Multivariate Statistical Approach for the Assessment of Water Quality of Mahanadi Basin,Odisha

Abstract

This research paper explores spatial and temporal water quality fluctuations to examine massive and complex water quality data sets used to quantify the influence of agricultural operations and household pollution sources on the Mahanadi River in Odisha. For a ten-year sampling study, (2008-2018), data sets containing 20 parameters were collected at 19 sampling sites along the river's length. The nineteen sampling locations were also split into three groups i.e. cluster 1 represents low polluted sites, Cluster 2 represents moderately polluted sites and cluster 3 depicts high polluted sites using the hierarchical clustering approach (HCA). It shows spatial and seasonal variations that are frequently symptomatic of contamination from rainfall or other sources. It yields positive results, with three separate groups of similarity across monitoring stations representing the river system's various water quality indicators. The FA/PCA identified the five most important factors, accounting for 93.899 percent of the total variance in the data matrix, allowing the selected parameters to be grouped based on common traits and the frequency of overall changes or variances within each group to be assessed. TSS, TKN, EC, TDS, B, SAR, and Fe have all been correlated with (loading >0.7) in the 1 PC, which accounted for 43.133 % of the total variance. COD, NH3-N, Free ammonia, and fluoride were all linked to the 2"d PC, which accounted for 23.055 % of the total variation, whereas the 3"d, 4th, and 5th PCs, while accounting for 12.866 %, 8.603 %, and 6.241% of the total variation, respectively. Five separate VFs with Eigenvalues > 1 were justified by maximum variance rotation of the PC (original). This illustrates that point source, non-point source and natural occurrences are the primary cause of changes in water chemical concentrations.