Introduction

Developing countries are facing immense problems of both quality and quantity of water due to unplanned urbanization and industrialization along with overexploitation and unethical/unscientific disposal of treated or untreated effluent (Simeonov et al. 2003; Brindha et al. 2014). Presently in India, 85 % of the water requirement for domestic use in rural areas, 55 % for irrigation and over 50 % for industrial and urban uses is met from groundwater (Herojeet et al. 2015). The increased dependency on groundwater has made water conservation gain top priority in water management studies. Groundwater resources in the alluvial regions are relatively more prone to contamination due to higher population densities, intensive agriculture and industrial activities in these areas (EPA 1993). Alluvial aquifers constitute a hydrological unit formed by the alluvial deposits and are characterized by a linear and shallow feature which make them highly vulnerable to contamination (Helena et al. 2000; Singh et al. 2006, 2014).

Water quality monitoring data consist of routine measurements of physiochemical variables that could help in understanding the structure and function of particular water body (Sargaonkar and Deshpande 2003; Singh et al. 2008; Murhekar 2012). For easy understanding and effective interpretation, water quality index (WQI) technique is used for demarcating groundwater quality and its suitability for drinking purposes (Tiwari and Mishra 1985; Singh 1992; Rao 1997; Mishra and Patel 2001). Various researchers studied groundwater monitoring with respect to water quality index (Singh et al. 2014; Ramesh and Seetha 2013; Rupal et al. 2012; Kushtagi and Srinivas 2012; Shivasharanappa et al. 2011; Reza and Singh 2010; Ramakrishnaiah et al. 2009). Groundwater chemistry is controlled by various complex factors comprising both natural and anthropogenic influences leading to complex interpretation of data (Sum and Gui 2015). The limitation of WQI can be overcome with the application of multivariate statistical approach (Prasanna et al. 2012; Singh et al. 2014). Multivariate analysis has been considered as a more trustworthy approach for environmental data reduction and revealing common and hidden factors that exist among a set observed variable (Singovszka and Balintova 2012; Okiongbo and Douglas 2015) Recently, several studies reported applications of various multivariate statistical techniques to analyze and interpret water quality data (Sghaier et al. 2011; Lu et al. 2012; Papazova and Simeonova 2012, 2013; Krishnaraj et al. 2013; Bhat et al. 2014; Venkatramanan et al. 2015).

Nalagarh valley belongs to the rapid industrial belt of Baddi, Barotiwala and Nalagarh (BBN) region. The valley has been rated as fastest in industrial growth in the last decade owing to the special packages of incentives granted by the central government (Herojeet et al. 2013b). As per Baddi Barotiwala Nalagarh Development Authority report 2007, around 72 % of industries in Nalagarh are processing without effluent treatment plants (ETPs) which is a major threat to water resources. With the surface water being no longer able to satiate the needs of the valley, groundwater becomes the major freshwater source for various purposes. Thus, groundwater depletion and contamination are major issues in the study area. The present study was carried out to characterize the probable factors controlling the hydrochemistry and discuss the suitability of groundwater for human consumption using environmetric techniques and water quality index.

Study area

The study area represents the southernmost expanse of Solan district which lies between Northern latitudes of 30°52′–31°04′ and Eastern longitudes of 76°40′–76°55′ with an aerial spread of 250 km2. The valley is having common border with Haryana towards south–east, i.e. Kalka-Pinjor area and with Punjab towards south–west, i.e. Ropar district (Fig. 1). Rainfall is the major source of recharge to groundwater. Other sources of the groundwater recharge body in the valley are chiefly affected through the influent stream seepage and percolation of surface precipitation and irrigation waters. The valley lacks proper drainage system along with waste disposal facilities where sewage, both treated or untreated, effluent is discharged into drainage system and other aquatic environments. Hence, surface water pollution is a major threat to groundwater regime, as 50 % of the area is shallow aquifer zone (CGWB 1998).

