4.7.2. Assessing the Swartkops River health using the South African Scoring System version 5 (SASS5)

4.6. Statistical analysis

12 Water Quality

4.7. Results

Dissolved oxygen

Electrical conductivity

pH 6.53 ± 1.11

Temperature (°C) 17.48 ± 5.46

Turbidity (NTU) 5.3 ± 2.22<sup>a</sup>

BOD5 (mg/l) 4.62 ± 1.45a

(mg/l)

(mS/m)

4.7.1. Water physico-chemical variables

One-way analysis of variance (ANOVA) was used to test for differences (p < 0.05) in the means of the analysed physico-chemical variables between the four sampling sites. When ANOVA indicated significant differences, a post hoc test, the Tukey's Honestly Significant Different (HSD) test was computed to indicate sites that differed. The basic assumptions of normality and homogeneity of variance were investigated using the Shapiro-Wilk test and the Levene's test, respectively. The nonparametric Kruskal-Wallis multiple comparison test was used to evaluate whether SASS5 scores, ASPT values and the number of taxa differed significantly between the biotope groups. ANOVA and Kruskal-Wallis multiple comparison tests were

Table 2 shows the mean, standard deviation and range of physico-chemical variables measured during the study period. With the exception of pH and temperature, the measured variables were statistically significantly different between the sampling sites (p < 0.05). The lowest value of DO and highest turbidity and BOD5 values were recorded at Site 3. The Tukey's HSD post hoc test revealed that the mean DO concentration was significantly lower at Site 3 than at Sites 1 and 2. Although pH and temperature were not statistically significantly different between the sampling sites, the highest mean pH and temperature values were at Sites 2 and 3, respectively, and the lowest at Sites 1 and 2, respectively. The Tukey's HSD post hoc test showed that the mean EC concentration was significantly lower at Site 1 than at the rest of the sampling sites and turbidity significantly higher at Site 3. The mean BOD5 concentrations were

> p value

4.81 ± 3.01ab (0.9–8.31)

7.27 ± 0.56 (6.31–8.01)

18.9 ± 4.14 (12.2–24.0)

259.63 ± 56.28<sup>b</sup> (171–354)

7.08 ± 8.06<sup>a</sup> (2.2–26)

11.77 ± 5.28bc (2.24–22.94)

F value

0.001 7.18

0.201 1.65

0.415 0.98

0.000 22.57

0.000 15.67

0.002 13.50

undertaken using the Statistica software package version 9.

significantly higher at Sites 3 and 4 than at Site 1 (Table 2).

6.99 ± 1.15a (4.73–9.5)

(4.69–7.75)

(7.31–24.0)

(3.0–10.1)

(2.16–6.86)

32.45 ± 17.74<sup>a</sup> (8.23–62.0)

Variable Site 1 Site 2 Site 3 Site 4

7.4 ± 1.52a (5.53–9.48)

7.37 ± 1.11 (5.69–8.99)

17.27 ± 7.17 (6.11–27.3)

160.75 ± 146<sup>b</sup> (30–460)

6.33 ± 2.44<sup>a</sup> (3.0–11.2)

8.25 ± 4.33ab (4.58–16.68)

test. The same superscript letter between sites per variable indicates no significant differences (p > 0.05).

Table 2. Mean ± standard deviation and range (in parenthesis) of the physico-chemical variables (n = 8) in the Swartkops River during the study period (August 2009–September 2012). p and F values are indicated by ANOVA. Different superscript letters per variable across sites indicate significant differences (p < 0.05) revealed by Tukey's HSD post hoc

3.19 ± 1.47b (1.81–6.36)

7.29 ± 0.42 (6.56–7.9)

20.88 ± 3.29 (14.3–25.2)

262.51 ± 76.14<sup>b</sup> (154.8–333)

72.7 ± 102.36b (10.5–320)

14.54 ± 3.57<sup>c</sup> (8.32–20.62)

The interpretation of the SASS5 results were based on the range of SASS5 scores and ASPT values reflecting ecological categories A, B, C, D and E/F indicative of natural, good, fair, poor and very poor water quality conditions, respectively (Table 1). The SASS5 scores and ASPT values revealed that the Swartkops river health conditions differed between the sampling sites. Seasonally, with the exception of the autumn and spring (2012) collections, SASS5 scores at Site 1 indicated the B ecological category indicative of good water quality condition (Figure 2). The ASPT values on the other hand, in all the sampling seasons, indicated the C ecological category for Site 1, suggesting that the water quality at Site 1 was fair (Figure 3). The numbers of taxa vary slightly between the sampling seasons at Site 1 with more taxa occurring in spring (2012) (Figure 4). Overall, the SASS5 score showed good water quality (ecological category B) for Site 1, but the ASPT value indicated that the water quality condition at the site was fair (ecological category C) (Figure 5).

