midland Naturalis. Vol. 127 pp 106 – 114.

Morgan MN (2002). Habitat Associations of Fish Assemblages in the Sulphur River, Texas. Michigan State University.

Munkittrick KR, Van der Kraak, McMaster ME, Portt CB (1992). Response of hepatic MFO activity and plasma sex steroids to secondary treatment of BKME and mill shut – down. Environmental toxicology and chemistry. Vol. 11 pp 1427 – 1439.

O’Brien GC (2003). An ecotoxicological investigation into the ecological integrity of a segment of the Elands River, Mpumalanga, South Africa. M.Sc. dissertation.

Pires AM, Cowx IG and Coelho MM (1999). Seasonal changes in fish community structure of intermittent streams in the middle reaches of the Guadaina basin. Journal of Fish Biology. Vol. 54 pp 235–249.

River Health Program (2005). State – of – Rivers Report: Monitoring and managing the ecological state of rivers in the Crocodile (West) Marico Water Management Area. Department of Environmental Affairs and Tourism, Pretoria.

Robinson RD, Carey JH, Solomon KR, Smith IR, Servos MR, Munkittrick KR (1994). Survey of receiving – water environmental impacts associated with discharge from pulp mills. 1. Mill

159 Chapter 5 characteristics, receiving – water chemical profiles and laboratory toxicity tests. Environmental toxicology and chemistry. Vol. 13 pp 1075 – 1088.

Schmutz S, Kaufmann M, Vogel B, Jungwirth M and Muhar S (2000). A multi-level concept for fish-based, river-type-specific assessment of ecological integrity. Hydrobiologia. Vol. 422/423 pp 279–289.

Shaw PJA (2003). Multivariate Statistics for the Environmental Science. Arnold Publishers, London.

Skelton PH (2001). A field guide to the freshwater fishes of Southern Africa. Southern publishers (Pty.) Ltd., Halfway House pp 1 – 395.

160

CHAPTER 6

MACROINVERTEBRATES

6.1 INTRODUCTION

With South Africa being a semi-arid country, the effective management and sustainable utilisation of water as a resource is becoming gradually more important. Management of this resource is ultimately based on resource monitoring (Roux, 2001). For this reason, development and application of various monitoring tools are important steps when trying to balance the protection of our water resources with social and economical development. Resident aquatic biota has increasingly formed part of these monitoring programs. The rationale is that aquatic biota integrates and reflects changes in their environment. Assessment of biological communities, like aquatic macroinvertebrates, can be used to present an integrated measurement of the integrity of an aquatic resource. Aquatic macroinvertebrates are aquatic insects (or the different life stages of insects), crabs, shrimps, leeches, etc. found in dames, lakes, streams, rivers, ponds, wetlands, etc. Resident aquatic macroinvertebrates are good, short-term indicators of ecological integrity because they integrate the effects of physical and chemical changes. They are adapted to live within certain environmental conditions and changes within this environment may adversely affect community composition and abundance. Integration of biological indicators (like aquatic macroinvertebrates) with chemical- and physical indicators will ultimately provide information on the ecological status of the river (Hill, 2005).

The Elands River consists of excellent riffle and rapid habitats (Roux, 2001). Along with this, water quality in the Elands River is in a relatively good state (chapter 3). The combination of excellent habitat availability and good water quality are two main contributing factors to the excellent biodiversity and abundance of aquatic macroinvertebrates in the Elands River. This biodiversity is, however, under constant threat due to possible changes in the water quality and flow within the study area. Agricultural activities, flow modifications, and rural settlements are just some activities that may cause these changes. The main concern remains the activities of the Ngodwana pulp and paper Mill. The major concerns regarding the pulp and paper Mill effluent, and more specifically the effluents they generate, are the salts and electrical conductivity associated with these effluents.

161 Chapter 6

The effects of salts on aquatic macroinvertebrates have extensively been studied (Palmer et al.; 1996, Zokufa et al., 2001; Williams et al., 2003; Kefford et al., 2004). It has become evident that exposure of aquatic macroinvertebrates to paper mill effluent result in toxic effects (Van Wijk and Hutchinson, 1995; Zokufa et al., 2001). This toxicity may cause mortalities and eventually alter the community structure. The Ngodwana Mill manages effluent discharge by irrigation onto pastures adjacent to the Elands River. Zokufa et al. (2001) showed that contaminated groundwater is the main route of exposure and is contributing to salinisation and may have an impact on the aquatic invertebrate communities in the Elands River.

The current index, being used to assess the status of riverine macroinvertebrates in South

The objective of this chapter was to assess the effect of anthropogenic activities on the invertebrate communities within the study area and to determine whether there were any temporal and spatial trends amongst these communities.

6.2 MATERIALS AND MEMETTTTHODSHODS

6.2.1 FIELD SURVEYS

Invertebrate samples were collected at the sites indicated in chapter 2 during both high flow (March 2005) and Low flow June (2005) by following the SASS 5 protocol

162 Chapter 6

(Dickens and Graham, 2002). This means that various biotopes in which macroinvertebrates may occur were sampled. There are three main biotopes. (1) Stones: which refer to stones in and out of the current and bedrock. These stones are usually found in areas where the movement of water prevents the settling of silt or sediment. (2) Vegetation: refers to the aquatic vegetation, whether it is marginal or submerged. (3) GSM (gravel, sand and mud) refers to fine stones, silt or sediment deposited over time as well as mud (Dickens and Graham, 2002). The biotopes sampled as well as the time spent sampling each biotope is presented in Table 6.1.

Table 6.1 Biotopes and sampling duration of the biotopes that were included in this study. Biotope Sampled Sampling time Stones in current 2.5 minutes Stones out current 1 minute Vegetation 2 m2

To maintain comparability between sites, GSM was not sampled as it does not occur at all of the sites selected. Throughout the study a point was made to sample as many habitat types as possible and these habitat types were based on both substrate and flow. Stones in current for example, was sampled were stones occur in rapids, slow flowing water, runs, glides, etc. The same applied to vegetation, where an attempt was made to sample vegetation occurring in fast flowing water and pools. As stipulated in the SASS 5 technique, isolated pools and backwater were avoided. After sampling each biotope, using the standard SASS net (1 mm mesh and dimensions of 30 cm x 30 cm x 30 cm), the samples were placed in an identification tray. Debris was removed and the macroinvertebrates were identified using an invertebrate guide (Gerber and Gabriel, 2002a). Identification took place on site for the set period of 15 minutes. If no new taxon was identified for 5 minutes, identification was stopped. The SASS score, number of taxa (NT) and the Average Score per Taxon (ASPT) was then calculated. These three biotic indices have been successful in the passed in reflecting patterns in macroinvertebrate community structure (Vos et al., 2002).

6.2.2 SPATIAL AND TEMPORAL ANALYSIS

The multivariate statistical assessment approach in this study was based on the approach and rational adopted by O’Brien (2003) during his study on the Elands River. Bray-Curtis similarity

163 Chapter 6 matrices, constructed from the abundances of the various aquatic macro-invertebrate taxa recorded at each site on the different sampling occasions, were subjected to group averaged clustering and two-dimensional non-metric Multidimensional Scaling (MDS) ordination (Clarke and Warwick, 1994). One-way Analysis of Similarities (ANOSIM) was used to determine the extent of the overall differences in macro-invertebrate composition among the various sites and flow conditions. For each ANOSIM test, the null hypothesis that there were no significant differences among groups was rejected when the significance level (P) was <5% (P<0.05). The extent of any significant differences produced by this test were determined using the R-statistic value (Clarke and Green, 1988), which can range from +1, i.e. all samples within each of the groups are more similar to each other than to any of the samples from other groups, down to approximately zero, when average similarities within and between groups are the same (i.e. the null hypothesis). Small negative values of R are possible by chance under the null hypothesis, but are not generally interpretable since they correspond to similarities between groups being smaller than within groups and the ANOSIM test for R is thus one-sided. When the pairwise comparisons in the ANOSIM test detected a significant difference in invertebrate compositions between sites and flow regimes, Similarity Percentages (SIMPER) was used to identify which species typified each of those habitat types. To determine which environmental variables were possibly responsible for the various groupings a Redundancy Analysis (RDA) was completed using Canoco version 4.5. RDA is derivative of PCA with one additional feature. The values entered into the analysis are not the original data but the best-fit values estimated from a multiple linear regression between each variable in turn and a second matrix of environmental data. Thus the PCA is constrained to optimize a fit to the environment data so that this technique is the canonical version of PCA. Interpretation of RDA is through biplots (Shaw, 2003) which is a map of the samples being analysed on a two dimensional bases, where the placements of the samples reflect the (dis)similarities between the samples; in this case the sampling sites. Indices of diversity and evenness were applied to describe the species abundance relationships among the macro invertebrate communities. The Shannon-Wiener diversity index was applied. This index incorporates both the species richness and equitability components (O’Brien, 2003). The species richness was expressed as Margalef’s index, while the evenness was expressed using Pielou’s evenness index. Margalef’s index measures the number of individuals present for a given number of species, incorporating both the total number of species and the total number of individuals. Dominance curves have been included in the study to indicate which sites reflect relative increased dominance of species. 164 Chapter 6

6.2.3 LABORATORY ANALYSIS

After identification, the invertebrate sample was placed in polyethylene jars (honey jars). Samples were then fixed using a 10 % neutrally buffered formalin solution and stained using Rose Bengal. In the laboratory the macroinvertebrates were removed from the formalin solution and the total number of individuals of each family in the sample was recorded.

6.3 RESULTS AND DISCDISCUSSIONUSSION

6.3.1 SPATIAL AND TEMPORAL ANALYSIS

Primer version 5 was implemented to complete a range of univariate diversity tests. When considering the results for the total number of species (Figure 6.1 A), polynomial lines indicate that, when compared to ER 1 and ER 2, there is a reduction in the total number of taxa found at site ER 3 and ER4 during both flow regimes. This is an indication that there is a definite change in the community structure as the river flows past the Ngodwana Mill. The change in this species diversity is reflected by a lower SASS score obtained at site ER 3. The polynomial trend lines show an improvement in the number of taxa present further downstream. Margalef’s species richness index is an indication of species diversity and abundance. Result for this index (Figure 6.1 B) shows the same trend as the total number of species. The polynomial trend lines for both high and low flow indicate that there was a reduction in the diversity and the abundance at sites ER 3 and ER 4. Both the total number of species (Figure 6.1 A) and Margalef’s species richness index (Figure 6.1 B) indicates that there is a loss of diversity and abundance at sites HR and NR 2 during both flow regimes and this is also reflected in the SASS 5 result for these sites. This lower diversity found in the Lupelule River (HR) was expected as this is a mountain stream and probably has much lower productivity than the Elands River. Any flow modification (e.g. low level bridge) may have a serious impact on the integrity of the river. The change in diversity at the Ngodwana River is, however, not brought about by natural causes. The construction of the dam wall, together with the change in the natural flow that the dam wall causes, has a major impact on the diversity at this site. This impact is largely in the form of flow reduction and loss of habitat (chapter 4). Result indicated that along with this loss in diversity, there is a decrease in the abundance of macroinvertebrates at this site.

165 Chapter 6 A

40 35 30 25 20 15 10 Total number of species of number Total 5 0 ER1 ER2 ER3 ER4 ER5 HR NR NR2 CR1 CR2 Sites

Margelef species richness species Margelef 1

0 ER1 ER2 ER3 ER4 ER5 HR NR NR2 CR1 CR2 Sites

Pielou’s evenness index is an indication of how the individuals in a sample are distributed over the various species that make up a community. The polynomial trend lines show that during both flow regimes there is a definite change in the distribution at site ER 3 and CR 2 (Figure 6.2 A). This lack of evenness is a good indication of the dominance of a family or families at this site. Tables 6.6 and 6.7 do show that the mollusc family Thiaridae dominates the abundances at

166 Chapter 6

ER 3 during both flow regimes. The Shannon-Wiener diversity index followed the same trend as Pielou’s evenness index. There was a reduction in the diversity at site ER 3 and a slight reduction at site CR 2 (Figure 6.2 B). The domination of the community by one taxon at ER 3 may be linked to habitat availability. This habitat availability in turn is linked to eutrophication that is occurring to a number of activities in this area. The loss of diversity is nevertheless, linked to the contamination of surface water (also see Figure 6.6). A

1 0.9 0.8 0.7 0.6 0.5 0.4 0.3

Pielou's eveness index eveness Pielou's 0.2 0.1 0 ER1 ER2 ER3 ER4 ER5 HR NR NR2 CR1 CR2 Sites

2.5

1.5

0.5 Shannon - Wiener diversity index diversity Wiener - Shannon 0 ER1 ER2 ER3 ER4 ER5 HR NR NR2 CR1 CR2 Sites

Figure 6.2 Univariate diversity indices indicating Pielou’s evenness index (A) and Shannon- Wiener diversity index (B). Red bars indicating high flow while blue bars indicate low flow. A polynomial trend line has been overlain on the graph (broken line indicates high flow and solid line low flow. 167 Chapter 6

Cluster analysis of the invertebrate communities sampled at the sites on the Elands and Crocodile Rivers show a clear separation of the HR and ER 3 sites during both flow regimes (Figure 6.3 A). The analysis further shows that high flow conditions separate from low conditions. This is a clear indication that there is definite seasonal variation among the invertebrate community in the study area. The only site that did not follow this trend was NR 2, with low flow result for this site grouping with high flow results from other sites. Sites above the Mill and below the Mill have been grouped together during both flow regimes. This is an indication that other than for site ER 3 the macro-invertebrate fauna community structures are very similar at the sites above and below the Mill. The disturbance in the natural flow and water quality at site ER 3 is the main causes for separating this site from the rest. HR and ER 3 have been grouped independently from other sites due to dominance of certain taxa. At ER 3 there was a definite dominance of the community by the mollusc family Thiaridae. This dominance is possibly due habitat availability. At the HR site, there was a definite dominance of the mayfly family Caenidae. This family is moderately sensitive to pollution and prefers slow flowing water with GSM as a habitat. The reason why this family dominates the community structure at this site is not clear, although it could be attributed to the high organic content of the sediment (chapter 4). What is clear is that there is a distinct alteration to the invertebrate communities in the Lupelule River. The NMDS (Figure 6.3 B) plot was completed following 30 iterations and showed a stress of 0.09. The stress of less the 0.1 is an indication that this is a good ordination with no real prospect of misleading interpretation. The NMDS plot indicates that there is a separation of sites HR and ER 3 with separate groupings of the high and low flow conditions. NR 2 during low flow has again been grouped with high flow sites. The influence of the dam wall on natural flow has already been discussed.

When considering the community structure of the different flow conditions separately, the cluster analysis and NMDS plots for high flow conditions show the same trend (Figures 6.4 A and 6.4 B). There is a distinct separation of the site directly below the Mill (ER 3) from the other sites in the study. The same analysis for low flow conditions indicate that more sites, apart from ER 3 separated from the rest of the sites. ER 3 again separated from the rest of the site based on the dominance of the mollusc family Thiaridae, while HR separated due to the dominance of the mayfly family Caenidae.

168 Chapter 6 A

Figure 6.3 Bray-Curtis similarity matrix-based cluster analysis (A) and two dimensional representation of the NMDS ordination (B) of the aquatic macroinvertebrates collected at the sites on the Elands and Crocodile Rivers during both high (sites without suffix) and low flow (sites with the suffix – L) regimes. The NMDS ordination was completed with 30 iterations and showed a stress of 0.09.

169 Chapter 6

A B

Figure 6.4 Bray-Curtis similarity matrix-based cluster analysis (A) and two dimensional representation of the NMDS ordination (B) of the aquatic macroinvertebrates collected at the sites on the Elands and Crocodile Rivers during high flow conditions. The NMDS ordination was completed with 30 iterations and showed a stress of 0.01.