Fig. 1
figure 1

Location map of the study area with sampling locations

Full size image

Geology and hydrogeology

The geology of the area is complex not from the stratigraphical point of view but for its tectonic complexities. Stratigraphically, the Nalagarh valley and its flanks are bounded by the tertiary formations and structurally they are highly disturbed (CGWB 1975). The rock types of the area can be broadly grouped into two tectonic zones striking and trending NW–SE direction. So, the direction of their tectonic zones position from North to South is as follows (Fig. 2):

Fig. 2
figure 2

Geological map showing thrust and fault of Nalagarh valley

Full size image
  1. (a)

    Belt of lower and middle tertiary occurring along the NE flank of the valley (para-autochthonous).

  2. (b)

    Belt of upper tertiary confined to the valley and along its SW flank (autochthonous). The contact of these zones is marked by a major fault (Nalagarh thrust).

Tectonically, the area is highly disturbed; the two major thrust trending NE–SW are Nalagarh and Sirsa thrusts. Nalagarh thrust is formed between Kasauli and middle Siwaliks whereas Sirsa thrust separates upper and middle Siwaliks. The major part of the Sirsa river basin is covered by alluvium soil with Holocene and Pre-Holocene deposits. The alluvium soil varies from 10 to 20 m thickness and is mostly of granular form. The upper and middle parts of the river basin are predominated by alternate beds of clay, cobbles, pebbles, gravel and sand. The sediments get finer and finer till it become clay in the downstream part of basin. The stratigraphical sequence of the basin is given in Table 1. The ground water occurs in porous unconsolidated alluvial formation (valley fills) comprising sand, silt, gravel, cobbles and pebbles. The thickness of such deposits is again restricted to 60–100 m below ground level (mbgl). Wells and tube wells are the main ground water abstraction structures. The depth of open dug wells and dug cum bored well in area ranges from 4.00 to 60.00 mbgl wherein depth to water level varies from near ground surface to more than 35 mbgl. Deeper semi-confined aquifers are being developed by tube wells ranging in depth from 65 to 120 mbgl tapping 25–35 m granular zones (CGWB 2007).

Table 1 Geological sequence of the Nalagarh valley
Full size table

Drainage

The study area lies within the catchment expense of Sirsa River drainage system. Sirsa river is a perennial river which flows south westerly in the area and joins Sutlej 10 km upstream of Ropar. It originates from the SW shoulder of Kasauli Dhar and flows across the valley up to its SW flanks. The master slope of the valley is towards NW direction up to the confluence of Sirsa river along Chikni nadi (small river). The valley is drained by numerous perennial and ephemeral streams emerging from the NE flank passing through industrial belt often loaded with industrial and sewage discharges and transverse flow across the valley, to join Sirsa river (Fig. 3). The important streams among them are Chikni nadi, Phula nadi, Ratta nadi, Balad nadi and Surajpur nadi. All the tributary nadis (small rivers) of Sirsa river flow in NE direction to SW direction and are almost parallel to each other. These streams are fed by numerous secondary streams which form a dendritic type of drainage pattern. The discharge in the streams fluctuates in accordance with the climatic conditions. During monsoon, the streams are flooded and carry enormous load of sediments and deposit them in the flood plains of the valley.

Fig. 3
figure 3

Drainage map of the study area

Full size image

Materials and methods

Groundwater samples were collected from 12 Tubewells, 7 Dugwells, 7 Handpumps, 4 Springs and 2 wells on the basis of different industrial unit and land use pattern during May 2012 and October 2012. Good quality, air tight plastic bottles with cover lock was used for sample collection to avoid unpredictable changes in characteristics as per standard procedures. Physical parameters like EC, pH, TDS were measured on the spot at the time of sample collection using portable water and soil analyzes kit. For major cations analysis samples were filtered through were filtered using Whatman filter paper no. 42 of diameter 125 mm and pore size 2.5 μm and preserved by acidifying to pH ~ 2 with HNO3 and kept at a temperature of 4 °C until analysis. Chemical analysis of major cations (Na+, K+, Ca2+, Mg2+) and anions (SO4 2−, Cl−, HCO3 2−, CO3 2− and NO3 2−) for the examination of water and wastewater using standard method (APHA 2002) was carried out. The determinations of immediate parameters were made within 2 days after sampling. Ca2+, Mg2+, CO3 2− and HCO3 − were analyzed by titration. Na+ and K+ were measured by flame photometry and NO3 −, SO4 2− and PO4 2− by UV Spectrophotometer. HCO3 − and Ca2+were analyzed within 24 h of sampling. Mean value was calculated for each parameters to understand the seasonal variation and standard deviation was used as an indication of the precision of each parameter. Maps were prepared using Mapinfo 6.5 and Piper trillinear diagram were plotted using RockWorks 15. The statistical software Minitab 16, and Microsoft Excel were employed for the calculations and data interpretation.

