Water quality assessment using NSFWQI, OIP and multivariate techniques of Ganga River system, Uttarakhand, India

Ganga River water is very much stressed with the rapidly increasing population, climate change and water pollution that increase domestic, agricultural and industrial needs. This study assesses the surface water quality of the River Ganga in India, using NSFWQI, OIP and multivariate techniques. During the current study, water samples from Ganga River were collected for the assessment of 19 physico-chemical determinants from 20 sampling locations. Water quality indices (WQIs) is used to classify the overall impact of different variables of water. Multivariate techniques were utilized to assess the water conditions for productive management of fresh water quality. The WQI results showed that surface water quality varied at the selected sampling sites among medium and good categories. The PCA generates the 6 principle components which highly contributes (80.3%) in influencing the hydro-chemistry of river water. Agricultural waste runoff, untreated effluents and many other anthropogenic activities were identified as main contributor in decreasing the water quality of the River Ganga. To maintain and protect this fresh water resources against contamination, the usage of stringent policies and rules are expected to preserve fresh water resources for people in the future.


Introduction
Rivers are one of the significant resources of freshwater to the people for the use in many purposes. India with 18% of world's population and 4% of world's water resources availability, is a water stressed nation with 1588 m 3 per capita per annum water supply. The Uttarakhand is water rich state of the country with 21% availability of surface water against inevitably increase in agriculture, industrial and domestic water needs (TERI 2017; Census of India 2011).
The importance of Ganga River is not only for its sociocultural perspective, but also for various livelihood and economic activities for the people living around the basin. The improvement in the river health is growing debate in India since last decade, mainly in terms of its quality and quantity (Matta and Uniyal 2017). The river is facing multi-dimensional issues like, flow disturbance, water pollution, climate influences, habitat and biodiversity conservation (Kaushal et al. 2019). The water pollution has become a global threat for developing countries due to the climate change, rapid urbanization, agricultural runoff and industrialization (Bhardwaj 2005;Matta et al. 2018c;Rana et al. 2017).
It is astute to save and control the surface water from pollution and to have solid data on surface water quality for viable management. In this way, regular assessment of the surface water quality is required to maintain and control the surface water from degradation (Barakat et al. 2016;Shukla et al. 2017b;Matta et al. 2015). For that purpose, water quality indices (WQIs) are the mathematical expressions to evaluate the water quality by process the dataset of different variables into simple numeric quality rating scales. (Abbasi and Abbasi 2012). WQIs are globally accepted approach to assess the water quality and each WQI uses a different set of variables (Bharti and Katyal 2011). It is used as communication tool for policy makers, researchers and different governmental bodies to report the health of water body to public in a reliable way (Sadiq et al. 2010). The WQIs uses different selection process, statistical transformation, weighing of parameters and interpretation of final numeric values (Lumb et al. 2011;Alobaidy et al. 2010). Many of the 206 Page 2 of 12 research work has been used various WQIs for the assessment and monitoring of different surface water quality (Shah and Joshi 2017;Matta et al. 2018a;Son et al. 2020;Matta et al. 2020).
Multivariate methods such cluster analysis (CA) and principle component analysis (PCA) are utilized in different studies to recognize potential pollutants (Fan et al. 2010;Massoud 2012;Misaghi et al. 2017). PCA is a multivariate technique, and in the situations where colossal measure of information is accessible, it is a reasonable way to diminish the data (Noori et al. 2009(Noori et al. , 2010. Moreover, CA is additionally one of the multivariate techniques used to evaluate relative similarity in the homogeneity of estimated parameters (Shrestha and Kazama 2007).
The goal of this study is to evaluating the physico-chemical properties of the river water and provide appropriate picture of present water quality of the River Ganga, through NSFWQI, OIP model and other environmetric techniques like PCA, CA so as to facilitate emerging issues.

Description of research area
The River Ganga basin cover four nations (India, China, Bangladesh and Nepal), spreads on land surface of about 1,080,000 Km 2 , of which 80% lies within India (Tare et al. 2015). About 43% of total inhabitants of India live along the River Ganga basin (Tare et al. 2013) and utilize river water directly or indirectly for various purposes. River Ganga with total run of approximately 2525 km long, is a mountainous river originates from Gangotri glacier of northern Himalayan region as 'Bhagirathi River' (known as one of major headwater stream) (NMCG 2019). The river name Ganga came to known where two head-stream (Bhagirathi and Alaknanda) confluence at Devprayag (Chakrapani and Saini 2009;Matta et al. 2018b).
The climatic condition of study area is subtropical to temperate with three seasons summer, monsoon and winter (Chauhan 2010). The research area lies between the coordinates from 30049′ N and 79010′ E to 29051′ N and 77053′ E. The state also has fully disparity between its hilly districts to plains and characterized by climatic variation from freezing cold (snow cover up to 10 feet from land surface) at upstream site to very warm condition (up to 450C during summer season) at downstream stations (CWC 2012). The most rainfall occurs in monsoon season (June-September) with an average annual rainfall range from 350 mm in west to 2500 mm in east of Ganga River Basin .
The sediment and water flux of Ganga River system annually supplies 400-500 Million tonne (Mt.) and 380 km 3 , respectively, to Bay of Bengal (Awasthi et al. 2018;Lupker et al. 2012). The river bed consists of boulders, rocks and cobbles, whereas water status is ultra-oligotrophic to oligotrophic in nature (Mathur and Kapoor 2015).
The sampling locations were selected on the basis of land use pattern, population, potential anthropogenic activities and socio-economic important of locations. The description of selected monitoring locations is given in Table 1 and study area map is shown in Fig. 1.

