Ecological Indicators 69 (2016) 533–539
Contents lists available at ScienceDirect
Ecological Indicators journal homepage: www.elsevier.com/locate/ecolind
The effect of a large-scale irrigation scheme on the fish community structure and integrity of a subtropical river system in South Africa W. Malherbe a,∗ , V. Wepener a , J.H.J. van Vuren b a Water Research Group, Research Unit for Environmental Science and Management, School of Biological Sciences, Potchefstroom Campus, North-West University, Private Bag X6001, Potchefstroom, 2520, South Africa b Department of Zoology, University of Johannesburg, PO 524, Auckland Park, 2006, South Africa
a r t i c l e
i n f o
Article history: Received 3 October 2015 Received in revised form 2 May 2016 Accepted 4 May 2016 Keywords: Vaalharts Fish response Multivariate Agriculture
a b s t r a c t Large scale irrigation schemes are vitally important for food security in developing countries. This is especially relevant in subtropical countries where there is pressure on their water resources. However, the potential impacts on the fish communities of the rivers associated with these irrigation systems are extensive and potentially devastating. Therefore, the aim of the study was to evaluate the impact of the Vaalharts Irrigation Scheme (VHIS) on the fish community of two rivers (Harts and Vaal rivers) in the subtropical region of South Africa. The fish community was assessed during a three year period from 2007 to 2009 together with environmental and habitat quality parameters. A multivariate approach together with a local biotic index was used to determine the present ecological state and the environmental drivers responsible for the fish community structure. The results indicated that the fish community was in a largely natural state at the start of the VHIS and increasingly became modified due to various environmental parameters being affected by the irrigation scheme. Annual variation in the fish community structures was high while nitrate, zinc and sulphates corresponded with changes in the fish community. The outcome of the study highlighted that a lack of long term monitoring of fish community structures together with environmental and habitat parameters are a major challenge in many developing countries that can potentially affect management of irrigation schemes and the fish communities associated with the aquatic ecosystems. © 2016 Elsevier Ltd. All rights reserved.
1. Introduction Irrigation schemes are a necessity in many countries as they supply the countries with sufficient water to produce food to meet their increasing food demand (Fanadzo, 2012). In South Africa, the Vaalharts Irrigation Scheme (VHIS) is a large irrigation scheme utilising water from the Vaal River via numerous gravity fed canals (Fig. 1). The return flow from the VHIS enters the Harts River (via numerous subsurface and surface drainage canals) (Fig. 1) where after it flows back into the Vaal River. The irrigation scheme is approximately 40 000 ha and supplies a wide range of agriculture products that include maize, wheat, cotton, soft fruit, citrus and some ground nuts. The irrigation scheme’s return flow runs into the Harts River and increases the salinity, nutrients (Malherbe et al., 2013), flow velocity (Malherbe et al., 2015) and potentially a variety of pesticides (Malherbe et al., 2013).
∗ Corresponding author. E-mail addresses:
[email protected],
[email protected] (W. Malherbe). http://dx.doi.org/10.1016/j.ecolind.2016.05.005 1470-160X/© 2016 Elsevier Ltd. All rights reserved.
Fish communities have been used as a biological indicator since Karr (1981) developed the Index of Biotic Integrity (IBI) to assess environmental degradation. Since then various fish indices have been used around the world and South Africa is no exception with the development of the Fish Assemblage Integrity Index (FAII) (Kleynhans, 1999), the weighted Sensitivity Index of Biotic Integrity (SIBI) (Kotze, 2001), and the Fish Response Assessment Index (FRAI) (Kleynhans, 2007). Fish indices of biological integrity generate a category for the integrity of the fish community based on a matrix of fish community attributes such as habitat, flow and water quality. The use of an index to determine the fish community status can often result in community structure patterns beinglost or overlooked due to the simplification of the data. Therefore, a multivariate statistical approach is used to determine community structure patterns. This has the ability to decrease the complexity of the matrices into a graphical display of the relationships between samples and species. The community structure patterns can be identified and potentially indicate changes in the ecosystem (Clarke and Warwick, 1994; Malherbe et al., 2010). The aim of this study was to apply selected indicators of fish community structure and integrity and evaluate whether the VHIS
534
W. Malherbe et al. / Ecological Indicators 69 (2016) 533–539
Fig. 1. (A) Map indicating regional location of the study area (inset) and the various sampling sites from the relative reference site (HR1) to the sampling sites on the Harts (HR2–HR5) and Vaal (VR1–VR2) Rivers. (B) Vaalharts Irrigation Scheme graphical depiction of scheme structure.
