sustainability

Article Impact on Water Quality of Nandoni Water Reservoir Downstream of Municipal Sewage Plants in Vhembe District, South

Jabulani Ray Gumbo 1,*, Ratshilumela Aaron Dzaga 2 and Nthaduleni Samuel Nethengwe 2

1 Department of Hydrology and Water Resources, University of , P/Bag x5050, 0950, 2 Department of Geography & Geo-Information Science, , P/Bag x5050, Thohoyandou 0950, South Africa; [email protected] (R.A.D.); [email protected] (N.S.N.) * Correspondence: [email protected] or [email protected]; Tel.: +27-15-962-8563; Fax: +27-86-517-6979

Academic Editor: Fausto Cavallaro Received: 10 May 2016; Accepted: 20 June 2016; Published: 24 June 2016

Abstract: The deterioration of water quality in our freshwater sources is on the increase worldwide and, in South Africa, mostly due to the discharge of municipal sewage effluent. Here we report on the use of principal component analysis, coupled with factor and cluster analysis, to study the similarities and differences between upstream and downstream sampling sites that are downstream of municipal sewage plants. The contribution of climatic variables, air temperature, humidity, and rainfall were also evaluated with respect to variations in water quality at the sampling sites. The physicochemical and microbial values were higher than the Department of Water Affairs and Forestry (DWAF) and World Health Organization (WHO) guidelines. The cluster analysis showed the presence of two clusters for each of the Mvudi, Dzindi, and Luvuvhu Rivers and Nandoni reservoir sampling sites. The principal component analysis (PCA) accounted for 40% of the water quality variation and was associated strongly with pH, electrical conductivity, calcium, magnesium, chloride, bromide, nitrate, and total coliform, and negatively with rainfall, which represented Mvudi downstream and was attributed to the Thohoyandou sewage plant. The PCA accounted for 54% of the variation and was associated strongly with electrical conductivity, sulfate; total dissolved solids, fluoride, turbidity, nitrate, manganese, alkalinity, magnesium, and total coliform represented Dzindi downstream, with inflows from the sewage plant and agriculture. The PCA accounted for 30% of the variation and was associated strongly with total dissolved solids, electrical conductivity, magnesium, fluoride, nitrate, sulfate, total coliform average air temperature, and total rainfall, and negatively associated with manganese and bromide represented Luvuvhu upstream and was associated with commercial agriculture. The PCA accounted for 21% of the variation and was associated strongly with turbidity, alkalinity, magnesium, chloride, fluoride, nitrate, and strongly negatively associated with rainfall, which represented Luvuvhu downstream, associated with inflows from Vuwani oxidation ponds, Elim and sewage plants, and agriculture. The PCA accounted for 14% of the variation and was moderately associated with rainfall and weakly associated with chloride and bromide and negatively associated with nitrate, which represented the natural Nandoni reservoir system. The continued discharge of effluent may render the raw water supply unsuitable for human consumption and lead to eutrophication due to nitrate enrichment and proliferation of harmful algal blooms and schistomiasis infections in the long term.

Keywords: pollution; water quality; municipal sewage plants; principal component analysis

Sustainability 2016, 8, 597; doi:10.3390/su8070597 www.mdpi.com/journal/sustainability Sustainability 2016, 8, 597 2 of 16

1. Introduction Sustainable development is development that meets the needs of the present without compromising the ability of future generations to meet their own needs [1]. One of the most severe environmental problems is the discharge of partially-treated sewage to freshwater bodies, thus reducing the quality of drinking water [2]. The effluents are rich in nutrients (nitrates, phosphates, ammonia, and nitrites), heavy metals and pathogens (bacteria, viruses, and protozoa) which are hazardous to the environment and may affect human health [3–5]. Hirji et al. [5] indicated that existing sewage plants were overloaded as a result of ever-increasing volumes of wastewater generated by the increasing urban population. Though there is legislation in place to manage water pollution, there are challenges in the enforcement and implementation of that legislation [5]. The results of a study conducted along Fez and Sebou rivers in the region of Fez in Morocco indicate that severe pollution occurred in most urbanized and industrial areas [6]. Water is the most important resource for both economic and social development for modern societies. In South Africa, water quality management is essential since it is predicted that, by the year 2025, the demand for water will outstrip its supply [7]. The deterioration of water quality, coupled with semi-arid conditions, may jeopardize the economic growth and development of South Africa. A study conducted by Bapela [8] indicates that poor agricultural practices, urban runoff, urbanization, and improper land use threaten the suitability of the water quality for domestic use in the Luvuvhu River basin. Bapela [8] asserted that regardless of contaminants from both point and non-point sources of pollution, in parts of the basin (which have changed the water quality to be unfit for different uses), Luvuvhu River is one of the main sources of water for drinking, agriculture, and other domestic uses in Thohoyandou, Makhado, and other areas in the Luvuvhu catchment. With a rising population, urban growth and the associated socio-economic developments, the deteriorating water quality is becoming a serious environmental challenge in the catchment area of . Principal component analysis (PCA), with cluster analysis (CA) and factor analysis (FA), is a multivariate statistical method for the interrogation of large, complex data now widely used in better understanding of water quality data coupled to climatic variables. A number of studies have used PCA and CA to categorize sampling sites and help identify underlying contamination by Zhou et al. [9] and Boyacioglu and Boyacioglu [10] in marine water sources in Southern Hong Kong and Tahtali river catchment in Turkey, respectively. The study of Mendiguchia et al. [11] showed how PCA and CA were used to identify four zones in a catchment and the main contaminates in the Guadalquivir river catchment in Spain. Zhang et al. [12] also used PCA and CA to characterize surface water quality in the Xiang river catchment, China, into clusters in order to design an optimum sampling strategy that reduces sampling sites and, thus, reducing operational costs. Kebede and Kebede [13] used principal component analysis in monitoring physicochemical monitoring in the Urgessa river catchment, Ethiopia, for organic pollution originating from coffee processing and its impact on macroinvertebrate communities. The aim of the present study was to assess the ecological impact of effluent discharged downstream of the municipal plants on water quality of Luvuvhu river catchment, South Africa. The principal component analysis with factor and cluster analysis was used to study the similarities and differences between upstream and downstream sampling sites linked to municipal sewage plants. The contribution of climatic variables, air temperature, humidity, and rainfall were also evaluated with respect to the variation in water quality at the sampling sites. Thus, the coupling of principal component analysis (PCA) with cluster analysis (CA) and factor analysis (FA) to water quality data and climatic variables is a novel approach. Sustainability 2016, 8, 597 3 of 16

