Introduction

The rapid increase of population is a cause of concern because it enhances the pressure on all natural resources. Groundwater is one of such indispensable natural resource which has vast uses. Globally, it is being exploited for agriculture, industries, irrigation and drinking purpose. It is widely known that the human interferences such as industrial and domestic waste disposal and excessive use of fertilizers contribute in contaminating the water. Rahim et al. (2010) discussed the impact of anthropogenic effects such as solid waste disposal and its effects on the shallow groundwater in west Malaysia. Elevated concentration of different elements as well as considerable ionic balance error was found in the groundwater. Hosono et al. (2009) studied the human impacts on groundwater flow as well as the contamination in South Korea. Multiple isotopes were used in the study to demonstrate the human-induced degradation of groundwater. Khalil et al. (2008) studied the heavy metal contamination and metal content in soil runoff and groundwater in the mining sites of South Morocco, while Takamatsu et al. (2010) worked on the similar aspect of heavy metal contamination in Kanto, Japan. Simeonov et al. (2003) and Kanchan and Ghosh (2012) discussed the impact of excessive withdrawal of groundwater and associated problems of lowering of water table, poorer quality of water and restricted availability of uncontaminated water. On the other hand, natural factors of leaching of minerals and chemical reactions of surface and subsurface rocks with water were studied by Oinam et al. (2011) in Bishnupur district of Manipur, India. Harvey et al. (2005) studied the health effects of groundwater arsenic contamination in the Ganges delta, while Mukherjee and Bhattacharya (2001) reported the high health risk of millions of people in Bangladesh due to groundwater arsenic contamination. According to the study by Kanchan and Ghosh (2011), presence of different elements, ions and minerals are valuable for human health, but at the same time if the presence of ion and minerals exceeds the permissible limit and is consumed for longer period of time then it may create serious health issues. The vital parameters of the groundwater are interdependent; hence, with the increase in the number of variables, their complexity also increases. Use of appropriate statistical techniques is one of the ways of reaching to probable generalization. Multivariate statistical analyses such as factor analysis (FA) and cluster analysis (CA) are some of the statistical tools, which are adopted to deal with large amount of data and number of parameters. According to the study of Simeonov et al. (2003), the multivariate statistical analyses helped in interpreting the underlying variable structure by deriving much simpler groups. Liu et al. (2003) performed a study in blackfoot diseased area of Taiwan and applied multivariate statistical analyses for the assessment of groundwater quality. Stüben et al. (2003) discussed about the geochemical characteristics of groundwater of Murshidabad district of West Bengal incorporating factor analysis. Oinam et al. (2011) studied spatio-temporal pattern of groundwater quality using cluster analysis in Bishnupur district of Manipur state, India, taking arsenic, iron and other parameters into considerations. Yammani et al. (2008) discussed the seasonal variability of different parameters and factors that controlled the groundwater quality in the hard terrain of Andhra Pradesh, India, using factor analysis. Giridharan et al. (2008) also worked on the similar line, and studied the seasonal pattern of geochemical parameters of groundwater quality in Chennai, India. Spatio-temporal pattern of Dianchi Lake basin water pollution was studied by Yang et al. (2010) using multivariate statistical analyses. Güler et al. (2002) examined the advantages and disadvantages of different types of clustering techniques along with the principal component analysis in parts of Southwestern USA. Yidana and Yidana (2010) have also applied factor analysis technique for the formulation of water quality indexing in Voltaian sedimentary aquifers of Ghana, while Singh et al. (2011) and Lake et al. (2003) attempted spatial distribution of groundwater quality using the statistical technique associated with the mapping technique for interpretation of groundwater quality. Techniques of quality index using statistical method, as well as conventional graphical scheme for the analysis of groundwater quality in the Volta region of Ghana was discussed by Yakubo et al. (2009).

