Analytica Chimica Acta 532 (2005) 15–25 Chemometric data analysis of pollutants in wastewater—a case study Kunwar P. Singh a,∗ , Amrita Malik a , Dinesh Mohan a , Sarita Sinha b , Vinod K. Singh a a Environmental Chemistry Section, Industrial Toxicology Research Centre, P.O. Box 80, MG Marg, Lucknow 226 001, India b National Botanical Research Institute, Rana Pratap Marg, Lucknow 226 001, India Received 3 April 2004; received in revised form 18 October 2004; accepted 18 October 2004 Available online 24 December 2004 Abstract In this study, chemometric techniques such as cluster analysis (CA), discriminant analysis (DA), principal component analysis (PCA) and partial least squares (PLS) were used to analyse the wastewater dataset to identify the factors which affect the composition of sewage of domestic origin, spatial and temporal variations, similarity/dissimilarity among the wastewater characteristics of cis- and trans-drains and discriminating variables. Samples collected from 24 wastewater drains in Lucknow city and from three sites on Gomti river in the month of January/February, May, August and November during the period of 5 years (1994–1999) were characterized for 32 parameters. The multivariate techniques successfully described the similarities/dissimilarities among the sewage drains on the basis of their wastewater characteristics and sources signifying the effect of routine domestic/commercial activities in respective drainage areas. Spatial and seasonal variations in wastewater composition were also determined successfully. CA generated six groups of drains on the basis of similar wastewater characteristic. PCA provided information on seasonal influence and compositional differences in sewage generated by domestic and industrial waste dominated drains and showed that drains influenced by mixed industrial effluents have high organic pollution load. DA rendered six variables (TDS, alkalinity, F, TKN, Cd and Cr) discriminating between cis- and trans-drains. PLS–DA showed dominance of Cd, Cr, NO 3 , PO 4 and F in cis-drains wastewater. The results suggest that biological-process based STPs could treat wastewater both from the cis- as well as trans-drains, however, prior removal of toxic metals will be required from the cis-drains sewage. Further, seasonal variations in wastewater composition and pollution load could be the guiding factor for determining the STPs design parameters. The information generated would be useful in selection of process type and in designing of the proposed sewage treatment plants (STPs) for safe disposal of wastewater. © 2004 Elsevier B.V. All rights reserved. Keywords: Wastewater; Sewage treatment; Multivariate analysis; Cluster analysis; Principal component; Discriminant analysis; Partial least squares 1. Introduction Safe disposal of domestic and industrial wastewater is an essential requirement under the Environmental Protec- tion Act, which establishes maximum permissible levels of various pollutants for discharge onto land and into the sur- face waters [1]. This emphasizes for proper treatment of sewage/wastewater prior to its final disposal. For selection of appropriate treatment process and design parameters to main- tain stable conditions for achieving optimum performance of the sewage treatment plants (STPs) round the year, a detailed ∗ Corresponding author. Tel.: +91 522 2508916; fax: +91 522 2228227. E-mail address: kpsingh 52@yahoo.com (K.P. Singh). information about the sources, composition and levels of pol- lutants in wastewater is needed. Lucknow, the State capital of Uttar Pradesh (India), is a city of over three and half million residents. The city spread over an area of about 400 km 2 is situated on the banks of Gomti river, a major tributary of the Ganga river. Out of the total about 550 million liter per day (mld) of the domestic and industrial sewage generated by the town, only a small fraction (about 40 mld) is currently treated by a recently established sewage treatment plant (STP) and major fraction of the rest of untreated sewage (510 mld) is discharged to the Gomti river through some 26 wastewater drains within a stretch of about 12 km. Among these, 14 drains collect sewage from the cis- side, while, another 12 from the trans-side of the river. The 0003-2670/$ – see front matter © 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.aca.2004.10.043