*3.3. Identification of Pollution Sources*

Pearson correlation analysis (Table 5) revealed significant relationship between the metals except Fe–Mn (0.584), Fe–Cr (0.632), Mn–Cu (0.567), Cu–Zn (0.619), Mn–Pb (0.592), Cu–Cr (0.555), Cu–Ni (0.622), Cr–Pb (0.572) and Pb–Co (0.608). However, Cd concentration at *p* < 0.05 and *p* < 0.01 was significantly correlated with all the other metals concentrations.

**Table 5.** Pearson correlation analysis among heavy metals and antioxidant enzymes.


Note: \*\* Correlation is significant at the 0.01 level (2-tailed), \* Correlation is significant at the 0.05 level (2-tailed), SOD = Superoxide dismutase, CAT = Catalase.

The similarity in terms of the pattern of distribution of heavy metals was established through the utilisation of principal component analysis (PCA); PCA pulls the data together into a form that can be managed easily and takes out a small number of latent factors to analyse the relationship between the variables observed [35,36]. PCA was applied here primarily to evaluate the source origin of heavy metals in the region.

The results for PCA are presented in Table 6. The high eigenvalue is an indication of pattern and to what extent the data is spread and this leads to consideration of high eigenvalue as the principal component. The components/factors associated with the results of PCA for heavy metals in sediments and freshwater mollusc had a total variation of 92.5%. Component 1 (PC1) recorded the highest eigenvalue of 13.82, accounting for 76.80% of the total variation, and dominate the other components with more significant variation and strong positive loadings (>0.90) of Mn, Cu, Cd, Cr, and Pb in the sediment. Component 2 (PC2) accounted for 8.81% of total variation and had moderate positive loadings (>0.50) of Cu and Pb in the sediment. Component 3 (PC3) accounted for 6.89% of the total variation having moderate positive loadings (>0.50) on Co.

**Table 6.** Principal component analysis (PCA) loadings for heavy metals in sediment (Sd) and *Bellamya unicolor* (S).


Note: Method of extraction: principal component analysis. Rotation method. Bold numbers indicate a strong loading value (>0.9); moderate loading value (>0.5).

Cluster analysis reveals the relationship between sampling sites (Figure 2a) based on heavy metals concentrations and also the relationship between the metals (Figure 2b). For the relationship between the sites, three (3) clusters were formed at a similarity level of 89.09. The first cluster presented only site S1 and the second cluster had only site 2 while the third cluster grouped sites S3, S4, and S5. Four (4) clusters were formed for the relationship between the metals at a similarity level of 99.57. The clusters were Cluster 1 (Fe), Cluster 2 (Mn, Cr, Cu, Cd, and Pb), Cluster 3 (Zn and Ni) and Cluster 4 (Co).

**Figure 2.** Cluster analysis for the relationship between (**a**) sites; (**b**) heavy metals.