Groundwater is one of the major sources of consumption for drinking as well as other purposes, in countries like India. Therefore, a detailed study is necessary to demarcate the zones of groundwater according to the level of contamination. The present study was carried in the central alluvial tract of Bengal plain, India, based on the analysis of ten geochemical groundwater parameters viz. arsenic (As), pH, total dissolved solids (TDS), electrical conductivity (EC), iron (Fe), chloride (Cl−), sulphate (SO −24 ), total hardness as calcium carbonate (will be used in the rest of the text as total hardness only), nitrite (NO2 −) and depth. In association with the understanding of geochemical characteristics, geology and hydrological setup, synthesis of multivariate statistical analyses such as factor and cluster analysis provided better opportunity to analyse these multifaceted relationship.

Materials and method

Study area

The study area extends between 23°43′30″ N to 24°50′20″ N latitude and 87°46′17″ E to 88°46′00″ E longitude with areal coverage of about 5,324 km2 in the central Bengal alluvial tract covering the entire Murshidabad district of West Bengal, India (District Census Hand Book 2001). Northeast and east of the study area shares the international boundary with Bangladesh, whereas, northern, western and southern boundaries are shared with Malda, Birbhum, Bardhaman and Nadia districts of West Bengal, respectively (Fig. 1). Northwestern region of the study area is demarcated by the state boundary between West Bengal and Jharkhand. Geomorphologically, the terrain is almost flat having elevation varying between 10 and 50 m above the mean sea level. (District Resource Map 2008)

Fig. 1
figure 1

Location map of the study area

Full size image

Hydrogeological setup

The study area falls in central alluvial tract of Bengal plain, India, and it comprises of unconsolidated sediments of Late Pleistocene to Late Holocene times. Stratigraphically, the area is mainly comprised of quaternary sediments belonging to Rampurhat, Kandi and Bhagirathi formations, whereas older basaltic rock occurs in the northwestern region covering a small patch of Rajmahal trap (District Resource Map 2008). The western part of the study area is dominated by the sandy and silty clay of Rampurhat formations and Kandi formations, which show the alternate layering of sand, silt and clayey sediments, extensively spread over the study region (Fig. 2). The Bhagirathi formation denoting the present day flood plain deposits and marked by the fine-grained sediments mainly silt and clay. Hydrological framework of the region is mainly controlled by river Ganga and its two distributaries, river Bhagirathi and river Bhairab. The river Ganga flows in northwest to southeast direction along the northwestern periphery of the Murshidabad district through Bhagirathi formation, while river Bhagirathi, which is draining in north–south direction, roughly midway of the region. River Bhairab also follows north–south direction in a highly meandered path on the east of river Bhagirathi. The river Ganga shows development of large bars and meandering in upper reach, whereas the lower reach is characterized by large meandering pattern with narrowed channel close to eastern part. As a part of Bengal plain, the study area comprises of three major aquifer systems. East of river Bhagirathi is associated with the thick unconfined aquifer, while the western tract is related to thick semi-confined aquifer (Fig. 3). Lithologically, eastern segment of river Bhagirathi is associated with recent alluvium, whereas older alluvium of Upper Tertiary period is found in the western segment of the region (Groundwater Information Booklet 2007). Only a small patch of unconfined aquifer consisting of basaltic rock of Upper Cretaceous period is situated in the northwestern part of the study area. Saturated groundwater zone is found as far down as 150 m in the eastern part of the region due to lack of any significant impermeable layer. Water table is generally found within 2–5 m below the surface. Groundwater potential is greater than 42 yield (L/s) in the eastern and southern part. (District Planning Map Series 2002; Deshmukh and Goswami 1973).