A B

Figure 6.5 Bray-Curtis similarity matrix-based cluster analysis (A) and two dimensional representation of the NMDS ordination (B) of the aquatic macroinvertebrates collected at the sites on the Elands and Crocodile Rivers during low conditions. The NMDS ordination was completed with 30 iterations and showed a stress of 0.03.

170 Chapter 6

The dominance of families at some sites during both flow regimes is clearly visible in Figures 6.6 A and 6.6 B. In the dominance plot for high flow (Figure 6.6 A), the dominance of Thiaridae is distinct at site ER 3, with this family contributing up to 90% of the community structure. During low flow there was a dominance of the community structure at three sites. This included the dominance of the Thiaridae family at ER 3, Caenidae family at HR and the dominance of the caddis fly family Hydropsychidae at CR 2.

The dominance curves (Figure 6.6) are in the form of ranked species abundance curves. Often one finds that where pollution is taking place the community structure is usually dominated by opportunistic species. These opportunistic species then dominate the biomass and abundance of the community (O’Brien, 2003). For the plots in Figure 6.6 the relative abundance of species at each site was expressed as a percentage of the total abundance at the particular sites. This is plotted against the cumulative abundance at each site and this cumulative plot is known as the k-dominance plot. This dominance of one invertebrate family downstream of the Mill was not present in the study of O’Brien (2003). The disturbance causing this domination is thus recent. One of the resource quality objectives (RQOs) set out by the comprehensive reserve determination is to ensure that no group consistently dominates the fauna (Hill, 2005). This resource quality objective is not being met.

A B

100 100 ER1 ER1 80 ER2 80 ER2 ER3 ER3 60 60 ER4 ER4 ER5 ER5 40 40 HR HR NR1 Cumulative Dominance% Cumulative Dominance% 20 NR2 20 NR2 CR1 CR1 0 CR2 0 CR2 1 10 100 1 10 100 Species rank Species rank

Figure 6.6 Ranked species K-dominance curves for the invertebrate communities collected at the sites on the Elands and Crocodile Rivers during high flow (A) and low flow conditions (B).

171 Chapter 6

Result of a SIMPER analysis for both flow regimes indicates the intergroup relationships between taxa (Table 6.2). The four groups identified by the Bray-Curtis analysis and NMDS (Figure 6.4 B) ordination are:

Group 1: ER 1, ER 2 ER 4, ER 5, NR 2, NR 2L, CR 1 and CR 2 Group 2: ER 1L, ER 2L, ER 4L, ER 5L, NR 1L, CR 1L and CR 2L Group 3: ER 3 and ER 3L Group 4: HR and HRL

The average similarity between taxa within group 1 was 40.7 %. Group 1 was all the high flow sites in the study apart from NR 2 during low flow conditions that has been grouped with these high flow sites. The families Baetidae and Hydropsychidae contribute to more then 50 % of the abundances within this group, with Baetidae being the more dominant taxon. The taxa that have a 50 % or higher contribution to the formulation of the group may be good indicators of any pollution taking place at this site. The high abundance of Baetidae and Hydropsychidae is, however, not an indication that pollution is taking place. These are just two of the more common families found in this resource unit (Godfrey and Roux, 2000). There was a high percentage of dissimilarity (group 1 and 2 = 77.03, group 1 and 3 = 92.03, group 1 and 4 = 66.94) between this group and other groups in this study. This is possibly an indication that there are different variables that are driving these groupings. The second grouping has a similarity of 50.2 %. This high similarity is an indication that similar taxa occurred at the sites that from this group. Group 2 are all the low flow sites that have been grouped together and Hydropsychidae and Baetidae families make up 52.61 % of the group. The difference between groups 1 and 2 is in the abundance of these families. As with group 1 there was a high percentage of dissimilarity between this group (group 2 and 1 = 77.03, group 2 and 3 = 80.94, group 2 and 4 =78.2) and the rest of the groups in the study. The driving variable that causes the difference between groups 1 and 2 is most likely flow.

Group 3 consisted of the site directly downstream of the Mill (ER 3) during both flow regimes. There was a high similarity amongst these sites (84.96 %). During both high and low flow the family Thiaridae made up 90.7 % of the community at this site. With this family being highly tolerant to changes within the environment it is possible that this site is polluted. Furthermore, this group shows a high dissimilarity (group 3 and 1 = 92.03, group 3 and 2 = 80.94, group 3 and 4 = 90.72) when compared to the other groups in this study. 172 Chapter 6

Table 6.2 The contribution of the various taxa to the similarity within groups (determined by using SIMPER). The groups were determined using Bray-Curtis cluster analysis and NMDS.

Average Abundance (per Average Contribution Cumulative % Taxon site) Similarity % contribution Baetidae 30.75 10.74 26.38 26.38 Hydropsychidae 31 10.68 26.23 52.61 Chironomidae 14 2.85 7 59.61 Leptophlebiidae 16.5 2.77 6.81 66.42 Group 1 Perlidae 9.38 2.17 5.34 71.76 Elmidae 12.13 1.77 4.34 76.09 Heptageniidae 12.75 1.76 4.32 80.41 Caenidae 7 1.26 3.1 83.51 Veliidae 4.38 1.06 2.6 86.11 Psephenidae 6 0.94 2.32 88.43 Ancylidae 3.88 0.86 2.11 90.54 Hydropsychidae 339.14 14.45 28.78 28.78 Baetidae 230.14 10.81 21.54 50.32 Elmidae 84.14 4.04 8.04 58.36 Leptophlebiidae 58.71 2.99 5.96 64.32 Simuliidae 53.14 2.53 5.04 69.37 Group 2 Tricorythidae 99.86 2.37 4.73 74.09 Caenidae 76.29 2.37 4.71 78.81 Heptageniidae 52 1.99 3.96 82.76 Chironomidae 47.43 1.96 3.91 86.67 Veliidae 19 0.94 1.87 88.54 Perlidae 16.14 0.85 1.69 90.23 Group 3 Thiardae 2000 77.91 91.7 91.7 Caenidae 97 21.38 39.84 39.84 Baetidae 31.5 8.81 16.41 56.25 Group 4 Athericidae 21 7.97 14.84 71.09 Chironomidae 16.5 5.03 9.38 80.47 Psephenidae 14.5 3.77 7.03 87.5 Hydropsychidae 11 2.1 3.91 91.41

The variable or variables that are causing the domination of Thiaridae at this site might be the variable that causes this dissimilarity of this group when compared to the other groups. The last grouping that was observed was that of the site on the Lupelule River (HR) during both flow regimes. The average similarity of the taxa within this group was 53.67. Nearly 40 % of the community consisted of the mayfly family Caenidae. There was a high dissimilarity (group 1 and 4 = 66.94, group 2 and 4 =78.2, group 3 and 4 = 90.72) between group 4 and the other groups. Thus the possibility exist that a different variable is driving the community structure at this site.

173 Chapter 6

To provide possible insight to the basis for the spatial groups of sites identified in the NMDS ordination the ANOISM procedure was applied to the original similarity matrix for the macroinvertebrates collected from the sites on the Elands and Crocodile rivers during both flow regimes. The ANOISM procedure is the acronym for analysis of similarities. The procedure compares all the sites in the study over both flow regimes to yield a test statistic and level of significance (Clark and Warwick, 2001). For the results of this test the R value is taken as the degree of similarity between sites. This value ranges between 1 and -1. The R value is a constant and has been chosen so that:

• R can never lie outside the range 1 to -1 • R = 1 only if all replicates within sites are more similar to each other then any replicates from different sites • R = approximately 0 if the similarities between and within sites will be the same on average.

The results of the ANOISM test performed on the groups determined by the Bray-Curtis similarity matrix derived cluster analysis and NMDS ordination indicate that there was a significant difference (p < 0.05; R = 0.807) between the two groups containing the sites during high and low flow conditions respectively. This is a further indication of the possible seasonal variation within the Elands River. There was also a significant difference between group 1 and group 3 (p < 0.05; R = 1) and group 1 and group 4 (p < 0.05; R = 0.478). Although seasonal variation may be the reason for the difference between high and low flow sites, water and/or habitat quality changes or the dominance of taxon may be the driving variables in the difference between group 1 and groups 3 and 4. As mentioned group 2 is significantly different from group 1, yet this group significantly differs (p < 0.05) from group 3 (R = 0.974) and group 4 (R = 0.987). The differences between these groups are caused by the dominance of taxon or changes to environment conditions. Apart from being significantly different from groups 1 and 2, groups 3 and 4 are also significantly different (p < 0.05) from each other (R= 1).

174 Chapter 6

The RDA biplot for both flow regimes indicate, as with as with the Bray-Curtis similarity matrices and related NMDS plots (Figure 6.3), that there is a distinct separation of high flow and low flow conditions. These groupings are significant (p = 0.006). The driving water quality variables responsible for the groupings (Figure 6.7) are oxygen and temperature. The increased oxygen availability may have been the cause of the seasonal variation observed in this study. The lower temperature is the cause of the higher oxygen availability. Dallas (2004a) and Soldner et al. (2004) also observed the effect that temperature change has on and oxygen availability. There is a clear separation of the site directly below the Mill (ER3) during both flow regimes (Figure 6.7). This is largely due to the increase in chloride concentrations and the corresponding increase in conductivity. To determine the effect that flow has on the invertebrate communities the flow regime as a variable was removed. The resultant RDA biplot (Figure6.8) also shows the dissimilarity of ER 3 compared to the other study sites. The driving variables still appears to be conductivity.

With the removal of flow regime as a variable nearly all the sites are grouped together (p = 0.006). The sites on the Lupelule River are grouped separately with nitrates being the driving variable during high flow and possibly turbidity during low flow (Figure 6.8). The separation of this site from the other study sites is a further indication of the unusual invertebrate assemblages at this site.

Figure 6.7 and 6.8 clearly show the relationship between SASS scores and habitat quality, further illustrating the importance of habitat as a driving variable in invertebrate community structures. A RDA was completed for the sites on the Elands River during both high flow and low flow with discharge added as a variable (Figures 6.9 and 6.10). It is evident from the results that during both flow regimes there is a separation of the sites above the Ngodwana Mill, from those below the mill. Sites ER 4 and ER 5 in turn show a distinct dissimilarity in comparison to the sites above the mill (ER 1 and ER 2) and ER 3 directly below the mill.

176 Chapter 6

ER 3 during both flow regimes is grouped separately from the other sites. Statistical analysis show that these groupings are not significant (p = 1). During both flow regimes a combination of variables were responsible for the dissimilarity of ER 3 compared to the other sites on the Elands River. It became evident that there is a strong relationship between discharge and conductivity. This occurrence was expected as the increase in discharge is directly related to the groundwater entering the system above ER 3. It has been mentioned that the groundwater is contaminated with the irrigated effluent (Claassen, 2005). The groundwater is the source of chlorides and thus the higher conductivity within the Elands River. It appears that during high flow conditions (Figure 6.9) there is a significant relationship between ortho – phosphates and nitrates. Both these variables are linked to the sewage released just above ER 3. The treated sewage released at this site is responsible for the unique community structure at this site. For both flow periods, it appears that a combination of conductivity and nutrients as well as discharge (these variables always fall between the ER3 and ER4/ER5 grouping) are responsible for this separation between the sites below the mill and those above the mill. The increased discharge is not only responsible for the increase in conductivity observed in the system, but also seems to be responsible for altering of available habitat due to sustained inundation of geomorphological structures.

6.3.2 BIOTIC INDICES

The recent comprehensive and intermediate reserve that has been completed for the Elands River state that the Present Ecological State (PES) for this resource unit is a B (Hill, 2005). This means that the invertebrate community structure for this section of river is largely in a good state. The proposed Ecological Management Class (EMC) that needs to be maintained to ensure that the Resource Quality Objectives are being met is a B. The alternative management scenario for this resource unit is set at a C class. Target indicator species that can be used to monitor whether targets are being met include, the water spec (Prospistomatidae). Other aquatic macroinvertebrates that are of importance are stoneflies (Perlidae), flat headed mayflies (Heptageniidae) and water pennies (Psephenidae) (Hill, 2005). SIMPER analysis (Table 6.2) indicates that the families Baetidae and Hydropsychidae made the largest contributions to the community structures at the sites on the Elands and Crocodile Rivers except at ER and HR where the Thiaridae and Caenidae families made the largest contribution. Despite this, the indicator taxon was present at most sites and should be kept as indicator taxon. The main reason for this is the fact that the indicators families require fast flowing, unpolluted water (Hill,

178 Chapter 6

2005). Graphical representation (Figure 6.11) of SASS 5 scores show that during the study higher SASS scores were obtained during low flow condition. In general higher scores are seen in high flow conditions, when the larger amount of water provides a higher dilution factor. River environments do, however, show variability with regards to flow, biotope availability and temperature (Dallas, 2004a). Variations in these factors may influence the abundance or distribution of aquatic macroinvertebrates. A study completed by Dallas (2004a) found that SASS scores in Mpumalanga did not vary significantly between seasons. Yet, there was an increase in the ASPT and the number of taxa recorded during winter low flow periods. Unlike the study conducted by Dallas (2004a), SASS scores showed a clear variation between flow regimes. The polynomial trend lines further indicate a clear reduction in SASS scores at site ER 3 (Figure 6.11). This is the site directly below the Ngodwana Mill and the springs where groundwater is discharged. It appears to be a clear indication that the activities of the Mill i.e. irrigation of effluent, is impacting upon the integrity of the invertebrate communities in the Elands River at this site. However, this is also the point where treated sewage is released by the Ngodwana sewage works, and where the Ngodwana River (with its poor water quality) enters the Elands River.

It becomes nearly impossible to separate the impacts of these activities. It needs to be stressed that regardless of which activity caused this impact; the integrity of the invertebrate community structure below the Ngodwana Mill is definitely compromised. SASS scores show some recovery of the invertebrate community integrity at the last site (CR 2), especially during low flow conditions. This is an indication that activities in the Elands River has little impact on the Crocodile River. The graphical representation (Figure 6.12) of the number of taxa recorded at each site in the study shows the same trend as SASS scores (Figure 6.11). There is a reduction in the number of taxa found at the site directly below the Mill during both flow regimes; there is however, a recovery when the river reaches the lower Crocodile River (CR 2). Dallas (2004a) found that there was an increase in the number of taxa recorded in Mpumalanga rivers during low flow conditions. This is largely due to the fact that low flow conditions in the province falls within winter. The lower temperatures in winter months increase the solubility of oxygen. The higher oxygen availability may increase the number of taxa present in the low flow conditions (Dallas, 2004a).

179 Chapter 6

250

200

150

100 SASS scores SASS

0 ER 1 ER 2 ER 3 ER 4 ER 5 HR NR 1 NR 2 CR 1 CR 2 Sites

40 35 30 25 20 15 Number of taxa of Number 10 5 0 ER 1 ER 2 ER 3 ER 4 ER 5 HR NR 1 NR 2 CR 1 CR 2 Sites

180 Chapter 6

Graphical representation of the ASPT scores (Figure 6.13) during both flow regimes show that there is no obvious trend during low flow conditions. The ASPT score directly below the Ngodwana Mill did show a reduction during high flow conditions. Although the SASS scores (Figure 6.11) and the number of taxa present at NR 2 was much lower in comparison to the other sites, the ASPT of these sites during both flow regimes was not affected. The main cause of changes to the invertebrate community integrity at this site was the loss of habitat, probably due to the changes in flow. Changes in flow may influence the habitat as a decrease in discharge may influence the wetted perimeter (Dallas, 2004a).

