Factor analysis and linear regression model (LRM) of metal speciation and physico-chemical characters of groundwater samples

Abstract

An approach is described for viewing the interrelationship between different variables and also tracing the sources of pollution of groundwater of north Chennai (India). The data set of 43 variables which include major ions, minor ions and trace metal speciation (Cu, Pb, Cd and Zn) collected during the pre-monsoon and post-monsoon seasons of the year 2000-2001, was subjected to R-mode factor analysis to comprehend the distribution pattern of the said variables. It was found that first factor measures salinity and hardness which explained 19.12% of the total variance (comprised of variables EC, TDS, Na+, K+, Ca2+, Mg2+, total hardness, Cl- and SO4(2-)) during pre-monsoon, while it was 25.08% during post-monsoon. The second and third factors were attributed to speciation of zinc and copper ions during both pre-monsoon and post-monsoon. Although there were two more factors, loaded with speciation parameters of lead and cadmium, the variance of them were less than 10%. From this study it is seen that sea water intrusion, municipal solid waste disposal are the identified sources of component of pollution. The importance of metal ions is taking a secondary role and the anthropogenic origin-industrial activity, is the reason in the evaluation of pollution status as they come in the second, third, fourth and fifth factors. As the trace metal speciation was grouped in separate factors, linear regression model (LRM) with correlation analysis was applied to check its validity for prediction of speciation and to apply LRM for rapid monitoring of water pollution.