Environ Monit Assess (2015) 187:170

DOI 10.1007/s10661-015-4382-0

Assessment of ground and surface water quality

along the river Varuna, Varanasi, India
Pardeep Singh & R. K. Chaturvedi & Ankit Mishra & Lata Kumari & Rishikesh Singh &
D. B. Pal & Deen Dayal Giri & Nand Lal Singh & Dhanesh Tiwary & Pradeep Kumar Mishra

Received: 2 April 2014 / Accepted: 17 February 2015

# Springer International Publishing Switzerland 2015

Abstract Multivariate statistical techniques were data from eight to three parameters and captured about
employed for monitoring of ground-surface water inter- 78.5 % of the total variance by performing varimax
actions in rivers. The river Varuna is situated in the Indo- rotation over the principal components. The varifactors,
Gangetic plain and is a small tributary of river Ganga. as yielded from PCA, were treated by CA which
The study area was monitored at seven sampling sites grouped them convincingly into three groups having
for 3 years (2010–12), and eight physio-chemical pa- similar characteristics and source of contamination.
rameters were taken into account for this study. The data Moreover, the loading of variables on significant PCs
obtained were analysed by multivariate statistical tech- showed correlations between various ground water and
niques so as to reveal the underlying implicit informa- surface water (GW-SW) parameters. The correlation
tion regarding proposed interactions for the relevant coefficients calculated for various physiochemical pa-
area. The principal component analysis (PCA) and clus- rameters for ground and surface water established the
ter analysis (CA), and the results of correlations were correlations between them. Thus, this study presents the
also studied for all parameters monitored at every site. utility of multivariate statistical techniques for evalua-
Methods used in this study are essentially multivariate tion of the proposed interactions and effective future
statistical in nature and facilitate the interpretation of monitoring of potential sites.
data so as to extract meaningful information from the
datasets. The PCA technique was able to compress the
Keywords Multivariate analysis . Varuna River . Cluster
analysis . Principal component analysis
P. Singh : L. Kumari : D. Tiwary
Department of Chemistry, Indian Institute of Technology
(Banaras Hindu University), Varanasi 221005, India

R. K. Chaturvedi : R. Singh
Introduction
Institute of Environment and Sustainable Development,
Banaras Hindu University, Varanasi 221005, India Surface water hydrochemistry depends on several natu-
ral factors like changes in precipitation, erosion and
A. Mishra : D. B. Pal : D. D. Giri : P. K. Mishra (*)
weathering of crustal materials as well as anthropogenic
Department of Chemical Engineering and Technology, Indian
Institute of Technology, (Banaras Hindu University), factors such as industrial and agricultural activities and
Varanasi 221005, India increased water consumption due to rapid urbanization.
e-mail: [email protected] Furthermore, urbanization has severely affected the wa-
ter resources by increasing the pressure on urban hy-
N. L. Singh
Department of Chemistry, Banaras Hindu University, drology (Mishra 2011). These factors have degraded the
Varanasi 221005, India quality of surface water for drinking purposes and are a
170 Page 2 of 10 Environ Monit Assess (2015) 187:170

