Identification and Apportionment of Pollution Sources to Groundwater Quality Using Receptor Models

Abstract

Characterizing groundwater quality and apportionment of pollution sources to groundwater pollution is important for managing water resources effectively. Owing to rapid industrialization and population growth in Bengaluru city, the groundwater quality is getting deteriorated. Application of source apportionment techniques to water quality problems, especially with respect to groundwater are limited in the Indian context. Therefore a scope exists for source apportionment of pollution sources to groundwater quality using receptor models. Multivariate statistical techniques (Cluster analysis, Discriminant analysis and Principal component analysis) and Receptor oriented source apportionment models were used to evaluate groundwater quality. To have first-hand information on the quality of groundwater, samples were collected and analyzed for 14 physico-chemical parameters from 67 sites distributed across the western half of the city region during the year 2013. It was revealed that overall groundwater quality in the study area is found to be less than desirable. To find out the possibility of heavy metals contamination also, groundwater quality data obtained from Karnataka State Pollution Control Board (KSPCB) on 20 parameters (physical, chemical and heavy metals) from 41 sampling stations (monthly data) was collected for the year 2015 and was used in the final analysis of this study for peenya industrial region. From the basic statistical analysis, it was observed that the average concentration of five groundwater quality parameters (turbidity, total hardness, iron, manganese chromium) considered for the study were exceeding permissible limit ,especially chromium which is known to be human carcinogen. Multivariate statistical techniques such as Cluster analysis (CA) was useful in classifying the 41 sampling sites into 3 main clusters as high pollution and low pollution areas. Discriminant analysis (DA) revealed that T-Hard, NO 3 , Ca, Mg, HCO 3 and TDS were the most significant parameters causing the temporal variations in groundwater quality and accounted for 94% assignation of seasonal cases. Fe, Cr, Cl, Mn, Cu and Cd were the most important parameters discriminating between the 3 clusters and accounted for 92% spatial assignation of cases thereby, delineating a few indicator parameters responsible for large variations in the groundwater quality. Principal component analysis (PCA) through varimax rotation achieved a simpler and more meaningful representation of the underlying factors by identifying 7 factors/sources with eigen value greater than one explaining 73.42% of the total variance. Receptor oriented source apportionment modeling using Absolute Principal Component Score Multi-Linear Regression (APCS-MLR), Unmix and Positive Matrix factorization provided apportionment of various sources responsible for the groundwater quality characteristics in the study area. The percentage contribution of the identified sources was determined. Results indicated that most variables were primarily affected by rock water interactions, seepage of sewage, geology of the area and industrial discharges especially different types of electroplating industries. It was also noticed that few parameters gained significant contribution from unidentified sources. Model performance was evaluated based on the ratio of estimated mean to measured mean (E/M). Results revealed that all three models provided good results regarding their ability to reproduce measured concentrations in most of the cases, with the APCS-MLR model showing better performance. This study concludes that these apportionment results provide useful help for policy and decision makers to enhance their ability to put in place effective policy and regulatory measures to reduce groundwater pollution.