2. Materials and Methods

2.1. The Study Area Nandoni dam was constructed by the Department of Water Affairs and Forestry (DWAF) in the Luvuvhu River to provide raw water supplies for 1.3 million people in the Vhembe and parts of Mopani districts, province, and controlled releases downstream to provide water supply for wildlife in [14]. The climatic condition in the catchment area of the dam is characterizedSustainability by hot, 2016, 8 humid, 597 summers and cool, dry winters. According to DWAF [143 ],of 16 the average ˝ ˝ summer and winterNandoni temperaturesdam was constructed are 23by theC Depart and 17mentC, of respectively.Water Affairs and The Forestry mean (DWAF) annual in precipitationthe ranges fromLuvuvhu 2000 River mm to in provide the mountainous raw water supplies areas for of1.3 themillion northwest, people in the to Vhembe 440 mm and at parts Limpopo of River confluence,Mopani with districts, a catchment Limpopo average province of, 800and controlled mm per annumreleases downstream [14]. There to are provide several water waterfront supply villages for wildlife in Kruger National park [14]. The climatic condition in the catchment area of the dam is around thecharacterized Nandoni by dam hot, arehumid typical summers examples and cool, ofdry developing winters. According rural to communities DWAF [14], the inaverage which there is poor servicesummer delivery, and winter a high temperatures rate of unemployment,are 23 °C and 17 °C, respectively. and poverty. The Themean Nandoniannual precipitation dam provides an opportunityranges for from these 2000 communities mm in the mountainous to obtain water areas of for the domestic northwest, purposes to 440 mm such at Limpopo as bathing, River washing, drinking,confluence, and cooking. with It a catchmenthas also become average of a 800 source mm ofperincome annum [1 to4]. many There unemployed are several waterfront menin the area villages around the Nandoni dam are typical examples of developing rural communities in which where theythere catch is poor fish service andsell delivery, them a to high local rate communities. of unemployment There, and are poverty. a number The Nandoni of tributaries dam of the Luvuvhuprovides River, namely an opportunity Dzindi, for Mudzwiri, these communities Madikatla, to obtain and water Mvudi, for domestic that contribute purposes such to inflows as to the Luvuvhubathing, River. washing, However drinking some, ofand the coo baseking. flows It has originatedalso become from a source municipal of income sewage to many plants that contributeunemployed to the Luvuvhu men in the River. area where These they activities catch fish are and linked sell them to to the local deterioration communities. ofThere the are quality a of the number of tributaries of the Luvuvhu River, namely Dzindi, Mudzwiri, Madikatla, and Mvudi, that dam water.contribute An excerpt to inflows from to the the environmental Luvuvhu River. impact However assessment some of the report base flows [14 ] originated indicated from the potential risk of eutrophicationmunicipal sewage and plants pollution that contribute by trace to metals the Luvuvhu and biological River. These pollutants activities are arising linked to from the upstream municipaldeterioration sewage discharges. of the quality Such of the pollutants dam water. causeAn excerpt the from deterioration the environmenta of waterl impact quality assessment by reducing the water suitabilityreport [14 for] indicated domestic the potential use. Consequently, risk of eutrophication recreational and pollution use, treatment, by trace metals and and conveyance biological of water pollutants arising from upstream municipal sewage discharges. Such pollutants cause the are also hampered.deterioration of water quality by reducing the water suitability for domestic use. Consequently, Thererecreational are five sewageuse, treatment plants, and located conveyance upstream of waterof are the also Nandoni hampered.dam; namely, Elim sewage plant, Waterval sewageThere plant,are five sewage Vuwani plants oxidation located upstream ponds, of and the Nandoni Vuwani dam and; namely Thohoyandou, Elim sewage plant, sewage plants (Figure1).Waterval The only sewage exception plant, Vuwani here oxidatio is then Elim ponds, sewage and Vuwani plant and because Thohoyandou it discharges sewage plants the effluent (Figure 1). The only exception here is the Elim sewage plant because it discharges the effluent into into the Muhohodithe Muhohodi River River (Doringspruit) (Doringspruit) that that flows flows into into the Albasini the Albasini dam situated dam situated upstream upstream of the of the Nandoni dam.Nandoni dam.

Figure 1. The location of the municipal sewage plants and the water sampling stations. Figure 1. The location of the municipal sewage plants and the water sampling stations. Sustainability 2016, 8, 597 4 of 16

The effluent from the Waterval sewage plant and Vuwani oxidation ponds are directly discharged into the Luvuvhu River through the Mudzwiri and Madikatla rivers, respectively, which enter the Luvuvhu River from the south side. Vuwani sewage plant effluents are discharged into the Dzindi River through the Dzondo rivulet that joins the river from the south. Thohoyandou sewage plant effluents are directly discharged into the Mvudi River that joins the Dzindi River just above the confluence point of the Luvuvhu and Dzindi rivers. Pollutants are finally discharged into the Nandoni dam through the Luvuvhu River, which enters the dam from the west side.

2.2. Water Sampling The water samples were collected from the Luvuvhu River and the following tributaries of Dzindi and Mvudi rivers, as well as from the Nandoni Dam (Table1). The water samples were collected on a monthly basis from July to December 2010. For chemical water samples, two-liter plastic sampling containers were rinsed with distilled water prior to sampling. Samples were also taken with the sampling container facing the stream current of flow. For microbiological water samples, 500 mL sterile sampling glass bottles were used. Samples were taken with the sampling bottle mouth facing the current flow of the stream. The sampling bottle was completely submerged into the water in order to avoid unwanted floating particles in the water. Samples were put inside a cooler box filled with ice and delivered to the laboratory within six hours from sampling time.

Table 1. Description of the sampling points at the Luvuvhu River basin and Nandoni Dam.

Sample # Sample Name Description

S1 Mvudi downstream To assess the nature and amount of pollutants before and after the Mvudi S2 Mvudi upstream River passes through the Thohoyandou sewage plant discharge point

S3 Dzindi downstream To evaluate the extent of the impact of the Vuwani sewage plant on the water quality of the Dzindi River. Dzindi upstream is before the discharge from municipal sewage plant and mainly subsistence farming, Dzindi irrigation S4 Dzindi upstream scheme and washing of motor cars and brick making.

S5 Luvuvhu downstream Assessment of the impact of the Elim and Waterfall sewage plants and the Vuwani oxidation ponds, as well as the commercial and Lotanyanda S6 Luvuvhu upstream subsistence farming areas

To verify the possible contribution to deterioration of water quality as a result S7 Nandoni reservoir of the upstream municipal sewage discharge activities

2.3. Physicochemical Water Quality Analysis For the physical-chemical analyses, the water quality tests were conducted within 24 h at the Chemistry Department, University of Venda, and the Department of Water Affairs and Forestry (DWAF) laboratories. At the sample point, onsite measurements, in triplicate, were pH and temperature (a pH meter (Multi 340/Set Instruments), supplied by Wissenschaftlich-Technische werkstätten GMBH: Weilheim, Germany) and turbidity (portable 2100P turbidity meter, supplied by LABTEC: Midrand, South Africa). The pH meter was calibrated according to the manufacturer’s guidelines using pH buffers of four and seven. In the laboratory, the ion chromatograph (Dionex Ion chromatograph, Supplier Thermo ARLabs, Gauteng, Switzerland) was used to determine the anions, chloride, bromide, fluoride, nitrate, phosphate, and sulfate. The metals, manganese, iron, potassium, calcium, and magnesium were determined by absorption spectrophotometer (Varian Spectra AA 110 Flame Atomic Absorption Spectrometer (220/880 series), Supplier SMM Instruments: Johannesburg, South Africa).

2.4. Microbiological Water Quality Analysis For the microbiological analyses, the water quality tests were conducted within 6 h from the sampling time of the first sample at the DWAF laboratory in Thohoyandou. One packet of colilert-18 Sustainability 2016, 8, 597 5 of 16 powder was added to a 100 mL water sample and was thoroughly mixed. The solution was then transferred into a quanti-tray. The quanti-tray was taken into a quanti-tray sealer and then transferred into an incubator at 35 ˝C for 18 h. After incubation, the quanti-tray was taken out of the incubator for total coliform and Escherichia coli observation. The yellow color indicated the presence of total coliform bacteria. Then, the quanti-tray was put under a ultra-violet (UV) light for E. coli observation. The numbers of E. coli and total coliform counts/100 mL were enumerated using the most probable (MPN) table. The interpolation method was used to count the amounts of total coliform and E. coli bacteria contamination.

2.5. Climate Data The climate data was obtained from the South African Weather Station (Thohoyandou observation station) which was specially requested from the Department of Environmental Affairs and Tourism. The Thohoyandou observation station is situated 15 km southwest of Nandoni dam.

2.6. Data Analysis A hierarchical method, average linkage cluster analysis, was applied to the mean values of the physical and chemical variables for each site using the IBM SPSS version 22 (IBM: Armonk, NY, USA) statistical package. Principal component analysis (PCA) was used to determine the water quality parameters that contribute to the water quality variation between the upstream and downstream sampling points. The data was subjected to factor analysis to determine suitability of PCA. The correction matrix showed that the loadings were above 0.3. Liu et al. [15] categorized that factor loading of strong, moderate, and weak were based on values of >0.75, 0.75–0.50, and 0.50–0.30, respectively. The Kaiser-Meyer-Olkin (KMO) was in the range of 0.4 to 0.5 which was slightly lower than the recommended value of 0.6 [16], and Bartlett’s Test of Sphericity was in the range of 0.000 to 0.001 [17], indicating the suitability of PCA.