has affected this community structure. The fish community indicators used included the use of multivariate statistical analyses and a biotic index, the Fish Response Assessment Index (FRAI) (Kleynhans, 2007), which makes use of biotic community metrics to assess present ecological state. 2. Materials and methods 2.1. Sampling site habitat Sampling sites comprised of seven sites on the Harts River (HR). Site HR1 was situated upstream of the VHIS and was chosen as a relative reference site for the VHIS. Habitat at site HR1 was composed of deep pools around bridge pylons, deep pool upstream of the bridge, rocky riffle section downstream (fast shallow habitat) with numerous smaller backwater areas. Marginal vegetation was present in the pool as well as in the riffle sections. Sites HR2–HR4 were situated along the length of the VHIS (Fig. 1) and the final site HR5 was situated downstream of the VHIS and the Spitskop Dam. Habitat at site HR2 to HR4 was comprised of slow deep sections upstream with marginal vegetation in the form of reeds while the downstream habitat was made up of fast shallow and slow deep habitats interspersed with marginal vegetation and overhanging trees. Additional habitat at site HR3 and HR4 included algal growth (mainly filamentous algae) in especially the deep water sections. Fish sampling surveys were also completed in the Spitskop Dam situated between sites HR4 and HR5. Site HR5 was located on a water monitoring weir with the upstream habitat thus mainly slow deep habitat while the downstream area were comprised of fast shallow habitat that flowed into slow deep areas with little marginal vegetation present. Filamentous algae were also dominant within the weir creating additional habitat. 2.2. Fish sampling The fish community in the Lower Harts River were sampled during one low flow season (July 2007) and two high flow seasons (January 2008 and February 2009). Sampling was carried out during all three surveys at five localities on the Lower Harts River (HR1–HR5) (Fig. 1).
Various sampling techniques were used to evaluate the fish community as recommended by river monitoring programmes in South Africa (DWAF, 1999; Kleynhans, 2007). All the caught fish were identified using the taxonomic keys in Skelton (2001). Fish were measured, counted and the data captured for analysis. Sampling of fast and shallow habitats was carried out using the electrofishing technique (Barbour et al., 1999). The 2007 and 2008 electrofishing surveys were carried out using a standard 220 V (2.5Kw) AC 50 Hz portable generator while the 2009 survey was carried out using a 12 V battery-operated SAMUS electrofishing apparatus (SAMUS 725 M Electrofisher, SAMUS Special Electronics, Poland). The battery operated electrofisher was used during the 2009 survey due to the added mobility but sampling times within each habitat was still the same as in the previous surveys. A medium (30 m; 22 mm mesh size) bagged seine net was used at sites with pools and slower flowing water. A gill net with various mesh sizes ranging from 28 mm to 100 mm were deployed for a set number of hours at Spitskop Dam and the two Vaal River sites. Electrofishing at these sites was ineffective due to the water depth. Medium sized (50 cm × 10m; 22 mm mesh size) fyke nets were used to sample fish at all the sites for a minimum of two hours per site during the 2007 and 2008 sampling surveys. The catch per unit effort was extremely low at most sites due to the depth.
2.3. Water and sediment analysis Water and sediment samples were collected to use as explanatory variables for the fish community structure analysis. The detailed water quality results can be found in Malherbe et al. (2013) and the sediment results have been published in Malherbe et al. (2015). Water and sediment samples were collected during each survey at each site for nutrient, ions, metal and pesticide analysis. Nutrient analysis was completed using a Merck Pharo 100 Spectroquant (Merck KGaA, Germany) and relevant test kits. Ions and metal concentrations were determined using an Ethos Microwave Digester with analysis on an Inductively Coupled Plasma Mass Spectrophotometer (ICP-MS). Pesticide analyses were completed in the sediment samples with analyses for synthetic pyrethroids, organochlorine and related pesticides using a GC-ECD method while organophosphorous pesticides were analysed using a GCFPD/NPD.
W. Malherbe et al. / Ecological Indicators 69 (2016) 533–539
535
Table 1 Reference species list for the different reaches within the Harts and Vaal rivers during the study (Kleynhans et al., 2007). References Species (HR1 – HR4)
Austroglanis sclateri (Boulenger, 1901) Barbus anoplus (Weber, 1897) Barbus paludinosus (Peters, 1852) Barbus trimaculatus (Peters, 1852) Clarias gariepinus (Burchell, 1822) * Gambusia affinis (Baird & girard, 1853) Labeo capensis (Smith, 1841) Labeo umbratus (Smith, 1841) Labeobarbus aeneus (Burchell, 1822) Labeobarbus kimberleyensis (Gilchrist & thompson, 1913) Pseudocrenilabrus philander (Weber, 1897) Tilapia sparrmanii (Smith, 1840)
Harts River
Vaal River
Reference Frequency of Occurrence (FROC)
Confidence
Relative abundance
Reference Frequency of Occurrence
Confidence
Relative abundance
– 3 3 3 4 2 3 2 3 – 4 3
– 3 4 4 4 3 4 3 4 – 4 4
– 2 2 2 1 2 2 2 2 – 2 2
1 3 3 3 4 2 3 2 3 2 4 3
3 3 4 4 4 3 4 3 4 3 4 4
1 2 2 2 1 2 2 2 2 1 2 2
FROC: 1 = Present at very few sites (<10% of sites); 2 = Present at few sites (>10–25%); 3 = Present at about >25–50% of sites; 4 = Present at most sites ( > 50–75%); 5 = Present at almost all sites ( > 75%). Relative abundance: 1 = 1–5 individuals; 2 = 6–50 individuals; 3 > 50 individuals. Confidence: 1 = Low confidence; 2 = Low to moderate; 3 = Moderate; 4 = Moderate to high; 5 = High. * Exotic species.
