chemometric analysis of hydro-chemical data of an alluvial river – a case study |
| |
| Authors: | kunwar p singh amrita malik vinod k singh |
| |
| Institution: | 1. Environmental Chemistry Section, Industrial Toxicology Research Centre, 80, MG Marg, Lucknow, 226 001, India
|
| |
| Abstract: | Hydrochemistry of an alluvial river was investigated employing the chemometric techniques such as cluster analysis (CA), principal component analysis (PCA), discriminant analysis (DA) and partial least square (PLS) with a view to extract information about the variables responsible for spatial and temporal variations in river hydrochemistry and water quality, the hidden factors explaining the structure of the hydro-chemical database of the river, factors/processes influencing the river hydro-chemistry. Analysis of spearman's correlation coefficient revealed non-significant correlation of the pollution indicator (BOD, COD, SO4, F, NH4-N, NO3-N) variables with season and significant correlation with site, indicating contribution of the site-specific anthropogenic sources in the catchments. Spatial CA clustered the monitoring sites (10nos.) into three groups of relatively non-polluted sites, moderately polluted sites, and highly polluted sites. Temporal CA differentiated among the samples of monsoon and non-monsoon months. PCA rendered considerable data reduction, in terms of eight parameters explaining about 71% of the total variance and evolved six PCs. PCA grouped samples belonging to different seasons and sites distinctly correlating them with natural and anthropogenic variables. Temporal and spatial DA rendered 97 and 92% correct assignations of the samples, respectively, and revealed that temperature, pH, BOD, DO, alkalinity and Ca are the most significant variables to discriminate between the different seasons and account for most of the expected temporal variations in hydrochemistry of the river, whereas, hardness, DO, BOD, COD, Ca and Mg were the most significant discriminating variables in space. Spatial and temporal groupings of the samples were successfully achieved through PLS modeling. PLS showed that the summer season samples are dominated by PO4, TDS, F, K, COD, BOD, Na, Cl, hardness and alkalinity, whereas, samples of winter season by DO, pH, NH4-N and coliforms. Furthermore, PLS indicated site-specific dominance of anthropogenic contaminants suggesting for their pollution sources in the corresponding catchments of these sites. |
| |
| Keywords: | |
| 本文献已被 SpringerLink 等数据库收录！ |
|