3. Results and Discussion

3.1. Sample Sites Analysis and Characteristics The dendrogram using average linkage (between groups) revealed two clusters. Cluster 1 was composed of Mvudi upstream activities, which involved subsistence and vegetable cultivation and the washing of motor cars using the waters of the Mvudi River. Cluster 2 was the Mvudi downstream section which received municipal effluent discharges from Thohoyandou sewage plant and the human activities associated here were subsistence farming, brick-making using clay materials close to the Mvudi River at this point and also received inflows from Mvudi upstream. The dendrogram using average linkage (between groups) revealed two clusters. Cluster 1 was the Dzindi downstream section which received municipal effluent discharges from Vuwani sewage plant and the human activities associated here were subsistence farming, brick-making using clay materials close the Mvudi River at this point, and also received inflows from Mvudi upstream, and this was more pronounced during dry months of July to September. Cluster 2 was composed of Dzindi upstream activities which involved subsistence and vegetable cultivation at the Dzindi irrigation scheme and the washing of motor cars using the waters of the Dzindi River, and this was more pronounced during the dry and wet periods of October to December. The dendrogram using average linkage (between groups) revealed two clusters. Cluster 1 was the Luvuvhu downstream and the Nandoni reservoir, which were more pronounced during the months of October to December, a dry and wet period. The Luvuvhu downstream activities included subsistence farming, brick-making using clay materials close the Mvudi River at this point, and also received inflows from Mvudi upstream. The Luvuvhu downstream activities involved receipt of municipal effluent discharges from Elim, Waterval sewage plants and Vuwani oxidation ponds, and the human activities associated here were commercial farming in Levubu, and subsistence and Sustainability 2016, 8, 597 6 of 16 vegetable cultivation along the Luvuvhu River. Cluster 2 was composed human activities of Luvuvhu upstream and the Nandoni reservoir, which were more pronounced during the months of July to September, a dry period. The Luvuvhu upstream is mainly affected by commercial agriculture and the natural riverine ecosystem. The Nandoni reservoir involved receipt of municipal inflows from Luvuvhu upstream and pollutants from Mvudi and Dzindi Rivers, fishing with nets in the Nandoni dam,Sustainability and brick-making 2016, 8, 597 along the river banks and Nandoni reservoir (Figure2). 4 of 16

(a) (b)

FigureFigure 2.2. TheThe makingmaking of of farm farm bricks bricks using using clay clay materials materials located located at theat the Nandoni Nandoni reservoir: reservoir: (a) raw (a) clayraw materialclay material and (andb) dry (b) bricks dry bricks waiting waiting to be to fired. be fired.

The effluent from the Waterval sewage plant and Vuwani oxidation ponds are directly The water quality data, mean value, standard deviation, and range for the Luvuvhu river discharged into the Luvuvhu River through the Mudzwiri and Madikatla rivers, respectively, which catchment are shown in Table2. The physicochemical and microbial values are in excess of the enter the Luvuvhu River from the south side. DWAF guideline values [18–20] for surface water sources. The discharge of excess nitrates, for Vuwani sewage plant effluents are discharged into the Dzindi River through the Dzondo example, is likely to lead to eutrophication and growth of harmful algal blooms [21]. A previous study rivulet that joins the river from the south. Thohoyandou sewage plant effluents are directly by Makhera et al. [22] in the Luvuvhu River showed the presence of microcystin-LR at 2 ppb which discharged into the Mvudi River that joins the Dzindi River just above the confluence point of the was higher than the 1 ppb safe limit suggested by the Word Health Organization for drinking and Luvuvhu and Dzindi rivers. Pollutants are finally discharged into the Nandoni dam through the recreational water [23]. The ecological impacts of municipal sewage discharges contributes to issues of Luvuvhu River, which enters the dam from the west side. harmful algal blooms and their cyanotoxins, as represented by microcystin, is real in the Luvuvhu River catchment2.2. Water Sampling for three reasons. There is the Nandoni water treatment plant, which draws its raw water supply from the Nandoni reservoir. During spring and summer the operators at the Nandoni water treatmentThe water plant resort samples to the were use collected of activated from carbon the Luvuvhuin order to River manage and taste the and following odor problems tributaries in the of drinkingDzindi and water, Mvudi probably rivers due, as to well the as presence from the of Nandonicyanobacteria Dam and (Table actinomycete 1). The water organisms samples [24 were,25]. Secondly,collected onthe a presence monthly of basis nitrates from and July essential to December metals—such 2010. Foras magnesium, chemical water a central samples, component two-liter of chlorophyll—encouragesplastic sampling containers the growthwere rinsed of aquatic with weedsdistilled that water promote prior the to proliferation sampling. Samples of schistosomiasis were also vectortaken withsnails the in the sampling shoreline container of the Nandoni facing the reservoir stream and current along of the flow. Luvuvhu For microbiological River downstream water ofsamples, the Nandoni 500 mL reservoir sterile sampling [26–28]. glass The studybottles of were Dzaga used. [29 Samples] has shown were a taken spike with in the the number sampling of schistosomiasisbottle mouth facing infections the current among flow waterfront of the stream. villagers The living sampling along thebottle shoreline was completely of Nandoni submerged reservoir. Thirdly,into the thewater Luvuvhu in order River to avoid flows unwanted into Kruger floating National particles Park, thusin the bringing water. Samples harmful were algal put blooms inside and a theircooler cyanotoxins box filled with to wildlife. ice and Thedelivered infrequent to the outbreaks laboratory of withi microcystinn six hours poisoning from sampling of wildlife time. has been recorded recently with wildlife fatalities in Kruger National Park [30]. Lastly, the high levels of E. coli in the surfaceTable waters 1. Description is worrisome of the sampling since the points rural at communities the Luvuvhu mayRiver resort basin and to the Nandoni use of Dam these. sources, theSample Nandoni # reservoir,Sample and Name river water, for human consumptionDescription because of erratic supply of municipal drinkingS1 waterMvudi [31,32]. downstream These high totalTo coliformassess the countsnature and may amount be associated of pollutants with before waterborne and after diseases the which may be caused by pathogenic organisms,Mvudi River such passes as Vibriothrough cholerae the Thohoyandouspp. and pathogenic sewage plantE. coli [33]. S2 Mvudi upstream discharge point To evaluate the extent of the impact of the Vuwani sewage plant on S3 Dzindi downstream the water quality of the Dzindi River. Dzindi upstream is before the discharge from municipal sewage plant and mainly subsistence S4 Dzindi upstream farming, Dzindi irrigation scheme and washing of motor cars and brick making. S5 Luvuvhu downstream Assessment of the impact of the Elim and Waterfall sewage plants and the Vuwani oxidation ponds, as well as the Levubu commercial S6 Luvuvhu upstream and Lotanyanda subsistence farming areas To verify the possible contribution to deterioration of water quality S7 Nandoni reservoir as a result of the upstream municipal sewage discharge activities Sustainability 2016, 8, 597 7 of 16

Sustainability 2016, 8, 597 7 of 16 Table 2. Water quality variables of the Luvuvhu river catchment from July to December 2010. Variables Mean Std Range Mean Std Range Mean Std Range pH 7.4 0.8 8.6–6.4 7.6 0.6 8.5–6.6 7.3 0.8 9–6.3 Sampling Stations Luvuvhu River System Dzindi River System Mvudi River System Turbidity (NTU) 8.3 4.5 14.3–2.4 11.7 4.1 15.5–5.4 23.2 18.4 67.9–8.1 Variables Mean Std Range Mean Std Range Mean Std Range Total Dissolved Solids (mg/L) 33.3 1.7 35.5–29.5 26.2 4.9 33.3–19.8 37.4 13.1 49.8–4.8 Electrical ConductivitypH (μS/cm) 7.44.9 0.80.6 8.6–6.45.6–4 7.64.2 0.61 8.5–6.66.2–3.2 7.34.5 0.81.3 9–6.37.5–3.3 Turbidity (NTU) 8.3 4.5 14.3–2.4 11.7 4.1 15.5–5.4 23.2 18.4 67.9–8.1 Alkalinity (mg/L) 82.4 14.1 101–65 75 16.4 98–47 101.8 14 134–78 Total Dissolved Solids (mg/L) 33.3 1.7 35.5–29.5 26.2 4.9 33.3–19.8 37.4 13.1 49.8–4.8 ElectricalManganese Conductivity (mg/L (µ)S/cm) 4.90.2 0.60 5.6–40.3-0.2 4.20.2 10.1 6.2–3.20.4–0.1 4.51.2 1.30.7 7.5–3.31.9–0.4 AlkalinityIron (mg/L (mg/L)) 82.40.1 14.10 101–650.2–0.1 750.1 16.40 98–470.2–0.1 101.80.2 140.1 134–780.3–0.1 PotassiumManganese (mg/L)(mg/L) 0.21 00.5 0.3-0.21.8–0.5 0.21.2 0.10.3 0.4–0.12–0.6 1.23.9 0.73 1.9–0.413–1.8 CalciumIron (mg/L) (mg/L) 0.114.4 04.1 0.2–0.122.0–8.2 0.123.3 013.1 0.2–0.138.0–1.0 0.236.3 0.115.6 58. 0.3–0.10–12.0 Potassium (mg/L) 1 0.5 1.8–0.5 1.2 0.3 2–0.6 3.9 3 13–1.8 MagnesiumCalcium (mg/L) (mg/L) 14.4135.3 4.15.3 143. 22.0–8.21–129.3 23.392.7 13.129.5 38.0–1.0138–56 36.3114.7 15.647.3 58.0–12.0198–68 MagnesiumChloride (mg/L (mg/L)) 135.349.3 5.310.4 143.1–129.362–36 92.750.5 29.517.8 138–5678–30 114.7122.9 47.388.2 198–68271–45 BromideChloride (mg/L)(mg/L) 49.30.9 10.40.7 62–362.6–0.3 50.50.9 17.80.6 78–302.2–0.3 122.90.4 88.20.1 271–450.8–0.3 FluorideBromide (mg/L)(mg/L) 0.936.8 0.72.2 2.6–0.339.1–32.5 0.919.4 0.619.9 2.2–0.339.3–0.3 0.418.8 0.119.4 0.8–0.339.2–0.2 Fluoride (mg/L) 36.8 2.2 39.1–32.5 19.4 19.9 39.3–0.3 18.8 19.4 39.2–0.2 NitrateNitrate (mg/L)(mg/L) 56.156.1 36.736.7 96.6–17.796.6–17.7 30.130.1 8.28.2 48.6–23.648.6–23.6 52.352.3 33.233.2 92–14.692–14.6 SulphateSulphate (mg/L)(mg/L) 3.83.8 1.81.8 6.3–1.06.3–1.0 2.42.4 1.31.3 4.2–1.14.2–1.1 3.73.7 2.12.1 7.4–1.47.4–1.4 Total Totalcoliform coliform (colony (colony forming 780780 420420 1011–351011–35 660660 438438 1011–211011–21 773773 430430 1011–241011–24 formingunits/100 units/100 mL) mL) E. coli (colony forming units/100 mL) 323 373 1011–36 390 274 914–37 572 427 1011–44 E. coli (colony forming units/100 mL) 323 373 1011–36 390 274 914–37 572 427 1011–44