Figure 2. Seasonal variations for the South African Scoring System version 5 (SASS5) score at the four sampling sites in the Swartkops River during the study period (August 2009–September 2012). The ecological categories: A (natural water quality), B (good water quality), C (fair water quality), D (poor water quality) and E/F (very poor water quality) are indicated on the bars.

At Site 2, SASS5 scores indicated the D ecological category, that is, poor water quality in spring (2009) and in autumn (2010), while in summer (2009), it revealed the C category indicative of fair water quality (Figure 2). During the rest of the sampling events, SASS5 scores revealed the E/F ecological category indicating very poor water quality. Although the SASS5 scores reflected other ecological categories in addition to the E/F for Site 2, the ASPT values consistently showed that Site 2 was in the E/F ecological category (Figure 3). Although the number of taxa did not vary significantly between the sampling seasons at Site 2, the highest number of taxa (20) was recorded during autumn (2010). At Sites 3 and 4, SASS5 scores and ASPT values revealed the E/F ecological category (very poor water quality) throughout the sampling seasons. The overall lowest number of taxa (8) in the river was recorded at Site 3 in winter 2010 (Figure 4).

Figure 3. Seasonal variations for the average score per recorded taxon (ASPT) at the four sampling sites in the Swartkops River during the study period (August 2009–September 2012). The ecological categories: C (fair water quality), D (poor water quality) and E/F (very poor water quality) are indicated on the bars.

Figure 4. Seasonal variations for the number of taxa at the four sampling sites in the Swartkops River during the study period (August 2009–September 2012).

4.7.3. Comparing SASS5 scores, ASPT values and the numbers of taxa between the sampling biotopes (stone, vegetation and GSM)

The vegetation and stone biotope had higher SASS5 scores, ASPT values and numbers of taxa than the GSM biotope at Site 1 (Figure 6). The Kruskal-Wallis multiple comparison test revealed that SASS5 scores were significantly higher for the vegetation than for the GSM biotope at Site 1 (p < 0.05; KW-H = 7.21). Similarly, at Site 2, SASS5 scores were significantly higher for the vegetation than for the GSM biotope (p < 0.05; KW-H = 10.13), and though the stone had higher SASS5 scores, they were not statistically higher than the scores recorded for the GSM biotope. The pattern described for Site 2 was similar to those observed for Sites 3 and 4 where the SASS5 scores were significantly higher for the vegetation biotope than the stone and GSM biotopes (Site 3: p < 0.05; KW-H = 40.44), (Site 4: p < 0.05; KW-H = 18.14).

taxa did not vary significantly between the sampling seasons at Site 2, the highest number of taxa (20) was recorded during autumn (2010). At Sites 3 and 4, SASS5 scores and ASPT values revealed the E/F ecological category (very poor water quality) throughout the sampling seasons. The overall lowest number of taxa (8) in the river was recorded at Site 3 in winter 2010

4.7.3. Comparing SASS5 scores, ASPT values and the numbers of taxa between the sampling biotopes

Figure 4. Seasonal variations for the number of taxa at the four sampling sites in the Swartkops River during the study

Figure 3. Seasonal variations for the average score per recorded taxon (ASPT) at the four sampling sites in the Swartkops River during the study period (August 2009–September 2012). The ecological categories: C (fair water quality), D (poor

water quality) and E/F (very poor water quality) are indicated on the bars.

The vegetation and stone biotope had higher SASS5 scores, ASPT values and numbers of taxa than the GSM biotope at Site 1 (Figure 6). The Kruskal-Wallis multiple comparison test

(Figure 4).

14 Water Quality

(stone, vegetation and GSM)

period (August 2009–September 2012).

Figure 5. Summary of the SASS5 scores, number of taxa and ASPT values at the four sampling sites in the Swartkops River during the study period (August 2009–September 2012). The overall ecological categories: B (good water quality), C (fair water quality), D (poor water quality) and E/F (very poor water quality) are indicated on the bars.

The average score per recorded taxon (ASPT) values were similar between the three biotopes at Site 1, but at Site 2, the ASPT values were significantly higher for the vegetation than the GSM biotope (p < 0.05; KW-H = 9.45). The vegetation had significantly higher ASPT values than the stone and GSM biotopes at Sites 3 (p < 0.05; KW-H = 26.9) and 4 (p < 0.05; KW-H = 14.25).

Figure 6. Median (small square), inter-quartile ranges (box), non-outlier ranges (bars) for SASS5 scores, numbers of taxa and ASPT values recorded per biotope at the four sampling sites in the Swartkops River during the study period (August 2009–September 2012).