threat to the existing water bodies (Singh et al. 2005a, b, et al. 2005a, b, c). These techniques effectively com-
c; Mishra 2010). press the data and reveal various correlations between
The discharge of waste water from municipal, indus- different parameters analysed. Studies have been con-
trial, and agricultural areas is an issue of serious concern ducted in the past to study GW-SW interactions by
as it affects river’s ecology (Varol et al. 2013). Surface evaluating their hydrochemical properties through such
run off from agriculture and urban areas varies mostly methods (e.g., Woocay and Walton 2008; Mencio and
with seasonal changes in precipitation, ground water Mas-Pla 2008). The previous studies have focused more
flow, water interception, and abstraction. These factors on surface water quality (e.g., Sundaray et al. 2006;
strongly affect the concentration of pollutants in river Sojka et al. 2008; and Selle et al. 2013) and
water which in turn deteriorates river water quality hydrochemistry. For example: Singh et al. (2014) have
(Vega et al. 1998). used ensemble learning approach to evaluate influences
The quality of river water at a point is a function of on seasonal variations and anthropogenic activities on
several developmental pressures including the lithology alluvial groundwater hydrochemistry; Isa et al. (2012)
of the basin, atmospheric and anthropogenic inputs, and evaluated extent and severity of groundwater contami-
climatic conditions (Yu et al. 2010). These developmen- nation based on hydrochemistry mechanism of sandy
tal pressures and increasing human population have tropical coastal aquifers. Various other studies have also
made the river of the study area vulnerable to sewage focused on aquifer dynamics and hydrochemistry,
flow, solid waste dumping, etc. and consequently exert (Stigter et al. 2006; Soulsby et al. 2002). Multivariate
pressure on the percolation and infiltration processes statistical analysis techniques facilitate to draw an ap-
responsible for the groundwater recharge (Kumar et al. propriate profile of the stream hydrology when there is
2012). Hence, these pressures affect the ground water insufficient data available or when the end members are
conditions which make it imperative to analyse the imprecise (Mencio and Mas-Pla 2008).
hydrochemical properties of both the stream and the In the present study river Varuna, a minor tributary of
aquifers in order to comprehend the interactions existent river Ganges was monitored. This river receives huge
between them. quantities of untreated sewage, agriculture runoff with
The interaction between ground water and surface pesticides, fertilizers, etc. from catchment areas which
water (GW-SW) has been studied focussing on different lead to degradation of water quality. A study by Singh
perspectives. For example: Sophocleous (2002) studied and Dwivedi (2007) and Kumar et al. (2012) asserted
this interaction focusing on geomorphological, that river Varuna receives sewage from twenty-two mu-
hydrogeological, and climatic control; Valett and nicipal drainages located on both sides of its 15-km-long
Sheibley (2009) studied the geomorphic and biological passage through Varanasi city. So there is an urgent
structure and how they influence the direction of GW- necessity to identify the pollution sources and estimate
SW interaction; Kikuchi et al. (2012) performed spatial their quantitative contributions, in order to achieve an
telescopic study for improved characterization of GW- effective pollution control. Therefore, the objective of
SW interaction; Lu et al. (2011) analysed interactions the present study is to establish correlations between
between atmosphere, surface, and ground water using various sites on the basis of CA and identify the major
EARTH model in the unsaturated zone; Anibas et al. hydrochemical factors based on the PCA. These results
(2011) adopted a simple thermal mapping method for will further help in the investigation of relevant human
studying seasonal spatial patterns of GW-SW interac- pressures contributing to degradation of the water qual-
tions; Harveya et al. (2006) modeled decadal time scale ity and also aid in concluding about the GW-SW inter-
GW-SW interactions; and Wade et al. (2006) studied actions. Hence, periodic monitoring and assessment of
nitrate dynamics in these systems. water quality helps to develop management strategies to
Multivariate statistical analysis techniques such as control surface water pollution inspite of increasing
principal component analysis (PCA) and cluster analy- urbanization and anthropogenic pressure on them.
sis (CA) have been applied to characterize and evaluate Multivariate statistical analysis was employed to ex-
surface freshwater quality, because they are useful in a m i n e G W- S W i n t e r a c t i o n s a n d t h e p o t e n t
verifying the temporal and spatial variations caused by hydrochemical factors responsible for altering the qual-
natural and anthropogenic factors linked to seasonality ity of the river water in Varanasi. This work is structured
(Helena et al. 2000; Shrestha and Kazama 2007; Singh as follows: (1) sampling and analysis of physiochemical
Environ Monit Assess (2015) 187:170 Page 3 of 10 170

parameters from different sites selected throughout the sediments of Pleistocene to recent age. The basin of this
city; (2) PCA analysis on the collected data to identify river consists of three main geological formations:
the extent of ground water or surface water contribution newer alluvium and older alluvium aging from upper
to pollution in the relevant site; and (3) a CA to show the Pleistocene to recent and from middle to upper
correlations between sites. Pleistocene, respectively (Raju et al. 2009). The rocks
associated with the newer alluvium formations are com-
posed mainly of unconsolidated sand, silt, and clay
Materials and methods while that associated with older alluvium consist mainly
of consolidated clay with kankar, fine to medium grain,
Hydrology and geology of Varuna River and some gravel. In the study area, the unconsolidated
sediments form a sequence of clays and sands of various
Varuna River basin is situated in Indo-Gangetic plain grades. The presence of clay and sand zones controls the
(Fig. 1) which is underlain by Quaternary alluvial availability of ground water in alluvial zones. The sand

Fig. 1 Geographical locations

and photographs of water quality
monitoring sites
170 Page 4 of 10 Environ Monit Assess (2015) 187:170