3.23.2.. Assessing the Impact of Thohoyandou Sewage Discharge on the Mvudi Downstream Sampling Site Ten items items of of the the Mvudi Mvudi sampling sampling sites, sites, upstream upstream and downstream, and downstream, were analyzed were analyzed with principal with principalcomponent component analysis (Table analysis3). The (Table principal 3). The component principal analysiscomponent showed analysis the first showed three the components first three componentsaccounted for accounted 83% of the for total 83% variation of the in total water variation quality in and water the scree quality plot and of eigenvalues the scree plot of the of eigenvaluescomponents of is the shown components in Figure is 3shown. The in principal Figure 3. component The princip 1al (PC1)component accounted 1 (PC1) for accounted 40% of the for 40%variation of the (Table variation4) and (Table was associated4) and was strongly associated with strongly pH, electrical with pH, conductivity, electrical conductivity, calcium, magnesium, calcium, magnesium,chloride, bromide, chloride, nitrate, bromide, and total nitrate coliform,, and and total negatively coliform with, and rainfall, negatively which with represented rainfall, Mvudiwhich representeddownstream. Mvudi There downst are severalream.mechanisms There are several that maymechanisms account that for themay interaction account for of the pH, interaction electrical ofconductivity, pH, electrical calcium, conductivity, magnesium, calcium, chloride, magnesium, bromide, chloride, nitrate, and bromide, total coliform. nitrate,First, and thetotal nitrate coliform. and Firsttotal, coliformthe nitrate in theand Mvudi total Rivercoliform downstream in the Mvudi of the River Thohoyandou downstream sewage of the plant Thohoyandou may be attributed sewage to plantsewage may discharge be attributed and or to subsistence sewage discharge farming. and or subsistence farming.

Figure 3. TheThe scree scree plot of eigenvalues for the Mvudi river system system..

Table 3. Water quality eigenvectors of the correlation matrix.

Parameter PC1 PC2 PC3 PC4 PC5 PC6 PC7 PC8 PC9 PC10 Sustainability 2016, 8, 597 8 of 16

Table 3. Water quality eigenvectors of the correlation matrix.

Parameter PC1 PC2 PC3 PC4 PC5 PC6 PC7 PC8 PC9 PC10 pH 0.039 0.401 0.004 0.178 0.014 0.304 0.223 0.395 0.156 Total dissolved solids 0.039 0.042 0.179 0.079 0.345 0.348 0.124 0.226 0.464 Electrical conductivity 0.401 0.042 0.263 0.002 0.096 0.014 0.017 0.225 0.201 Calcium 0.004 0.179 0.263 0.085 0.001 0.203 0.310 0.195 0.007 Magnesium 0.178 0.079 0.002 0.085 0.008 0.112 0.000 0.047 0.278 Chloride 0.014 0.345 0.096 0.001 0.008 0.082 0.061 0.321 0.040 Bromide 0.304 0.348 0.014 0.203 0.112 0.082 0.326 0.470 0.160 Nitrate 0.223 0.124 0.017 0.310 0.000 0.061 0.326 0.011 0.348 Total coliform 0.395 0.226 0.225 0.195 0.047 0.321 0.470 0.011 0.100 Total rainfall 0.156 0.464 0.201 0.007 0.278 0.040 0.160 0.348 0.100

Table 4. The Eigenvalues of the component matrix of water quality variables of Mvudi upstream and downstream.

Parameter PC1 PC2 PC3 Eigenvalue 4.035 2.735 1.561 % variance explained 40.350 27.346 15.612 %Cumulative variance 40.350 67.696 83.308 pH 0.521 ´0.590 0.512 Total dissolved solids 0.194 0.730 ´0.455 Electrical conductivity 0.737 0.412 ´0.421 Calcium 0.673 ´0.673 0.009 Magnesium 0.910 0.346 0.080 Chloride 0.866 ´0.334 0.126 Bromide 0.565 ´0.031 ´0.472 Nitrate 0.723 0.516 0.349 Total coliform 0.313 0.634 0.568 Total rainfall ´0.462 0.574 0.475 Note: three components extracted.

These research findings suggest the downstream sample point may be associated mostly with discharge from the Thohoyandou sewage plant. The sewage plant receives mostly sewage influent mainly associated with domestic with minimal industry in the Thohoyandou urban and Shayandima industrial areas. The presence of chloride in the sewage effluent indicates the origin is from sodium chloride, which is mostly associated with human consumption of table salt [34]. The research findings are in agreement with the study of Sudevi and Lokesh [35], who found that the sewage plant discharge affected the water quality of the downstream sample point. The association with rainfall was strongly negative and this may imply that rainfall did not contribute to washing of animal waste and/or nitrates from subsistence farming and soil erosion into the Mvudi River at the downstream section. Thus, the discharge of nitrates and total coliform to the Mvudi River originated from the Thohoyandou sewage plant and not from subsistence farming. The presence of bromide in Mvudi River downstream may be attributed to the Thohoyandou sewage plant. Here, the bromide probably enters from the stormwater drains that are connected to the sewage plant. The stormwater drains are located in the Thohoyandou commercial business district, which is close to Mvudi River upstream. The bromide originated from urban runoff when ethylene dibromide was added to petrol [36]. The average pH of the Mvudi River downstream was 7.3 and almost neutral. Thus, most of the anions and cations were in solution, thus contributing to electrical conductivity. PC2 accounted for 27% of the variation (Table4) and was associated strongly with total dissolved solids, electrical conductivity, magnesium, nitrate, total coliform, and rainfall, which represented Mvudi upstream. The association with rainfall was strongly positive and this may imply that rainfall did contribute to the washing of fertilizers, animal waste, and soil erosion into the Mvudi River at the upstream section. The Mvudi upstream is associated with subsistence farming (Figure4) and transects a major tarred road (R524). The research findings are corroborated by the study of De Wit et al. [37] who found that rainfall had a positive effect of washing of road surfaces and discharging the stormwater Sustainability 2016, 8, 597 9 of 16 and discharging the stormwater wastes into rivers. At this particular sample point, there is a major

Sustainabilitytarred road2016 (R524), 8, 597 and the stormwater waste may account for contaminates at the Mvudi upstream9 of 16 sample point. The presence of nitrates may be due from diffuse pollution from farming activities, which is supported by the study of Muriithi and Yu [38]. The stormwater may also account for high wastestotal dissolved into rivers. solids At and this, correspondingly particular sample, with point, high thereelectrical is a conductivity major tarred, that road were (R524) observed and the at stormwaterthe Mvudi upstream waste may point. account This for is supported contaminates by the at the stu Mvudidy of Singh upstream et al. sample[34] who point. found The high presence levels of total nitrates dissolved may be solids due fromand electrical diffuse pollutionconductivity from were farming directly activities, linked to which stormwater is supported intrusion. by the studyPC3 of Muriithiaccounted and for Yu 15.6% [38]. of The the stormwater variation (Table may also4) and account was associated for high totalstrongly dissolved with pH, solids nitrate, and, correspondingly,total coliform, and with rainfall high, electrical and weak conductivity, association that with were chloride observed, which at the represented Mvudi upstream the natural point. ThisMvudi is supportedriver system by and the studythe subsistence of Singh et farming al. [34] whothat is found practiced high levelsin this of section total dissolved of the river solids (Figure and electrical4). The presence conductivity of nitrates were may directly be due linked from to diffuse stormwater pollution intrusion. from farming activities [38].