Sample collection and analysis
A total of 120 water samples were collected from 20 monitoring locations during winter: November-February, summer: March-June, monsoon: July-August and post-monsoon season September-October, for the year 2015-2016. The samples were gathered bimonthly, from 15 cm depth of superficial layer in triplicate manner at each site, in a preacid washed polypropylene Nalgene Wide-Mouth Natural HDPE bottles of 1000 ml. The samples were stored in ice bucket by maintaining its temperature up to 4 °C, transported to laboratory and refrigerated to avoid any contamination and the effect of temperature and light.

National sanitation foundation water quality index (NSFWQI)
This mathematical expression provides a uniform and meaningful method for assessing the overall quality of freshwater. In this method, a total of 9 parameters were assigned and calculation is expressed as following equation: (1): where q i represents the assigned curve based sub-index value for ith variable which is ranged from 0 to 100, w i is the weighting coefficient for ith parameter with range from 0 to 1, and described in Eq. (2), n is the number of total variables considered, the NSFWQI rating scale is also divided in five quality classes and considered as very bad (0-25), bad (25-50), medium (50-75), good (70-90) and excellent (90-100) (Brown et al. 1970).

Overall index of pollution (OIP)
The OIP was proposed by Sargaonkar and Deshpande (2003) for the assessment of surface water quality, specifically under Indian condition. The calculation of OIP was performed by considering turbidity, TDS, BOD, hardness, chloride and sulfate concentrations of water samples. The observed concentrations were put into assigned mathematical expressions for individual parameter, given in Table 2 to find pollution index score (P i ). The final OIP value was estimated by following Eq. (3): where, Pi = pollution index value for the ith parameter, n = number of parameters. The final numeric index score was used to classify water quality into different pollution classes given in Table 3.

Multivariate techniques
The multivariate technique like principle component analysis (PCA) was analyzed with the support of Kaiser-Meyer-Olkin (KMO) and Bartlett tests. To recover the association among data sources and essential factors similarly as to accomplish better goals, the Varimax technique was used (Noori et al. 2010;Ouyang 2005). To observe cluster analysis (CA), averages of nineteen parameters for 20 sampling sites were evaluated. The comparability among groups and isolating similar clusters were resolved dependent on Euclidean distance. In the present study, hierarchical CA was executed for a set of standardized statistics with Ward's Method.