9 8 7 6 5

ASPT 4 3 2 1 0 ER 1 ER 2 ER 3 ER 4 ER 5 HR NR 1 NR 2 CR 1 CR 2 Sites

ER 1

Table 6.3 shows the method for assigning ecological classes applied in the study based on the relationship between SASS 5 scores and ASPT scores. In certain cases, where the SASS score assigns one category and the ASPT another, “half categories” were assigned e.g. B/C. During both high and low flow conditions the invertebrate communities at this site were in an A/B class (Tables 6.4 and 6.5). This means that the community structure was largely natural, with little anthropogenic impacts. In chapter 3 it was demonstrated that there are some impacts on the water quality above this site. These impacts seem to have little impact on the integrity of the

181 Chapter 6 community structure at this site. The available aquatic invertebrate habitat seems to be in a satisfactory state when looking at the IHAS scores for both high and low flow conditions (Tables 6.4 and 6.5). IHAS is commonly used with SASS to determine the quality of the habitat available to macroinvertebrates (Dickens and Graham, 2002).

Table 6.3 The invertebrate scoring system (SASS 5) results and the method applied in assigning ecological classes in this study. SASS 5 Score ASPT Condition Class Colour > 140 >7 Natural/unmodified A 100 – 140 5 – 7 Minimally modified B 60 – 100 3 – 5 Moderately modified C 30 – 60 2 – 3 Largely modified D <30 <2 Seriously modified E

The indicator species stoneflies, flat headed mayflies and water pennies where all present at this site during both high and low flow conditions (Tables 6.6 and 6.7). There was an improvement in the abundance and diversity of aquatic macroinvertebrates during low flow conditions when compared to high flow conditions. This is not the norm, as the community structure during high flow conditions should be less affected by pollutants due to the higher dilution factor. It is important, however, to note that during most high flow conditions sampling for aquatic macroinvertebrates are complicated by stronger flow conditions. It has been mentioned in the comprehensive reserve determination that the highly sensitive mayfly, the water spec (Prospistomatidae), should occur in the Elands River. A study completed by James and Barber (1991) also showed that this invertebrate occurs in the study area. This invertebrate was, however, not sampled in this survey or the earlier study by O’Brien (2003). It is possible that the water quality has changed over the last decade to such an extent that this family of macroinvertebrates do not occur in the study area any more. Another highly sensitive family of mayflies, Oligoneuridae was sampled, but it’s possible that this family of macroinvertebrates are more sensitive to changes in flow then water quality.

182 Chapter 6

Table 6.4 Result for SASS 5 (SASS score, number of taxa found and the average score per taxa) as recorded during high flow conditions. IHAS scores are calculated with 100 as a maximum score. ER 1 ER 2 ER 3 ER 4 ER 5 HR NR 1 NR 2 CR 1 CR 2

Stones 120 107 48 100 98 66 NS 31 111 97 SASS 5 score Vegetation 44 59 40 41 70 26 NS 45 40 -

TOTAL 136 143 62 107 122 77 NS 66 134 97

Stones 16 16 9 14 12 12 NS 9 15 14 Number of Vegetation 9 10 8 8 12 6 NS 9 6 - taxa

TOTAL 19 23 12 17 18 15 NS 12 19 14

Stones 7.5 6.7 5.3 7.1 8.7 5.5 NS 4.3 7.4 6.3 ASPT Vegetation 4.9 5.9 5 5.3 5.83 4.3 NS 5 6.7 -

TOTAL 7.2 6.2 5.2 6.3 6.7 5.2 NS 5.5 7 6.3

IHAS 75 81 75 66 73 69 - 55 79 66

Category A/B B B/C B B B/C B/C A/B B/C

ER 2

Although results show that there was degradation of the habitat quality during low flow conditions, the ecological category at this site improved from a B to and A class (Tables 6.4 and 6.5). This is an indication that the invertebrate community is largely in a natural state. There are almost no impacts at this site. The study completed by O’Brien (2003) found this site to be in a largely unmodified state during high flow conditions. During low flow conditions, however, the same study showed that there was a change in the integrity of the invertebrate communities. Although stone flies (Perlidae) were not sampled during high flow conditions, all other indicator families were present (Table 6.6). During low flow conditions all indicators of flow and water quality were sampled (Table 6.7). The mayfly of the family Polymitarcyidae was collected at this site. It is seldom collected, not only due to its sensitivity to water quality changes, but also due to its burrowing nature, making it difficult to collect.

183 Chapter 6

Table 6.5 Result for SASS 5 (SASS score, number of taxa found and the average score per taxa) as recorded during low flow conditions. ER 1 ER 2 ER 3 ER 4 ER 5 HR NR 1 NR 2 CR 1 CR 2

Stones 196 186 139 152 165 99 177 72 198 164 SASS 5 score Vegetation 67 74 69 75 38 52 103 15 103 63

TOTAL 208 230 166 207 177 123 203 75 228 195

Stones 28 25 23 26 23 14 25 9 29 24 Number of Vegetation 10 14 13 12 7 9 17 3 16 11 taxa

TOTAL 31 32 24 32 26 19 30 11 34 29

Stones 7 7.4 6 5.9 7.2 7 7.1 8 7.1 6.8 ASPT Vegetation 6.7 5.3 5.3 6.25 5.5 5.7 6.1 5 6.4 5.7

TOTAL 6.7 7.2 6.3 6.3 6.8 6.47 7 6.8 6.7 6.7

IHAS 75 60 71 68 73 78 83 57 82 65

Category A/B A A/B A/B A/B B A B/C A/B A/B

ER 3

SASS 5 results at this site indicate that the invertebrate communities at this site was in a B/C class during high flow conditions (Table 6.4). Did this improve during low flow and the invertebrate community was in an A/B state (Table 6.5). The ASPT for this site during high flow conditions was lower in comparison to reference sites and was the lowest of all the sites in the study. The fact that the IHAS score was still good is an indication that this change was due to a change in water quality. The principle that needs to be considered is that where habitat is poor the SASS score will decrease but the ASPT will be less affected (Dickens and Graham, 2002).

184 Chapter 6

Table 6.6 Abundances of invertebrate families collected during high flow conditions (S – Stones; VG – vegetation). Families indicated in grey represent indicator families. ER1 ER2 ER3 ER4 ER5 HR NR2 CR1 CR2 S VG S VG S VG S VG S VG S VG S VG S VG S VG Aeshnidae - - 4 - 4 1 - - - - 1 ------Ancylidae - - 3 - - - 9 - 9 2 2 - - - 5 - - - Athericidae ------1 - 23 - 1 - - 1 13 - Baetidae 38 6 39 39 34 1 7 21 3 31 18 3 7 16 3 11 9 - Belostomatidae ------1 - 2 - - - - - 10 1 - Caenidae 2 2 27 - 49 - 7 1 1 - 44 7 8 1 7 - - - Calopterygidae - - - - - 2 ------1 - - - - Ceratopogonidae ------1 ------Chironomidae 49 4 4 4 4 - 2 23 1 1 4 8 11 1 1 - 4 - Chlorolestidae - - - - 1 - 1 ------Coenagrionidae ------1 1 1 4 - - - 9 - 1 - - Corbiculidae 4 1 1 - 3 1 ------5 - Corixidae ------1 - - Ecnomidae 3 - 1 - - - 1 ------Elmidae 9 2 1 8 - - 2 54 3 2 7 1 1 - - - 13 - Gerridae - - 1 ------Gomphidae 1 1 2 - - 1 - - - - 3 - 1 - 2 - - - Gyrinidae - 2 1 - - - 1 - - - - 1 - - 1 1 - - Heptageniidae 28 - 18 9 4 - 4 2 3 - - - - - 3 35 - - Hydropsychidae 55 4 21 7 8 - 24 5 39 4 5 - - 2 8 - 53 - Hydroptilidae 2 ------Leptoceridae ------Leptophlebiidae 8 - 54 1 50 - 48 - 7 - 1 - 8 - 1 - 2 - Leptoceridae ------1 - - - Lestidae ------Libellulidae 2 - - 1 ------1 1 - 1 - 4 - Naucoridae - - - 3 - - - - - 5 - - - - - 8 1 - Nepidae - - - - - 1 ------Notonectidae ------1 - - Oligochaeta - - - - 10 5 - - - - 2 1 2 - - - 5 - Oligoneuridae 3 - 2 ------1 - - - Perlidae 19 - - - - - 5 - 13 1 - - - - 14 - 23 - Philopotamidae - - - - 1 ------Planorbidae - - - 1 ------Pleidae ------1 ------Polymitarcyidae - - 1 ------Potamonautidae - - 2 - 2 - 2 - 1 - 2 - 4 1 2 - - - Protoneuridae ------Psephenidae 12 - 20 - 18 - 10 - 1 1 9 - - - 2 - 1 - Psychomidae - - - 2 ------Pyralidae ------1 ------6 1 - Simuliidae 1 7 - - - - - 3 - - 1 - - 6 2 1 - - Tabanidae 1 - - - - - 1 - 1 - 2 - - - - - 1 - Tipulidae 1 - 2 - - - 2 ------Thiardae - - - - 1000 1000 1 4 - - 6 ------Tricorythidae 10 2 1 1 2 ------2 - 1 - 3 - Veliidae 6 5 - 1 - - - - - 7 - 2 - 9 - 7 - -

185 Chapter 6

Table 6.7 Abundances of invertebrate families collected during low flow conditions (S – Stones; VG – vegetation). Families indicated in grey represent indicator families. EL1 EL2 EL3 EL4 EL5 HR NR1 NR2 CR1 CR2 S VG S VG S VG S VG S VG S VG S VG S VG S VG S VG Aeshnidae 2 - 7 - 13 - 1 - - - 1 - 3 ------Ancylidae 3 - 9 3 23 4 29 1 91 12 3 2 4 3 3 - 27 2 - - Athericidae 11 1 3 - - - 1 - - 1 11 8 2 - 2 - 1 1 18 - Baetidae 297 15 57 6 128 38 204 47 188 380 31 11 35 110 11 5 71 19 166 15 Belostomatidae ------2 Caenidae 24 - 38 3 41 2 53 1 8 1 135 8 289 52 - - 20 17 25 3 Ceratopogonidae 2 - - 2 - - 2 - 2 - - - 2 1 - - 2 2 2 - Chironomidae 7 11 - 3 65 12 44 25 74 28 4 17 45 10 7 - 8 13 44 18 Chlorocyphidae - 1 2 ------1 3 1 2 - - 2 1 - - Chlorolestidae ------1 - - 3 ------Coenagrionidae - 2 - 2 - 2 - - - 34 - 8 - - - - 2 2 - 2 Corbiculidae 1 - 4 - 24 19 1- - 3 ------Corduliidae ------1 2 - Corixidae 1 - 1 - 5 - - 1 5 - - - - 2 - - 8 11 - - Ecnomidae 2 - - 1 - 1 2 - 3 7 ------4 - - - Elmidae 25 11 30 - 78 5 99 1 222 1 - - 50 3 2 - 54 2 88 3 Gerridae 11 - 2 - - 3 ------1 - - 15 Gomphidae 2 - 3 1 22 - 2 - 2 1 - - 28 1 - - 1 2 5 - Gyrinidae 3 - 5 2 2 - 6 - - 2 - - 2 - 1 - 1 - 5 2 Helodidae - - - - - 1 - - - 1 ------1 - - - Heptageniidae 131 2 78 4 30 1 14 - 4 - - - 108 - - - 8 4 7 4 Hydracarina ------2 - - - - - 2 - - - 2 - - - Hydrophilidae ------1 ------Hydropsychidae 147 7 103 4 192 8 350 2 356 109 12 5 333 10 21 5 81 10 830 32 Hydroptilidae 1 - - - - 2 2 - 1 - - - - 2 ------Leptoceridae - 1 ------1 ------1 Leptophlebiidae 48 - 121 - 111 - 89 - 72 5 4 - 18 3 3 - 8 - 47 - Leeches ------1 - Libellulidae 1 1 8 - 1 - 1 - 1 - - - 6 - - - 3 2 2 - Naucoridae ------5 5 - - - 1 ------Oligochaeta - - 2 2 2 2 ------3 8 8 - Perlidae 9 - 4 - 8 - 17 - 14 2 - - 1- - - - 35 - 22 - Philopotamidae 6 - 6 1 12 - 7 - 5 1 3 - 22 1 2 - 2 A 5 - Planorbidae - 2 - 3 - - - - - 1 - 2 - 2 - - - 1 - - Pleidae 2 ------Polymitarcyidae - - 9 ------1 - 1 - Porifera 2 ------11 ------Potamonautidae 1 - 2 - 9 1 3 - 1 - 5 - 6 2 - - 1 - - - Protoneuridae 2 ------Psephenidae 14 1 22 - 30 2 4 - 6 - 19 1 17 - 1 - 12 - 2 - Pyralidae 1 - - - 2 - - - 10 ------Simuliidae 7 5 5 2 13 - 72 16 28 65 4 - 47 20 1 - 62 8 30 5 Tabanidae 1 - 4 - 3 - 5 - 8 - 3 - 29 - - - 4 - 9 - Tipulidae - - 1 - 2 - 2 - 1 - - - 1 ------Thiardae - - 1 - 1000 1000 ------Turbellaria 9 - 5 - 3 - 8 - 8 - 4 - 10 - - 1 4 - 8 - Tricorythidae 18 - 44 - 3 ------375 7 - - 111 1 143 - Veliidae 13 3 6 - 11 1 12 - 45 5 1 2 21 2 - - 9 2 15 -

186 Chapter 6

O’Brien (2003) found that there was a change in the diversity of macroinvertebrates at this site and the present study shows this trend during high flow conditions. There was a definite reduction in the number of taxa found at this site (Table 6.6). Table 6.7 shows that the same trend occurred during low flow conditions. Stone flies (Perlidae) were not found at this site during high flow conditions (Table 6.6) and even though the other indicator families were present they occurred in low numbers. All indicator families were present during low flow conditions (Table 6.7). During both flow regimes large numbers of molluscs of the family Thiaridae occurred at this site. During this study a large amount of submerged aquatic vegetation was found at this site. Although this vegetation has not been identified, it was clear that it was true aquatic macrophytes and not filamentous algae. This is the only site in this study that had true aquatic vegetation present. Although the Thiaridae family is highly tolerant to pollution, they probably occur at this site because of the excellent habitat the aquatic vegetation provides. The high numbers of the families Baetidae, Chironomidae and Simuliidae that occurred at this site (Table 6.7) is an indication that eutrophication is taking place. Chironomidae are especially good indicators of organic pollution (Buss et al., 2002). Chapter 3 indicated that the treated sewage released at this site has an impact on the water quality. The increase in abundance of the families mentioned above further supports the assumption that the sewage being released is impacting the ecological integrity of the Elands River at this site.

ER 4

SASS 5 results indicate that during high flow conditions the integrity of the invertebrate community at this site was in a class B. This is an indication that this site is in a natural state (Table 6.4). As with the sites further upstream there was an improvement in this integrity during low flow conditions (Table 6.5). The integrity of the aquatic invertebrate community at this site was a class A/B during low flow. IHAS result during both flow regimes indicated some impacts on the habitat available to the aquatic macroinvertebrates (Tables 6.4 and 6.5). This change in habitat was mainly due to the presence of algae on the cobbles and rocks at this site. The presence of algae is linked to eutrophication taking place further upstream. Although Tables 6.6 and 6.7 shows that all indicator families were present during both flow regimes, the abundance of these families generally increased during low flow conditions. The high abundance of Baetidae, Chironomidae and Simuliidae during low flow conditions is an indication that eutrophication is

187 Chapter 6 definitely taking place at this site (O’Brien, 2003). Macroinvertebrates of the caddis fly family (Hydropsychidae) occurred in large numbers at this site. This family has a high preference for fast flowing water, and the large abundance may be an indication of the flow modifications identified in chapter 4 and 5 at this site.