Fig. 2
figure 2

Geological setup of the study area (Source: District Resource Map 2008)

Full size image
Fig. 3
figure 3

Hydrological setup of the study area (Source: Groundwater Information Booklet 2007)

Full size image

Sample collection and geochemical analysis

78 groundwater samples were collected from the Bengal alluvial tract during the pre-monsoon period of May 2011. Sampling locations were marked using GPS of Garmin e-Trex Vista make. Later, the data from GPS were downloaded to computer, using “Mapsource” software. Depths of the hand pumps were known with the help of local authorities as well as Public Health Engineering Departments (PHED). Parameters of the water quality were chosen on the basis of their significance for attaining groundwater condition. pH value is a measure of acidity or alkalinity of water, and changes in level convey the geochemical reactions, which may change the characteristics of groundwater (Yammani et al. 2008). TDS, EC, iron, total hardness, chloride and sulphate are the natural constituents of water and needed to be monitored, as the primary portability in terms of drinking and for other purposes depended upon these parameters (Mathes and Rasmussen 2006). The parameters such as arsenic and nitrite are very hazardous, as a minute concentration above the permissible limit can create serious health issues (Oinam et al. 2011; Mishima et al. 2010).

Concentration of pH, total dissolved solids and electrical conductivity was examined on field using digital pH metre (Hanna, Model No. HI-9827), TDS tester (Hanna, Dist 1, Model No. HI 98300) and digital EC metre (Hanna, Dist 4, Model No. HI 98303). Before collection of samples, hand pumps were mechanically pumped continuously for 10 min to flush out the water in the upper part of the pipe of the hand pumps. Samples were collected in 500-ml PET (Polyethylene terephthalate) bottles, and pH level of water was maintained below 2 using HCl. The samples were stored at low temperature (4° C) until further chemical analyses were undertaken.

Concentration of arsenic, iron and nitrite was analysed through spectrophotometric techniques using Molybdenum Blue Complex method (Jeffery et al. 1989), 1/10 Phenanthroline method (APHA 1989) and Cadmium Reduction technique (APHA 1989), respectively, by Elico Double Beam UV–Vis Spectrophotometer (Model sl-210). Total Hardness was determined through EDTA Titration (APHA 1989), chloride by Argentometric method using silver nitrate (APHA 1989), and sulphate by Iodometric titration method (APHA 1989), respectively. The results from the analyses were tabulated in Table 1.

Table 1 Values of geochemical parameters
Full size table

Data treatment

All data were subjected to statistical analysis using statistical package SPSS 19. Shapiro–Wilk (W) test showed that data set had a positive skewness, which indicated non-normal distribution of data. To ensure normality, standardization had been done by ‘z-score’ (Liu et al. 2003). Standardized data set was further used for factor analysis (FA). Ward’s linkage method was applied for hierarchal cluster analysis (HCA) using squared euclidian distance.

Spatial interpolation and mapping

Arc GIS 10® software was used for mapping of groundwater parameters and different factors. All 78 GPS locations were plotted on the base map of the study area. Factor scores collected from the statistical analysis were tabulated with respect to their locations in the attribute table. Inverse Distance Weighting (IDW) method was applied for generation of isolines. Class interval of 1, both positive and negative, was used to identify the high as well as low potential contamination zones. Graded shades were applied to indicate the values of the factors wherein, lighter shades were used to indicate lower concentration and darker shades for higher concentration of parameters. These maps were superimposed and final composite picture was drawn. For each of the factors, individual maps were generated. Cluster numbers of each variable analysed from cluster analysis were tabulated according to their respective locations and applied for mapping of different clusters. For comparison with the factor distribution in GIS environment, individual parameters were also taken into considerations and interpolated maps were prepared using IDW technique.

Results and discussions

General geochemical properties and spatial distribution

The arsenic concentration in the groundwater in the study area varied between Below Detection Limit (BDL) and 0.98 mg/L (Table 1) with mean value 0.10 mg/L indicating a critical condition according to the permissible limit (0.05 mg/L), set by Bureau of Indian Standard (BIS). Although standard deviation was as low as 0.19, but the skewness and kurtosis both are highly positive (Table 2). The concentration of arsenic in the eastern segment of river Bhagirathi showed higher values exceeding the permissible limit of BIS. In the western segment of the region, concentration of arsenic was considerably low and in most of the places, it was even below permissible limit. Patches of higher concentration were also observed in the northwestern tip of the region (Fig. 4a).