Water quality index (WQI) computation

Water quality index is one of the most effective tools to communicate information on the quality of any water body. WQI is a mathematical equation used to transform large number of water quality data into a single number (Stambuk-Giljanovic 1999). Water quality index (WQI) is defined as a technique of rating that provides the composite influence of individual water quality parameter on the overall quality of water. It is calculated from the view point of human consumption. Water quality status and its suitability for drinking purpose can be assessed by determining its quality index. For computing WQI, Bureau of Indian standards (BIS 1991) has considered water standard of each chemical parameter in mg/l. The index developed by Tiwari and Mishra (1985) was used. In the present study ten water quality parameters, namely, pH, EC, TDS, TH, Ca2+, Mg2+, HCO3 −, Cl−, NO3 − and SO4 2− were considered for computing WQI.

The calculation involves the following steps:

The first step involved computing the relative weight (wi) of each parameter using Eq. (1). The unit weight (wi) for various water quality parameters is assumed to be inversely proportional to the recommended standards for the corresponding parameter.

$$\begin{aligned} {\text{wi }}({\text{unit weight}}) \propto 1/S_{n} \hfill \\ \Rightarrow {\text{wi}} = K/S_{\text{n}} \hfill \\ \end{aligned}$$
(1)

where K (constant) = \(\frac{1}{{{\raise0.7ex\hbox{$1$} \!\mathord{\left/ {\vphantom {1 {V_{{s_{1} }} }}}\right.\kern-0pt} \!\lower0.7ex\hbox{${V_{{s_{1} }} }$}} + {\raise0.7ex\hbox{$1$} \!\mathord{\left/ {\vphantom {1 {V_{{s_{2} }} }}}\right.\kern-0pt} \!\lower0.7ex\hbox{${V_{{s_{2} }} }$}} + {\raise0.7ex\hbox{$1$} \!\mathord{\left/ {\vphantom {1 {V_{{s_{3} }} }}}\right.\kern-0pt} \!\lower0.7ex\hbox{${V_{{s_{3} }} }$}} + \cdots + {\raise0.7ex\hbox{$1$} \!\mathord{\left/ {\vphantom {1 {V_{{s_{n} }} }}}\right.\kern-0pt} \!\lower0.7ex\hbox{${V_{{s_{n} }} }$}}}}\) and S n  = standard permissible value. The calculated unit weight of each parameters based on the Indian drinking water standards (IS: 10500, 1991) are also given in Table 2.

Table 2 Unit weightage of parameters based on the BIS (IS: 10500, 1991)
Full size table

In the second step, water quality rating (qi) was computed for each parameter using Eq. (2).

$${\text{qi}} = 100 \times \frac{{(V_{\text{a}} {-}V_{\text{i}} )}}{{(V_{\text{s}} {-}V_{\text{i}} )}}$$
(2)

where V a = actual value present in the water sample, V i = ideal value (0 for all parameters except pH which is 7.0), and V s = standard value.

Third, the summation of these sub-indices in the overall index. The water quality index (WQI) is then calculated as follows using Eq. (3).

$${\text{Water quality index (WQI}}) = \sum\limits_{i - 1}^{n} \left( {Q_{i} W_{i} } \right)/ \sum\limits_{i - 1}^{n} W_{i}$$
(3)

where Q i is the sub-index of ith parameter, W i is the unit weightage for ith parameter and n is the number of parameters considered.

According to this, water quality index is classified into five different water quality status with maximum permissible value is 100 and above this value is unfit for human consumption.

Multivariate statistical analysis

Multivariate statistical techniques have been applied by many researchers to characterize and evaluate freshwater, marine water and sediment quality (Lin et al. 2003; Reghunath et al. 2002; Shrestha and Kazama 2007). Environmetrics, also called multivariate statistical methodologies, such as principal component analysis/PCA and cluster analysis/CA were employed with the objective to group the sampling site according to water quality characteristics and identify the probable factors influencing the water chemistry.