ER 5

SASS 5 results at this site indicate that during both flow regimes the ecological integrity of the invertebrate community was in a class B (Tables 6.4 and 6.5). There were also fewer impacts on the habitat availability. In the past some unusual features of the invertebrate fauna have been observed at this site. Weddepohl et al. (1991) found that this was possibly due to local disturbances and not to the activities of the Ngodwana Mill. The abundances of Baetidae, Chironomidae and Simuliidae were extremely high at this site, indicating that eutrophication may be taking place (Table 6.7). Although the treated sewage being released in the vicinity of the Ngodwana Mill may be the main cause of this impact, the local community that uses the river at this site to fulfill basic human needs is also contributing to the nutrient loading. Compared to other sites in this study the abundances of macroinvertebrates of the family Elmidae, Hydropsychidae and Ancylidae were high. These families have different preferences in flow, habitat and water quality. This is an indication that there are some unknown local attributes or impacts at this site that may be driving the community structure. All indicator families were present during both flow regimes (Tables 6.6 and 6.7).

SASS 5 result shows that this site was in a B/C state during high flow (Table 6.4). These changes may be related to both changes in habitat and water quality. Low flow result show that there was an improvement in the integrity to a B class (Table 6.5). During both flow regimes, both stone flies (Perlidae) and flat headed may flies (Heptageniidae) were absent from the site. The family Psephenidae was sampled during both high and low flow conditions (Tables 6.6 and 6.7). These families have the same preferences for physical and flow habitats, but differ in there preference for water quality. The loss of stone flies (Perlidae) and flat headed may flies (Heptageniidae) may be a further indication that there is deterioration in the water quality and

188 Chapter 6 habitat at this site. A large number of individuals from the family Caenidae were sampled at this site. This family of macroinvertebrates has a high preference for muddy areas as a habitat and the domination of the community structure by this family and the loss of indicator families might be an indication of habitat loss, although this is not fully supported by IHAS results (Gerber and Gabriel, 2002b).

NR 1

After high flow sampling it became evident that the reference site ER 1 might be subjected to anthropogenic impacts from further upstream. NR 1 was thus included in the study as possible substitute reference conditions for this study. Therefore only low flow data are available for this site. SASS 5 results for this site supports the inclusion of this as a reference site in the study. The results show that there are very little impacts on the habitat availability and water quality and that the invertebrate community at this site is in a natural state and were thus classed as a class A (Table 6.5). All the indicator families were present at this site. There was a higher abundance of the family Gomphidae, Caenidae and Tabanidae at this site (Gerber and Gabriel, 2002b). These families have a high preference for GSM as a habitat. Although these families were collected in the stones in current habitat there is a high availability of GSM compared to other sites in the study. This higher abundance of these families could thus be expected. A large number of individuals of the family Tricorythidae were also sampled at this site (Table 6.7). This family prefers cobbles with fast flowing water as a habitat (Gerber and Gabriel, 2002b).

NR 2

SASS 5 results for this site signifies that the ecological integrity of the invertebrate community is definitely modified. During both high and low flow conditions this site was classed in a B/C class. However, during low flow conditions the ASPT score was still acceptable, indicating that habitat availability contributed to the lower SASS score. It is well known that apart from changes in water quality, degradation in habitat may influence the invertebrate community structure (Nedeau et al., 2003). During low flow conditions the SASS score and ASPT score was affected. This possibly points to a degradation in both habitat and water quality. During a survey conducted in 2000, (Godfey and Roux, 2000) it was found that this site was in an E

189 Chapter 6 class. This indicated that the invertebrate fauna was in a seriously modified state. During the study in 2000 only 10 taxa were found at this site. Although the ecological category for this site is below the recommended class of a B, it does fall within the alternative management class that has been set as a C class for this resource unit. Only one individual of the indicator family Psephenidae (water penny) was found at this site during low flow conditions (Table 6.7). None of the other indicator families were present during either low or high flow conditions. Apart from the family Hydropsychidae during low flow, none of the families found at this site showed appreciable abundances during the entire study.

CR 1

SASS 5 results for this site indicated that during both high and low flow the ecological integrity of the macroinvertebrates at this site was in an A/B class. This is an indication that anthropogenic impacts at this site are not altering the invertebrate community structures. Macroinvertebrates are, nevertheless, short-indicators of ecological integrity. Analysis of the fish communities at this site (chapter 5), did indicate that the flow modifications caused by the Kwena Dam has an impact on the integrity of the system. IHAS scores indicate that habitat availability is good (Tables 6.4 and 6.5) All indicator species were present during both flow regimes (Tables 6.6 and 6.7). There was a high abundance of the families preferring GSM as a habitat, during low flow conditions (Table 6.5). There are signs of sedimentation at this site caused largely by the modification of the natural flow. The presence of GSM habitat at this site is the main reason why these families were present. This is a sign of sedimentation and a further indication of flow modification.

CR 2

Although SASS results indicate that this site is in a B/C class during high flow conditions (Table 6.4) the water quality was not seriously degraded. Results show that even though the SASS score for this site was low, the ASPT score was unaffected. This is a good indication that habitat availability was affected. IHAS results for this site during high flow do show some signs of habitat degradation. Low flow results indicated an improvement in the ecological category to an A/B (Table 6.5), although there were signs of habitat related impacts. The water quality is of

190 Chapter 6 major concern as the effect the Elands River has on the Crocodile River is measured at this site. The good ASPT scores during both flow regimes is an indication that water quality from the Elands River has little effect on the invertebrate community integrity in the Crocodile River.

All indicator families were present at this site except for flat headed mayflies (Heptageniidae) during high flow conditions (Tables 6.6 and 6.7). The abundance of Baetidae, Chironomidae and Simuliidae families at this site during low flow conditions was high, indicating that eutrophication is taking place (Table 6.7). As with the upper Crocodile site, there was a high abundance of individuals of the Tricorythidae families at this site (Table 6.7). An unusually high abundance of macroinvertebrates of the family Hydropsychidae was present at this site during low flow conditions. It has been mentioned that this family has a particularly high preference for fast flowing water. This may be an indication that the flow modifications caused by the Kwena Dam have an impact on the community structures at this site.

6.4 CONCLUSION AND RRECOMMENDATIONSECOMMENDATIONS

The main objective of this chapter was to the describe and quantify the invertebrate communities at various sites on the Elands and Crocodile rivers, as well as to determine how these communities respond to changes to their environment and more specifically water quality. Linking community structure and environmental variables are possible when enough data is available. For this particular study, enough site specific data was available to make such an assessment.

It has become clear that the invertebrate communities within the Elands and Crocodile rivers are generally in a good state with minimum modification. During both flow regimes most sites in the study were within the recommended ecological category of a class B as determined in the comprehensive reserve determination for this resource unit, with the exception of ER 3, HR during high flow and NR 2 during both flows. During both flow regimes all sites fell within the alternative ecological category of a C class as determined by the ecological reserve. SASS 5 results indicate that there is a change in the integrity of the invertebrate communities at ER 3. It appears that the changes were brought about largely due to water quality changes (chapter 3), but changes to flow and habitat also made a contribution to the change in integrity. The change in flow caused by the construction of the dam wall above NR 2 has caused a serious loss of 191 Chapter 6 habitat. In combination with a degradation in water quality the integrity of the invertebrate community at this site is affected. Ultimately the Ngodwana River flows into the Elands River and the poor water quality within the Ngodwana River may have a compounding impact at ER 3. Although the Lupelule River is a mountain stream with low productivity, there are definite signs of anthropogenic impacts at site HR. These changes are reflected in the altered community structures observed at this site. It is recommended that the cause of these changes to the community structure in is investigated.

SASS 5 has proven to be effective when assessing the ecological integrity of the invertebrate communities in the study area. As with most indices that reduce complicated data to a single value to make them more accessible to non-specialists, some important data is lost when only looking at these results. A problem still remains with the assessment of the quality and quantity of habitat available to aquatic macroinvertebrates. The IHAS index has proved ineffective in the assessment of habitat quality, especially if small changes have occurred. With habitat being one of the major factors affecting invertebrate communities (along with water quality), the lack of an effective method for the assessment of habitat quality ultimately has a impact on the confidence in the interpretation of SASS results.

Where SASS results indicated whether water quality changes have taken place, multivariate and univariate statistics have proven more effective in determining the effect of these changes. From the data obtained through the suite of statistical analysis applied in the study it has become clear that the community structures at ER 3 and HR are definitely altered, with a domination of taxon at these sites. As mentioned the reason why there is a domination of the mayfly family Caenidae at this site is unknown. It is thus recommended that this occurrence should be further investigated. The domination of the community by the mollusc family Thiaridae at ER 3 is possibly due to the availability of habitat for this invertebrate family. This habitat (vegetation) may be present due to the combination of chlorides and conductivity from the pulp and paper Mill and the nutrients from the sewage works.

In general the community structure at ER 3, HR and NR 2 are dominated by tolerant invertebrate families, compared to the presence of the highly sensitive families in the upper sites of the study area. There is also a loss of diversity at these sites. The loss in diversity and sensitive families are attributed to various anthropogenic impacts including the activities of the Mill and sewage works at ER 3. Poor water quality and a loss of habitat caused these changes

192 Chapter 6

(loss in diversity and sensitive families) at NR 2. The reason why diversity and sensitive species is lost at HR is still unknown. The most important observation is the recovery of the Elands River, with the invertebrate community at CR 2 showing little impacts of the activities at ER 3.

Some of the RQOs set out by the reserve are not being met. There is domination of the community by a single taxon at some sites, perriniality and natural flow is lost at some sites and biodiversity of macroinvertebrates is also not being maintained at all sites.

The mayfly family Prosopistomatidae (water spec) was not sampled during the study or the study by O’Brien in 2003. It has been recorded in previous studies and forms part of the reference list for this resource unit. Although this may be due to seasonal patterns in the reproduction of this invertebrate it is still a cause of concern. The loss of this highly sensitive family may be an indication of changes to the water quality in the Elands River and this should be investigated. The families identified in the reserve as indicator taxon should be kept as indicators of flow and water quality. The abundance of Baetidae, Chironomidae and Hydropsychidae should, however, be monitored to determine whether they are indicators of organic pollution in these rivers. The contribution that the water quality in the Elands River makes to community structures within the Crocodile River must be further investigated, should the reserve determination of the Crocodile River materialise.

6.5 REFERENCES

Buss DF, Baptista DF, Silveira MP, Nessimian JL and Dorville, LFM (2002). Influence of water chemistry and environmental degradation on macroinvertebrate assemblages in a river basin in south-east Brazil. Hydrobiologia. Vol. 481 pp 125 – 136.

Clark KR and Green RH (1988). Statistical design and analysis for a ‘biological effects’study. Marine Ecology progress Series. Vol. 46 pp 213 – 226.

Clark KR and Warwick RM (1994). Change in marine communities: an approach to statistical analysis and interpretation. Manual for the PRIMER statistical programme. Natural environment research council.

193 Chapter 6

Dickens CWS and Graham PM (2002). The South African Scoring System (SASS), Version 5, Rapid bioassessment method for rivers. African Journal of Aquatic Science. Vol. 27 pp 1 – 10.

Dallas HF (2004b). Seasonal variability in macroinvertebrate assemblages in two regions of South Africa: implications for aquatic bioassessment. African Journal of Aquatic Science. Vol. 29 (2) pp 173 – 184.

Gerber A and Gabriel MJM (2002a). Aquatic macroinvertebrates of South African Rivers: Illustrations. Version 2. Institute for Water Quality Studies (IWQS). Department of Water Affairs and Forestry, Pretoria.

Gerber A and Gabriel MJM (2002b). Aquatic macroinvertebrates of South African Rivers: Field Guide. First Edition. Institute for Water Quality Studies (IWQS). Department of Water Affairs and Forestry, Pretoria.

Hill L (2005). Elands Catchment Comprehensive Reserve Determination Study, Mpumalanga Province, Ecological Classification and Ecological Water Requirements (quantity) Workshop Report, Contract Report for Sappi-Ngodwana, Submitted to the Department Water Affairs and Forestry, by the Division of Water Environment and Forestry Technology, CSIR, Pretoria. Report No. ENV-P-C 2004-019 pp 1 – 98.

194 Chapter 6

James NPE and Barber HM (1991). A survey of the fishes of the Elands and Crocodile Rivers in the vicinity of the Sappi Kraft pulp and paper Mill at Ngodwana, Eastern Transvaal. South African Institute of Aquatic Biodiversity. Investigational report No. 37.

Kefford BJ, Palmer CG, Pakhomova L and Nugegoda D (2004). Comparing test sytems to measure salinity tolerances of freshwatermacroinvertebrates. Water SA. Vol. 30 (4) pp 499 – 506.

Nedeau EJ, Merrit RW, Kaufman, MG (2003). The effect of an industrial effluent on an urban stream benthic community: water quality vs. habitat quality. Environmental Pollution. Vol. 123 pp 1-13.

O’Brien GC (2003). An ecotoxicological investigation into the ecological integrity of a segment of the Elands River, Mpumalanga, South Africa. M.Sc. dissertation.

Palmer CG, Goetsch PA, O’Keeffe JH (1996). Development of a recirculating artificial stream system to investigate the use of macro –macroinvertebrates as water quality indicators. WRC report No 475/1/96. Water Research Commission, Pretoria.

Roux DJ (2001). Development of the procedures for the implementation of the national River Health Programme in the province of Mpumalanga. WRC report No 850/1/01. Water Research Commission, Pretoria.

Shaw PJA (2003). Multivariate Statistics for the Environmental Science. Arnold Publishers, London.

Soldner M, Stephen I, Ramos L, Angus R, Wells, NC, Grosso A and Crane M (2004). Relationship between macroinvertebrate fauna and environmental variables in small streams of the Dominican Republic. Water Research. Vol. 38 pp 863 – 874.

195 Chapter 6

Van Wijk DJ Hutchinson TH (1995). The ecotoxicity of chlorate to aquatic organisms: a critical review. Ecotoxicology and environmental safety. Vol. 32 pp 244 – 253.

Vos P, Wepener V, Cyrus DP (2002). Efficiency of the SASS 4 rapid bioassessment protocol in determining river health: A case study on the Mhlathuze River, Kwazulu – Natal, South Africa. Water SA. Vol. 28 pp 13 – 22.

196

CHAPTER 7

ECOCLASSIFICATION

7.1 INTRODUCTION

The freshwater resources of our country will not be able to support the needs of humans and living organisms alike indefinitely. Environmental degradation and public pressure have led to the need for accurate assessment, both of the damage caused by human activities and of the improvement due to management practises (Norris and Hawkins, 2000). Because research has shown that our rivers are under constant threat and the fact that our government assumes responsibility for our freshwater resources, new procedures are always being developed to assess the “health” of our rivers. The most direct and effective measure of this health is the state of the life in the water (Karr and Chu, 2000). Based on the latter, the DWAF has developed procedure for river EcoClassification and EcoStatus determination for application in reserve determination and biomonitoring (Kleynhans and Louw, 2005). EcoClassification - the term used for Ecological Classification - refers to the determination and categorisation of the Present Ecological State (PES) of various attributes of rivers compared to the natural or close to natural reference condition. EcoStatus on the other hand refers to the determination of the integrated ecological category (EC) of rivers and implies some form of integration of the ecological categories of all the components (fish, invertebrates, geomorphology, water quality, etc.) that comprise the overall EcoStatus.