2.4. Statistical analysis Multivariate statistical analyses were used to evaluate spatial and temporal differences between the various sites based upon their fish community structure. The multivariate technique implemented was ordination which determines differences, if any, in composition of various sites or samples (Van den Brink et al., 2003). The best fit values, derived with multiple linear regressions between each variable in turn, were used in the analysis together with environmental data as a second matrix instead of the original data (Shaw, 2003). The assumption is also made that one of the sets of environmental variables can be considered “independent” while the other set is considered “dependant” (Ter Braak and ˇ Smilauer, 2002). The significance of the distributions was tested using Monte Carlo permutation testing (499 unrestricted permutations). The environmental variables that indicated signficance (p < 0.05) were included in the data interpretation. All the data were log transformed prior to analysis. The specific ordination analysis used was redundancy analysis (RDA). Canoco Version 4.5 was used for the redundancy analysis (RDA) to determine which environmental variables were responsible for the grouping of the different sites.
2.5. Fish Response Assessment Index (FRAI) The current fish biotic ndex of choice in South Africa to assess the Present Ecological Status (PES) of fish communities is called the Fish Response Assessment Index (FRAI) (Kleynhans et al., 2007). This index was developed as part of a suite of tools used in the EcoClassification process (Kleynhans et al., 2005) to determine the Reserve, which is defined as the amount of water available within the water resource that are available for the future supply of water for human use as well as ecological requirements of the water resource (NWA, 1998). This method is exclusively for riverine systems and whas not used for the Spitskop Dam. The fish assessment results for all three surveys were combined to determine the PES. The FRAI index is a habitat-based cause and effect model based on the attributes of fish species (environmental intolerances and preferences) and the responses of the fish community to changes of the ecosystem drivers (water quality, geomorphology and hydrology). The PES is derived from the integration of the environmental requirements, the responses to modified habitat conditions and changes in fish community structure from reference conditions
(Kleynhans et al., 2005). The various environmental requirements and habitat drivers are incorporated in the metrics of the FRAI index. The various metrics are then weighted and the percentage change from the reference condition is calculated. This value is then expressed as a percentage score and an ecological category of either A (100%: Unmodified), B (80–99%: Largely natural), C (60–79%: Moderately modified), D (40–59%: Largely modified), E (20–39%: Seriously modified) or F (0–19%: Critically modified). The FRAI index are only applicable for riverine systems and the calculation of the FRAI index was therefore only for the riverine sites. The Spitskop Dam fish community was however included in the statistical analysis as it is potentially a major variable within the overall functioning of the system. The FRAI index is generally robust (Kleynhans, 2007) but some uncertain in the FRAI index results are present as many of the intolerance and preferences used within the database are expert derived information rather than empirical data (Kleynhans, 2007). The reference list for the fish community was derived from previous surveys on the Harts River as well as the Frequency of Occurrence Database (Kleynhans et al., 2007). The list is provided in Table 1 together with the frequency of occurrences and relative abundances of each fish expected to occur in the systems. The reference fish community were used to complete the FRAI index and compare the observed fish community to the expected fish community.
3. Results 3.1. Sampling results Results for the 2007 low flow survey indicated that seven indigenous species and one exotic species, Gambusia affinis, were caught (Table 2). The results of the 2008 high flow fish sampling indicated that the total species count for the survey was 13 species of which two species were exotic, namely the Common carp, Cyprinus carpio, and mosquitofish, G. affinis (Table 1) The results of the fish survey in the high flow survey during 2009 yielded nine different species of which one was an exotic species (Table 2). The Largemouth Yellowfish, Labeobarbus kimberleyensis, is listed on the IUCN Redlist (IUCN, 2010) as being near threatened. This species was only found once (in 2008) during the three surveys at site VR2 and only two individuals were sampled. The lowest species count was observed in the Vaal River during the 2009 survey were fish
536
W. Malherbe et al. / Ecological Indicators 69 (2016) 533–539
Table 2 Fish sampling results and sampling techniques carried out at the various sampling sites from 2007 to 2009 on the Harts (HR) and Vaal Rivers (VR). Sampling site
HR1