Principal component analysis

Principal component analysis (PCA) is one of the best multivariate statistical techniques for extracting linear relationships among a set of variables (Simeonov et al. 2003). PCA is most widely used as an analytical technique whereby a complex dataset containing variables is transformed to a smaller set of new variables, which maximize the variance of the original dataset. PCA provides information on the significant parameters with minimum loss of original information (Singh et al. 2004). This is achieved by transforming to a new set of variables which are uncorrelated, and which are ordered so that the first few retain most of the variation present in all the original variables. Therefore standardization (z scale) was made on each chemical parameter prior to statistical analysis (Simeonov et al. 2004), to eliminate biasness by any parameter of different units with high concentration and renders the data dimensionless. The principal components are generated in a sequentially ordered manner with decreasing contributions to the variance, i.e. the first principal component (PC1) explains most of the variations present in the original data, and successive principal components account for decreasing proportions of the variance (Pires et al. 2009; Vieira et al. 2012). Principal components corresponding to absolute loading values “strong”, “moderate” and “weak”, corresponding to absolute loading values of >0.75, 0.75–0.50 and 0.50–0.30, respectively (Liu et al. 2003).

Cluster analysis

Cluster analysis (hierarchical clustering), on the other hand, is a useful method of objectively organizing a large dataset into groups on the basis of a given set of characteristics. The primary objective of CA is to identify relatively homogenous groups or clusters of sampling sites based on their similarities/dissimilarities (Wai et al. 2010). The grouping of similar objects occurs first and eventually, as the similarity decreases, all subgroups are merged into a single cluster. In this study, cluster variable (CA) was performed on the standardized dataset by means of the ward’s method using squared Euclidean distance as a measure of similarity to obtain dendrogram. The dendrogram provides a visual summary of the clustering processes, presenting a picture of the groups and their proximity, with a dramatic reduction in dimensionality of the original data. The seasonal variability of groundwater quality was determined from CA using the linkage distance, reported as D link/D max, which represents the quotient between the linkage distances for a particular case divided by the maximal linkage distance. The quotient is then multiplied by 100 as a way to standardize the linkage distance represented on the y-axis (Wunderlin et al. 2001; Simeonov et al. 2003; Singh et al. 2004, 2005).

Results and discussion

The physiochemical composition of groundwater samples are presented in Table 3 and also show the critical parameters exceeding the BIS (1991) desirable and permissible limits. The ionic dominance pattern is in the order of Mg2+>Ca2+>Na+>K+ among cations and HCO3 −>SO4 2−>NO3 −>PO4 2− among anions during pre-monsoon and post-monsoon, respectively. During the period of investigation, mean value of all parameters, except EC, TDS, Na+ and K+, respectively, were more pronounced in post-monsoon than pre-monsoon. The pH values in the study area were well within the prescribed limits for both seasons except for 1 sample above permissible limit in both the seasons. EC accounted for 68.75 and 50 % above desirable limit with mean values 908.59 and 807.59 µS/cm during the period of investigation. TDS values above desirable limit were more in pre-monsoon (68.75 %) as compare to post-monsoon (46.875 %) which may be attributed to over extraction of groundwater and declining of water table during pre-monsoon. TH 3.125 and 9.375 % samples were found exceeding permissible limits for pre-monsoon and post-monsoon, respectively. The higher degree of hardness in the study area can definitely be attributed to higher soluble concentration of Ca2+, Mg2+ and HCO3 − ions in aquifers system. The calcium concentrations above desirable limit were more prominent during post-monsoon (53.375 %) than pre-monsoon (40.625 %). Magnesium accounted for 3.125 and 15.625 % samples were found to be unsafe for both seasons and also majority of samples were above desirable limits which required attention and regular monitoring. Dissolved magnesium exceeds calcium in water once calcium precipitates after reaching super saturation and accounts for higher magnesium concentrations than calcium (Hem 1991). The higher concentrations of Mg2+ in the study area are due to combination of weathering of minerals magnesite, sandstone, dolomite and also from various industrial effluents, as these locations are very close to a cluster of several industries (Herojeet et al. 2013a). Mean value of HCO3 − was more significant during post-monsoon due to dissolution of carbonate rocks, weathering of feldspar by carbonic acids and oxidation of NO3 − and SO4 2− with organic matter as compared to pre-monsoon (Awadh and Ahmed 2013). The parameters like Na+, K+, Cl−, NO3 −, SO4 2− and PO4 2− were well within desirable limits. At least one or more parameters such as EC, TDS, TH, Ca2+ and Mg2+ accounted for water samples above desirable limits of more than 50 % of 32 samples examined which gives us caution for deterioration of water quality in near future and being unfit for drinking and other purposes.