751.00±0.33

845.00±0.32

764.00±0.67

1008.77±0.19

716.67±0.00

780.00±0.00
849.11±1.07
660.17±0.29

750.33±0.58

530.33±0.58

1122.11±3.95

319.78±1.71

392.22±1.17

620.89±0.84
beds with or without kankar in the study area forms the
main aquifer zones of the multi-tier aquifer system
(Shukla and Raju 2008). Near surface groundwater oc-

EC
curs under water table condition, while deeper aquifers
occur in semi-confined to confined conditions.
Extraction of ground water is through both dug wells

Table 2 Mean and standard deviation of water quality parameter studied both on ground (G) and surface (S) at different sites on Varuna River during winter season

324.26±1.48

317.33±0.00
323.89±0.51

350.78±0.19

296.67±0.33

296.22±3.08

293.67±0.58

361.45±0.39
305.45±0.69

239.56±0.19

356.61±0.67

371.78±0.63

280.52±0.63

350.26±0.36
and deep bore wells. The rainfall is the main source of

HARD
recharge to groundwater. The average hydraulic gradi-
ent is 0.35 m/km which indicates porous nature of near
surface formations of the area (Pandey 1993).

215.93±0.19

43.63±0.21
42.40±0.22
73.15±0.12
35.42±0.23
56.72±0.13
60.35±0.01
62.78±0.16

122.04±0.27
60.35±0.01

85.27±0.01
80.40±0.33
46.26±0.16
55.61±0.38
Sampling and qualitative analysis

Cl-
Seven sampling sites were randomly selected across the
width of the river for the assessment of water quality.

2.90±0.03

3.44±0.13

3.18±0.08

2.39±0.10

3.49±0.15
19.80±0.33
1.95±0.15

2.86±0.03
29.95±0.33

41.63±0.32

31.91±0.34

42.61±0.48

43.86±0.02

39.62±0.00
Each sampling site is in close proximity to a ground

PO43-4−3
water pumping station to make the analysis more cred-
ible. These samples were collected during winter (2010–
12) across a period of 3 months to monitor changes
caused by anthropogenic as well as natural sources.

21.00±0.50

10.21±0.12

5.41±0.12

18.11±0.42

5.33±0.09

8.30±0.09

3.21±0.02
9.85±0.51
10.00±0.19

6.78±0.38

4.10±0.37

13.82±0.08

10.78±0.19

6.04±0.00
Sampling, preserving, and transportation of the water
samples to the laboratory were done as per the standard
NO3−

methods (APHA 1998). Samples were qualitatively

analysed for eight different physiochemical parameters
and were tabulated in Table 1. Triplicate samplings were
440.12±0.84

310.33±0.67

350.22±0.38
330.60±0.51

910.45±0.69

320.56±0.51

250.44±0.51
443.89±0.51
417.45±0.83

527.22±0.51

400.22±0.51

603.33±0.00

387.34±0.00

450.11±0.19
done at each selected site at an interval of 1 month for
three consecutive months as mentioned above. The pH,
TDS

temperature, and electrical conductivity (EC) were de-

termined immediately at the collection site by a portable
pH meter, to minimize errors with time due to biological
7.66±0.13

7.73±0.06
7.40±0.12

7.42±0.08
7.68±0.16

7.29±0.05

7.94±0.05
7.80±0.06
7.89±0.12

7.73±0.03

7.63±0.06

8.01±0.05

7.91±0.07

7.73±0.07
and chemical reactions between the atmosphere and the
sample (Hutton 1983). Mean and standard deviation of
PH

water quality parameter studied both on ground (G) and

surface (S) at different sites on Varuna River during
winter season is given in Table 2.
19.78±0.35
19.70±0.03

20.17±0.44

19.50±0.17
20.00±0.38
19.45±0.25

20.22±0.38
19.40±0.00

20.08±0.10

20.14±0.18

20.47±0.29

20.91±0.14

21.41±0.13

20.86±0.05
TEMP

Table 1 The associated water quality parameters, with abbrevia-

tion and their units used in study

Parameters Abbreviations Units

(G) Mean±SD
(G) Mean±SD

(G) Mean±SD

(G) Mean±SD

(G) Mean±SD
(G) Mean±SD

(G) Mean±SD
(S) Mean±SD

(S) Mean±SD

(S) Mean±SD

(S) Mean±SD

(S) Mean±SD

(S) Mean±SD

(S) Mean±SD

pH pH
Temperature TEMP °C
Electric conductivity EC μS cm−1
Phosphate PO43−4−3 Mg l−1
Chloride Cl− Mg l−1
Nitrate NO3− Mg l−1
Chaukaghat