Seasonal variation in physico-chemical parameters of Ganga River water
The observed analytical ranges of different parameters of water testing are represented in Table 4    Dissolved oxygen (DO) plays an important role in the oceanic biological system and to assess the level of freshness of surface water. It also helps to evaluate the quality and natural contamination in the surface water (Wetzel and Likens 2006). For this study the average dissolved oxygen concentration of surface water was noticed 8.89 ± 1.36, 8.89 ± 0.98, 9.29 ± 1.18 and 9.18 ± 1.33 mg/L in winter, summer, monsoon and post-monsoon season, respectively. The total suspended solids were recorded 269.92 ± 323.89, 283.04 ± 276.33, 321.40 ± 259.97 and 350.52 ± 326.1 mg/L during winter, summer, monsoon and post-monsoon season, respectively. Similarly, Rani et al. (2014) observed the TSS variation within the limit of 100-190 mg/L while observing the quality of Ganga Canal water during Kanwar Mela at Haridwar. High amount of TSS may be due to suspension of clay and soil particles (Daphne et al. 2011).
The average biological oxygen demand of all seasons was recorded 1.88 ± 0.55, 1.88 ± 0.58, 1.76 ± 0.57 and 1.83 ± 0.55 mg/L for winter, summer, monsoon and postmonsoon season, respectively. Aenab (2013) observed that the BOD (6.5 mg/L) value was out of the WHO standard limit. The average concentration of chemical oxygen demand (COD) was evaluated 6.15 ± 1.71, 6.43 ± 1.46, 6.50 ± 1.75 and 6.35 ± 1.74 mg/L for winter, summer, monsoon and post-monsoon season, respectively. The seasonally average free CO2 of surface water was recorded1.03 ± 0.37 mg/L, 1.11 ± 0.48, 0.92 ± 0.24 and 0.89 ± 0.26 mg/L for winter, summer, monsoon and post-monsoon season, respectively. Joshi et al. (2009) reported the COD range between 4.58 and 13.72 mg/L and free carbon dioxide in the River Ganga was perpetually present consistently. The seasonally average value of total hardness was recorded 83.06 ± 44.17, 88.32 ± 48.67, 90.06 ± 51.29 and 84.28 ± 35.16 mg/L during winter, summer, monsoon and post-monsoon season, respectively. The hardness represents the usability of water for drinking, industrial and domestic purposes.
Concentration of Alkalinity in water is the assurance of the ability to deactivate a solid corrosive. The bases, for instance, hydroxides, bicarbonates, phosphates, nitrates, carbonates, borates, silicates, etc., are in charge of alkalinity of water. It provides a thought of characteristic salts in water (Sharmila and Rajeswari 2015). The seasonally average value of alkalinity was recorded 88.21 ± 43.21, 85.06 ± 44.23 89.78 ± 53.53 and 87.81 ± 51.10 mg/L for winter, summer, monsoon and post-monsoon season, respectively. Acidity is vital parameter to evaluate the water quality. Dissolved CO 2 is the important and main factor which is responsible for acidity in unpolluted water. In this study the seasonally average value of acidity was recorded 50.03 ± 9.42, 52.10 ± 12.50, 51.85 ± 10.06 and 55.42 ± 10 mg/L during winter, summer, monsoon and post-monsoon season, respectively.
Use of detergents may escalation the phosphate concentration in water to great extent. The man-made additions of phosphorus to the surface water have a considerable impact on the quality of the river water. Such phosphorus is determined for the most part from domestic sewage and the overflow from rural territories. During the study the phosphorus content was also recorded 0.06 ± 0.02, 0.08 ± 0.05, 0.09 ± 0.05 and 0.08 ± 0.04 mg/L in winter, summer, monsoon and post-monsoon season, respectively. A very slight variation in average concentration of Cl − was found as 5.74 ± 1.63, 5.39 ± 1.40, 5.68 ± 1.87 and 5.75 ± 1.74 mg/L during the study period. The ingestion of water containing higher dose of Cl − can cause hypertension, osteoporosis, rental stones and asthma (McCarthy 2004). The phosphate concentration has positively impact on photosynthetic activities and chlorophyll-a content of aquatic flora, whereas the TKN concentration negatively affect the chlorophyll-a pigment (Tare et al. 2003). The average concentration of P was ranged from 0.06 ± 0.02 mg/L in winter to 0.09 ± 0.05 mg/L in monsoon season and the mean level of TKN was varied from 0.06 ± 0.01 mg/L during winters to 0.07 ± 0.01 mg/l during monsoon season. The concentration of SO 4 was also reported with a slight variation from 19.61 ± 0.95, 19.74 ± 1.40, 20.06 ± 1.22 and 20.43 ± 1.44 mg/L during winter, summer, monsoon and post-monsoon season, respectively.

Multivariate statistical analysis
The variability of original average dataset was presented by screen plot (Fig. 2) which show that the first Six components, contributes for 80.3%. Therefore, the details about the water quality at different 20 sites can be performed in reduced data set of six variables. On the basis of Eigen value criteria, six principal components (PCs) were noticed important as their eigenvalues are greater than one. Now the number of clusters can be selected according to the reported number of principle components to categorize the sampling sites. In this study, it is observed from screen plot that only six clusters are necessary to group the monitoring sites.
The factor weights values of observed factors related to each principle components are represented in Table 5. The PCs are classified as 'weak', 'moderate' and 'strong' on the basis of their factor loading values (> 0.75, 0.75-0.50 and 0.50-0.30, respectively) (Liun et al. 2003). It was determined that the variables turbidity, BOD, free CO 2 had a negative factor weight, -0.357, -0.335 and -0.328, respectively in the development of PC1 with 24.9% of variation. This component indicates organic source of pollution from natural processes and agro-based industries (Sharma et al. 2015). The PC2 was associated with a negative weight with DO (-0.422) only with 18.8% variation. The weight of the variables Hardness (0.363) and acidity (0.400) was positive and Cl (-0.393) had a negative factor weight in the formation of PC3 with 11.8% variation. Positive weight in PC4 had LI (-0.401), alkalinity (-0.355), TKN (0.359), and SO 4 -(-0.353) had a negative weight with 10.3 variation. For PC5 the variables temperature (0.451) had positive weight while only alkalinity (-0.463) had a negative weight with 8.3% variation. The formation of moderate PC6 includes a positive weight with TS (0.559) and TKN (0.506) with total variation of 6.3%.  The grouping between the similar monitoring sites was detected by CA and presented by a dendrogram in Fig. 3. The dendrogram represents the six groups of monitoring site, in each group all sampling locations without any specific trends. Cluster I is generated by sites 1, 3, 6 and 4, Cluster II of site number 7, 9 and 12, Cluster III by 2, 5 and 11, Cluster IV shows two sites 8 and 10, Cluster V is generated by sites 13, 15, 14 and 16 and the last Cluster VI include 17, 19, 20 and 18. These clusters receive domestic sewage, sewage musk from hydro-electric power plants, effluents from electroplating, automobile and pharmaceutical industries (Sharma et al. 2015).