Figure 4. Subsistence agricultural activities practiced at Mvudi upstream.upstream.

3.3. Assessing the Impact of Vuwani Sewage Discharge on Dzindi Sampling Sites PC3 accounted for 15.6% of the variation (Table4) and was associated strongly with pH, nitrate, totalEleven coliform, items and rainfall, from Dzindi and weak sampling association sites, with upstream chloride, and which downstream, represented were the analyzed natural Mvudi with riverprincipal system component and the analysis subsistence (Table farming 5). The that principal is practiced component in this analysis section of showed the river the (Figure first three4). Thecomponents presence ofaccounted nitrates may for be86% due of from the diffuse total variation pollution in from water farming quality activities and the[38 ]. scree plot of eigenvalues of the components is shown in Figure 5. PC1 accounted for 54% of the variation (Table 3.3.6) and Assessing was associated the Impact strongly of Vuwani with Sewage electrical Discharge conductivity, on Dzindi Samplingsulfate, total Sites dissolved solids, fluoride, turbidity,Eleven nitrate, items from manganese, Dzindi alkalinity, sampling sites, magnesium upstream, and and total downstream, coliform, represent were analyzeding Dzindi with principaldownstream. component The presenc analysise of ions (Table (cations5). The and principal anions) component may account analysis for the showed electrical the conductivity first three componentsand total dissolved accounted solids. for 86%These of theresearch total variationfindings insuggest water origins quality anddownstream the scree plotmay ofbe eigenvalues associated ofmostly the components with the Vuwani is shown sewage in Figure plant5 .and PC1 agriculture accounted [3 for7– 54%39]. ofThe the sewage variation plant (Table receiv6) andes mostly was associatedsewage influent strongly mainly with electrical associated conductivity, with domestic sulfate, totaluse from dissolved Dzwerani, solids, fluoride, Mathule, turbidity, and Khumbe nitrate, manganese,villages, Khumbe alkalinity, poultry magnesium, farm, and and thetotal Vuwani coliform, military representing base. There Dzindi was no downstream. association Thewith presence rainfall ofat the ions Dzindi (cations River and at anions) the downstream may account section. for thePC2 electrical accounte conductivityd for 20% of the and variation total dissolved (Table 6) solids. and Thesewas associated research findings strongly suggest with origins turbidity, downstream manganese, may total be associated coliform mostly, and weak withly the associated Vuwani sewage with plantchloride and, whi agriculturech represented [37–39 ].Dzindi The sewage upstream. plant receives mostly sewage influent mainly associated with domestic use from Dzwerani, Mathule, and Khumbe villages, Khumbe poultry farm, and the Table 5. Water quality eigenvectors of the correlation matrix. Vuwani military base. There was no association with rainfall at the Dzindi River at the downstream section. PC2Parameter accounted forPC1 20% ofPC2 the variationPC3 PC4 (Table PC56) andPC6 was PC7 associated PC8 stronglyPC9 PC10 with turbidity,PC11 manganese,Turbidity total coliform, and weakly0.004 0.028 associated 0.165 with0.000 chloride, 0.466 which0.209 represented0.063 0.065 Dzindi0.055 upstream.0.000 Total dissolved solids 0.004 0.000 0.147 0.020 0.138 0.414 0.002 0.003 0.001 0.012 Electrical conductivity 0.028 0.000 0.021 0.035 0.043 0.442 0.000 0.002 0.000 0.114 Alkalinity Table0.165 5. Water0.147 quality0.021 eigenvectors 0.179 of the0.030 correlation 0.435 matrix.0.001 0.356 0.008 0.441 Manganese 0.000 0.020 0.035 0.179 0.481 0.263 0.070 0.035 0.040 0.007 MagnesiumParameter 0.466 PC10.138 PC20.043 PC3 0.030 PC4 0.481 PC5 PC6 PC70.339 PC80.008 PC9 0.212 PC10 0.021 PC11 0.397 ChlorideTurbidity 0.209 0.414 0.0040.442 0.028 0.435 0.165 0.263 0.000 0.4660.339 0.209 0.0630.436 0.065 0.165 0.055 0.472 0.000 0.455 FluorideTotal dissolved solids0.063 0.0040.002 0.000 0.000 0.001 0.147 0.070 0.020 0.1380.008 0.4140.436 0.002 0.0030.040 0.001 0.000 0.012 0.264 Electrical conductivity 0.028 0.000 0.021 0.035 0.043 0.442 0.000 0.002 0.000 0.114 Nitrate 0.065 0.003 0.002 0.356 0.035 0.212 0.165 0.040 0.007 0.071 Alkalinity 0.165 0.147 0.021 0.179 0.030 0.435 0.001 0.356 0.008 0.441 SulfateManganese 0.055 0.0000.001 0.0200.000 0.035 0.008 0.179 0.040 0.4810.021 0.2630.472 0.0700.000 0.035 0.007 0.040 0.007 0.210 Total coliformMagnesium 0.000 0.4660.012 0.1380.114 0.043 0.441 0.030 0.007 0.481 0.397 0.3390.455 0.0080.264 0.212 0.071 0.021 0.210 0.397 Chloride 0.209 0.414 0.442 0.435 0.263 0.339 0.436 0.165 0.472 0.455 Fluoride 0.063 0.002 0.000 0.001 0.070 0.008 0.436 0.040 0.000 0.264 Nitrate 0.065 0.003 0.002 0.356 0.035 0.212 0.165 0.040 0.007 0.071 Sulfate 0.055 0.001 0.000 0.008 0.040 0.021 0.472 0.000 0.007 0.210 Total coliform 0.000 0.012 0.114 0.441 0.007 0.397 0.455 0.264 0.071 0.210 Sustainability 2016, 8, 597 10 of 16 Sustainability 2016, 8, 597 10 of 16

Figure 5. The scree plot of eigenvalues for Dzindi river system.system.

This is is mainly mainly due due to to subsistence subsistence agriculture agriculture that that is practiced is practiced at Dzindi at Dzindi upstream, upstream and the, and Dzindi the irrigationDzindi irrigation scheme and scheme the washing and the of washingmotor cars of at motor the Dzindi cars irrigation at the Dzindi canal. PC3 irrigation accounted canal. for PC3 12% ofaccounted the variation for 12% (Table of6 ) the and variation was associated (Table strongly 6) and with was chloride, associated and strongly negatively with associated chloride, with and nitrate,negative whichly associated represented with thenitrate natural, which Dzindi represented river system. the natural Dzindi river system.

Table 6.6. TheThe eigenvalues eigenvalues of of the the component component matrix matrix of ofwater water quality quality variables variables of Dzindi of Dzindi upstream upstream and and downstream. downstream.

