Source Apportionment of Groundwater Pollution using Unmix and Positive Matrix Factorization

Abstract

Receptor models are used to understand the attributes of groundwater contaminants by recognizing their sources and evaluating the contribution from each source to receptor concentrations. Two receptor models, Unmix and Positive matrix factorization (PMF), were applied to the data obtained from 41 sampling points on 20 parameters, in order to identify and apportion the pollution sources to groundwater quality in Peenya region of Bengaluru. Overall six and seven sources were identified by Unmix and PMF models, respectively. Most groundwater quality variables were found to be influenced primarily by pollution from chromium electroplating, sewage, geology of the area and lead acid battery manufacturing units located in the study area. The models could recognize significant sources adding to groundwater quality in the region with most of them being anthropogenic due to the presence of industrial activity. It was observed that both models gave good outcomes with regard to their capacity to repeat measured concentrations with fundamentally the same slopes in a large portion of the cases, yet with the PMF model demonstrating the best correlation and the nearest slope to unit. Receptor models are regularly applied to distinguish source contributions. The dissimilarities among the results of various models are essential to better interpret source apportionment.