Old Bridge

Hukulganj
Dhobighat
Nakhighat

Kachahari

Total hardness HARD Mg l−1

Rajghat

Total dissolved solid TDS Mg l−1

Environ Monit Assess (2015) 187:170 Page 5 of 10 170

Statistical analysis to extract significant PCs and to further reduce the

contribution of variables with minor significance; these
PCA extracts meaningful information from raw data by PCs were subjected to varimax rotation (raw) generating
utilizing the existing correlations among various param- VFs. Consequently, a small number of factors will usu-
eters which are being monitored and hence successfully ally account for approximately the same amount of
aid the interpretation of meaningful information (e.g., information as does the much larger set of original
Chaturvedi and Raghubanshi 2015). Hierarchical CA observations.
groups various sites into different clusters based on the In CA, unsupervised pattern of the data is used to
similarity in various studied parameters. uncover intrinsic structure or underlying behavior of a
The statistical analysis was done using IBM SPSS data set without any prior assumption in classification of
Statistics software package. PCA technique extracts the objects into categories based on similarity (Vega et al.
eigenvalues and eigen vectors from the covariance ma- 1998). The most commonly used hierarchical clustering
trix of original variables to find association between the technique is CA in which clusters are formed sequen-
variables such that it reduces the dimensionality of the tially, starting with the most similar pair of objects
data set. The principal components (PCs) obtained from followed by forming higher clusters step by step. The
PCA are the uncorrelated (orthogonal) variables, obtain- Euclidean distance usually gives the similarity between
ed by multiplying the original correlated variables with two samples, and a Bdistance^ can be represented by the
the eigenvector (loadings or weightings). The eigen- Bdifference^ between analytical values from both the
values of the PCs are the measure of their associated samples (Otto 1998). This technique was performed on
variance, the participation of the original variables in the the normalized data set by means of the Ward’s method,
PCs is given by the loadings, and the individual trans- using Euclidean distances as a measure of similarity.
formed observations are called scores (Singh et al. This method uses the analysis of variance approach to
2005a, b, c) (Arslan 2013). Thus, the PCs are weighted evaluate the distances between clusters, attempting to
linear combinations of the original variables. These PCs minimize the sum of squares of any two clusters that can
extract meaningful information by capturing maximum be formed at each step. CA has been applied to assess
variance and hence facilitate the compression of data the quality of river water in relation to its spatial vari-
(Vega et al. 1998; Helena et al. 2000). PCA is designed ability at selected sites spread over the river stretch. The
to transform the original variables into new, uncorrelat- linkage distance is reported as Dlink/Dmax=(quotient
ed variables (axes), called the principal components, between the linkage distance for a particular case/
which are linear combinations of the original variables. maximal distance)×100 to standardize the linkage dis-
The new axes lie along the directions of maximum tance and represented on y-axis (Wunderlin et al. 2001;
variance. So, PCA provides an objective way of finding Simeonova et al. 2003).
indices of this type such that the variation in the data can
be accounted for as concisely as possible. This purpose
can be achieved by rotating the axis defined by PCA, Results and discussion
according to well-established rules, and constructing
new variables, also called varifactors (VF). PC is a linear Principal component analysis
combination of observable water quality variables,
whereas VF can include unobservable, hypothetical, PCA was performed on the combined data set of 3 years
and latent variables (Vega et al. 1998; Helena et al. (between the sites and parameters) and PCs with eigen-
2000). PCA of the normalized variables was performed value >1 were retained. Results from this analysis in-
cluding the loadings, eigenvalue, and variance were
Table 3 The PCA axes explaining complete variance of the data explained by each PC. The extracted three PCs explain
78.5 of the total variance as shown in Table 3, VF1
Axis % of variance Cum % of variance explains 46 % of total variance, has strong positive
1 46.096 46.096 loadings of TDS, chloride, and EC both on ground water
2 18.674 64.770 as well as surface water. VF2 explains 18.7 % of total
3 13.698 78.468 variance, has strong positive loadings on PO4−3, on
ground and surface water and strong negative loadings
170 Page 6 of 10 Environ Monit Assess (2015) 187:170