Year Barbus paludinosus (BPAU) Pseudocrenilabrus philander (PPHI) Tilapia sparrmanii (TSPA) Barbus trimaculatus (BTRI) Labeobarbus aeneus (BAEN) Gambusia affinisa (GAFF) Labeo capensis (LCAP) Barbus palidus (BPAL) Cyprinus carpioa (CCAR) Clarias gariepinus (CGAR) Labeo umbratus (LUMB) Labeobarbus kimberleyensis (BKIM) Austroglanis sclateri (ASCL) Barbus anoplus (BANO)
07 8 53
08 13 76
09 31 61
07 – 9
08 13 68
09 28 14
07 8 –
08 17 110
09 11 38
07 – 6
08 13 18
09 56 51
07 1 16
08 1 55
09 – –
07 – –
08 – 41
09 1 35
07 7 18
08 23 51
09 – –
07 28 20
08 1 15
09 21 8
44 1 31 – – – – – – –
31 6 1 – 9 – 4 1 – –
31 – 3 – 1 – – – – –
4 13 38 7 – – – – – –
9 – 9 5 3 – 1 4 – –
11 – – – 2 – – 2 – –
5 4 – 1 2 – – – – –
1 – 37 12 13 – 1 2 – –
5 3 4 4 2 – – 3 – –
– – – – – – – – – –
1 – 3 – 11 – – 3 1 –
11 – 3 1 – – – 2 – –
14 16 1 26 5 – – – – –
– 13 8 – 14 – – 2 – –
87 – – – – – – 3 – –
– – – – – – – – – –
3 – – – 5 – 4 1 2 –
17 – – 2 – – – – – –
10 9 – 1 17 1 – – – –
10 – 4 3 51 – – – – –
10 – – – – – – – – –
20 – – 16 1 – – – – –
22 5 1 100 2 – – – 3 2
15 – – 10 – – – – – –
– –
– 37
– –
– –
– –
– –
– –
– –
– –
– –
– –
– 2
– –
– –
– –
– –
– –
– –
– –
2 –
– –
– –
– –
– –
45 3 – –
40 3 – –
75 – – –
100 – – 2
40 – – 2
60 – – –
70 3 – –
60 2 – –
50 – – –
50 3 – 1
40 2 – –
65 – – –
70 – – –
55 – – –
20 – – –
– 3 – –
40 3 3 –
60 – – –
85 3 – –
45 – 2 –
5 – – –
30 3 – –
20 – 2 –
30 – – –
Technique Electrofishing (min) Fyke nets (hours) Gill nets (hours) Medium seine (drags) a
HR2
HR3
HR4
HR5
SD
VR1
VR2
Exotic species.
Table 3 Results of the FRAI assessment for the sampling surveys from 2007 to 2009 for the Harts and Vaal River sites. Sites
FRAI score (%)
Ecological category
Ecological category description
HR1 HR2 HR3 HR4 HR5 VR1 VR2
86.57 86.4 80.19 62.66 63.26 76.56 62.4
B B B/C C C C C
Largely natural with few modifications. A small change in natural habitats and biota may have taken place but the ecosystem functions are essentially unchanged. Moderately modified. A loss and change of natural habitat and biota have occurred but the basic ecosystem functions are still predominantly unchanged.
sampling was extremely difficult as the river was in flood. Only limited electroshocking was possible on the sides of the river at site VR1, while sampling at site VR2 was confined to an area where the river flooded its banks. The Southern Mouthbrooder, Pseudocrenilabrus philander, and the Banded Tilapia, Tilapia sparmanii, was the most abundant fish species occurring in the system. None of the species caught are additionally listed as endangered by the IUCN or local conservation departments. However, the minnow, Barbus palidus, does not have a wide distribution range within South Africa. The RDA triplot (Fig. 2) presents the site groupings based on species diversity and abundances together with the environmental variables responsible for driving changes in the fish community structures. Monte Carlo permutations were used to determine the significant environmental variables and only these are presented in Fig. 2. Temporal variation was indicated as sites HR1–HR4 during the 2009 survey grouped together (Fig. 2), potentially due to higher abundances of T. sparrmanii (TSPA) and Barbus paludinosus (BPAU). The copper concentrations were also a significant explanatory variable that was identified with the Monte Carlo permutations at these sites. Spatially, only site HR1 during the 2007 and 2008 surveys grouped separately from the other sites in the study area during the various surveys. This could relate to the changes in P. philander abundances and possibly elevated nitrate concentrations. Zinc concentrations were found to be higher during the 2008 surveys at sites HR2, HR3 and HR5 while sulphate concentrations were higher at site HR5 (2007 and 2009) and in the Spitskop Dam. The fish species Labeo umbratus, Barbus trimaculatus, Labeo capensis and Cyprinus carpio were more abundant during the 2007 and 2008 surveys. Fish habitat and sediment physical properties were used in Fig. 3 as explanatory variables to determine if the habitat changes at
Fig. 2. RDA triplot for the fish community results of the 2007–2009 sampling surveys (HR = Harts River; VR = Vaal River; SD = Spitskop Dam). Full names of the species abbreviations are presented in Table 2 The triplot explains 29.8% of the variance with 17.5% explained on first axis and a further 12.3% on the second axis (Site code: River; Site Number; Year).
the sites were having an effect on the fish community. The general trend identified in Fig. 3 was that sites HR1–HR4 and site VR1 grouped together due to increased diversity of flow velocity depth classes presented at the site i.e. slow shallow (SS), slow deep (SD),
W. Malherbe et al. / Ecological Indicators 69 (2016) 533–539
Fig. 3. RDA triplot with supplementary habitat variables for the fish community results of the 2007–2009 sampling surveys (HR = Harts River; VR = Vaal River; SD = Spitskop Dam). Full names of the species abbreviations are presented in Table 2. The triplot explains 37.4% of the variance with 21.7% explained on first axis and a further 15.7% on the second axis (Site code: River; Site Number; Year). VFS: very fine sand; CS: coarse sand; VCS: very coarse sand; FD: fast-deep; FS: fast-shallow; SS: slow-shallow; SD: slow-deep.
