A well approached sample technique was adopted to collect the surface water sample along the river stretch of the study area. Total of thirty six water samples, eighteen samples during pre-monsoon (June 2019) and eighteen samples during post - monsoon (September 2019) were collected from river Pārbati and its tributaries. Samples were collected in1000 ml best quality, sealed shut polyethylene bottles with cover lock. Physical parameters like EC, pH, TDS were measured on the field after sample collection using portable water and soil analysis kit. Major cations (Ca2+, Mg2+, Na+, K+) and anions (HCO3-, Cl-, SO42-, NO3-) were analyzed by following standard procedure given by American Public Health Association [30]. The accuracy of the chemical analysis was checked by Charge Balance Error (CBE) [31].
Charge Balance Error (CBE) = (∑Cations - ∑Anions)/(∑Cations+∑Anions)*100% Eq.1
Majority of the analyzed samples showed CBE around ± 10%
Calcium, Magnesium, Bicarbonates, Chloride, Total alkalinity and Total hardness were determined using EDTA titrimetric method. Sodium and Potassium using flame photometric method, Nitrate and Sulphate by UV spectrophotometric method respectively. Mean value were calculated for each parameters to understand the seasonal variation as an indication of the precision of each parameter. Maps were prepared using ArcGIS 10.8 software and Piper trillinear, Durov were plotted using Grapher. The statistical software SPSS, Grapher, Aquachem and Microsoft excel were employed for the calculations and data interpretation.
3.1 Appraisal of Water Quality Index (WQI)
WQI is an effective mathematical tool to evaluate the suitability of water for drinking purpose [32]. It is calculated by adopting the weighted arithmetical index method [33]. In the present study physiochemical parameters namely pH, EC, TDS, TH, Ca2+, Mg2+, Cl-, SO42-, NO3- were considered for computing WQI for river Parbati. Tiwari and Mishra (1985) developed an equation to calculate WQI, the calculation involves following steps (Eq. 2-5):
WQI = ∑qnWn / Wn Eq. 2
Where, Wn= unit weight of nth parameters, is calculated by the equation
Wn=K/Sn Eq. 3
and K, is the proportionality constant obtained from,
K = [1/ (∑ni=0 1/Si)] Eq. 4
Where, Sn and Si are the BIS standard values of the water quality parameter
qni= 100 × 𝑉𝑎 –𝑉𝑖/𝑉𝑠 −𝑉𝑖 Eq. 5
qni is the quality rating of the ith parameter for a total of n water quality parameters
Where, Va = value of the water quality parameter obtained from laboratory analysis, Vi = ideal value (for pH = 7 and 0 for other parameters) and Vs = BIS standard value of water quality parameters.
3.2 Evaluation of hydrogeochemical facies of surface water
To comprehend the hydro geochemical attributes of surface water, different plots were utilized specifically, Piper, Durov and Gibbs plot [34-36]. These plots represent the graphical relationship characterizing different geochemical marks in surface water samples. The grapher 12.0 was used to prepare the Piper diagram and Durov, while Gibbs plot was prepared by aquachem software.
3.3 Multivariate statistical analysis
Multivariate Statistical techniques have been used to organize and simplify datasets and characterize freshwater, marine water and sediment quality [37-39]. In recent years, water quality assessment has been widely done using multivariate statistical techniques [40, 41]
Correlation analysis is a technique which determines the correlation coefficient between variables. The relationship between two variables can be measured by the strength and significance of the variables. The strength is indicated by the correlation (r), whereas the significance is expressed in probability levels (p values). Larger the correlation coefficient, stronger the relationship, whereas smaller the p level, more significant the relationship.
Statistical extraction of linear relationship from a given set of variables is performed by applying Principal Component Analysis (PCA) technique [22]. PCA allows to gain insight into the data without significant loss of information in the process [42, 43]. Principal components generated during the analysis are arranged in such a manner that they correspond to decreasing contribution of variance, i.e., principal component 1 (PC1) explains the highest amount of variance in the original data [44, 45] classified the factor loadings as ‘‘strong’’, ‘‘moderate’’ and ‘‘weak’’, corresponding to absolute loading values of 0.75, 0.75–0.50 and 0.50–0.30, respectively. However, loading reflects the relative importance of a variable within the component and does not reflect the importance of the component itself [46].