Table 4 Loading of variables on significant principal component of surface and ground water data

variable VF1 VF2 VF3

TEMP (G) −0.6615 −0.3687 −0.4092

pH (G) 0.2866 −0.7308 0.3998
TDS (G) 0.9082 0.2919 −0.1262
NO3- (G) 0.6797 0.2028 0.0161
PO43-(G) −0.5518 0.7283 −0.1331
Cl- (G) 0.8962 −0.1536 −0.3264
HARD (G) −0.2554 0.1119 −0.7608
EC (G) 0.8521 0.4080 0.2224
TEMP (S) −0.0523 −0.6383 −0.1712
pH (S) 0.8108 −0.2478 −0.2152
TDS (S) 0.7723 0.4783 0.0376
NO3- (S) 0.7643 −0.1961 −0.3671
PO43- (S) 0.5399 −0.3168 0.7293
Cl- (S) 0.9445 −0.2203 −0.0352
HARD (S) −0.3199 0.6286 0.5050
EC (S) 0.7371 0.4224 −0.3095

Significant loading is shown in italic font

on pH, whereas, it has moderate positive loadings on small dyeing industries, laundries, small scales foundry,
total hardness. VF3 (13.7 % of total variance) has strong battery industries, and haphazard dumping of municipal
negative loadings on total hardness on both ground and solid waste along the river (Barai and Kumar 2013).
surface water as shown in Table 4.
From PCA (Fig. 2), we observed that Chaukaghat, Clustering of the sites based on spatial similarity
Hakulganj, Nakhighat, and Kachahari are affected by
total hardness and PO43-4−3, whereas, Rajghat site is Cluster analysis (CA) was applied to detect spatial sim-
influenced by TDS, EC, and NO3-. The pollutants at ilarity among sites under the monitoring network. It
these sites are contributed by anthropogenic sources like rendered a dendrogram as shown in (Fig. 3), grouping

Fig. 2 Result of principal

component analysis (PCA) for the
site and physiochemical
parameters of surface and ground
water
Environ Monit Assess (2015) 187:170 Page 7 of 10 170

Fig. 3 Dendrogram showing the cluster of sampling sites on Varuna River (linkage method: Ward’s method, distance measure: Euclidean
(Pythagorean) percent chaining=85.71)

all the seven sampling sites on the river into three Correlation analysis
statistically significant clusters at (Dlink/Dmax)×100≤
75. The clustering procedure generated three groups of As the raw data are not normally distributed, the corre-
sites in a very convincing way, as the sites in these lation among the eight variables was determined by
groups have similar characteristic features and same estimating Spearman’s ranked correlation coefficient
source of contamination, and their range of correlation (r), which is a non-parametric test of the degree of
coefficient are shown in Table 5. Cluster 1 (Nakhighat, correlation (Shrestha and Kazama 2007). Because of
Hukulganj, Old Bridge, Kachahari, Dhobighat), cluster the known association among variables used in the
2 (Chaukaghat), and cluster 3 (Rajghat) are the clusters study and the fairly large number of observations (42),
that correspond to contamination of water. It is evident the critical value of r, which is 0.259 at p<0.05 was
that this technique is useful in offering reliable classifi- replaced by 0.424 at p<0.001, as the critical value for
cation of surface and ground water in the whole region. this study. This is expected to highlight the stronger
So instead of monitoring eight sites, only three sites positive or negative correlations that are likely to be of
could be selected in future spatial sampling strategy more utility (Helena et al. 2000). It is also of interest to
without affecting the result (Fig. 2). The reduction in see how these correlations are reflected in the multivar-
sampling sites in the monitoring network will be more iate statistics, especially PCA, and subsequently in the
economical without losing any significance of the out- VFs after varimax rotation of the PCs. Data in Table 4
come (Singh and Singh 2010; Singh et al. 2005b, c). provide the correlation matrix of the quality parameters

Table 5 The range of correlation coefficient of cluster analysis

Nakhighat Chaukaghat Old Bridge Rajghat Kachari Dhobighat

Chaukaghat 0.9920
Old Bridge 0.9896 0.9747
Rajghat 0.9382 0.9562 0.9110
Kachari 0.9279 0.9143 0.9682 0.8253
Dhobighat 0.9691 0.9571 0.9897 0.8727 0.9858
Hukulganj 0.9965 0.9869 0.9957 0.9250 0.9504 0.9808
170 Page 8 of 10 Environ Monit Assess (2015) 187:170