537
Fig. 5. RDA triplot with supplementary variables that control for the sampling site for the fish community results of the 2007–2009 sampling surveys (HR = Harts River; VR = Vaal River; SD = Spitskop Dam). Full names of the species abbreviations are presented in Table 2. The triplot explains 37.4% of the variance with 21.7% explained on first axis and a further 15.7% on the second axis (Site code: River; Site Number; Year).
3.2. Fish Response Assessment Index (FRAI)
Fig. 4. RDA triplot with supplementary variables that control for the sampling year for the fish community results of the 2007–2009 sampling surveys. (HR = Harts River; VR = Vaal River; SD = Spitskop Dam). Full names of the species abbreviations are presented in Table 2. The triplot explains 37.4% of the variance with 21.7% explained on first axis and a further 15.7% on the second axis (Site code: River; Site Number; Year).
fast deep (FD) and fast shallow (FS). Site HR5 and VR2 generally had a lower abundance of flow velocity classes present while the gravel sediment size fraction tended to be higher at these sites. Sites HR1–HR4 indicated medium sand and very fine sand (VFS) as the predominant grain size fractions. Sites VR2, VR1 and HR5 during the 2009 survey grouped separately due to the mud fraction being the highest. Another RDA plot was constructed (Fig. 4) to identify temporal changes in the fish community with the aid of supplementary variables to control for the sampling year. This indicated that each sampling year was different; however, Monte Carlo permutation testing indicated that this difference was not significant. The final RDA plot in Fig. 5 used supplementary variables to control for the sampling site to identify any spatial difference between the sites. It is evident here that site HR1 was dissimilar while site HR4 did group more towards site HR1 than the other sites on the Hats River (HR2 and HR3). Site VR2 and HR5 were found to be similar. Site HR3 grouped separately with site HR2 and VR1 grouping in between the HR1/HR4 and VR2/HR5 grouping.
The fish community assessment was carried out using the FRAI index to determine the present ecological state at the various sites on the Harts River and Vaal River (Table 3). Although fish were collected during three surveys, the FRAI results were only calculated once for the three year period so that accidental miss sampling and natural fish movement were taken into account. The fish reference community in Table 1 provides a comparison for the fish that were collected during the surveys in Table 2. It is evident that most of the reference fish community were collected during the field surveys. The expected relative abundances from the reference fish community (Kleynhans et al., 2007) were also found to be similar during the field surveys from 2007 to 2009. The relative reference site HR1 situated before the VHIS was found to be largely natural while site HR2, situated within the first few kilometers of the VHIS, was also largely natural (Table 3). This indicates that minimal changes have occurred in the natural habitat as well as the biota present. The ecological category indicated a slight decrease at site HR3 from a Category B (largely natural) to a Category B/C (Table 3). The result of the FRAI at sites HR4 and HR5 indicated the fish community was moderately modified (Category C) that could be attributed to a loss of habitat and biota. The FRAI results for the Vaal River at sites VR1 and VR2 were found to be moderately modified (Category C); however, the actual FRAI score was at the lower range of the ecological category compared to site VR1. 4. Discussion The VHIS can potentially impact the fish community of the Harts River due to flow changes, habitat changes and various pollutants in the form of pesticides, nutrients, salts and metals. Therefore, multivariate statistical analyses were used to distinguish which potential impacts were present within the system while also controlling for any upstream impacts from other land uses on the Harts River. The use of the biological integrity index, FRAI, was not able to distinguish between the individual impacts but provided an integrated result on the PES of the fish community within the Lower Harts River. The water quality parameters that did affect the community were shown to be strongly dependent on spatial and temporal changes within the system. Previously, probabilistic risk assessments indicated that some selected pesticides used within the VHIS are a risk to the aquatic environment (Malherbe et al., 2013);
538
W. Malherbe et al. / Ecological Indicators 69 (2016) 533–539
however, no pesticides were obtained within the sediment or fish tissue within the system during 2007–2009. Metal concentrations in the sediment indicated that cadmium, manganese, cobalt and zinc posed a medium to high risk of causing effects within the VHIS (Malherbe et al., 2015). However, when evaluated against the fish community structure in this study it was observed that most of these variables were not a significant contributor of the fish community. The exception was zinc that was highlighted as a potential water quality driver, especially during the 2008 surveys at sites HR2, HR3 and HR5 which are situated downstream of the VHIS. Zinc is a necessary trace metal and fulfils physiological processes but when present at abnormally high concentrations it can be toxic to fish (Wepener et al., 2001). High zinc concentrations can disrupt ion regulation, acid-base regulation, gill tissue and lead to hypoxia (Murugan et al., 2008). Zinc sources in the aquatic environment could occur from natural weathering process, industrial activities and agricultural activities in the form of insecticides (DWAF, 1996). Instream evidence of excess nutrients in the form of filamentous algal growth was present at downstream sites HR3 to HR5 during all of the surveys. The presence of increased nitrate concentrations could have led to the algal growth. A study using remote sensing techniques recently indicated that over the last ten years that the Spitskop Dam (downstream of the VHIS and in between HR4 and HR5) has experienced eutrophication and cyanobacterial blooms (Matthews et al., 2015). This proliferation of algae can potentially decrease habitat availability for selected fish species and in turn allow specific fish species to dominate. Most of the fish communities sampled were generally tolerant to environmental water quality and only some species showed a preference to specific