Water quality variation and source apportionment using multivariate statistical analysis

Abstract

Rivers are the primary source of potable water and other residential uses. Since wastewater from the nearby industries and urban areas degrades the quality of the rivers, it is essential to monitor their water quality and quantity continuously. This study determines the pollution level in a river in accordance with drinking water standards. The entropy-weighted water quality index (EWQI) indicates the level of river pollution, and the principal component analysis (PCA) and Pearson correlation matrix were used to identify pollution sources. Twenty parameters were considered for the study, which includes alkalinity, biochemical oxygen demand (BOD), calcium (Ca⁺²), cadmium (Cd), chloride (Cl^-), copper (Cu), dissolved oxygen (DO), electrical conductivity (EC), fluoride (F^-), hardness, iron (Fe), lead (Pb), magnesium (Mg⁺²), manganese (Mn), nitrate (NO₃^-), pH, potassium (K⁺), sodium (Na⁺), sulphate (SO₄^2-), total dissolved solids (TDS) and zinc (Zn). The EWQI varied between 68.93 (good) to 259.91 (extremely poor). Post-monsoon EWQI values are higher due to the fact that monsoon-filled sewers continue to drain into the river even after the rainy or monsoon season has ended. Winter, pre-monsoon, and monsoon all produced five components (PCs) that explained 79%, 74%, and 80% of the overall variance, respectively and post-monsoon produced 6 PCs that explained 83% of the overall variance. PCA and Pearson correlation matrix indicate pollution from runoff sources during the pre-monsoon and monsoon season along with the pollution sources existing in the monsoon and post-monsoon season, domestic and metal sources.

Keywords:

• Entropy weighted water quality index
• information entropy
• principal component analysis
• Pearson correlation matrix
• water quality

Availability of data and materials

The datasets used and/or analysed during the current research are available in a manuscript.