Table 3 Seasonal variation in physiochemical of groundwater
Full size table

Hydrochemical diagram

Hydrochemical diagram aims at facilitating interpretation of evolutionary trends, particularly in conjunction with distribution maps and hydrochemical sections of water regime. The graphical representations of facies are useful in identifying chemical processes and detecting the effects of mixing water within the different lithological frameworks (Todd 1980). Various workers namely Collins (1923), Piper (1944, 1953), Black (1960), Walton (1970) and Chadha (1999) proposed the concept of graphical methods of representation of chemical analysis of water.

Piper trillinear diagram

Piper (1944) diagrams were employed for hydrochemical characterization of groundwater. On the basis of Piper diagram (Fig. 4; Table 4) of groundwater during pre-monsoon and post-monsoon, all groundwater samples belong to alkaline earth elements (Ca2++Mg2+) and exceed the alkali elements (Na++K+) where Mg2+ is leading cation in the study area. Also weak acids (CO3 2−+HCO3 −) exceed the strong acids (SO4 2− and Cl−) during both the seasons indicating HCO3 − is the principal anion in groundwater.

Fig. 4
figure 4

Piper classification diagram illustrating the chemical composition of groundwater

Full size image
Table 4 Groundwater samples characterization based on Piper diagram
Full size table

The cation triangle shows that Mg2+ (87.5 %) is dominating ion in both the seasons. However, 9.25 % samples are in Ca2+ type for pre-monsoon and the remaining 3.13 and 12.5 % samples fall in no dominant cation zone in both the seasons. The anion triangle exhibits that HCO3 − (93.75 and 90.63 % samples) is the dominant ion and (6.25 and 9.37 % samples) falls in no dominant zone during pre-monsoon and post-monsoon. This indicates that Na+, K+ and Cl−, SO4 2− ions have insignificant dominance in both seasons. Majority of groundwater samples (93.75  and 90.63 %) fall in the field of Ca2+–Mg2+–HCO3 − water types indicating temporary hardness; the groundwater of the study area is of Mg2+–HCO3 − facies having secondary salinity exceeds 50 % indicating inverse or reverse ion exchange, leaching process of dolomites, limestones and gypsum responsible for controlling groundwater chemistry (Davis and Dewiest 1966; Tay 2012). The remaining samples (6.25 and 9.37 %) are Ca2+–Mg2+–Cl−–SO4 2− water types which exhibit mixed Ca2+–Mg2+–Cl− facies where groundwater cannot be classified as either cation or anion dominant hydrochemical facies for both pre-monsoon and post-monsoon seasons (Todd and Mays 2005).

Water quality index

Water quality index computation shows variation in the quality status related to the suitability for human consumption during pre-monsoon and post-monsoon seasons is given in Table 5. A perusal of Table 5 reveals that 93.76 and 81.25 % samples in pre-monsoon and post-monsoon seasons, respectively, fall in poor to unfit class whereas the percentages of sample falling in good class in pre-monsoon and post-monsoon seasons were 6.25 and 18.75 %, respectively. The decline in poor to unfit class and rise in good class in post-monsoon season as compared to pre-monsoon season may be attributed to the dilution due to recharge of aquifer in post-monsoon season. The groundwater samples falling in poor to unfit class in both the seasons may be due to high water table, lesser soil filter media, valley fill aquifer formation and lack of proper drainage system. The calculated WQI for individual samples is represented in Table 6 and Fig. 5 shows the variation of WQI in the study area which indicates fluctuation in water class that require proper attention and regular monitoring.