habitat conditions (Skelton, 2001; Kleynhans, 2007). The most vulnerable species found was the endangered Largemouth Yellowfish (L. kimberleyensis) and that was only once at one site on the Vaal River. The changes in water quality and flow patterns due to the VHIS and the various impoundments could be a reason for this species not occurring in this reach of the Harts River anymore. Personal communication with local fisherman indicated thatthat L. kimberleyensis still occurs within the Taung Dam but this was not confirmed during the current study. Pesticides can have numerous effects on fish, depending on the exposure as well as the type of pesticide. Certain fish species will also be more affected than others while different age classes may also show different effects. Currently, research on the effects on fish communities is scant, limited to either accidental spills or intentional spills (Relyea, 2005). A study in Mexico assessed the fish community of a reservoir following a spill event (Favari et al., 2002). Another study done on trout in the USA by Gormley et al. (2005) before and after an agricultural runoff event of azinphosmethyl indicated that brook trout (Salvelinus fontinalis) mortality was higher than the mortality for rainbow trout (Oncorhynchus mykiss) directly after the runoff event. The younger age classes for both fish species were also more affected than the older age classes. Sampling was repeated approximately one year after the event and revealed the fish community was still skewed towards rainbow trout as the brook trout population had not recovered. Habitat for fish is also an important variable that can be affected by the VHIS. The sediment characteristics and velocitydepth classes were used as explanatory variables and this indicated that the habitat diversity was generally higher at sites HR1 to HR4 indicating the VHIS was not necessarily affecting the habitat diversity. Sediment physical characteristics indicated that the larger grain sizes dominated downstream of the VHIS (Malherbe et al., 2015). This is possibly a result of sediment transportation in the system due to increased return flows from the VHIS. Habitat diversity within the system is also similar from HR1 to HR4 due to the upstream Taung Dam and downstream Spitskop Dam. The construction of impoundments has been known to alter stream flow
volume, flow regimes, habitat, and nutrients (Postel and Richter, 2003). In the United States, Poff et al. (2007) found impoundments to alter stream flow, making them more homogenous and resulting in the increased abundance of tolerant species. A study by Rashleigh et al. (2009) on the fish assemblage pattern in the Olifants River catchment did not indicate a loss of species diversity due to impoundments, but rather a shift in the community composition with a specific increase in species tolerant to flow and physico-chemical modifications. Similar results were found on the Crocodile (West) River when using land use like impoundments or agricultural fields to predict river integrity (Amis et al., 2007). The fish distribution in the system was also influenced by the habitat and flow velocities present at the different sites. The Harts River showed an increase in depth and width as the VHIS return water enters the river at sites HR3 and HR4 (where the river is significantly wider and deeper with the presence of steep banks ˙ due to erosion). A study by Wyzga et al. (2009) showed that anthropogenic impacts, specifically flow regulation, can lead to homogenous flows and decreased habitat availability which ultimately results in decreased biotic diversity. However, in this case, the habitat diversity (specifically at HR3) was maintained even though the width and depth increased and road crossings were present. A decrease in habitat diversity was, however, seen at HR4 and HR5 which partly led to the decreased fish ecological category as calculated with FRAI. Furthermore, the presence of weirs and bridges at these sites increased the flow disruption immediately downstream of these sites. The presence of the hardy fish community ensures that numerous fish species occurred throughout the sampling sites as well as the different surveys. In general, those expected fish species not sampled during the 2007 surveys were sampled during the 2008 and 2009 surveys. There was also an increased fish abundance noted during 2008 and 2009. The flow conditions during February 2009 were higher than normal and this influenced the fish abundances and results during this survey. The results of the fish community cleared showed temporal variation with each survey resulting in different species and abundance of species present. The water temperature fluctuation in the system during the 2007 winter survey was extreme with the temperatures dropping to below 10 ◦ C, while summer temperatures were around 20 ◦ C. Fish therefore moved into the deeper pooled areas of the river to areas where the temperature remained stable and did not fluctuate with the outside temperatures. Dörgeloh (1994) completed a study on food selection in a few man-made impoundments and found that a minimum of 8 ◦ C and maximum of 22 ◦ C was not optimal for C. gariepinus. Most fish can survive in colder water indicating that this range should protect the majority of species. Once the temperatures increase, the fish starts migrating into the shallower areas of the river to feed and spawn (King et al., 1998). The FRAI index implemented in this study indicated that the fish community at sites HR1 to HR3 are still in a largely natural condition even though abundances of certain fish species have decreased from relative reference conditions. The decreased FRAI scores at sites HR4 and HR5 are the result of the cumulative impacts of the VHIS as the habitat availability decreases due to incised channels ˙ and sedimentation (Wyzga et al., 2009). The major driver of these changes in the fish community is the flow conditions within the system together with the salinity, nutrients and sedimentation from the VHIS. Dubey et al. (2012) completed a study on a tropical river in India to determine the influence of habitat on fish community structure and found that flow velocity, depth, temperature, TDS and EC together with land use, were the driving variables of the fish community. Other studies in tropical countries identified EC as a major driver of fish community change, with decreased fish diversity noted when EC increased (Tongnunui et al., 2009; Rosso and Quiros, 2010).
