EVALUATION OF WATER POLLUTION LEVELSUSING MULTIVARIATE WATER QUALITY PARAMETERS

Abstract

The environmental sustainability and human health depend greatly on water quality measurement especially under the background of mounting anthropogenic pressure on water bodies. This study offers a detailed assessment of the degree of water pollution in terms of multivariate water quality parameters that are formed by combining statistical analysis and machine learning methods. The physicochemical indicators that have been included in the analysis are pH, turbidity, temperature, dissolved oxygen (DO) and bio-demand of oxygen (BOD), as well as the level of heavy metal such as lead, mercury and arsenic. The descriptive statistical analysis showed that there was a high variation in all parameters which indicated that the environmental conditions were not homogenous. The correlation analysis revealed that turbidity (r = 0.276) and BOD (r = 0.233) showed a positive relation with the level of pollution, and DO was negatively correlated (r = -0.118) which indicated that oxygen depletion occurred in the presence of the pollutant. Principal Component Analysis (PCA) has shown that few components are used to explain a large portion of the variance, which is why it is effective to perform dimensionality reduction and detect the factors of pollution that are dominant. Additionally, a random forest model was used to determine the predictive value of each of the parameters. The findings suggest that turbidity and BOD are most significant predictors, with heavy metals being the next significant contributors, whereas pH, DO, and temperature have relatively minor contributions. These results highlight predominance of particulate matter and organic pollution in the degradation of water quality. Overall, it can be concluded that the combination of multivariate statistical methods and machine learning offers a strong and confident framework of water quality assessment. The research provides useful information on the activities of pollution and contributes to the creation of evidence-based policies regarding effective monitoring and sustainable water resources management.