ParameterParameter PC1PC1 PC2 PC3 PC3 EigenvalueEigenvalue 5.9595.959 2.208 1.266 1.266 % variancevariance explained explained 54.17054.170 20.073 11.513 11.513 %Cumulative variance variance 54.1754.17 20.07 85.756 85.756 ´ Electrical conductivity conductivity 0.9440.944 −0.176 Sulfate 0.926 ´0.299 Total dissolvedSulfatesolids 0.9150.926 − 0.1460.299 Total dissolvedFluoride solids 0.8940.915 ´0.1460.389 0.181 TurbidityFluoride 0.7500.894 − 0.5990.389 0.181 0.182 TurbidityNitrate 0.7450.750 0.599 ´0.1820.534 Manganese 0.725 0.532 0.135 AlkalinityNitrate 0.6100.745 ´0.487 −0.534 0.322 MagnesiumManganese 0.4840.725 ´0.5320.701 0.135 TotalAlkalinity coliform 0.5700.610 − 0.6770.487 ´0.3220.173 MagnesiumChloride 0.484 − 0.2770.701 0.866 Total Note:coliform three components0.570 were extracted.0.677 −0.173 Chloride 0.277 0.866 3.4. Assessing the Impact of UpstreamNote: Municipal three components Sewage Discharge were extracted on Luvuvhu. Sampling Sites Fifteen items from Luvuvhu sampling sites, upstream and downstream, and the Nandoni reservoir were3.4. Assessing analyzed the with Impact principal of Upstream component Municipal analysis Sewage (Table Discharge7). on Luvuvhu Sampling Sites Fifteen items from Luvuvhu sampling sites, upstream and downstream, and the Nandoni reservoir were analyzed with principal component analysis (Table 7). Sustainability 2016, 8, 597 11 of 16

Table 7. Water quality eigenvectors of the correlation matrix.

Parameter PC1 PC2 PC3 PC4 PC5 PC6 PC7 PC8 PC9 PC10 PC11 PC12 PC13 PC14 PC15 Turbidity 0.086 0.105 0.025 0.276 0.090 0.000 0.215 0.105 0.327 0.441 0.387 0.016 0.493 0.023 Total dissolved solids 0.086 0.027 0.000 0.125 0.163 0.393 0.164 0.231 0.334 0.169 0.032 0.002 0.143 0.017 Electrical conductivity 0.105 0.027 0.156 0.103 0.349 0.127 0.059 0.457 0.041 0.235 0.107 0.002 0.218 0.419 Alkalinity 0.025 0.000 0.156 0.218 0.154 0.362 0.321 0.150 0.329 0.208 0.074 0.034 0.019 0.000 Manganese 0.276 0.125 0.103 0.218 0.230 0.184 0.273 0.189 0.133 0.088 0.018 0.163 0.073 0.070 Magnesium 0.090 0.163 0.349 0.154 0.230 0.004 0.026 0.000 0.101 0.466 0.094 0.437 0.425 0.429 Chloride 0.000 0.393 0.127 0.362 0.184 0.004 0.197 0.006 0.369 0.499 0.067 0.293 0.366 0.466 Bromide 0.215 0.164 0.059 0.321 0.273 0.026 0.197 0.033 0.000 0.013 0.006 0.326 0.437 0.423 Fluoride 0.105 0.231 0.457 0.150 0.189 0.000 0.006 0.033 0.139 0.354 0.109 0.425 0.455 0.491 Nitrate 0.327 0.334 0.041 0.329 0.133 0.101 0.369 0.000 0.139 0.001 0.006 0.356 0.472 0.423 Sulfate 0.441 0.169 0.235 0.208 0.088 0.466 0.499 0.013 0.354 0.001 0.067 0.260 0.321 0.169 Total coliform 0.387 0.032 0.107 0.074 0.018 0.094 0.067 0.006 0.109 0.006 0.067 0.013 0.394 0.010 Average air temperature 0.016 0.002 0.002 0.034 0.163 0.437 0.293 0.326 0.425 0.356 0.260 0.013 0.070 0.113 Average humidity 0.493 0.143 0.218 0.019 0.073 0.425 0.366 0.437 0.455 0.472 0.321 0.394 0.070 0.005 Total rainfall 0.023 0.017 0.419 0.000 0.070 0.429 0.466 0.423 0.491 0.423 0.169 0.010 0.113 0.005 Sustainability 20162016,, 8,, 597597 12 of 16

The principal component analysis showed the first five components accounted for 87% of the total The variation principal in water component quality analysis and the showed scree plot the offirst eigenvalues five components of the accountedcomponents for is 87% shown of the in totalFigure variation 6. PC1 accounted in water qualityfor 30% and of the the variation scree plot (Table of eigenvalues 8) and was of a thessociated components strongly is with shown total in Figuredissolved6. PC1 solids, accounted electrical for conductivity, 30% of the magnesium, variation (Table fluoride,8) and nitrate, was associated sulfate, total strongly coliform with average total dissolvedair temperature solids,, and electrical total conductivity,rainfall, and negatively magnesium, associated fluoride, with nitrate, manganese sulfate, total and coliform bromide average, which airrepresented temperature, Luvuvhu and total upst rainfall,ream (S6). and There negatively was strong associated association with manganese with rainfall and atbromide, the Luvuvhu which representedRiver at the Luvuvhuupstream upstream section, indicating (S6). There the was washing strong association of animal withwastes rainfall and leaching at the Luvuvhu of fertilizers River atfrom the commercial upstream section, agriculture indicating in the theLevubu washing tropical of animal area [40]. wastes The and Luvuvhu leaching upstream of fertilizers section from is commerciallocated close agriculture to the source in the of Levubu the Luvuvhu tropical River area and [40 ]. is The a high Luvuvhu rainfall upstream area in the section Soutpansberg is located closeMountains to the source[40]. of the Luvuvhu River and is a high rainfall area in the Soutpansberg Mountains [40].

Figure 6. TThehe scree plot of eigenvalues for Luvuvhu river system and the Nandoni reservoir.reservoir.

Table 8. The Eigenvalues of the component matrix of water quality variables of Luvuvhu upstream Table 8. The Eigenvalues of the component matrix of water quality variables of Luvuvhu upstream and downstream and the Nandoni reservoir. and downstream and the Nandoni reservoir. Parameter PC1 PC2 PC3 PC4 PC5 ParameterEigenvalue PC14.516 PC23.267 PC32.124 1.882 PC4 1.218 PC5 % varianceEigenvalue explained 4.51630.108 3.26721.283 2.12414.160 12.547 1.882 8.123 1.218 %%Cumulative variance explained variance 30.10830.108 21.28351.891 14.16066.051 78.598 12.547 86.721 8.123 %Cumulative variance 30.108 51.891 66.051 78.598 86.721 TurbidityTurbidity ´0.212−0.212 0.8170.817 0.0700.070 ´−0.3090.309 0.290 0.290 TotalTotal dissolved dissolved solids solids 0.7540.754 ´0.276−0.276 0.1770.177 0.281 0.139 0.139 ElectricalElectrical conductivity conductivity 0.5360.536 ´0.221−0.221 ´0.583−0.583 0.198 0.189 0.189 AlkalinityAlkalinity ´0.704−0.704 0.3490.349 ´0.447−0.447 ´−0.2430.243 0.189 0.189 Manganese ´0.479 0.126 ´0.053 0.584 ´0.545 Manganese −0.479 0.126 −0.053 0.584 −0.545 Magnesium 0.421 0.729 0.133 0.486 ´0.106 ChlorideMagnesium 0.1950.421 0.7600.729 0.3350.133 ´0.4860.046 −0.106 0.451 BromideChloride ´0.6410.195 ´0.4570.760 0.3360.335 −0.046 0.167 0.451 0.289 FluorideBromide 0.377−0.641 0.722−0.457 0.2210.336 0.167 0.478 0.289´0.165 NitrateFluoride 0.6200.377 0.3340.722 ´0.4440.221 ´0.4780.356 −0.165´0.323 Sulfate 0.534 0.000 ´0.279 ´0.535 ´0.281 Total coliformNitrate 0.8240.620 0.1320.334 ´0.052−0.444 ´−0.3560.107 −0.323 0.242 Average airSulfate temperature 0.5890.534 0.1320.000 ´0.052−0.279 ´−0.5350.107 −0.281 0.242 AverageTotal humidity coliform 0.2610.824 ´0.1120.132 0.738−0.052 ´−0.1070.419 0.242´0.231 AverageTotal air rainfall temperature 0.5990.589 ´0.4210.132 0.565−0.052 ´−0.1070.124 0.242´0.062 Average humidityNote: five components0.261 −0.112 were extracted. 0.738 −0.419 −0.231 Total rainfall 0.599 −0.421 0.565 −0.124 −0.062 Note: five components were extracted.