W. Malherbe et al. / Ecological Indicators 69 (2016) 533–539
This study used water and sediment quality parameters, habitat and different sampling years to assess the impact of a large-scale irrigation scheme on the fish community structure. Very little fish monitoring data was available for comparison in the catchment for comparison with the present study. Data from the national river monitoring programme were only available for certain years, were not expressed into a fish biological index and had no concurrent water, sediment or habitat data making it impossible to link fish community changes to individual environmental parameters (RHP, 2003). It has long been accepted that the ability to link long term community changes to specific water quality parameters aids management programmes and ultimately can improve river systems by creating sustainable utilisation and management systems (Parker et al., 2015). This lack of data on fish communities linked to environmental quality is a major problem in South Africa as well as in most developing countries due to a lack of implementation as well as a lack of sufficient funds. Long term monitoring data of fish communities together with environmental and habitat quality measurements are vitally important for proper management programmes especially in systems that are highly utilised. Long term monitoring programmes are able to characterise the inter-annual variation present in fish and environmental parameters (Parker et al., 2015). The variation in fish communities can be driven by multiple factors like climate, rainfall, flow volumes, water quality and potential anthropogenic activities (Parker et al., 2015). If this variation is not adequately accounted for, it makes decision making more problematic as the uncertainty is increased and it can often result in poor decision making. Annual variation was evident in this study as the fish community varied during each survey. Therefore, it is recommended that this study can serve as a baseline for future monitoring activities of the fish community within the Harts River associated with the VHIS. Acknowledgements The authors would like to acknowledge the National Research Foundation (NRF) and the University of Johannesburg for funding. Opinions expressed and conclusions arrived at, are those of the author and not necessarily to be attributed to the NRF. Dr M Ferreira, L Ferreira, Z Visser, Dr R Gerber, Dr N Degger and Dr K Malherbe are thanked for sampling and analysis during the project. References Amis, M.A., Rouget, M., Balmford, A., Thuiller, W., Kleynhans, C.J., Day, J., Nel, J., 2007. Predicting freshwater habitat integrity using land-use surrogates. Water SA 33 (2), 215–221. Barbour, M.T., Gerritsen, J., Snyder, B.D., Stribling, J.B., 1999. Rapid Bioassessment Protocols for Use in Streams and Wadeable Rivers: Periphyton, Bentic Macroinvertebrates and Fish, second edition. EPA 841-B-99-002. U.S. Environmental Protection Agency; Office of Water, Washington D.C. Clarke, K.R., Warwick, R.M., 1994. Change in marine communities: an approach to statistical analysis and interpretation. In: Plymouth. Plymouth Marine Laboratory, pp. 144. Dörgeloh, W.G., 1994. Diet and food selection of Barbus aeneus, Clarias gariepinus and Oncorhynchus mykiss in a clear man-made lake, South Africa. Water SA 20 (1), 91–98. DWAF (Department of Water Affairs and Forestry), 1996. South African water quality guidelines. Aquatic Ecosystems, vol 7. Department of Water Affairs and Forestry, Pretoria, South Africa. DWAF (Department of Water Affairs and Forestry), 1999. Resource directed measures for protection of water resources. River Ecosystems Version 1.0., vol 3. Department of Water Affairs and Forestry, Pretoria (DWAF Report No. N/28/99). Dubey, V.K., Sarkar, U.K., Pandey, A., Sani, R., Lakra, W.S., 2012. The influence of habitat on the spatial variation in fish assemblage composition in an unimpacted tropical River of Ganga basin, India. Aquat. Ecol. 46, 165–174.
539
Fanadzo, M., 2012. Revitalisation of smallholder irrigation schemes for poverty alleviation and household food security in South Africa: a review. Afr. J. Agric. Res. 7 (13), 1956–1969. Favari, L., Lopez, E., Martinez-Tabche, L., Diaz-Pardo, E., 2002. Effect of insecticides on plankton and fish of Ignacio Ramirez reservoir (Mexico): a biochemical and biomagnifications study. Ecotoxicol. Environ. Saf. 51, 177–186. Gormley, K., Teather, K.L., Guignion, D.L., 2005. Changes in salmonid communities associated with pesticide runoff events. Ecotoxicology 14, 671–678. IUCN., 2010. Red List of Threatened Species Version 2010.1.