Indexing approach
The all values of NSFWQI are given in Table 6, which presents the surface water quality status during the study period was monitored to differ among medium and good range. The average value of NSFWQI for all seasons is introduced in Table 6. Table 6 shows the good water quality at Gangotri and Uttarkashi in winter (NSFWQI: 77 and 75) and summer season (NSFWQI: 84 and 81), while the lowest was drawn at Jatwara Bridge, Haridwar (NSFWQI: 41) in monsoon season (Fig. 4). As to average, the most noticeably worst water quality was found in Monsoon season. During the 1-year study period, depend on quality of surface water index values (Table 6), the quality of water in the monsoon and post-monsoon was average and was good in summer and winter season. At Gangotri, Badrinath, Rudraprayag, Srinagar, Mayapur (Haridwar) and Jatwara Bridge (Haridwar), the quality of River water was acceptable in the summer season, and in other season, the quality was normal, which was the consequence of illegal wastewater release in river. Some other responsible factors, for example, anthropogenic activities, street development, farming activities and different factors unfavorably influence the surface water quality of these sampling locations.
For this study, OIP was additionally used for the assessment of water quality indexing. The seasonally values of OIP of all the sampling sites of selected physico-chemicals parameters is presented in Table 6. The result shows that the water quality status of Ganga River was in the range class C2 (1-2) acceptable in summer and C3 (2-4) scarcely polluted in winter, Monsoon and post-monsoon season, respectively (Fig. 5). Shukla et al. (2017a) examined the water quality of Upper Ganga River basin using OIP and observed that because of high stream flow, water quality at Uttarkashi was acceptable (class C2: 1-2) from 2001 to 2012, while slightly contaminated (class C3: 2-4) in monsoon period during 2006.
The highest value (C3 class) of OIP was observed in monsoon season with value varied from 2.01 at Trivenighat, Rishikeshto 3.97 at Jatwara Bridge, Haridwar with an mean value of 2.45 ± 0.713 (Table 6). During this season, the water quality status of Ganga River was slightly contaminated. This might be because of heavy overflow of water sediments, agricultural runoff and converging of sewage water from various point sources. Different analysts were additionally observed the comparable results of pollution status in monsoon period, for example, Gupta et al. (2017) in Narmada River, Jindal and Sharma (2010) in Sutlej River, Sebastian and Yamakanamardi (2013) in Kapila and Cauvery River.

Conclusions
Based on the results obtained from the present study, the water quality of Ganga River system in Himalayan region of Uttarakhand, India, was vulnerable to pollution during 2015-2016. The higher level of turbidity made the water quality unsuitable for drinking purposes, while compared to the Indian standards. The WQI and multivariate statistical models were proved to be beneficial approaches to classify the spatial and temporal variation in river water quality, by transforming the dataset into commensurate unit data and numeric index values. NSFWQI and OIP elicit the water quality in good or acceptable class for upstream sites (Gangotri and Uttarkashi) during winter and summer season, whereas slightly pollution was reported for downstream site (Jatwara Bridge, Haridwar) during monsoon season. The PCA analysis elucidates the relationship among important parameters by reducing number in the form of PCs and their contribution in variation. It also helps in pollution source identifications.
There were six components recognized that can attribute to organic pollution (agricultural waste), rainfall runoff and riverbed erosion. According to CA, similar pollution causing monitoring stations were grouped into six clusters. The CA findings will help to reduce the number of sampling stations and the sampling expanses for further assessment. These findings essentially provide information on river water quality management and pollution control measures.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes 100 S1 S2 S3 S4 S5 S6 S7 S8 S9 S10 S11 S12 S13 S14 S15 S16 S17 S18 S19 S20 NSFWQI Sampling sites 5.00 S1 S2 S3 S4 S5 S6 S7 S8 S9 S10 S11 S12 S13 S14 S15 S16 S17 S18 S19 S20 were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creat iveco mmons .org/licen ses/by/4.0/.