Table 6 Correlation coefficient for the various physiochemical parameters of surface and ground water

pH TDS NO3− PO43− Cl− Hard EC

TEMP 0.315 ns −0.145 ns −0.167 ns 0.542a −0.171 ns 0.020 ns 0.007 ns

pH 0.188 ns 0.056 ns −0.520a 0.271 ns −0.350b 0.354b
TDS 0.469a −0.170 ns 0.813a −0.015 ns 0.866a
NO3− −0.124 ns 0.472a −0.374b 0.465a
PO43− 0.034 ns 0.437a −0.020 ns
Cl− −0.026 ns 0.548a
Hard −0.187 ns

ns p > 0.05
a
p < 0.01
b
p < 0.05

obtained from the PCA. Among the various physio- water are quite evident by studying loading of variables
chemical parameters of surface and ground water, only on significant principal components of ground and sur-
few parameters exhibited significant correlations face water.
(Table 6). Strong negative correlations have been ob- The previous studies over establishing this kind of
served between pH and PO43-, and NO3- and hardness interactions were done in different environment but used
(Table 3). We observed significant positive relationship the same methodology. Mencio and Mas-Pla (2008)
of temperature with PO (r=0.54, p<0.01). It has been performed similar kind of analysis in urbanized
reported that the solubility of calcium phosphate min- Mediterranean streams, and Woocay and Walton
erals increases with increasing temperature (House (2008) also followed the same methodology and suc-
1999). It can implicitly be concluded that the degree of cessfully exhibited the interactions which have hypoth-
contamination is also dependent on the season of sam- esized to establish for this study area. Furthermore,
pling. There exists strong positive correlation of TDS different studies have been conducted for analysis of
with Cl- (r=0.81, p<0.01) and EC (r=0.87, p<0.01), ground water using multivariate statistical techniques
and a moderate correlation with NO3− (r=0.47, p<0.01) but only a few of them proposed for any kind of
which is clearly comprehensible as TDS accounts for interactions.
both organic and inorganic impurities. The increase in
impurities enhances number of free ions in the water
leading to increased EC. Correlation analysis showed Conclusions
the degree of positive and negative correlation between
the parameters used for analysis and made necessary The present study took into account only a particular
conclusions about the interdependence of free ions on season spanning across 3 years so as to make the nec-
EC. NO3− was also positively correlated with the EC essary conclusions. The data for different sampling sites
(r=0.47, p<0.01). The correlation coefficients of vari- collected across the aforementioned time period suc-
ous parameters have been presented in Table 3. cessfully established an interaction between ground wa-
Multivariate statistical techniques (PCA and CA) ter and surface water. These techniques further aided in
effectively evaluated the variations in surface water finding out relevant parameters for monitoring the task
and ground water quality of river Varuna. PCA helped at every site. In conclusion, there is an interaction as
in identifying the factors or sources responsible for hypothesized in our proposal which was exhibited by
water quality variations explaining about 78.5 % of the the statistical technique. Moreover, this analysis will
total variability. CA reduced sampling sites into three help in future water control management program as it
groups based on the similarity in the water quality, and has outlined the parameters contributing to pollution for
this could be helpful in minimizing the sampling loca- every site. This will make the future monitoring more
tion for the future monitoring. The interactions between economical and also easier to comprehend. It is there-
hydrochemical parameters of ground water and surface fore, needful, to develop a comprehensive river water
Environ Monit Assess (2015) 187:170 Page 9 of 10 170