Ecological Classification is important to create an understanding of the PES and ecological functioning of the Elands River and, based on this, to set realistic ecological aims and objectives. For the present study the EcoClassification approach was followed and only some of the indices used to determine ecological categories were applied. These indices were the Macro Invertebrate Response Assessment Index (MIRAI) and the Fish Response Assessment Index (FRAI). Both these indices are based on the comparison of reference conditions to the PES by taking into account the various factors that may influence biotic assemblages. These “factors” are known as metrics. Metrics are systems of parameters or ways of quantitative

197 Chapter 7 assessment of a process that is to be measured, along with the processes to carry out such measurement. Metrics define what is to be measured.

They are usually specialised by the subject area, in which case they are valid only within a certain domain and cannot be directly benchmarked or interpreted outside it (Kleynhans and Louw, 2005). The MIRAI is used to determine the Invertebrate EC. It integrates the ecological requirements of the invertebrate taxa in a community or assemblage and their response to modified habitat conditions (Thirion, 2005). Although the MIRAI can be determined using information collected during a standard South African Scoring System (SASS) survey (Dickens and Graham 2002), it can also be determined using more detailed information such as diversity and abundances (as applied in the present study). FRAI assess the ecological integrity of the fish communities in a river. This procedure determines the ecological integrity of fish through an integration of ecological requirements of fish species in an assemblage and their derived or observed response to modified habitat conditions.

The fact that new indices have been developed is not an indication that the current indices used in the RHP are not efficient. SASS 5 and the FAII, for example, still form the basis of the RHP and various biomonitoring or management programs. The new indices have been developed due to the fact that they are more adaptable to types of information, types of rivers and are applicable in the RHP and reserve determinations that require a higher level of confidence. Although SASS 5 and the FAII are very effective in assessing the ecological integrity of invertebrate and fish communities respectively, they do have certain short comings. SASS, for example, is a good measure of water quality, but does not take into account the effect of changes like habitat or flow changes (Dickens and Graham 2002). The FAII on the other hand does not take into account the effect that exotic species has on fish assemblages. The presence of exotic fish is usually interpreted as a possible reason for a decrease in the ecological integrity of fish communities (Kleynhans, 1999). Although these aspects are not all that important for the RHP, it forms an integral part of the reserve.

The objective of this chapter was to use the new indices (made available through EcoClassification) to determine the PES of the aquatic macroinvertebrate and fish communities within the study area. As mentioned, EcoClassification refers to the determination and

198 Chapter 7 categorisation of the PES of various attributes of rivers compared to the natural or close to natural reference condition. The reference conditions used in the assessment of the biotic communities in the present study was obtained from both the intermediate and rapid reserve for the Elands River (Godfrey and Roux, 2000; Hill, 2005), as well as selected reference sites.

7.2 MATERIAL AND METMETHODSHODS

7.2.1 MIRAI

The invertebrate data applied in the MIRAI for this study was collected during the SASS survey (chapter 6). However, instead of using the abundances as determined by the SASS protocol, the diversity and abundances used for multivariate analysis was entered into the MIRAI data sheet. This data was then entered into the data sheet for MIRAI (see appendix A). After the ranking and weighting of the metrics were completed, each metric was rated according to the change from reference conditions. Reference conditions for the MIRAI index were obtained from the comprehensive reserve determination for the Elands River (Hill, 2005). The metrics that make up the MIRAI are: flow modification, habitat, water quality, connectivity and seasonality (see appendix A for examples). The stepwise procedure to completing the MIRAI is as follows:

1. Determine the reference conditions. 2. Complete the data sheet. 3. Fill in the “Season” column if applicable. 4. Rank and weight the flow modification metrics. 5. Sort the data according to the >0.6m/s velocity category. 6. Compare the observed (present) taxa to the expected (reference) taxa. 7. Rate the metric accordingly, indicating the reason for the rating in the comment block. 8. Repeat the process (3-6) for the other metrics and metric groups. 9. Rank and weight the metric groups.

7.2.2 FRAI

The fish data applied in the FRAI was obtained from the data collected during sampling for completion of the FAII (chapter 5). For the FRAI the number of species expected for the 199 Chapter 7 reference condition should be compared with the observed (sampled) data to determine deviation from the reference situation. Where sampling is not representative (not all habitats were sampled, for instance) or effective (difficult conditions to employ a particular sampling method), some generally common species may be absent. In the Elands River, for example, L. polylepis and A. mossambica were not sampled at all the sites due to the fact that these fish are difficult to sample but they are expected to occur at most of the sites. In such a situation the species likely to be present (based on habitat, presence of closely related species and other environmental conditions) may be used to supplement the list of “observed” species. If such an approach is used (as in the present study) it is essential that this be indicated explicitly, as it will have an influence on the confidence of the fish ecological integrity determination. The metrics assessed in the FRAI were: velocity-depth preferences cover preferences, flow requirements, migration, physico-chemical preferences and the effect of introduced species (see appendix A for examples).

7.2.3 RATING, RANKING AND WEIGHTING

For all the indices applied as part of EcoClassification and EcoStatus a rating system is used to determine changes to the ecological integrity of biotic (responders) and abiotic (drivers) components. A six-point rating system is followed, where metrics of the drivers and biological responses are scored in terms of the degree to which they have changed compared to the natural or reference (if necessary, half points such as 1.5 and so on can also be used).

0 = No discernable change from reference/close to reference 1 = Small modification from reference 2 = Moderate modification from reference 3 = Large modification from reference 4 = Serious modification from reference 5 = Extreme modification from reference

In the case of the FRAI, a modified approach is followed where changes in some metrics are interpreted in terms of an increase or decrease (Kleynhans, 2005).

200 Chapter 7

The principle of following a ranking-weighting approach is that not all driver or biological response metrics have the same relative ecological significance in all types of rivers. That is, a particular metric may be seriously modified but it may be of relatively low significance in terms of the functioning and integrity of the river. In another river (or a different section of the same river) this metric may, however, be of very high ecological importance. Thus, the ranking- weighting process is done separately from the rating and should not be influenced by it. The metric of the component (driver or biological response) that is considered to be most important in influencing the EC of the component if it changed is ranked as 1. The metric with a rank of 1 is awarded a weight of 100%. The weight of the metric with a rank of 2 is considered relative to its importance when compared to the metric with a rank of 1, and this can be any percentage lower than 100%. Where all metrics (or metric-groups) are ranked as 1, they will all receive a weight of 100%.

7.3 RESULTS AND DISCDISCUSSIONUSSION

The EC determined by MIRAI and FRAI range from A – F (>89=A; 80-89=B; 60-79=C; 40- 59=D; 20-39=E; <20=F) with A representing an invertebrate community that is natural with almost no modifications and F representing a completely modified community. The so-called ‘half categories’, e.g. B/C, are also used in cases where there are uncertainties as to whether the category is, for example, a definite B or a definite C.

7.3.1 MIRAI

MIRAI results for both flow regimes (Table 7.1) indicate that invertebrate community structures at most sites in the study fell within the recommended EC of a B class (Hill, 2005). All the sites in the study did prove to be within the alternative EC of a C class. MIRAI results (Table 7.1) show that ER 3 during high flow conditions was in a C class. The main cause of concern at this site was the poor habitat in the form of vegetation and flow modification that were occurring. The flow modifications were largely due to the increase in amount of water at this site, originating from the springs. The increase in the amount of groundwater in the system is directly linked to the irrigation practises of the mill.

201 Chapter 7

Table 7.1 MIRAI results (EC and ecological class) for the site on the Elands and Crocodile Rivers for both flow regimes. ER 1 ER 2 ER 3 ER 4 ER 5 HR NR 1 NR 2 CR 1 CR 2 EC for high 86.5 87.4 73.8 80.3 80.6 70.26 - 65.8 79.1 77.2 flow conditions Ecological class for high B B C B B C - C B/C C flow EC for low 91.5 91 88.7 87.2 87.9 77.5 90.2 63.7 88.7 88.1 flow conditions Ecological class for low A A A/B B B C A C A/B A/B flow

MIRAI results further indicate a large change to the community structure in terms of invertebrate families that prefer good water quality. This is a clear indication that the water quality at this site is impacted (appendix A). CR 2 during high flow was not within the recommended EC of a B class. There were impacts at this site in the form of flow modifications, habitat degradation and water quality. Habitat degradation was largely in the form of loss of marginal vegetation as a habitat, while water quality is due to the activities in the Elands River. The flow modification at this site is probably due to the presence of the Kwena Dam further upstream in the Crocodile River. The flow modifications in the Elands River may have a further compounding effect at this site (appendix A).

MIRAI results obtained for the tributaries indicate that both HR and NR 2 fell outside the recommended EC of a B class during both flow regimes. Although there were signs of water quality degradation at HR it is unlikely due to the fact that the river is a mountain stream and there are almost no impacts upstream of this site. What was evident was the lack of marginal vegetation at this site and this was reflected in the result for this site. Mountain streams often show a lack of marginal and aquatic vegetation (Dallas, 2004). The flow modifications caused a loss in families with a preference for fast and moderately fast flowing water. Low flow MIRAI results (Table 7.1) show an improvement in the ecological integrity, although flow modifications and a loss of marginal vegetation were still evident at this site. The importance of reference conditions became evident during the assessment of this site. The reference condition to which this site was compared was in the Elands River. Should this site be compared to a

202 Chapter 7 similar site on a different mountain stream in the study area, the results may have been different. Results for NR 2 during both flow regimes (Table 7.1) indicate that there are impacts on the water quality, flow and habitat at this site. The construction of the dam has been largely responsible for changes to the natural flow at this site with the loss in natural flow causing changes to the habitat that is available to aquatic invertebrates. Although the poor water quality at NR 2 is largely due to the little flow, run-off from the Ngodwana Mill is entering the Ngodwana River due to the proximity if the mill to this river. This run-off contributes to the deteriation in water quality at this site (appendix A).

7.3.2 FRAI

FRAI results indicate that all the sites in the study area fell within the recommended ecological category of a B (Hill, 2005), although not all sites were approaching natural (unmodified) conditions (Table 7.2). The upper Ngodwana appears to be in a natural state, further supporting the selection of this reach of river as a possible sanctuary for the critically endangered C. bifurcus. FRAI results did indicate that there were definite flow modifications at CR 1. This flow modification in turn altered the habitat at this site in terms of velocity-depth habitats. Results from CR 2 indicate that the flow modifications at CR 1 may have an impact further downstream in the Crocodile River. The flow modifications, in the section of the Crocodile River that forms part of this study, are not related only to the Kwena Dam. There are several agricultural activities in the area that are dependent on irrigation. The abstraction of water for these irrigation purposes and the return flow associated with this also affects the natural flow within the river. The impacts that these flow modification have on the community structure as well as sampling efficiency have been discussed in chapter 5.

Table 7.2 FRAI results (FRAI % and ecological class) for the site on the Elands and Crocodile Rivers for both flow regimes. ER 1 ER 2 ER 3 ER 4 ER 5 HR NR 1 NR 2 CR 1 CR 2 FRAI % 86.95 88.58 87.18 85.45 89.21 82.24 90.94 - 77.92 81.37 Ecological class B A/B B B A/B B A - B/C B

203 Chapter 7

7.4 CONCLUSION AND RRECOMMENDATIONSECOMMENDATIONS

The degradation of our countries’ water resources has long been a concern for the public and scientists alike. The management and assessment of the effects that human society has on these resources have thus become a focus point over the last decade. Many indices and techniques have been developed to assess the integrity of aquatic environments with assessments based largely on ecological principals. The latest indices to be developed proved highly effective in using natural attributes of biotic communities in the assessment of the ecological integrity of our rivers. Not only do they provide a clear and concise ecological category, but also offer possible reasons as to why the specific ecological category has been assigned.

Results obtained from the MIRAI indicate that the invertebrate community within the Elands River is largely in a good state. There are some impacts on the water quality at ER 3 that may be linked to the sewage treatment works and the paper mill activities at this site. The construction of the Kwena and Ngodwana dam has caused flow modifications at CR 2 and NR 2 respectively, altering the invertebrate community structures at these sites. The main concern regarding the invertebrate communities within the study area is the availability of marginal vegetation as a habitat. MIRAI results have indicated that there is a loss of families with a high preference for marginal vegetation at ALL sites that form part of the study area. Although there is some marginal vegetation in the form of reeds, this does not provide habitat of a good quality. The loss of marginal vegetation and the impact it has on the invertebrate communities should be further investigated. As with the SASS 5 protocol the invertebrate communities sampled in the study were only identified to family level. It is evident that the differences in sensitivities amongst species may influence results. Any further studies on the invertebrate communities of the study area should include identification of the aquatic invertebrate fauna up to at least genus level.

As with MIRAI results, FRAI result indicated that the fish communities within the study area are largely in a natural to minimally modified state. The only noticeable changes occurred at CR 1, largely in the form of flow modifications caused by the Kwena Dam. The results of sampling surveys were combined for the FRAI, so the differences between flow regimes could not be determined. Not all the species expected to occur in the study area were sampled. If no apparent

204 Chapter 7 reason can be given as to why as certain species has not been sampled, it is added to the samples list for the FRAI. This has an impact on the confidence level of FRAI results.

The new indices (MIRAI and FRAI) have proved highly effective in the assessment of anthropogenic impacts within the study area. Both indices provided a good indication of the ecological integrity of the riverine biota and made available reasons for any changes to biotic communities. The indices have a possible application in the prediction of changes to the biotic communities caused by possible potential impacts. Based on all the advantages of these indices, they should in future form a basic part of most biomonitoring programs. There are, however, future research needs regarding the tolerance of invertebrates and fish to the various factors that influence there community composition (flow, habitat, etc.).

7.5 REFERENCES

Dallas HF (2004). Spatial variability in macroinvertebrate assemblages: comparing regional and multivariate approaches for classifying reference sites in South Africa. African Journal of Aquatic Science. Vol. 29 (2) pp 161 – 171.

Dickens CWS and Graham PM (2002). The South African Scoring System (SASS), Version 5, Rapid bioassessment method for rivers. African Journal of Aquatic Science. Vol. 27 pp 1 – 10.

Hill L (2005). Elands Catchment Comprehensive Reserve Determination Study, Mpumalanga Province, Ecological Classification and Ecological Water Requirements (quantity) Workshop Report, Contract Report for Sappi-Ngodwana, Submitted to the Department Water Affairs and Forestry, by the Division of Water Environment and Forestry Technology, CSIR, Pretoria. Report No. ENV-P-C 2004-019 pp 1 -98.

205 Chapter 7

Karr JR and Chu EW (2000). Sustaining living rivers. Hydrobiologia. Vol. 422/423 pp 1 – 14.

Kleynhans CJ (1999) the development of a fish index to assess the biological integrity of South African rivers. Water SA. Vol. 25 (3) pp 265 – 278.

Norris RH and Hawkins CP (2000). Monitoring river health. Hydrobiologia. Vol. 435 pp 5–17.

206 Chapter 8 CHAPTER 8

GENERAL CONCLUSION AND RECOMMENDATIONS

8.1 CONCLUSION

It has become evident that the Elands and Crocodile River, along with their tributaries, are of great importance to the environment and the public alike. These systems support a rich variety of terrestrial and aquatic fauna and are a source of freshwater for several towns, cities and important industries. There are several activities in the area that pose a potential threat to the “health” of these rivers. This includes the numerous agricultural activities, sewage treatment works, informal settlements and pulp and paper mill activities. The latter is the greatest concern in the study area, especially in the Elands River. The pulp and paper industry is a large consumer of water. With South Africa being a water scarce nation the volume of water used and the quality of the water released needs to be constantly monitored. A few regrettable incidents over the years have given the industry a reputation as a major water polluter. The industry’s management of water is, however, of world class and every attempted is made to manage the environment in a sustainable manner.