Stone and vegetation biotopes supported significantly higher numbers of taxa than the GSM biotope at Site 1 (p < 0.05; KW-H = 11.89), but at Site 2, only the vegetation supported significantly higher numbers of taxa than the GSM (p < 0.05; KW-H = 7.23). More taxa were recorded on the vegetation and GSM biotopes than on the stone biotopes at Site 3. The Kruskal-Wallis multiple comparison test indicated that the numbers of taxa for the stone biotope were significantly lower than the taxa recorded for the vegetation and GSM (p < 0.05; KW-H = 40.44) at Site 3. At Site 4, the stone and vegetation supported more taxa, but only the numbers of taxa supported by the vegetation biotope were significantly higher than the values recorded for the GSM (p < 0.05; KW-H = 16.27).

### 4.8. Discussion

Stone and vegetation biotopes supported significantly higher numbers of taxa than the GSM biotope at Site 1 (p < 0.05; KW-H = 11.89), but at Site 2, only the vegetation supported significantly higher numbers of taxa than the GSM (p < 0.05; KW-H = 7.23). More taxa were

Figure 6. Median (small square), inter-quartile ranges (box), non-outlier ranges (bars) for SASS5 scores, numbers of taxa and ASPT values recorded per biotope at the four sampling sites in the Swartkops River during the study period (August

2009–September 2012).

16 Water Quality

The ecosystem approach takes into account biodiversity conservation and therefore prioritises the protection of biodiversity as well as the sustainable use of water resources and the associated ecosystems. In the case study provided, the South African Scoring System version 5 (SASS5) was used to evaluate the health of the Swartkops River. In South Africa, SASS5 is one of the tools that contribute ecological information for the determining the ecological reserve and setting Resource Quality Objectives (RQOs). The SASS5 results indicated that water quality in the Swartkops River was critically modified at Sites 3 and 4 throughout the sampling period and the numbers of taxa occurring at these sites were significantly reduced compared to those occurring at Sites 1 and 2. Sites 3 and 4 were situated downstream of a WWTWs, which influenced the health of the river. The values of the measured physico-chemical variables at these sites, that is, Sites 3 and 4, provided evidence for negative impact arising from the discharges of wastewater effluents. For example, at Sites 3 and 4, higher values of turbidity and EC concentrations and lower DO concentrations were recorded. Since highly sensitive taxa have higher scores in the SASS5 sheet, oxygen depletion could easily affect the occurrence and distribution of these taxa. Therefore, it was expected that sites with low concentration of DO would experience the disappearance of sensitive taxa and the dominance of tolerant taxa, and hence, the critically modified health conditions recorded at Sites 3 and 4. In addition to lower DO concentrations at the downstream sites, the elevated turbidity level recorded at Site 3 could be detrimental to oxygen-sensitive biota as decomposition of solids with high organic content could lead to oxygen depletion, as was evident at Sites 3 and 4. The majority of the highly sensitive taxa on the SASS5 sheet use external gills for respiration. Highly turbid water is likely to impact on the breathing apparatus of external gill-bearing organisms, which can then lead to clogging [41]. The river health condition at Site 2, which was upstream of the effluent discharge point, but still situated within the urban and industrial town of Uitenhage, was mostly in the range of fair and very poor conditions. Diffuse pollution sources on the river catchments were the main contributors to deteriorating river health recorded at this site. Site 1, which was used as the control site, had conditions mostly in the good and fair categories. The implication was that the control site had some sensitive taxa, which had disappeared from the impacted sites.

The number of taxa, SASS5 scores and ASPT values were highest mostly in the stone and vegetation biotopes and differed significantly between the three biotopes. These differences could be attributed to differences in hydraulic, substrate and thermal conditions between the three biotopes. The stone and vegetation biotopes are morphologically complex and more stable than the GSM biotope and are therefore more likely to support more food and space resources, and thus more macroinvertebrate families leading to increased SASS5 scores and ASPT values. These results are in agreement with those of Dallas [42] who reported that the stone and vegetation biotopes supported more macroinvertebrate families and higher SASS5 score and ASPT values than the GSM biotope. It is therefore important to sample all available biotopes to capture a wider range of biodiversity when undertaken aquatic biomonitoring.

In summary, the deteriorating environmental water quality in the Swartkops River has impacted on the macroinvertebrate assemblages particularly at the downstream sites. This was expected because of the ranges of impacts these sites receive which include industrial and sewage effluent discharges, run-off from informal settlement and agricultural activities such as livestock farming. Water quality at Site 1 which was used as the control site in this study was indicated as good and fair by the SASS5 score and ASPT value, respectively. This is a cause for concern as the results showed that macroinvertebrates at this site were experiencing noticeable impacts. Overall, both the physico-chemical variable analysis and the biotic index results revealed that the Swartkops River was deteriorating in quality as it flowed downstream, indicating the need for an urgent management intervention.