Table 2 Summery statistics of the 78 groundwater samples
Full size table
Fig. 4
figure 4

Spatial distribution of parameters a arsenic, b pH, c TDS, d EC, e iron, f chloride, g sulphate, h TH as CaCO3, i nitrite

Full size image

The pH value of the groundwater varied between the minimum and maximum value of 6 and 8.5 with a mean value of 7.49 and low standard deviation of 0.57 indicating slightly alkaline condition. The skewness of pH was moderately negative (−0.47), and kurtosis was positive with a value (+0.33). The pH value of the groundwater in the north-central region was found between 7 and 7.5 except for one continuous patch from south central to east where pH value was between 7.5 and 8 (Fig. 4b).

The total dissolved solids were in the range from 253.40 to 900 mg/L with a mean of 514.95 mg/L, and standard deviation of 152.01 indicated a considerable variability in concentration of throughout the region. Both skewness (+0.53) and kurtosis (+0.41) were slightly positive. The higher concentration of total dissolved solids was observed along the river Bhagirathi (Fig. 4c). With the increasing distance from the river, the concentration of total dissolved solids decreased considerably and reached to its minimum level in the western end.

The electrical conductivity showed a similar pattern of total dissolved solids ranging from 0.39 × 10−3 to 1.41 × 10−3 S/cm−1 having a mean value of 0.81 × 10−3 S/cm−1 with standard deviation of 0.23 (Fig. 4d).

Figure 4e revealed the concentration of iron and varied between 0.40 and 38.0 mg/L with a standard deviation of 7.04 and mean value of 5.98 mg/L. Both values of skewness and kurtosis were high and positive, i.e. +3.25 and +10.92, indicating a wide variation in the concentration in the study area. The major concentration of iron was observed in the eastern and southern segment of the studied region. Concentration of iron was found below 5 mg/L in rest of the region.

The concentration of chloride has been shown in Fig. 4f. The variation was from 10 to 125 mg/L with standard deviation of 35.12 and mean value 55.51 mg/L indicating substantial variation of concentration throughout the region. Skewness showed a moderate value of +0.85, while the kurtosis indicated −0.49. The concentration of chloride was observed in a continuous patch from northwest to south-central portion of the study area. In the western segment of the region, the concentration of the chloride was relatively less. Similarly, the sulphate concentration widely varied between 20 and 1,100 mg/L with a mean value of 354.18 mg/L and standard deviation 321.17. The skewness was observed positive (+0.85) and kurtosis was having a value of −0.49. Sulphate concentration was extended spatially from the central portion to the western region of the area (Fig. 4g).

Total Hardness was determined by EDTA titration and the its value was observed in the range from 30 to 850 mg/L and mean value of 370.41 mg/L. Very high standard deviation of 255.61 and skewness of +0.12 and kurtosis −1.35 were observed. The hardness was widely spread in the western end and in the northwestern tip of the region with considerably higher value, while the rest of the area revealed the low concentration values of total hardness (Fig. 4h).

Nitrite concentration, which is responsible for the ‘Blue Babies syndrome’ (Mishima et al. 2010), was also analysed and found to be in the range from 6.25 to 71.42 mg/L with an average of 27.45 mg/L and standard deviation of 14.28 with skewness and kurtosis as +1.40 and +1.48, respectively. Higher concentration of nitrite in the central part of the study area gradually decreased in all directions (Fig. 4i).