The main aim of this study was to determine the impact anthropogenic activities has on the Elands River. From the results obtained in the study it has become obvious that the anthropogenic impacts within the study area have little impact on the aquatic fauna within both the Elands and Crocodile Rivers. Yet, some sites in the study area do show signs of human impacts. The stretch of river directly below the Ngodwana pulp and paper mill shows a definite deteriation in water quality, mostly in the form of high conductivity and chloride concentrations. Along with this, there are indications that nutrient loading is taking place. These activities appear to have little impact on the community structures of aquatic invertebrates and fish communities after the confluence with the Crocodile River. Directly below the source of pollution it has, however, become apparent that the invertebrate community is dominated by gastropods from the family Thiaridae. These snails are highly tolerant to pollution. Although this is a sign of a degradation in water quality, the combination of habitat availability and nutrient loading does contribute to this problem.

207 Chapter 8

With the exception of ER 4, the fish communities in the study area are not responding to the activities at ER 3. The increase in the volume of water entering the system at ER 3 has caused the loss of certain flow related habitats at ER 4 which coincided with the loss of some fish species. Apart from this, the pulp and paper mill activities appear to have little impact on the ecological integrity of the system. The irrigation of effluent is proving to be highly affective in the mitigation of water quality pollution. The problem is that the volume of groundwater appears to be increasing and this may lead to the drowning of riffle and rapid habitats downstream of the springs.

The agricultural activities within the Elands River appear to have no impact on the ecological integrity of the instream fauna. Although there are signs of erosion no sedimentation is taking place. It is quit clear that the flow within the river is sufficient enough to transport these sediments. In cases of extreme low flow, sedimentation and the impact it has on the river environment may become a concern. The affect that the agricultural activities within the Crocodile River have on the system is not yet known. Although there might be problems related to abstraction and return flow of water, it will probably be masked by the flow modification caused by the Kwena Dam. The agricultural activities in the study area are effecting the riparian vegetation. The increase in agricultural activities has coincided with encroachment of exotic species.

Flow modification within the study area is a major concern. The effect of the increasing volume of groundwater on flow in the system has been mentioned. Apart from this, the Kwena Dam causes flow modifications in the Crocodile River that appears to be affecting both the community structure of fish and the effective sampling of aquatic fauna. The Ngodwana Dam has caused serious flow modifications. These flow modifications have, intern, caused habitat degradation and water quality ultimately altering the invertebrate community structure below the dam wall. The section of river above the dam appears to be unimpacted.

The rural settlements near the Elands River appear to have no impact on the ecological integrity of the river. Although there are some impacts on water quality this is not reflected in the community structures of both fish and invertebrates.

208 Chapter 8

It is obvious that the ecological integrity of the instream biotic communities is not seriously affected by any of the activities in the area. Below the Ngodwana pulp and paper mill there is domination of the invertebrate community by a single family. It appears, however, that this is due to a combination of factors (activities). The Elands River does show recovery and there are almost no signs of water quality impacts after the confluence with the Crocodile River related to the activities in the Elands River. The few alterations to the community structures of the aquatic fauna, at both the upper and lower section of the Crocodile Rivers, appear to be related to flow modifications. There are three major concerns regarding the aquatic fauna in the study area. The first is the low numbers of Chiloglanis bifurcus that has been sampled. This critically endangered species will be lost should something not be done in the near future. Secondly, the highly sensitive family of mayfly Prospistomatidae has not been sampled in studies that have been completed over the past five years. This may be an indication that water quality has deteriorated to such an extent that it does not occur within the study area any more. Finally, there is a large number of exotic species occurring within the riparian zone at all the study sites.

8.2 RECOMMENDATIONS

• Further investigate the loss of marginal vegetations as a habitat for aquatic invertebrates and the effect it has on the invertebrate communities. • Further investigate the effect that the flow modifications, caused by the increase in the volume of groundwater, has on habitat quality and the resultant biotic community structure within the Elands River. • To determine the extent of flow modifications within the Crocodile River caused by the Kwena Dam and return flow from irrigation activities and ultimately, the effect it has on the biotic communities. • Determine the tolerances of selected aquatic invertebrates within the Elands River to pollution on a genus, but preferably a species level. • Development of an index, or refinement of the current indices available, to assess the availability and quality of habitat for aquatic macroinvertebrates.

209 Chapter 8

• Incorporation of SASS 5 on quarterly bases by the Ngodwana Mill, along with water quality to determine the effect of potential water quality changes on invertebrate communities. The samples should be collected at the sites where water samples are taken.

• The focus should shift from assessing the state of C. bifurcus, to an active breeding and release program.

210

APPENDIX A

ECOCLASSIFICATION SHEETS

MIRAI SPREADSHEETS

Final ecological category table

INVERTEBRATE EC METRIC GROUP SCORE GROUP GROUP GROUP GROUP GROUP GROUP METRIC METRIC METRIC METRIC WEIGHT WEIGHT RANK OF RANK SCORE OFSCORE WEIGHTED WEIGHTED CALCULATED CALCULATED CALCULATED %WEIGHT FOR FOR %WEIGHT FLOW MODIFICATION FM 90.7 0.351 31.8202 1 100 HABITAT H 90.5 0.263 23.8026 3 75 WATER QUALITY WQ 91.4 0.316 28.8474 2 90 CONNECTIVITY & SEASONALITY CS 100.0 0.070 7.01754 4 20 285 INVERTEBRATE EC 91.4877 INVERTEBRATE EC CATEGORY A

Flow modification metric

FLOW MODIFICATION METRICS. WITH REFERENCE TO VELOCITY PREFERENCES, WHAT ARE THE CHANGES TO THE FOLLOWING RATING RATING

OBSERVED OR EXPECTED TO BE? Weight % METRICS METRICS RANKING OF RANKING

Presence of taxa with a preference for very fast flowing water 0.5 2 90 Abundance and/or frequency of occurrence of taxa with a 0.5 1 100 preference for very fast flowing water Presence of taxa with a preference for moderately fast 0.5 2 90 flowing water Abundance and/or frequency of occurrence of taxa with a 0.5 3 75 preference for moderately fast flowing water Presence of taxa with a preference for slow flowing water 0.5 3 75 Abundance and/or frequency of occurrence of taxa with a 0.5 3 75 preference for slow flowing water Presence of taxa with a preference for standing water 1 4 60 Abundance and/or frequency of occurrence of taxa with a 1 4 60 preference for standing water 8 Overall % change in flow dependence of assemblage 78.125 625

Habitat modification metric

HABITAT MODIFICATION METRICS. WITH REFERENCE TO INVERTEBRATE HABITAT PREFERENCES, WHAT ARE THE CHANGES TO THE

FOLLOWING OBSERVED OR EXPECTED TO BE? RATING METRICS METRICS %WEIGHT %WEIGHT RANKING OF RANKING Has the occurrence of invertebrates with a preference for bedrock/boulders 0.5 4 60 changed relative to expected? Has the abundance and/or frequency of occurrence of any of the taxa with a 0.5 4 60 preference for bedrock/boulders changed? Has the occurrence of invertebrates with a preference for loose cobbles 0.5 1 100 changed relative to expected? Has the abundance and/or frequency of occurrence of any of the taxa with a 0.5 2 90 preference for loose cobbles changed? Has the occurrence of invertebrates with a preference for vegetation 1 2 90 changed relative to expected? Has the abundance and/or frequency of occurrence of any of the taxa with a 0.5 2 90 preference for vegetation changed? Has the occurrence of invertebrates with a preference for sand, gravel or 1 3 75 mud changed relative to expected? Have the abundance of any of the taxa with a preference for sand, gravel or 0.5 3 75 mud changed relative to expected? Has the occurrence of invertebrates with a preference for the water column 0.5 3 75 or water surface changed relative to expected? Has the abundance and/or frequency of occurrence of any of the taxa with a 0.5 3 75 preference for the water column/water surface changed?

10 Overall % change in flow dependence assemblage 79 790

Connectivity and seasonality metric

Based on observed and derived data, with reference to migration and seasonality, how did the following change? RATING % Weight Weight % METRICS METRICS RANKING OF RANKING

Impact on distribution of migratory taxa 0 1 100

Impact on abundance and/or frequency of occurrence of 0 2 90 migratory taxa Impact on occurrence of taxa with seasonal distribution

Impact on abundance and/or frequency of occurrence of taxa

with seasonal distribution 2

Overall % change in flow dependence of assemblage 95 190

Water quality metric

WATER QUALITY METRICS. WITH REFERENCE TO WATER QUALITY REQUIREMENTS, WHAT ARE THE CHANGES TO THE FOLLOWING OBSERVED EIGHT

OR EXPECTED TO BE? RATING METRICS METRICS % W % RANKING OF RANKING Have the number of taxa with a high requirement for unmodified physico- chemical conditions changed? 0.5 2 90 Have the abundance and/or frequency of occurrence of the taxa with a high requirement for unmodified physico-chemical conditions changed? 0.5 2 90 Have the number of taxa with a moderate requirement for unmodified physico-chemical conditions changed? 0.5 2 90 Have the abundance and/or frequency of occurrence of the taxa with a moderate requirement for modified physico-chemical conditions changed? 0.5 3 75 Have the number of taxa with a low requirement for unmodified physico- chemical conditions changed? 0.5 3 75 Have the abundance and/or frequency of occurrence of the taxa with a low requirement for unmodified physico-chemical conditions changed? 0.5 3 75 Have the number of taxa with a very low requirement for unmodified physico-chemical conditions changed? 1 4 60 Have the abundance and/or frequency of occurrence of the taxa with a very low requirement for unmodified physico-chemical conditions changed? 0.5 4 60 How does the total SASS score differ from expected? 0.5 2 90 How does the total ASPT score differ from expected? 0.5 1 100 10 Overall change to indicators of modified water quality 80.5 805

Data sheet

0.1- 0.3- Taxon Ref SASS SASS <0.1 0.3 0.6 >0.6 BEDROCK COBBLES VEG GSM WATER QUALITY Baetidae >2spp 50 - 100 312 2 2 2 2 2 2 2 2 1 HIGH Oligoneuridae <10 0 0 0 1 5 2 3 1 1 1 HIGH Ephemeridae 2 2 3 2 0 1 0 4 0 HIGH Amphipoda 1 2 3 2 0 2 2 3 0 HIGH Pyralidae 1 1 3 2 0 2 3 0 0 HIGH Helodidae <10 0 2 2 2 1 0 2 3 0 0 HIGH Hydropsalpingidae 0 1 3 4 2 3 2 0 0 HIGH Barbarochthonidae 0 2 3 1 2 3 2 0 0 HIGH Sericostomatidae 0 1 3 2 0 3 2 0 0 HIGH Perlidae 10 - 50 9 1 1 1 5 1 4 1 0 0 HIGH Notonemouridae 1 1 2 4 1 4 1 0 0 HIGH Telagonodidae 0 0 2 4 1 4 1 0 0 HIGH Prosopistomatidae 1 1 2 3 1 4 1 0 0 HIGH Heptageniidae 10 - 50 133 1 1 3 2 1 4 1 0 0 HIGH Hydropsychidae >2spp 10 - 50 154 0 1 2 4 2 3 1 0 0 HIGH Polycentropodidae 0 0 3 4 4 3 0 0 0 HIGH Blepharoceridae 0 0 3 4 2 3 0 0 0 HIGH Gyrinidae <10 3 1 2 2 3 0 0 0 0 5 LOW Naucoridae <10 0 2 2 3 0 1 1 1 1 4 LOW Hydrophilidae 0 2 2 0 0 0 3 2 2 LOW Dytiscidae 4 2 1 0 1 2 3 1 2 LOW Baetidae 1sp 2 2 2 2 2 2 2 2 1 LOW Baetidae 2spp 2 2 2 2 2 2 2 2 1 LOW Haliplidae 3 4 1 1 1 1 4 1 1 LOW Pleidae 4 1 0 0 0 0 4 0 1 LOW Gomphidae 10 - 50 2 0 2 3 0 0 1 0 5 0 LOW Corbiculidae <10 1 2 3 1 0 0 2 0 4 0 LOW Unionidae 2 3 1 1 0 1 0 4 0 LOW Sialidae 4 3 1 0 0 1 0 4 0 LOW Caenidae 10 - 50 24 3 2 1 1 0 2 1 3 0 LOW Tipulidae <10 0 3 4 1 1 1 2 0 3 0 LOW Tabanidae <10 1 2 3 1 0 0 2 0 3 0 LOW Ceratopogonidae <10 2 2 2 2 4 2 3 2 2 0 LOW Leptoceridae <10 1 0 1 3 2 2 2 2 2 0 LOW Coenagrionidae <10 2 1 2 3 1 0 1 4 1 0 LOW Hydroptilidae <10 1 0 3 2 2 1 2 3 1 0 LOW Empididae 0 0 2 4 1 4 1 1 0 LOW Libellulidae <10 2 1 2 3 1 1 4 0 1 0 LOW Viviparidae 3 2 0 0 1 2 3 0 0 LOW Simuliidae 10 - 50 12 0 2 2 4 2 3 2 0 0 LOW Hydropsychidae 1sp 0 1 2 4 2 3 1 0 0 LOW Hydropsychidae 2spp 0 1 2 4 2 3 1 0 0 LOW Porifera <10 2 2 2 2 2 3 2 1 0 0 LOW Ancylidae <10 3 1 2 2 1 3 2 1 0 0 LOW Dixidae 3 2 2 0 0 0 0 0 5 MODERATE Veliidae 10 - 50 16 5 1 1 0 0 0 0 0 5 MODERATE Gerridae <10 11 4 1 0 0 0 0 0 0 5 MODERATE Hydrometridae 4 1 0 0 0 0 2 0 4 MODERATE Hydraenidae 2 2 3 2 0 1 3 1 2 MODERATE Hydracarina <10 0 0 2 2 0 1 1 2 3 1 MODERATE Dipseudopsidae 4 1 0 0 0 1 0 4 0 MODERATE Calamoceratidae 4 1 0 0 0 2 2 3 0 MODERATE Corduliidae 2 3 1 0 0 2 1 3 0 MODERATE Limnichidae 2 3 2 0 0 2 0 3 0 MODERATE Polymitarcyidae <10 0 2 2 2 3 0 1 0 3 0 MODERATE

MIRAI RESULTS ER 3 during high flow conditions

FLOW MODIFICATION METRICS. WITH REFERENCE TO VELOCITY PREFERENCES, WHAT ARE THE CHANGES TO THE FOLLOWING RATING RATING

OBSERVED OR EXPECTED TO BE? Weight % METRICS METRICS RANKING OF RANKING

Presence of taxa with a preference for very fast flowing water 3 2 90 Abundance and/or frequency of occurrence of taxa with a 1 1 100 preference for very fast flowing water Presence of taxa with a preference for moderately fast flowing 2 2 90 water Abundance and/or frequency of occurrence of taxa with a 1 3 75 preference for moderately fast flowing water Presence of taxa with a preference for slow flowing water 3 3 75 Abundance and/or frequency of occurrence of taxa with a 1 3 75 preference for slow flowing water Presence of taxa with a preference for standing water 3 4 60 Abundance and/or frequency of occurrence of taxa with a 0.5 4 60 preference for standing water 8 Overall % change in flow dependence of assemblage 78.125 625

HABITAT MODIFICATION METRICS.