. Downloaded on 8 October. Karr, J.R., 1981. Assessment of biotic integrity using fish communities. Fisheries 6 (6), 21–27. King, J., Cambray, J.A., Impson, N.D., 1998. Linked effects of dam-released floods and water temperature on spawning of the Clanwilliam yellowfish Barbus capensis. Hydrobiologia 384, 245–265. Kleynhans, C.J., Louw, M.D., Thirion, C., Rossouw, N., Rowntree, K., 2005. River Ecoclassification: Manual for EcoStatus Determination. Resource Quality Services, Department of Water Afffairs and Forestry, Pretoria. Kleynhans, C.J., Louw, M.D., Moolman, J., 2007. Module D Volume 2: Reference frequency of occurrences of fish in South Africa (version 2). Joint Water Research Commission and Department of Water Affairs and Forestry report. WRC Report No. TT 329/08. Kleynhans, C.J., 1999. The development of a fish index to assess the biological integrity of South African rivers. Water SA 25 (3), 265–278. Kleynhans, C.J., 2007. Module D Volume 1: Fish Response Assessment Index (FRAI) (version 2). Joint Water Research Commission and Department of Water Affairs and Forestry report. WRC Report No. TT 329/08. Kotze, P.J., 2001. The ecological integrity of the Kli River and the development of a sensitivity weighted fish index of biotic integrity (SIBI). Unpublished (PhD thesis), University of Johannesburg, South Africa. Malherbe, W., Wepener, V., van Vuren, J.H.J., 2010. Anthropogenic induced spatial and temporal changes in the aquatic macro-invertebrate community of the lower Umvoti River, KwaZulu-Natal. Afr. J. Aquat. Sci. 35 (1), 13–20. Malherbe, W., van Vuren, J.H.J., Wepener, V., 2013. Preliminary risk assessment of common-use pesticides using PRIMET and PERPEST pesticide risk models in a semi-arid subtropical region. Water SA 39 (5), 599–610. Malherbe, W., Wepener, V., van Vuren, J.H., 2015. The contribution of physical and chemical sediment characteristics to environmental risk from an irrigation scheme in South Africa. J. Soils Sediments 15 (4), 1005–1018. Matthews, M.W., Bernard, S., Matthews, M.W., 2015. Eutrophication and cyanobacteria in South Africa’s standing water bodies: a view from space. S. Afr. J. Sci. 111 (5/6). Murugan, S.S., Karuppasamy, R., Poongodi, K., Puvaneswari, S., 2008. Bioaccumulation pattern of zinc in freshwater fish Channa punctatus (Bloch.) after chronic exposure. Turk. J. Fish. Aquat. Sci. 8 (1), 55–59. NWA (National Water Act) (Act No.36, 1998), 1998. Part 3, the Reserve Government Gazette19182, 26 August (Chapter 3). Parker, J., Epifanio, J., Casper, A., Cao, Y., 2015. The effects of improved water quality on fish assemblages in a heavily modified large river system. River Res. Appl., http://dx.doi.org/10.1002/rra.2917. Poff, N.L., Olden, J.D., Merritt, D.M., Pepin, D.M., 2007. Homogenization of regional river dynamics by dams and global biodiversity implications. Proc. Natl. Acad. Sci. 104 (14), 5732–5737. Postel, S., Richter, B., 2003. Rivers for Life. Island Press, Washington, USA, pp. 220. RHP (River Health Programme), 2003. State-of-Rivers Report: Free State Region River Systems. Department of Water Affairs and Forestry, Pretoria. Rashleigh, B., Hardwick, D., Roux, D., 2009. Fish assemblage patterns as a tool to aid conservation in the Olifants River catchment (East) South Africa. Water SA 35 (4), 517–524. Relyea, R., 2005. The impact of insecticides and herbicides on the biodiversity and productivity of aquatic communities. Ecol. Appl. 15 (2), 618–627. Rosso, J.J., Quiros, R., 2010. Patterns in fish species composition and assemblage structure in the upper Salado River lakes, Pampa Plain Argentina. Neotrop. Ichthyol. 8, 135–144. Shaw, P.J.A., 2003. Multivariate Statistics for the Environmental Sciences. Hodder Arnold Publishers London, UK. Skelton, P.H., 2001. A Complete Guide to the Freshwater Fishes of Southern Africa. Southern Publishers (Pty.) Ltd., Halfway House. ˇ Ter Braak, C.J.F., Smilauer, P., 2002. CANOCO Reference Manual and CanoDraw for Windows User’s Guide: Softeware for Canonical Community Ordination (version 4.5). Microcomuter Power, New York, USA, pp. 500. Tongnunui, S., William, F., Beamish, H., 2009. Habitat and relative abundance of fishes in small rivers in eastern Thailand. Environ. Biol. Fishes 85, 209–220. Van den Brink, P.J., Van den Brink, N.W., Ter Braak, C.J.F., 2003. Multivariate analysis of ecotoxicological data using ordination: demonstrations of utility on the basis of various examples. Australas. J. Ecotoxicol. 9, 141–156. Wepener, V., van Vuren, J.H.J., Du Preez, H.H., 2001. Uptake and distribution of a copper, iron and zinc mixture in gill, liver and plasma of a freshwater teleost Tilapia sparrmanii. Water SA 27 (1), 99–108. ˙ Wyzga, B., Amirowicz, A., Radecki-Pawlik, A., Zawiejska, J., 2009. Hydromorphological conditions, potential fish habitats and the fish community in a mountain river subjected to variable human impacts, The Czarny Dunajex, Polish Carpathians. River Res. Appl. 25 (5), 517–536.