Table 5 Water quality index classification
Full size table
Table 6 Calculation of WQI for individual water samples
Full size table
Fig. 5
figure 5

Variation of WQI in groundwater

Full size image

Environmentric model

Principal component analysis (PCA)

PCA was applied on the analyzed dataset to draw the latent factors influencing the hydrochemistry. An eigenvalue gives a measure of the significance of the factor and factors with eigenvalues >1.0 are considered significant (Shrestha and Kazama 2007). In this study, PCA was performed on the standardized dataset among the variables and the PC scores extracted by the scree plot method as shown in Fig. 6a, b and obtained four PCs which explained 80.1 and 75.7 % of the total variance for pre-monsoon and post-monsoon, respectively, through varimax rotation.

Fig. 6
figure 6

Scree plot extracting principal component

Full size image

Principal components (PCs) corresponding to absolute loading values of >0.75 (marked bold in the table) and additionally second level of interpretation (bold plus italics) are taken into consideration as statistically significant in the interpretation by PCA. The calculated component loadings, cumulative percentage, percentages of variance and communality explained by each factor are listed in Table 7.

Table 7 Varimax rotated component matrix of analysed groundwater samples
Full size table

During pre-monsoon, PC1 explained 28.0 % of the total variance was strong positive correlation with TH, Mg2+ and SO4 2− and moderately with EC, TDS and Na+ reflects lithogenic factor. This condition can be associated with silicate weathering in addition to dolomite, limestone and pyrite dissolution (Okiongbo and Douglas 2015). The high-positive loading of Mg2+ and SO4 2− with TH indicates permanent hardness. Sulphate ions are associated with higher concentration of Mg2+ and Na+ acts as a laxative and may cause gastric disorders (Montgomery 1985; Herojeet et al. 2015). PC2 (25.9 % of total variance) has strong positive loading with EC, TDS, Cl−, moderate positive loading NO3 − and inversely proportional with pH could be conditionally named as anthropogenic factors. The positive loading of Cl− and NO3 − may be due to industrial and domestic wastes and agricultural runoff (Dinkaa et al. 2015; Ghosh et al 2010). The applications of fertilizer such as anhydrous ammonium chloride have main elemental components of N and Cl. The negative loading of pH is associated with organic matter oxidation related to anthropogenic activities (Krishnaraj et al. 2011). PC3 was strongly negative weighted on Ca2+ and moderate positive score on Na+, K+ and HCO3 − accounts for 16.1 % of the total variance related to natural factor. For instance, a strong negative loading of Ca2+ reflects the cation exchange reactions leading to adsorption of Ca2+ on clay minerals and simultaneous releasing of Na+ and K+ ions. The strong loading of HCO3 − ions with alkali earth metals supports the view of natural weathering sources (Srivastava and Ramanathan 2008). Additional 10.1 % of the total variance explained PC4 having strong negatively weighted on PO4 2− and weak negative loading with NO3 −. This may be attributed to anthropogenic activities such as irrigated agricultural activities and domestic wastes.

Also, the PCA of data obtained for post-monsoon extracted four PCs which accounted 75.7 % of the total variance. PC 1 explained 34.3 % of the total variance with strong positive loading on Ca2+, Cl− and moderate positive loading with Na+ and SO4 2− is associated with various hydro geochemical processes resulting in high TDS (mineralized water). It indicates natural weathering of more soluble minerals like calcite, feldspar and chloride and various ion exchange processes in the groundwater system. PC 2 explained 17.1 % of the total variance shows strong negative loading of TH and Mg2+, and may be attributed to natural factor contributing from the weathering of magnesium bearing minerals. PC 3 (10.1 % of the total variance) was explained by moderate positive score with PO4 2−, HCO3 − and K+. PC 4 was responsible for 9.0 % of the total variance have strong positive loading with NO3 − and moderate loading with pH. Both PC3 and PC4 may be attributed to anthropogenic activities. It indicates the leaching of agricultural runoff (NPK fertilizer) and seepage of wastewater due to lack of proper sewerage system.

