Assessment of variations in metal concentrations of the Ganges River water by using multivariate statistical techniques

Worldwide, metal pollution of river waters is a societal problem as human civilization thrives on the bank of rivers, warrants identification of sources and evaluation of possible toxicity to formulate strategies for pollution abatement and sustainable management of water resources. The present study was conducted to document the metal concentrations of surface water samples along the Ganges River using multivariate statistics and to generate information for comparison of water quality. Ba, Cu, Fe, Li, Na and Sr showed significant variations (P < 0.05) both at spatial and temporal scales. Contamination factor and metal index demonstrated that the water in the middle segment stretching from Kanpur to Varanasi is more contaminated and vulnerable to anthropogenic stress. Principal component analysis (PCA) generated four principal components (PCs) with eigenvalues > 1 and these PCs explain the 87.4% of variation in metal concentration. The first two PCs accounted for 52.2% of the total variance and showed a strong correlation with Fe, Li, Mn, Na and Mg. The hierarchical cluster analysis (HCA) shows three clusters based on seasonal sampling at the four locations along the Ganges River. The first two discriminant functions (DFs) explained 99.7% of the variance in metal concentrations among Narora, Kanpur, Varanasi and Bhagalpur sampling locations. Mn, Sr and Na were most significant in the distinction of water samples to their original location with a cross-validation classification accuracy of 63.9%. In addition to longterm monitoring programs, the information generated on the variations of metal concentrations can be used to solve the problems of metal pollution of the Ganges River water.