Factor analysis (FA)

Factor analysis is a multivariate statistical analysis, which helps in detecting the similarity among the variables. The major aim of this analysis is to reduce the dimensionality of the data and reproduce set of related variables without losing any information (Farnham et al. 2002). This statistical approach had been adopted by Helena et al. (2000) to interpret the interrelated complex processes, which were controlling the general water chemistry. Factor analysis technique was used in the present study to extract the factors using “Kaiser Criterion” where eigenvalues greater than the unity (1) were taken into consideration (Davis 1986). Scree test was applied though decreasing order of the eigenvalues in respect to the factors. Break in the scree plot represented the number of factors to be considered. To ensure the maximum variability, “varimax rotation” was applied. With extracted factors, the interrelationships among the variables were analysed in much efficient way as the number of variables was grouped into lesser number.

In the present analysis, four significant factors were extracted using varimax rotation with eigenvalues more than 1 explaining 66.57 % of the total variability of the data (Table 3). Higher factor loading value for factor 1 showed that total dissolved solids (+0.955) and electrical conductivity (+0.957) accounted for 26.22 % of the total variation among the sample. Factor 2 had 16.80 % of variance with higher positive loadings on arsenic (+0.668) and iron (+0.685), while negative loadings on depth (−0.609) parameter. Third and fourth factors had been observed having 12.84 and 10.70 % variability, respectively. The third factor depicted positive loadings on chloride (+0.855) and nitrite (+0.739). The fourth factors showed higher positive loadings on pH (+0.609) and sulphate while negative loading on total hardness (−0.634). The last two factors depicted relatively lower percentage of variance, which indicated more local effects than the first two factors. Factor loadings of the four factors of the data set are listed in the Table 4.

Table 3 Factor analysis
Full size table
Table 4 Rotated component matrix
Full size table

Interdependence between the factors

Inter-factorial relationship can be better interpreted through the scatter plot of factors, which is essential for understanding the importance of each factor and also the interdependence between the factors. High positive loadings were observed in all the parameters of factor 1, i.e. total dissolved solids and electrical conductivity and factor 2 having iron and arsenic excluding depth as shown in Fig. 5a. Positive loadings of iron and arsenic showed a relationship with total dissolved solids and electrical conductivity. The negative loading on depth indicated an inverse relationship which controls the concentration of arsenic and iron. A high positive loading of iron in both unconfined and semi-confined aquifer is a better indication of dissolution of iron oxides under reducing conditions (Akai et al. 2004; Ravenscroft et al. 2001). Fig. 5b showed that the iron and arsenic are mostly concentrated in the shallow aquifer with higher total dissolved solids and electrical conductivity. Positive loadings on both factor 1 (total dissolved solids and electrical conductivity) and factor 3 (chloride and nitrite) showed strong control on both the factors (Fig. 5b). Control of arsenic, iron and depth on pH is indicated by high positive loadings and determined through interrelationship between factor 2 and factor 4 (Fig. 5c). Inverse relation was noticed between arsenic, iron (Factor 2) and sulphate, total hardness and pH (factor 4).

Fig. 5
figure 5

Interfactoral relationship. a factor 1 and factor 2, b factor 1 and factor 3, c factor 2 and factor 4

Full size image

Hierarchical cluster analysis (HCA)

Among several clustering techniques, HCA is widely applied in earth science studies (Davis 1986). It is one of the key techniques, which is widely used for grouping of hydrochemical data having similar characteristics and applied by several workers (Mencio and Mas-Pla 2008; Forina et al. 2002). The grouping of variables was done by the Euclidian distance method and linking up on the basis of overall similarity. The data set can be classified easily by HCA and direct way with representation through dendrogram (Davis 1986). In the present study, Ward’s linkage agglomeration schedule coefficient was applied for the identification of number of classes. Dendrogram represented the actual pattern of data (Fig. 6), and 78 samples were clustered into four groups (Table 5). As per Table 5, cluster 1 was distinguished from others on the basis of presence of higher electrical conductivity (0.75 S/cm−1 × 10−3), chloride (52.83 mg/L) and total hardness with 55 % of the total sampling locations. Figure 7 showed the cluster 2 in the central and southern part of the district with alkaline condition (7.61), moderate total dissolved solids (405.68 mg/L) and high sulphate (714.78 mg/L) concentration. Cluster 3, though observed in only 6 % locations, but is problematic because of very high concentration of arsenic (0.75 mg/L) and iron (10.96 mg/L) in shallow aquifers. The amount of sulphate (342 mg/L) was moderate and that of chloride (24.80 mg/L) was low in this cluster. Fourth cluster showed higher amount of total dissolved solids (636.12 mg/L), nitrite (49.59 mg/L) and electrical conductivity (1.01 S/cm−1 × 10−3) while the sulphate (615.28 mg/L) concentration was moderately present.