Sustainability 2016, 8, 597 13 of 16

PC2 accounted for 21% of the variation (Table 8) and was associated strongly with turbidity, alkalinity, magnesium, chloride, fluoride, nitrate, and strong negative association with rainfall, which represented Luvuvhu downstream (S5). The negative association with rainfall at the Luvuvhu downstream point implies that the rainfall was not involved and that the pollutants probably originated from the municipal sewage plants and agriculture return flows [36–39]. These research Sustainabilityfindings suggest2016, 8, 597 origins downstream may be associated mostly with municipal sewage plants, such13 of 16 as Vuwani oxidation ponds, Elim and Waterval sewage plants, and agriculture. The sewage plant receives mostly sewage influent mainly associated with domestic use from Ha-Manavhela, Mambedi,PC2 accounted , for Mabodo, 21% of and the Mpheni variation villages (Table, 8and) and small was towns associated of Vuwani, strongly Mambedi with turbidity,lodge, alkalinity,Waterval magnesium,, and Elim and chloride, the Elim fluoride, hospital nitrate,, and no and industry strong, negativeas shown association by the negative with rainfall,association which representedwith manganese. Luvuvhu The presence downstream of chloride (S5). in The the negativesewage effluent association indicates with the rainfall origin is at from the sodium Luvuvhu downstreamchloride, which point is implies mostly thatassociated the rainfall with wastable not salt involved and human and thatdiets the [38,3 pollutants9]. The agriculture probably originated return fromflows the are municipal due to intensive sewage plants irrigation and agriculture that is practiced return at flows the Levubu [36–39]. tr Theseopical research area (Figure findings 7) withsuggest originsirrigation downstream coming from may bethe associated Luvuvhu mostlyRiver, Albasini with municipal dam, and sewage groundwater plants, such sources as Vuwani [37,38]. oxidation Even ponds,the subsistence Elim and Watervalfarming operations sewage plants, in rural and areas agriculture., such Valdezia The sewage and Mambedi plant receives, also pump mostly water sewage influentfrom the mainly Luvuvhu associated River. The with study domestic of Zha useng et from al. [1 Ha-Manavhela,2] in China also showed Mambedi, that Valdezia, the presence Mabodo, of andnitrates Mpheni in villages, freshwater and bodies small aretowns attributed of Vuwani, to sewage Mambedi plants, lodge, point Waterval, sources and of Elim pollution and the and Elim hospital,agriculture, and as no nonpoint industry, sources as shown of pollution by the negativewhich are association difficulty to withcontrol. manganese. The presence of chloridePC3 in accounted the sewage for effluent 14% of indicates the variation the origin (Table is 8 from) and sodium was moderately chloride, whichassociated is mostly with associatedrainfall and weakly associated with chloride and bromide, and negatively associated with nitrate, which with table salt and human diets [38,39]. The agriculture return flows are due to intensive irrigation represented the natural Nandoni reservoir system. The villages that are close to the Nandoni that is practiced at the Levubu tropical area (Figure7) with irrigation coming from the Luvuvhu reservoir are Dididi, Makovha, and Mulenzhe, and the communities obtain some of their water for River, Albasini dam, and groundwater sources [37,38]. Even the subsistence farming operations in human and livestock watering here [22]. PC4 accounted for 13% of the variation (Table 8) and was rural areas, such Valdezia and Mambedi, also pump water from the Luvuvhu River. The study of moderately associated with manganese, magnesium, and fluoride, and negatively associated with Zhangturbidity et al. and [12 nitrate] in China, which also represented showed that the the Luvuvhu presence River of nitrates downstream. in freshwater PC5 accounted bodies are for attributed 8% of tothe sewage variation plants, (Table point 8) sources and was of weakly pollution associated and agriculture, with chloride as nonpoint, and negative sourcesly ofassociated pollution with which aremanganese difficulty, towhich control. represented the natural Luvuvhu River downstream.

(a) (b)

Figure 7. The practice of commercial agriculture: (a) maize growing and (b) banana plantations in the Figure 7. The practice of commercial agriculture: (a) maize growing and (b) banana plantations in the Levubu tropical area. Levubu tropical area. 4. Conclusions PC3 accounted for 14% of the variation (Table8) and was moderately associated with rainfall The use of principal component analysis, coupled with factor and cluster analysis, was used to and weakly associated with chloride and bromide, and negatively associated with nitrate, which study the similarities and differences between upstream and downstream sampling sites that are represented the natural Nandoni reservoir system. The villages that are close to the Nandoni reservoir downstream of municipal sewage plants. Cluster analysis showed the presence of two clusters in are Dididi, Makovha, and Mulenzhe, and the communities obtain some of their water for human and each of the sampling sites. The principal component analysis identified nitrate and total coliforms as livestock watering here [22]. PC4 accounted for 13% of the variation (Table8) and was moderately major pollutants in the downstream sections of the sampling sites and attributed this to effluent associateddischarge withfrom manganese,municipal sewage magnesium, plants. Thus and fluoride,, the ecological and negatively impact of municipal associated sewage with turbidity discharge and nitrate,contributes which to representedproliferation theof cyanobacteria Luvuvhu River due downstream. to available nitrates PC5 accounted and aquatic for weeds 8% of to theprovide variation a (Tablehabitat8) andfor schistomiasis was weakly associatedvector snails. with This chloride, may render and negativelythe raw water associated supply withunsuitable manganese, for human which representedconsumption the without natural Luvuvhuconventional River water downstream. treatment processes and an added cost on the use of

4. Conclusions The use of principal component analysis, coupled with factor and cluster analysis, was used to study the similarities and differences between upstream and downstream sampling sites that are downstream of municipal sewage plants. Cluster analysis showed the presence of two clusters in each of the sampling sites. The principal component analysis identified nitrate and total coliforms as major pollutants in the downstream sections of the sampling sites and attributed this to effluent discharge from municipal sewage plants. Thus, the ecological impact of municipal sewage discharge contributes Sustainability 2016, 8, 597 14 of 16 to proliferation of cyanobacteria due to available nitrates and aquatic weeds to provide a habitat for schistomiasis vector snails. This may render the raw water supply unsuitable for human consumption without conventional water treatment processes and an added cost on the use of activated carbon to reduce cyanotoxins in the final treated water supply. The other ecological impact is the increase in schistomiasis infections among the waterfront villagers along the Nandoni water reservoir and Luvuvhu River downstream of Nandoni dam.

Acknowledgments: The University of Venda is acknowledged for their financial support for research study (IS05). The climate data was provided by the South African Weather Services from their Thohoyandou weather station. Author Contributions: All the authors were involved in the study. Jabulani Ray Gumbo (Hydrology and Water Resources) was involved in data collection, analysis, writing and proof reading the manuscript. Ratshilumela Aaron Dzaga (Geography and Geo-Information Science) was involved in data collection, analysis, writing of initial draft manuscript. Nthaduleni Samuel Nethengwe (Geography and Geo-Information Science) was involved in data collection; analysis and writing draft the manuscript. Conflicts of Interest: The authors declare no conflict of interest. The founding sponsors had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, and in the decision to publish the results.

Abbreviations The following abbreviations are used in this manuscript: DWAF Department of Water Affairs and Forestry PCA Principal component analysis CA cluster analysis FA factor analysis MPN most probable number LD linear dichroism

References

1. Brundtland Commission. Our common future, Chapter 2: Towards sustainable development. In World Commission on Environment and Development (WCED); United Nation: Geneva, Switzerland, 1987. 2. Lemarchand, K.; Lebaron, P. Occurrence of Salmonella spp. and Cryptosporidium spp. in a French coastal watershed: Relationship with fecal indicators. FEMS Microbiol. Lett. 2003, 218, 203–209. [CrossRef][PubMed] 3. Herzog, F.; Prasuhn, V.; Spiess, E.; Richner, W. Environmental cross-compliance mitigates nitrogen and phosphorus pollution from Swiss agriculture. Environ. Sci. Policy 2008, 11, 655–668. [CrossRef] 4. Jonnalagadda, S.B.; Mhere, G. Water Quality of the Odzi River in the Eastern Highlands of . Water Res. 2001, 35, 2371–2376. [CrossRef] 5. Hirji, R.; Johnson, P.; Maro, P.; Matiza Chiuta, T. Defining and Mainstreaming Environmental Sustainability in Water Resources Management in Southern Africa. Available online: http: //www-wds.worldbank.org/external/default/WDSContentServer/WDSP/IB/2004/03/11/000090341_ 20040311143729/Rendered/PDF/280300vol101001stainability0Summary.pdf (accessed on 17 June 2016). 6. Koukal, B.; Dominik, J.; Vignati, D.; Arpagaus, P.; Santiago, S.; Ouddane, B.; Bennaabidate, L. Assessment of water quality and toxicity of polluted rivers Fez and Sebou in the region of Fez (Morocco). Environ. Pollut. 2004, 131, 163–172. [CrossRef][PubMed] 7. Oberholster, P.J.; Botha, A.M.; Cloete, T.E. Biological and chemical evaluation of sewage water pollution in the Rietvlei nature reserve wetland area, South Africa. Environ. Pollut. 2008, 156, 184–192. [CrossRef] [PubMed] 8. Bapela, M.H. Water Quality and Associated Problems in the Luvuvhu Basin. Master’s Thesis, University of Venda, Thohoyandou, South Africa, 2001. 9. Zhou, F.; Guo, H.; Liu, Y.; Jiang, Y. Chemometrics data analysis of marine water quality and source identification in Southern Hong Kong. Mar. Pollut. Bull. 2007, 54, 745–756. [CrossRef][PubMed] 10. Boyacioglu, H.; Boyacioglu, H. Water pollution sources assessment by multivariate statistical methods in the Tahtali Basin, Turkey. Environ. Geol. 2008, 54, 275–282. [CrossRef] Sustainability 2016, 8, 597 15 of 16