quality monitoring program all over the world (Sharma Kumar, S., Mishra, P.K., Singh, N.L., Singh, K.K. & Srivastava, P.
(2012). Water Quality of River Varuna in Varanasi City,
and Kansal 2011).
Uttarpradesh, India. Asian Journal of Biochemical and
Pharmaceutical Research (2).
Acknowledgments Pardeep Singh and Rishikesh Singh thank Lu, X., Zhou, Z., & Wang, J. (2011). Analysis on interactions
University Grants Commission and Council of Scientific and among atmosphere, surface and ground water using EARTH
Industrial Research, Govt. of India for funding support. Pardeep model in the unsaturated zone. Procedia Environmental
Singh also thanks Indian Institute of Technology (Banaras Hindu Sciences, 8, 134–139.
University) for funding support. R.K. Chaturvedi thanks Council Mencio, A., & Mas-Pla, J. (2008). Assessment by multivariate
of Scientific and Industrial Research, Govt. of India for funding analysis of groundwater-surface water interactions in urban-
support. D.D. Giri thanks University Grants Commission, Govt. of ized Mediterranean streams. Journal of Hydrology, 352(3–4),
India for funding support. Authors also thank the two anonymous 355–366.
reviewers for reviewing the manuscript and offering helpful Mishra, A. (2010). Assessment of water quality using principal
suggestions. component analysis: a case study of the river Ganges.
Journal of Water Chemistry and Technology, 32(4), 227–234.
Mishra, A. K. (2011). Impact of urbanization on the hydrology of
Ganga Basin (India). Water Resources Management, 25(2),
705–719.
References Otto, M. (1998). Multivariate methods. In R. Kellner, J. M.
Mermet, M. Otto, & H. M. Widmer (Eds.), Analytical chem-
istry. Weinheim: WileyeVCH.
Anibas, C., Buis, K., Verhoeven, R., Meire, P., & Batelaan, O. Pandey, D. S. (1993). Groundwater pollution studies in urban
(2011). A simple thermal mapping method for seasonal spa- settlements of Varanasi city, UP, Annual work programme
tial patterns of ground water-surface water interaction. report, 1992–1993, (p 35). Allahabad: Central Groundwater
Journal of Hydrology, 397(1–2), 93–104. Board.
APHA (1998) Standard methods for the examination of water and Raju, N. J., Ram, P., & Dey, S. (2009). Groundwater quality in the
wastewater, American Public Health Association Lower Varuna River Basin, Varanasi District, Uttar Pradesh.
Washington D.C., Ed.19. Journal of the Geological Society of India, 73(2), 178–192.
Arslan, H. (2013). Application of multivariate statistical tech- Selle, B., Schwientek, M., & Lischeid, G. (2013). Understanding
niques in the assessment of groundwater quality in seawater processes governing water quality in catchments using prin-
intrusion area in Bafra Plain, Turkey. Environmental cipal component scores. Journal of Hydrology, 486, 31–38.
Monitoring and Assessment, 185(3), 2439–2452. Sharma, D., & Kansal, A. (2011). Water quality analysis of river
Barai, S. R., & Kumar, S. (2013). Evaluation of the physio- Yamuna using water quality index in the national capital
chemical characterestics of river Varuna at Varanasi, India. territory, India (2000–2009). Applied Water Science, 1,
Journal of Environmental Biology, 34, 259–265. 147–157.
Chaturvedi, R. K., & Raghubanshi, A. S. (2015). Assessment of Shrestha, S., & Kazama, F. (2007). Assessment of surface water
carbon density and accumulation in mono-and multi-specific quality using multivariate statistical techniques: a case study
stands in Teak and Sal forests of a tropical dry region in India. of the Fuji river basin. Environmental Modelling & Software,
Forest Ecology and Management 339, 11–21. 22(4), 464–475.
Harveya, W. J., Newlina, T. J., & Krupab, L. S. (2006). Modeling Shukla, U. K., & Raju, J. N. (2008). Migration of the Ganga river
decadal timescale interactions between surface water and and its implication on hydro-geological potential of Varanasi
ground water in the Central Everglades, Florida, USA. area, U.P., India. Journal of Earth System Science, 117(4),
Journal of Hydrology, 320, 400–420. 489–498.
Helena, B., Pardo, R., Vega, M., Barrado, E., Fernandez, J. M., & Simeonova, P., Simeonov, V., & Andreev, G. (2003).
Fernandez, L. (2000). Temporal evolution of groundwater Environmetric analysis of the Struma River water quality.
composition in an alluvial aquifer (Pisuerga River, Spain) by Central European Journal of Chemistry, 2, 121–126.
principal component analysis. Water Research, 34(3), 807– Singh, S. J., & Dwivedi, A. K. (2007). Numerical interdependence
816. in pH, acidity and alkalinity of a polluted river water. Journal
House, W. A. (1999). The physico-chemical conditions for the of Environmental Biology, 30(5), 773–775.
precipitation of phosphate with calcium. Environmental Singh, S., & Singh, K. N. (2010). Physio-chemical analysis of
Technology, 20(7), 727–733. sewage discharged into Varunariver at Varanasi. Current
Hutton, L. G. (1983). Field testing of water in developing coun- World Environment, 5(1), 201–203.
tries. Medmenham, England: Water Research Center. Singh, K. P., Malik, A., & Sinha, S. (2005a). Water quality
Isa, N. M., Aris, A. Z., & Sulaiman, W. A. N. W. (2012). Extent assessment and apportionment of pollution sources of
and severity of groundwater contamination based on Gomti river (India) using multivariate statistical tech-
hydrochemistry mechanism of sandy tropical coastal aquifer. niques a case study. Analytica Chimica Acta, 538(1–2),
Science of the Total Environment, 438, 414–425. 355–374.
Kikuchi, C. P., Ferre, T. P. A., & Welker, J. M. (2012). Spatially Singh, K. P., Malik, A., Singh, V. K., Mohan, D., & Sinha, S.
telescoping measurements for improved characterization of (2005b). Chemometric analysis of groundwater quality data
ground water-surface water interactions. Journal of of alluvial aquifer of Gangetic plain, North India. Analytica
Hydrology, 446–447, 1–12. Chimica Acta, 550(1–2), 82–91.
170 Page 10 of 10 Environ Monit Assess (2015) 187:170