Fig. 6
figure 6

Cluster dendrogram

Full size image
Table 5 Mean of clusters of parameters
Full size table
Fig. 7
figure 7

Cluster distribution map

Full size image

Categorization of groundwater quality

Factor 1 score distribution map is associated with high positive loadings on total dissolved solids and electrical conductivity (Fig. 8a). Therefore, this factor can be assigned as the turbidity factor, which might be originated from the point sources such as agricultural wastes and runoff water with high load of solids that mixed with the groundwater through percolation. High scores of factor 1 are mostly associated with the thick unconfined aquifer situated in the eastern part of the study area. Influence of river Ganga on total dissolved solids and EC concentration is clearly confirmed in Fig. 8a by the continuous stretch throughout the northern peripheral region. The factor scores were substantially higher along a narrow trail extended from north to south and followed the path of river Bhagirathi. Groundwater mixing with the river water might be one of the controlling factors in this respect. The increased concentration of solids was observed at the confluence of river Bhagirathi with river Ganga. The scores were decreasing gradually towards western part where the tributaries of river Ganga were scanty. The hand pump no. 1, 12, 38 and 47, which are located along the river, had higher total dissolved solids as well as EC.

Fig. 8
figure 8

Factor score distribution maps. a factor 1 (TDS and EC), b factor 2 (depth, arsenic and iron), c factor 3 (chloride and nitrite), d factor 4 (sulphate, pH and TH as CaCO3), e composite factor

Full size image

Figure 8b represents factor 2, which is associated with variables of depth, arsenic and iron. The higher positive scores of arsenic and iron and negative loading on depth were observed in the eastern end and southern segment. The presence of higher concentration of arsenic and iron in the shallower depth depicted a typical redox condition largely because of reduction dissolution (Chapagain et al. 2010). In this region, the factor score showed more than 1 which indicated an elevated concentration of the parameters. Over pumping of groundwater might be the governing factor as it is the major source of drinking water as well as irrigation in that region. Over extraction introduces excess amount of oxygen, which helps in greater amount of oxidation of the minerals and release of arsenic in the groundwater (Liu et al. 2003). The eastern side of the river Bhagirathi is composed of thick unconfined aquifer. There is no significant intervening layer between the surface and subsurface, and this leads to percolate the water directly into the aquifers through these surface and subsurface layer. This might be the governing factor of arsenic release into the groundwater (Ghosh and Kanchan 2011). The hand pumps no. 25, 26 and 27 are some of the locations located in the eastern part of the region, with very high concentration of arsenic in the shallower depth with considerably high concentration of iron. On the other hand, location no. 40, 41 and 43 showed considerably low arsenic concentration below permissible limit in the higher depth with low iron concentration. The result indicated a definite relationship between the parameters.