11. Mendiguchía, C.; Moreno, C.; Galindo-Riaño, M.D.; García-Vargas, M. Using chemometric tools to assess anthropogenic effects in river water: A case study: Guadalquivir River (Spain). Anal. Chim. Acta 2004, 515, 143–149. 12. Zhang, Q.; Zhongwu, L.; Guangming, Z.; Jianbing, L.; Yong, F.; Qingshui, Y.; Yamei, W.; Fangyi, Y. Assessment of surface water quality using multivariate statistical techniques in red soil hilly region: A case study of Xiangjiang watershed, China. Environ. Monit. Assess. 2009, 152, 123–131. [CrossRef][PubMed] 13. Kebede, Y.K.; Kebedee, T. Application of principal component analysis in surface water quality monitoring. In Principal Component Analysis—Engineering Applications; Sanguansat, P., Ed.; InTech: Rijeka, Croatia, 2012; pp. 83–100. 14. Department of Water Affairs and Forestry (DWAF). Sustainable Utilization Plan for the Nandoni Dam in the Thohoyandou District of the Limpopo Province; Department of Water Affairs and Forestry/Van Riet and Louw Landscape Architects: Pretoria, South Africa, 2003. 15. Liu, C.W.; Lin, K.H.; Kuo, Y.M. Application of factor analysis in the assessment of groundwater quality in a Blackfoot disease area in Taiwan. Sci. Total Environ. 2003, 313, 77–89. [CrossRef] 16. Kaiser, H.F. An index of factorial simplicity. Psychometrika 1974, 39, 31–36. [CrossRef] 17. Bartlett, M.S. A note on the multiplying factors for various χ2 approximations. J. R. Stat. Soc. Ser. B Methodol. 1954, 16, 296–298. 18. Department of Water Affairs and Forestry (DWAF). South Africa Water Quality Guidelines, Domestic Uses, 2nd ed.; DWAF: Pretoria, South Africa, 1996; Volume 1. 19. Department of Water Affairs and Forestry (DWAF). South African Water Quality Guidelines, Recreational Water Use, 2nd ed.; DWAF: Pretoria, South Africa, 1996; Volume 2. 20. Department of Water Affairs and Forestry (DWAF). South Africa Water Quality Guidelines, Aquatic Ecosystems, 1st ed.; DWAF: Pretoria, South Africa, 1996; Volume 7. 21. Mulaudzi, D.A.; Gumbo, J.R. The Reduction of Microcystin congeners in Raw Water Source with the use of Ceramic Water Filters. Transdiscipl. E-J. (TEJ) Inaug. Ed. 2014, 1, 77–105. 22. Makhera, M.; Gumbo, J.R.; Chigayo, K. Monitoring of microcystin-LR in Luvuvhu River catchment: Implications for human health. Afr. J. Biotechnol. 2013, 10, 405–412. 23. World Health Organization (WHO). Guidelines for Drinking-Water Quality; WHO: Geneva, Switzerland, 2011. 24. Mukhari, R.K.B. Chemical Characterization of Drinking Water Sludge: A Case Study of Nandoni Water Treatment Works. Master’s Thesis, University of Venda, Thohoyandou, South Africa, 2013. 25. Srinivasan, R.; Sorial, G.A. Treatment of taste and odor causing compounds 2-methyl isoborneol and geosmin in drinking water: A critical review. J. Environ. Sci. 2011, 23, 1–13. [CrossRef] 26. Ndimele, P.E.; Kumolu-Johnson, C.A.; Anetekhai, M.A. The invasive aquatic macrophyte, water hyacinth {Eichhornia crassipes (Mart.) Solm-Laubach: Pontedericeae}: Problems and prospects. Res. J. Environ. Sci. 2011, 5, 509–520. [CrossRef] 27. McCartney, M.; Boelee, E.; Cofie, O.; Amerasinghe, F.; Mutero, C. Planning and management to improve the benefits and reduce the environmental and health costs of agricultural water development in sub-Saharan Africa: Desk. In Review of Literature Report, Volume 2–Full Component Reports; International Water Management Institute: Colombo, Sri Lanka, 2004; p. 140. 28. Mathye, C. Disease threatens villagers. Rise and Shine Newspaper, 17–23 October 2014; p. 5. 29. Dzaga, R.A. An Assessment of the Impact of Catchment activities on Water Quality: A Case study of Nandoni Dam, Vhembe District, Limpopo Province. Unpublished Master’s Thesis, University of Venda, Thohoyandou, South Africa, 2012. 30. Oberholster, P.J.; Myburgh, J.G.; Govender, D.; Bengis, R.; Botha, A.M. Identification of toxigenic Microcystis strains after incidents of wild animal mortalities in the Kruger National Park, South Africa. Ecotoxicol. Environ. Saf. 2009, 72, 1177–1182. [CrossRef][PubMed] 31. Foss-Kankeu, E.; Jagals, P.; du Preez, H. Exposure of rural households to toxic cyanobacteria in container-stored water. Water SA 2008, 34, 631–636. 32. Rananga, H.T.; Gumbo, J.R. Willingness to Pay for Water Services in Two Communities of Local Municipality, South Africa: A Case study. J. Hum. Ecol. 2015, 49, 231–243. 33. Bessong, P.O.; Odiyo, J.O.; Musekene, J.N.; Tessema, A. Spatial Distribution of Diarrhoea and Microbial Quality of Domestic Water during an Outbreak of Diarrhoea in the Tshikuwi Community in Venda, South Africa. J. Health Popul. Nutr. 2009, 5, 652–659. [CrossRef] Sustainability 2016, 8, 597 16 of 16

34. Singh, K.P.; Malik, A.; Mohan, D.; Sinha, S. Multivariate statistical techniques for the evaluation of spatial and temporal variations in water quality of Gomti River (India)—A case study. Water Res. 2004, 38, 3980–3992. [CrossRef][PubMed] 35. Sudevi, B.; Lokesh, K. Evaluation of Cauvery River Water Quality at Srirangapatna in Karnataka using Principal Component Analysis. Int. J. Eng. Sci. 2012, 1, 6–12. 36. Davis, S.N.; Whittemore, D.O.; Fabryka-Martin, J. Uses of chloride/bromide ratios in studies of potable water. Ground Water 1998, 36, 338–350. [CrossRef] 37. De Wit, M.; Meinardi, C.; Wendland, F.; Kunkel, R. Modelling water fluxes for the analysis of diffuse pollution at the river basin scale. Hydrol. Process. 2000, 14, 1707–1723. [CrossRef] 38. Muriithi, F.K.; Yu, D. Understanding the Impact of Intensive Horticulture Land-Use Practices on Surface Water Quality in Central Kenya. Environments 2015, 2, 521–545. [CrossRef] 39. Rezaei, A.; Sayadi, M.H. Long-term evolution of the composition of surface water from the River Gharasoo, Iran: A case study using multivariate statistical techniques. Environ. Geochem. Health 2015, 37, 251–261. [CrossRef][PubMed] 40. Water Research Commission (WRC). State of Rivers Report: Letaba and Luvuvhu River Systems; WRC Report, Number TT 165/01; WRC: Pretoria, South Africa, 2001.

© 2016 by the authors; licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC-BY) license (http://creativecommons.org/licenses/by/4.0/).