Cluster analysis (CA)

CA is used to detect the similarity group between 32 sampling points for pre-monsoon and post-monsoon. CA rendered a dendrogram in Fig. 7a, b, grouping all the sampling points into four cluster significance at (D link/D max) × 100 < 90 for pre-monsoon and (D link/D max) × 100 < 80 for post-monsoon. The clustering procedure generated four groups in a very convincing way, as the sampling locations at each groups have similar characteristic features.

Fig. 7
figure 7

Hierarchical dendogram of sampling locations of groundwater

Full size image

However, three clusters (C1, C2, C3) for pre-monsoon and four clusters (C1, C2, C3, C4) for post-monsoon were characterized by mixed features where sampling locations were included randomly in each cluster without any specific trends. The pattern formed by each clusters cannot be sought for explanation only considering certain special parameters. Therefore, the specific tracers for each pattern representing individual clusters can be clarified by calculating the average values of each chemical parameter belonging to individual clusters. The results are presented in Table 8.

Table 8 Average values for the water quality parameters for each cluster
Full size table

During pre-monsoon, Cluster 1 was characterized by highest levels of pH and higher tracer of Ca2+ and Cluster 2 has no specific tracers which depict natural water quality influenced by rock water interaction. Cluster 3 has highest tracers on Ca2+, NO3 −, PO4 2− and higher value of K+ which can be marked by anthropogenic activities contributing from agricultural and domestic sources. Cluster 4 was characterized by single sampling location having highest level of EC, TDS, TH, Mg2+, Na+, K+, HCO3 − and Cl− may be concluded mixed factor (natural and anthropogenic sources) influencing the hydrochemistry.

The dataset obtained in post-monsoon also statistically displayed four clusters. Cluster 1 was detected as the biggest group of sampling sites where the average values for all the water quality parameters are less than the other remaining clusters except pH reflects natural quality water. Cluster 2 was characterized by highest tracers of K+, HCO3 − and PO4 2−. Cluster 3 confirmed the highest levels of Ca2+, Cl− and NO3 − and higher level of PO4 − indicates anthropogenic activities. The sampling location group in Cluster 2 and 3 was almost similar with Cluster 3 of pre-monsoon. Finally, Cluster 4 formed a pattern of highest level of EC, TDS, TH, Mg2+, Na+, Cl− and SO4 2− and lowest pH reflects natural and anthropogenic sources controlling the groundwater chemistry. High concentrations of Cl− and SO4 2− have negligible concentration of HCO3 − resulting low pH in groundwater (Hounslow 1995).

However, it was observed that some sampling locations for both seasons clustered under different group indicate different sources of pollution from point or nonpoint sources localized from the surrounding area or microsite effect of constituent bearing rock minerals. The findings obtained from PCA and CA indicate water chemistry were strongly controlled by natural factors like weathering of minerals, ion exchange and rocks water interaction and anthropogenic factors mainly agricultural, domestic and industrial wastes. Even though there are some differences between the CA and PCA results, a good agreement between the two statistical techniques is evident in all the datasets analyzed.

Conclusion

The present study highlighted the usefulness of multivariate statistical techniques, WQI and conventional graphical representations for interpreting or understanding the factors controlling water chemistry, geochemical and hydrochemical evolution of groundwater system. The overall water samples are suitable for drinking and domestic purposes, except for some selected sites where parameters like pH, TH and Mg2+ were above the permissible limit of BIS. The Piper plot classified 93.75 and 90.83 % of samples for both seasons under Ca2+–Mg2+–HCO3 − water types indicating temporary hardness and remaining 6.25 and 9.37 % under Ca2+–Mg2+–Cl−–SO4 2− mixed type of water. The plot also highlighted the influence of inverse or reverse ion exchange and dominance of alkaline earth metals over the alkalis and weak acidic anions over strong acidic anions in the study area. The WQI shows groundwater samples falling in poor to unfit class in both seasons due to high water table, valley fill aquifer formation, lesser soil filter media and lack of proper drainage system and there is also rise in good class in post-monsoon which may be attributed to recharge of aquifer during monsoon period. PCA and CA identify the groundwater chemistry was strongly controlled by natural factors namely rock water interaction, ion exchange and leaching of parent materials with minor anthropogenic factors like agricultural runoff and seepage of industrial and domestic wastes. An integrated institutional system for groundwater conservation and recharging measures will be affirmative solution for sustainable water resources management.