Whole of the northern region showed dominance of factor 3. Higher positive loading on both chloride and nitrite is observed in the north-central part of the study area (Fig. 8c) with factor score more than 1. The origin of chloride from local natural sources has moderate concentration, while, nitrite is one of the important components of fertilizers. Excessive use of fertilizers and untreated sewage contamination might have caused the concentration of nitrogen bearing elements in the groundwater (Chanakya and Sharathandra 2008; Schmoll et al. 2006). Nitrite, having a positive relationship with the wastewater discharge and pesticides, excessively used in agriculture, contributed to anthropogenic local pollution (Kumazawa 2002; Gupta et al. 2008). Higher concentration of chloride and nitrite had been observed from hand pump no. 14, 15 and 16. On the other hand, the hand pump nos. 1, 19 and 20, which are considerably away from the river, had shown a low amount of chloride and nitrite. This is an indication of a probable relation with the river water.

Fourth factor is associated with high positive loadings on pH, sulphate and high negative loadings on total hardness (Fig. 8d). Variation in pH of water was considered to be a critical parameter as it triggers hydrochemical reactions in groundwater. Throughout the region, pH value varied between 6 and 8.5. Total hardness showed a range from 30 to 850 mg/L, which was considerably high. According to Sawyer and McCarthy (1967), this might be due to the weathering of sedimentary rocks, calcium bearing elements and use of excessive lime in agricultural land. Hand pump nos. 1, 14, 15 and 16 are some of the locations, which also follow the above said conditions. Moreover, factor 3 and factor 4 are showed the influence of agricultural activities.

Composite results after superimposition of the analysed factors, groundwater zones were categorized. Major contaminated zones were identified in the north central and in the eastern part of the study area where factor score showed very high value of more than 2. However, the higher factor scores were extended from central part to the eastern end of the region (Fig. 8e). The contact zone between thick unconfined and semi-confined aquifer has an important role to play in this regard. As per the results, the entire thick unconfined aquifer situated in the eastern part of river Bhagirathi is a matter of concern in terms of groundwater contamination. Hand pump nos. 35, 34, 13, 14 and 15 showed the higher scores, which indicated the highest contaminated zones. Entire region of the eastern part had positive composite factor score seemed to be the most problematic zone because of the highest composite factor of more than 2 in the eastern end region. Gradual decrease in factor scores was observed from east to western part. Here the factor scores are negative indicating the less contaminated groundwater regions.

Conclusion

In the present study, geochemical analysis of different parameters coupled with statistical analysis ascribed the contamination zones of groundwater in the central alluvial tract of Bengal plain of India. Result showed that the rivers are playing an important role in controlling the geochemical properties of groundwater of the region. In areas closer to the recharge zones of river Ganga, its tributaries and distributaries, higher concentration of total dissolved solids and electrical conductivity was observed. Higher concentration of arsenic and iron in shallower aquifer was found in the eastern end, while it decreased considerably in the western segment. Southern part showed higher values of pH in the water samples, while the groundwater in the central part extended up to the southern end of the studied area that had higher concentration of Chloride. The concentration of sulphate was spread from the central to the western part. However, hardness of groundwater was noted in the entire region of the central alluvial tract except eastern part and nitrite concentration in the north-central portion of the region.

The factor scores ascribed the abundance of higher scores from eastern segment of river Bhagirathi to the western segment in a gradually decreasing trend. Turbidity factor related to TDS and EC showed higher concentration along the path of river Bhagirathi. It was observed that the eastern most segment of the studied area had hazardous elements viz., arsenic and iron in the shallower depth and agricultural factor i.e. nitrite and chloride was concentrated in the north-central portion. Hardness factor comprising of total hardness, pH and sulphate was spread over almost whole eastern segment of river Bhagirathi. Cluster analysis also confirmed the presence of arsenic and iron in the eastern most part in the shallower depth and higher concentration of TDS and total hardness at the confluence of river Bhagirathi and river Ganga in the north-central part. The present results concluded the need of the groundwater quality categorization to ensure availability of groundwater that can be used for drinking and other purposes. Further, characterization of groundwater quality in different seasons may depict the seasonal pattern of groundwater processes. For the identification of subsurface hydrological characteristics, subsurface lithological modelling accompanied by the geochemical data and statistical analysis may be useful.