Singh, K. P., Malik, A., Sinha, S., Singh, V. K., & Murthy, R. C. river-estuarine system (India) a case study. Environmental
(2005c). Estimation of source of heavy metal contamination Geochemistry and Health, 28, 317–330.
in sediments of Gomti River (India) using principal compo- Valett H.M., & Sheibley, R.W. (2009).Ground water and surface
nent analysis. Water, Air, and Soil Pollution, 166(1–4), 321– water interactions. Reference Module in Earth Systems and
341. Environmental Sciences, from Encyclopedia of Inland
Singh, K. P., Gupta, S., & Mohan, D. (2014). Evaluating influ- Waters, 691–702.
ences of seasonal variations and anthropogenic activities on Varol, M., Gokot, B., Bekleyen, A., & Sen, B. (2013).
alluvial groundwater hydrochemistry using ensemble learn- Geochemistry of the Tigris River basin, Turkey: spatial and
ing approaches. Journal of Hydrology, 511, 254–266. seasonal variations of major ion compositions and their con-
Sojka, M., Siepak, M., Ziola, A., Frankowski, M., Murat- trolling factors. Quaternary International, 304, 22–32.
Blazejwska, S., & Siepak, J. (2008). Application of multivar- Vega, M., Pardo, R., Barrado, E., & Deban, L. (1998). Assessment
iate statistical techniques to evaluation of water quality in the of seasonal and polluting effects on the quality of river water
Mala Welna River (Welna Poland). Environmental by exploratory data analysis. Water Research, 32(12), 3581–
Monitoring and Assessment, 147, 159–170. 3592.
Sophocleous, M. (2002). Interactions between ground water and Wade, J. A., Butterfield, D., & Whitehead, G. P. (2006). Towards
surface water: the state of the science. Hydrogeology Journal, an improved understanding of the nitrate dynamics in low-
10, 52–67. land, permeable river-systems: applications of INCA-N.
Soulsby, C., Gibbins, C., Wade, A. J., Smart, R., & Helliwell, R. Journal of Hydrology, 330(1–2), 185–203.
(2002). Water quality in the Scottish uplands: a hydrological Woocay, A., & Walton, J. (2008). Multivariate analysis of water
perspective on catchment hydrochemistry. Science of the chemistry: surface and ground water interactions. Ground
Total Environment, 294(1–3), 73–94. Water, 46(3), 437–449.
Stigter, T. Y., Carvalho Dill, A. M. M., Ribeiro, L., & Reis, E. Wunderlin, D. A., Diaz, M. P., Ame, M. V., Pesce, S. F., Hued, A.
(2006). Impact of the shift from groundwater to surface water C., & Bistoni, M. A. (2001). Pattern recognition techniques
irrigation on aquifer dynamics and hydrochemistry in a semi- for the evaluation of spatial and temporal variations in water
arid region in the south of Portugal. Agricultural Water quality. A case study: Suquia river basin (Cordoba,
Management, 85, 121–132. Argentina). Water Research, 35, 2881–2894.
Sundaray, K. S., Panda, C. U., Nayak, B. B., & Bhatta, D. (2006). Yu, Y. J., Guan, J., Ma, Y. W., Yu, S. X., Guo, H. C., & Bao, L. Y.
Multivariate statistical techniques for the evaluation of spatial (2010). Aquatic environmental quality variation in Lake
and temporal variations in water quality of the Mahanadi Dianchi watershed. Procedia Environmental Sciences 2, 76–81.