Evaluation of Monitoring Sites for Protection of Groundwater in an Urban Area Aabha P. Sargaonkar 1 *, Apurba Gupta 1 , Sukumar Devotta 2 ABSTRACT: Monitoring for seasonal variation and changes in ground- water is a costly project. Assessing groundwater at selected monitoring sites and for site-specific indicators may reduce the cost of subsequent monitoring. In this context, the present study developed a method to assess groundwater using a combination of multivariate and univariate statistical techniques to identify critical sites of contamination. The sample data used describes the groundwater quality in Allahabad, India. The factor analysis brings out the observable parameters for groundwater pollution. Finally, univariate techniques such as analysis of variance (ANOVA) and Bonferroni t-test identify the critical sites of groundwater pollution. The first factor indicated high loading (.0.6) of total dissolved solids, Cl, Na, Mg, conductivity, SO 4 , and hardness. This represented overall pollution status of groundwater from human habitation, waste disposal, and agricultural activities in Allahabad. Iron, Mn, and Zn showed loading on distinct factors and indicated local contamination. Univariate techniques ANOVA and Bonferroni t-test for Zn concentration in handpump samples revealed heavy metal contamination at Hasimpur and Beniganj in India. Thus, initial monitoring followed by statistical analysis can help identify critical sampling locations and important indicators. Water Environ. Res., 80, 2157 (2008). KEYWORDS: groundwater quality, multivariate analysis, factor analysis, univariate analysis, ANOVA, t-statistics. doi:10.2175/106143008X304695 Introduction The chemical composition of groundwater, a natural resource for rural and urban areas, dictates its suitability for various uses. It can also be an indicator of overall ecosystem health. In general, because groundwater occurs in sand, gravel, and rock formations beneath earth’s surface, it is relatively clean unless contaminated by seepage of polluted surface water or by in-situ contaminants. Various pollutants from power stations, waste disposal sites, and agricultural lands potentially can contaminate groundwater. Discharge of industrial effluents and municipal waste in surface waterbodies are another significant concern in many cities and industrial clusters in India. Therefore, central and state governments regularly monitor and assess the quantity and quality of groundwater. A 1995 survey by the Central Pollution Control Board (CPCB) identified 22 sites in 16 states in India as critical for groundwater pollution, primarily from industrial effluents (Kumar and Shah, 2006). Although many cities are well aware of the potential threats to groundwater posed by urbanization, they often do not have groundwater monitoring programs in place nor a vision for its protection and management. Because the quality of groundwater depends on several factors— including climate, soil characteristics, manner of circulation through rock types, topography, saline water intrusion in coastal areas, and human activities—assessment of groundwater quality is complex. Studies are developing measures such as a pollution potential index or vulnerability index to assess groundwater. However, these approaches are good for overall assessment on a watershed basis and have little practical value on a local scale (Aller et al., 1986; Baalousha, 2006). The conventional way to interpret groundwater quality data is to plot concentrations of different ions or pairs of ions to understand the similarity between samples or variables. However, the complexity of the groundwater system makes the interpretation of results difficult. For urban groundwater quality management, many agree that studies should focus on selected monitoring locations and site-specific geoindicators, or important contaminants. A few important, first-order and second-order parameters can be used in most circumstances to assess significant trends in urban pollution. First-order pollutants include Cl, HCO 3 , and dissolved organic carbon (DOC, as measure of products of biological pro- cesses). Second-order pollutants comprise B, a fatty acid salt that is completely biodegradable (where detergents are used), hydro- carbons, and organic solvents (Geoindicators, 2006). Change detec- tion in groundwater quality is an important geoindicator that can help reduce the number of sampling locations and, consequently, monitoring cost. It is, therefore, necessary to develop tools for decision-making regarding change-detection in groundwater qual- ity. Evaluation of monitoring needs and monitoring strategies— such as the number and location of monitoring sites, frequency of monitoring, and variables to be monitored—can help develop guidelines for successful management of an aquifer. In this context, multivariate and univariate techniques form useful tools. Multivariate techniques such as factor analysis, principal component analysis (PCA), and cluster analysis can help understand overall groundwater chemistry (Reghunath et al., 2002; Liu et al., 2003; Vega et al., 1998). Univariate techniques mostly include trend-detection studies and focus on analyzing the data using a single variable [U.S. Environmental Protection Agency (U.S. EPA), 1989; Mac Berthouex and Brown, 2002]. This paper describes an approach wherein application of multi- variate techniques is followed by univariate tests to help identify 1 Environmental Systems Design and Modelling Division, National Environmental Engineering Research Institute, Nehru Marg, Nagpur, India. 2 Ex Director, National Environmental Engineering Research Institute, Nehru Marg, Nagpur, India. 1 * Environmental Systems Design and Modelling Division, National Environmental Engineering Research Institute, Nehru Marg, Nagpur, India 440 020; e-mail: ap_sargaonkar@neeri.res.in. November 2008 2157