WITH REFERENCE TO INVERTEBRATE HABITAT S

PREFERENCES, WHAT ARE THE CHANGES TO THE HT GOF %WEIG RATING RATING RANKIN FOLLOWING OBSERVED OR EXPECTED TO BE? METRIC Has the occurrence of invertebrates with a preference for bedrock/boulders 1 4 60 changed relative to expected? Has the abundance and/or frequency of occurrence of any of the taxa with a 0.5 4 60 preference for bedrock/boulders changed? Has the occurrence of invertebrates with a preference for loose cobbles 2 1 100 changed relative to expected? Has the abundance and/or frequency of occurrence of any of the taxa with a 1 2 90 preference for loose cobbles changed? Has the occurrence of invertebrates with a preference for vegetation changed 3 2 90 relative to expected? Has the abundance and/or frequency of occurrence of any of the taxa with a 1 2 90 preference for vegetation changed? Has the occurrence of invertebrates with a preference for sand, gravel or 2 3 75 mud changed relative to expected? Have the abundance of any of the taxa with a preference for sand, gravel or 0.5 3 75 mud changed relative to expected? Has the occurrence of invertebrates with a preference for the water column 4 3 75 or water surface changed relative to expected? Has the abundance and/or frequency of occurrence of any of the taxa with a 0.5 3 75 preference for the water column/water surface changed? 10 Overall % change in flow dependence assemblage 79 790

WATER QUALITY METRICS. WITH REFERENCE TO WATER QUALITY REQUIREMENTS, WHAT ARE THE CHANGES TO THE FOLLOWING OBSERVED

OR EXPECTED TO BE? RATING METRICS METRICS % WEIGHT WEIGHT % RANKING OF RANKING Have the number of taxa with a high requirement for unmodified physico- chemical conditions changed? 2 2 90 Have the abundance and/or frequency of occurrence of the taxa with a high requirement for unmodified physico-chemical conditions changed? 2 2 90 Have the number of taxa with a moderate requirement for unmodified physico-chemical conditions changed? 2 2 90 Have the abundance and/or frequency of occurrence of the taxa with a moderate requirement for modified physico-chemical conditions changed? 0.5 3 75 Have the number of taxa with a low requirement for unmodified physico- chemical conditions changed? 3 3 75 Have the abundance and/or frequency of occurrence of the taxa with a low requirement for unmodified physico-chemical conditions changed? 1 3 75 Have the number of taxa with a very low requirement for unmodified physico-chemical conditions changed? 2 4 60 Have the abundance and/or frequency of occurrence of the taxa with a very low requirement for unmodified physico-chemical conditions changed? 2 4 60 How does the total SASS score differ from expected? 2 2 90 How does the total ASPT score differ from expected? 2 1 100 10 Overall change to indicators of modified water quality 80.5 805

INVERTEBRATE EC METRIC GROUP SCORE GROUP GROUP WEIGHT RANK OF RANK SCORE OF SCORE WEIGHTED WEIGHTED CALCULATED CALCULATED CALCULATED %WEIGHT FOR FOR %WEIGHT METRIC GROUP GROUP METRIC GROUP METRIC METRIC GROUP METRIC FLOW MODIFICATION FM 71.6 0.316 22.6184 2 90 HABITAT H 74.7 0.263 19.6579 3 75 WATER QUALITY WQ 70.1 0.351 24.5789 1 100 CONNECTIVITY & SEASONALITY CS 100.0 0.070 7.01754 4 20 285 INVERTEBRATE EC 73.8728 INVERTEBRATE EC CATEGORY C

CR 2 during high flow conditions

FLOW MODIFICATION METRICS. WITH REFERENCE TO VELOCITY PREFERENCES, WHAT ARE THE CHANGES TO THE FOLLOWING RATING RATING

OBSERVED OR EXPECTED TO BE? Weight % METRICS METRICS RANKING OF RANKING

Presence of taxa with a preference for very fast flowing water 3 2 90 Abundance and/or frequency of occurrence of taxa with a 0.5 2 90 preference for very fast flowing water Presence of taxa with a preference for moderately fast 2 1 100 flowing water Abundance and/or frequency of occurrence of taxa with a 1 3 80 preference for moderately fast flowing water Presence of taxa with a preference for slow flowing water 3 3 70 Abundance and/or frequency of occurrence of taxa with a 1 3 70 preference for slow flowing water Presence of taxa with a preference for standing water 3 4 50 Abundance and/or frequency of occurrence of taxa with a 2 5 40 preference for standing water 8 Overall % change in flow dependence of assemblage 73.75 590

HABITAT MODIFICATION METRICS. WITH REFERENCE TO INVERTEBRATE HABITAT PREFERENCES,

WHAT ARE THE CHANGES TO THE FOLLOWING OBSERVED OR OF RATING

EXPECTED TO BE? METRICS RANKING RANKING %WEIGHT %WEIGHT Has the occurrence of invertebrates with a preference for 0.5 5 40 bedrock/boulders changed relative to expected? Has the abundance and/or frequency of occurrence of any of the 0.5 5 40 taxa with a preference for bedrock/boulders changed? Has the occurrence of invertebrates with a preference for loose 2 1 100 cobbles changed relative to expected? Has the abundance and/or frequency of occurrence of any of the 1 2 90 taxa with a preference for loose cobbles changed? Has the occurrence of invertebrates with a preference for vegetation 2 2 90 changed relative to expected? Has the abundance and/or frequency of occurrence of any of the 1 4 60 taxa with a preference for vegetation changed? Has the occurrence of invertebrates with a preference for sand, 2 1 100 gravel or mud changed relative to expected? Have the abundance of any of the taxa with a preference for sand, 1 3 70 gravel or mud changed relative to expected? Has the occurrence of invertebrates with a preference for the water 4 3 70 column or water surface changed relative to expected? Has the abundance and/or frequency of occurrence of any of the 1 3 70 taxa with a preference for the water column/water surface changed?

WATER QUALITY METRICS. WITH REFERENCE TO WATER QUALITY REQUIREMENTS, WHAT ARE THE CHANGES TO THE FOLLOWING OBSERVED

OR EXPECTED TO BE? RATING RANKING RANKING % WEIGHT WEIGHT % OF METRICS OFMETRICS Have the number of taxa with a high requirement for unmodified physico- chemical conditions changed? 1 2 90 Have the abundance and/or frequency of occurrence of the taxa with a high requirement for unmodified physico-chemical conditions changed? 1 2 90 Have the number of taxa with a moderate requirement for unmodified physico-chemical conditions changed? 3 2 90 Have the abundance and/or fequency of occurrence of the taxa with a moderate requirement for modified physico-chemical conditions changed? 2 3 75 Have the number of taxa with a low requirement for unmodified physico- chemical conditions changed? 2 3 70 Have the abundance and/or frequency of occurrence of the taxa with a low requirement for unmodified physico-chemical conditions changed? 1 3 70 Have the number of taxa with a very low requirement for unmodified physico-chemical conditions changed? 3 4 50 Have the abundance and/or frequency of occurrence of the taxa with a very low requirement for unmodified physico-chemical conditions changed? 1 4 50 How does the total SASS score differ from expected? 0.5 2 90 How does the total ASPT score differ from expected? 0.5 1 100 10 Overall change to indicators of modified water quality 77.5 775

INVERTEBRATE EC METRIC GROUP SCORE GROUP GROUP WEIGHT WEIGHT RANK OF RANK SCORE OFSCORE WEIGHTED WEIGHTED CALCULATED CALCULATED CALCULATED CALCULATED %WEIGHT FOR FOR %WEIGHT METRIC GROUP GROUP METRIC GROUP METRIC GROUP METRIC FLOW MODIFICATION FM 72.4 0.321 23.2634 2 90 HABITAT H 76.2 0.250 19.05 3 70 WATER QUALITY WQ 77.9 0.357 27.8214 1 100 CONNECTIVITY & SEASONALITY CS 100.0 0.071 7.14286 4 20 280 INVERTEBRATE EC 77.2777 INVERTEBRATE EC CATEGORY C

HR during high flow conditions

FLOW MODIFICATION METRICS. WITH REFERENCE TO VELOCITY PREFERENCES, WHAT ARE THE CHANGES TO THE FOLLOWING RATING RATING

OBSERVED OR EXPECTED TO BE? Weight % METRICS METRICS RANKING OF RANKING Presence of taxa with a preference for very fast flowing 3 2 90 water Abundance and/or frequency of occurrence of taxa with a 2 1 100 preference for very fast flowing water Presence of taxa with a preference for moderately fast 2 2 90 flowing water Abundance and/or frequency of occurrence of taxa with a 2 3 75 preference for moderately fast flowing water Presence of taxa with a preference for slow flowing water 3 3 75 Abundance and/or frequency of occurrence of taxa with a 1 3 75 preference for slow flowing water Presence of taxa with a preference for standing water 3 4 60 Abundance and/or frequency of occurrence of taxa with a 2 4 60 preference for standing water 8 Overall % change in flow dependence of assemblage 78.125 625

HABITAT MODIFICATION METRICS.

WITH REFERENCE TO INVERTEBRATE HABITAT S

PREFERENCES, WHAT ARE THE CHANGES TO THE HT G OFG %WEIG RATING RANKIN FOLLOWING OBSERVED OR EXPECTED TO BE? METRIC Has the occurrence of invertebrates with a preference for bedrock/boulders 0.5 4 60 changed relative to expected? Has the abundance and/or frequency of occurrence of any of the taxa with 0.5 4 60 a preference for bedrock/boulders changed? Has the occurrence of invertebrates with a preference for loose cobbles 2 1 100 changed relative to expected? Has the abundance and/or frequency of occurrence of any of the taxa with 3 2 90 a preference for loose cobbles changed? Has the occurrence of invertebrates with a preference for vegetation 4 2 90 changed relative to expected? Has the abundance and/or frequency of occurrence of any of the taxa with 0.5 2 90 a preference for vegetation changed? Has the occurrence of invertebrates with a preference for sand, gravel or 2 3 75 mud changed relative to expected? Have the abundance of any of the taxa with a preference for sand, gravel 0.5 3 75 or mud changed relative to expected? Has the occurrence of invertebrates with a preference for the water column 3 3 75 or water surface changed relative to expected? Has the abundance and/or frequency of occurrence of any of the taxa with 1 3 75 a preference for the water column/water surface changed? 10 Overall % change in flow dependence assemblage 79 790

WATER QUALITY METRICS. WITH REFERENCE TO WATER QUALITY REQUIREMENTS, WHAT ARE THE CHANGES TO THE

FOLLOWING OBSERVED OR EXPECTED TO BE? RATING METRICS METRICS % WEIGHT WEIGHT % RANKING OF RANKING Have the number of taxa with a high requirement for unmodified physico-chemical conditions changed? 3 2 90 Have the abundance and/or frequency of occurrence of the taxa with a high requirement for unmodified physico-chemical conditions changed? 1 2 90 Have the number of taxa with a moderate requirement for unmodified physico-chemical conditions changed? 2 2 90 Have the abundance and/or fequency of occurrence of the taxa with a moderate requirement for modified physico-chemical conditions changed? 2 3 75 Have the number of taxa with a low requirement for unmodified physico-chemical conditions changed? 2 3 75 Have the abundance and/or frequency of occurrence of the taxa with a low requirement for unmodified physico-chemical conditions changed? 1 3 75 Have the number of taxa with a very low requirement for unmodified physico-chemical conditions changed? 3 4 60 Have the abundance and/or frequency of occurrence of the taxa with a very low requirement for unmodified physico-chemical conditions changed? 0.5 4 60 How does the total SASS score differ from expected? 3 2 90 How does the total ASPT score differ from expected? 2 1 100

INVERTEBRATE EC METRIC GROUP SCORE GROUP GROUP GROUP GROUP GROUP METRIC METRIC METRIC WEIGHT WEIGHT RANK OF RANK SCORE OFSCORE %WEIGHT WEIGHTED WEIGHTED FOR METRIC FORMETRIC CALCULATED CALCULATED CALCULATED CALCULATED FLOW MODIFICATION FM 65.0 0.316 20.5263 2 90 HABITAT H 71.6 0.263 18.8289 3 75 WATER QUALITY WQ 68.1 0.351 23.8947 1 100 CONNECTIVITY & SEASONALITY CS 100.0 0.070 7.01754 4 20 285 INVERTEBRATE EC 70.2675 INVERTEBRATE EC CATEGORY C

HR during low flow conditions

FLOW MODIFICATION METRICS. WITH REFERENCE TO VELOCITY PREFERENCES, WHAT ARE THE CHANGES TO THE FOLLOWING RATING RATING

OBSERVED OR EXPECTED TO BE? Weight % METRICS METRICS RANKING OF RANKING

Presence of taxa with a preference for very fast flowing water 2 2 90 Abundance and/or frequency of occurrence of taxa with a 1 1 100 preference for very fast flowing water Presence of taxa with a preference for moderately fast 3 2 90 flowing water Abundance and/or frequency of occurrence of taxa with a 0.5 3 75 preference for moderately fast flowing water Presence of taxa with a preference for slow flowing water 2 3 75 Abundance and/or frequency of occurrence of taxa with a 0.5 3 75 preference for slow flowing water Presence of taxa with a preference for standing water 3 4 60 Abundance and/or frequency of occurrence of taxa with a 1 4 60 preference for standing water 8 Overall % change in flow dependence of assemblage 78.125 625

HABITAT MODIFICATION METRICS. WITH REFERENCE TO INVERTEBRATE HABITAT

PREFERENCES, WHAT ARE THE CHANGES TO THE OF

FOLLOWING OBSERVED OR EXPECTED TO BE? RATING METRICS METRICS RANKING RANKING %WEIGHT %WEIGHT Has the occurrence of invertebrates with a preference for bedrock/boulders 0.5 4 60 changed relative to expected? Has the abundance and/or frequency of occurrence of any of the taxa with a 0.5 4 60 preference for bedrock/boulders changed? Has the occurrence of invertebrates with a preference for loose cobbles 2 1 100 changed relative to expected? Has the abundance and/or frequency of occurrence of any of the taxa with a 1 2 90 preference for loose cobbles changed? Has the occurrence of invertebrates with a preference for vegetation 2 2 90 changed relative to expected? Has the abundance and/or frequency of occurrence of any of the taxa with a 0.5 2 90 preference for vegetation changed? Has the occurrence of invertebrates with a preference for sand, gravel or 3 3 75 mud changed relative to expected? Have the abundance of any of the taxa with a preference for sand, gravel or 0.5 3 75 mud changed relative to expected? Has the occurrence of invertebrates with a preference for the water column 4 3 75 or water surface changed relative to expected? Has the abundance and/or frequency of occurrence of any of the taxa with a 0.5 3 75 preference for the water column/water surface changed? 10 Overall % change in flow dependence assemblage 79 790

WATER QUALITY METRICS. WITH REFERENCE TO WATER QUALITY REQUIREMENTS, WHAT ARE THE CHANGES TO THE FOLLOWING OBSERVED

OR EXPECTED TO BE? RATING METRICS METRICS % WEIGHT WEIGHT % RANKING OF RANKING Have the number of taxa with a high requirement for unmodified physico- chemical conditions changed? 3 2 90 Have the abundance and/or frequency of occurrence of the taxa with a high requirement for unmodified physico-chemical conditions changed? 0.5 2 90 Have the number of taxa with a moderate requirement for unmodified physico-chemical conditions changed? 2 2 90 Have the abundance and/or fequency of occurrence of the taxa with a moderate requirement for modified physico-chemical conditions changed? 1 3 75 Have the number of taxa with a low requirement for unmodified physico- chemical conditions changed? 3 3 75 Have the abundance and/or frequency of occurrence of the taxa with a low requirement for unmodified physico-chemical conditions changed? 0.5 3 75 Have the number of taxa with a very low requirement for unmodified physico-chemical conditions changed? 3 4 60 Have the abundance and/or frequency of occurrence of the taxa with a very low requirement for unmodified physico-chemical conditions changed? 0.5 4 60 How does the total SASS score differ from expected? 1 2 90 How does the total ASPT score differ from expected? 0.5 1 100 10 Overall change to indicators of modified water quality 80.5 805

INVERTEBRATE EC METRIC GROUP SCORE GROUP GROUP GROUP GROUP GROUP GROUP METRIC METRIC METRIC METRIC WEIGHT WEIGHT RANK OF RANK SCORE OFSCORE WEIGHTED WEIGHTED CALCULATED CALCULATED CALCULATED CALCULATED %WEIGHT FOR FOR %WEIGHT FLOW MODIFICATION FM 74.6 0.316 23.5658 2 90 HABITAT H 76.5 0.263 20.1316 3 75 WATER QUALITY WQ 76.4 0.351 26.7895 1 100 CONNECTIVITY & SEASONALITY CS 100.0 0.070 7.01754 4 20 285 INVERTEBRATE EC 77.5044 INVERTEBRATE EC CATEGORY C

NR 2 during high flow conditions

FLOW MODIFICATION METRICS. WITH REFERENCE TO VELOCITY PREFERENCES, WHAT ARE THE CHANGES TO THE FOLLOWING RATING RATING

OBSERVED OR EXPECTED TO BE? Weight % METRICS METRICS RANKING OF RANKING

Presence of taxa with a preference for very fast flowing water 4 2 90 Abundance and/or frequency of occurrence of taxa with a 2 1 100 preference for very fast flowing water Presence of taxa with a preference for moderately fast 2 2 90 flowing water Abundance and/or frequency of occurrence of taxa with a 2 3 75 preference for moderately fast flowing water Presence of taxa with a preference for slow flowing water 4 3 75 Abundance and/or frequency of occurrence of taxa with a 0.5 3 75 preference for slow flowing water Presence of taxa with a preference for standing water 3 4 60 Abundance and/or frequency of occurrence of taxa with a 2 4 60 preference for standing water 8 Overall % change in flow dependance of assemblage 78.125 625

HABITAT MODIFICATION METRICS. WITH REFERENCE TO INVERTEBRATE HABITAT

PREFERENCES, WHAT ARE THE CHANGES TO THE OF

FOLLOWING OBSERVED OR EXPECTED TO BE? RATING METRICS METRICS RANKING RANKING %WEIGHT Has the occurrence of invertebrates with a preference for bedrock/boulders 0.5 4 60 changed relative to expected? Has the abundance and/or frequency of occurrence of any of the taxa with a 0.5 4 60 preference for bedrock/boulders changed? Has the occurrence of invertebrates with a preference for loose cobbles 3 1 100 changed relative to expected? Has the abundance and/or frequency of occurrence of any of the taxa with a 2 2 90 preference for loose cobbles changed? Has the occurrence of invertebrates with a preference for vegetation 3 2 90 changed relative to expected? Has the abundance and/or frequency of occurrence of any of the taxa with a 0.5 2 90 preference for vegetation changed? Has the occurrence of invertebrates with a preference for sand, gravel or 3 3 75 mud changed relative to expected? Have the abundance of any of the taxa with a preference for sand, gravel or 1 3 75 mud changed relative to expected? Has the occurrence of invertebrates with a preference for the water column 4 3 75 or water surface changed relative to expected? Has the abundance and/or fequency of occurrence of any of the taxa with a 2 3 75 preference for the water column/water surface changed? 10 Overall % change in flow dependanceof assemblage 79 790

WATER QUALITY METRICS. WITH REFERENCE TO WATER QUALITY REQUIREMENTS, WHAT ARE THE CHANGES TO THE FOLLOWING OBSERVED

OR EXPECTED TO BE? RATING METRICS METRICS % WEIGHT WEIGHT % RANKING OF RANKING Have the number of taxa with a high requirement for unmodified physico- chemical conditions changed? 4 2 90 Have the abundance and/or frequency of occurrence of the taxa with a high requirement for unmodified physico-chemical conditions changed? 2 2 90 Have the number of taxa with a moderate requirement for unmodified physico-chemical conditions changed? 3 2 90 Have the abundance and/or fequency of occurrence of the taxa with a moderate requirement for modified physico-chemical conditions changed? 2 3 75 Have the number of taxa with a low requirement for unmodified physico- chemical conditions changed? 3 3 75 Have the abundance and/or frequency of occurrence of the taxa with a low requirement for unmodified physico-chemical conditions changed? 2 3 75 Have the number of taxa with a very low requirement for unmodified physico-chemical conditions changed? 3 4 60 Have the abundance and/or frequency of occurrence of the taxa with a very low requirement for unmodified physico-chemical conditions changed? 1 4 60 How does the total SASS score differ from expected? 3 2 90 How does the total ASPT score differ from expected? 1 1 100 10 Overall change to indicators of modified water quality 80.5 805

INVERTEBRATE EC METRIC GROUP SCORE GROUP GROUP WEIGHT RANK OF RANK SCORE OF SCORE WEIGHTED WEIGHTED CALCULATED CALCULATED CALCULATED CALCULATED %WEIGHT FOR FOR %WEIGHT METRIC GROUP GROUP METRIC GROUP METRIC GROUP METRIC FLOW MODIFICATION FM 61.8 0.316 19.5197 2 90 HABITAT H 67.9 0.263 17.8684 3 75 WATER QUALITY WQ 61.1 0.351 21.4386 1 100 CONNECTIVITY & SEASONALITY CS 100.0 0.070 7.01754 4 20 285 INVERTEBRATE EC 65.8443 INVERTEBRATE EC CATEGORY C

NR 2 during low flow conditions

FLOW MODIFICATION METRICS. WITH REFERENCE TO VELOCITY PREFERENCES, WHAT ARE THE CHANGES TO THE FOLLOWING RATING RATING

OBSERVED OR EXPECTED TO BE? Weight % METRICS METRICS RANKING OF RANKING Presence of taxa with a preference for very fast flowing 2 2 90 water Abundance and/or frequency of occurrence of taxa with a 1 1 100 preference for very fast flowing water Presence of taxa with a preference for moderately fast 2 2 90 flowing water Abundance and/or frequency of occurrence of taxa with a 2 3 75 preference for moderately fast flowing water Presence of taxa with a preference for slow flowing water 4 3 75 Abundance and/or frequency of occurrence of taxa with a 2 3 75 preference for slow flowing water Presence of taxa with a preference for standing water 5 4 60 Abundance and/or frequency of occurrence of taxa with a 2 4 60 preference for standing water 8 Overall % change in flow dependence of assemblage 78.125 625

HABITAT MODIFICATION METRICS. WITH REFERENCE TO INVERTEBRATE HABITAT

PREFERENCES, WHAT ARE THE CHANGES TO THE OF

FOLLOWING OBSERVED OR EXPECTED TO BE? RATING METRICS METRICS RANKING RANKING %WEIGHT %WEIGHT Has the occurrence of invertebrates with a preference for bedrock/boulders 0.5 4 60 changed relative to expected? Has the abundance and/or frequency of occurrence of any of the taxa with a 0.5 4 60 preference for bedrock/boulders changed? Has the occurrence of invertebrates with a preference for loose cobbles 3 1 100 changed relative to expected? Has the abundance and/or frequency of occurrence of any of the taxa with a 2 2 90 preference for loose cobbles changed? Has the occurrence of invertebrates with a preference for vegetation 4 2 90 changed relative to expected? Has the abundance and/or frequency of occurrence of any of the taxa with a 2 2 90 preference for vegetation changed? Has the occurrence of invertebrates with a preference for sand, gravel or 5 3 75 mud changed relative to expected? Have the abundance of any of the taxa with a preference for sand, gravel or 5 3 75 mud changed relative to expected? Has the occurrence of invertebrates with a preference for the water column 4 3 75 or water surface changed relative to expected? Has the abundance and/or frequency of occurrence of any of the taxa with a 1 3 75 preference for the water column/water surface changed? 10 Overall % change in flow dependence assemblage 79 790

WATER QUALITY METRICS. WITH REFERENCE TO WATER QUALITY REQUIREMENTS, WHAT ARE THE CHANGES TO THE FOLLOWING OBSERVED

OR EXPECTED TO BE? RATING METRICS METRICS % WEIGHT WEIGHT % RANKING OF RANKING Have the number of taxa with a high requirement for unmodified physico- chemical conditions changed? 3 2 90 Have the abundance and/or frequency of occurrence of the taxa with a high requirement for unmodified physico-chemical conditions changed? 2 2 90 Have the number of taxa with a moderate requirement for unmodified physico-chemical conditions changed? 3 2 90 Have the abundance and/or fequency of occurrence of the taxa with a moderate requirement for modified physico-chemical conditions changed? 2 3 75 Have the number of taxa with a low requirement for unmodified physico- chemical conditions changed? 3 3 75 Have the abundance and/or frequency of occurrence of the taxa with a low requirement for unmodified physico-chemical conditions changed? 1 3 75 Have the number of taxa with a very low requirement for unmodified physico-chemical conditions changed? 4 4 60 Have the abundance and/or frequency of occurrence of the taxa with a very low requirement for unmodified physico-chemical conditions changed? 2 4 60 How does the total SASS score differ from expected? 3 2 90 How does the total ASPT score differ from expected? 0.5 1 100 10 Overall change to indicators of modified water quality 80.5 805

UP INVERTEBRATE EC METRIC GROUP SCORE GROUP GRO GROUP GROUP GROUP METRIC METRIC METRIC METRIC WEIGHT WEIGHT RANK OF RANK SCORE OF SCORE WEIGHTED WEIGHTED CALCULATED CALCULATED CALCULATED CALCULATED %WEIGHT FOR FOR %WEIGHT FLOW MODIFICATION FM 63.0 0.316 19.8947 2 90 HABITAT H 55.9 0.263 14.7105 3 75 WATER QUALITY WQ 63.0 0.351 22.1053 1 100 CONNECTIVITY & SEASONALITY CS 100.0 0.070 7.01754 4 20 285 INVERTEBRATE EC 63.7281 INVERTEBRATE EC CATEGORY C

FRAI SPREADSHEETS

Final ecological category table

FISH EC METRIC GROUPs GROUP GROUP GROUP GROUP GROUP: GROUP: RATING RATING METRIC METRIC METRIC METRIC RANK OF RANK % WEIGHT WEIGHT % WEIGHTED WEIGHTED RATING FOR FOR RATING FOR METRIC METRIC FOR CALCULATED CALCULATED VELOCITY-DEPTH METRICS 76.50 19.87 1 100 COVER METRICS 82.40 14.98 4 70 FLOW MODIFICATION METRICS 83.00 20.48 2 95 MIGRATION METRICS 100.00 2.60 6 10 PHYSICO-CHEMICAL METRICS 85.50 19.99 3 90 IMPACT OF INTRODUCED SPP 0.00 0.00 5 20 (NEGATIVE) 5.00 385.00 FRAI (%) 77.92 EC: FRAI C BOUNDARY EC

Velocity-depth metric

CHANGES IN COMMONNESS OF SPECIES WITH HIGH TO VERY HIGH PREFERENCE FOR VELOCITY DEPTH CLASSES

VELOCITY-DEPTH CLASSES METRICS BASED ON OBSERVED AND DERIVED DATA, AND WITH REFERENCE TO

VELOCITY-DEPTH CLASS PREFERENCES, HOW DID THE FOLLOWING RANK RATINGS RATINGS

CHANGE? %WEIGHT

Response of species with high to very high preference for FAST-DEEP conditions -1.0 2 90 10 Response of species with high to very high preference for FAST-SHALLOW conditions -1.0 1 0 Response of species with high to very high preference for SLOW-DEEP conditions -2.0 3 50 Response of species with high to very high preference for SLOW-SHALLOW conditions -3.0 4 60 Absolute sum 4 Absolute overall weighed % assemblage change 23.50

Cover metric

CHANGE IN COMMONNESS OF SPECIES WITH PREFERENCE FOR SPECIFIC COVER FEATURES

COVER METRICS: WITH REFERENCE TO CHANGES IN FISH COVER FEATURES, WHAT ARE

THE CHANGES TO THE FOLLOWING OBSERVED OR EXPECTED TO BE? RANK RATINGS RATINGS %WEIGHT %WEIGHT

Response of species with a very high to high preference for overhanging vegetation -2.0 2 70 Response of species with a very high to high preference for undercut banks and root wads -1.0 3 40 Response of species with a high to very high preference for a particular substrate type -2.0 1 100 Response of species with a high to very high preference for instream vegetation -1.0 4 20 Response of species with a very high to high preference for the water column -1.0 3 40 Absolute sum 5 Absolute overall % assemblage change 17.6

Flow modification metric

FLOW MODIFICATION METRICS: IMPACT ON SPECIES WITH DIFFERENT LEVELS OF FLOW DEPENDANCE

FLOW DEPENDANCE METRICS: BASED ON OBSERVED AND DERIVED DATA, AND WITH WITH

REFERENCE FLOW DEPENDANCE, HOW DID THE FOLLOWING RANK RATINGS RATINGS

CHANGE? %WEIGHT

Response of species intolerant of no-flow conditions -1.0 1 100 Response of species moderately intolerant of no-flow conditions 0.0 3 20 Response of species moderately tolerant of no-flow conditions -2.0 2 60 Response of species tolerant of no-flow conditions -2.0 2 60 Absolute sum 4 Absolute overall % assemblage change 17.0

Migration metric

IMPACT ON SPECIES WITH DIFFERENT LEVELS OF MIGRATORY REQUIREMENTS

MIGRATION METRICS: BASED ON OBSERVED AND DERIVED DATA, AND WITH REFERENCE TO

CHANGES IN SYSTEM CONNECTIVITY, HOW DID THE FOLLOWING RANK RATINGS RATINGS

CHANGE? %WEIGHT

Response in terms of distribution/abundance of spp with catchment scale movements 0.0 3 20 Response in terms of distribution/abundance of spp with requirement for movement between reaches or fish habitat segments 0.0 1 100 Response in terms of distribution/abundance of spp with requirement for movement within reach or fish habitat segment 0.0 2 70 Absolute sum 3 Absolute overall % change in assemblage longitudinal continuity 0.0

Physico-chemical condition metric

IMPACT ON SPECIES WITH DIFFERENT INTOLERANCE LEVELS TO CHANGE IN PHYSICO- CHEMICAL CONDITIONS

PHYSICO-CHEMICAL METRICS: BASED ON OBSERVED AND DERIVED DATA, AND WITH REFERENCE TO INTOLERANCE TO MODIFIED PHYSICO- CHEMICAL CONDITIONS, HOW DID THE FOLLOWING RESPOND IN RANK RATINGS RATINGS TERMS OF FISH HEALTH AND CONDITION? %WEIGHT Response of species intolerant of modified physico-chemical conditions -1.0 1 100 Response of species moderately intolerant of modified physico-chemical conditions -2.0 4 20 Response of species moderately tolerant of modified physico-chemical conditions -1.0 2 80 Response of species tolerant of modified physico-chemical conditions -1.0 3 70 Absolute sum 4 Absolute overall % impact on assemblage 14.5

Introduced species condition metric

INTRODUCED SPECIES IMPACT INTRODUCED SPECIES METRICS: WITH REFERENCE TO THE TYPES OF INTRODUCED SPECIES, THE CHARACTERISTICS OF THE HABITAT AND THE NATIVE

SPECIES, WHAT IS THE FOLLOWING OBSERVED OR EXPECTED RANK RATINGS RATINGS

TO BE? %WEIGHT The impact/potential impact of introduced competing/predaceous spp? 0.0 1 100 How widespread (frequency of occurrence) are introduced competing/predaceous spp? 0.0 2 90 The impact/potential impact of introduced habitat modifying spp? 0.0 3 70 How widespread (frequency of occurrence) are habitat modifying spp? 0.0 3 70 Absolute sum 4 Absolute overall potential % assemblage change 0.0