Water quality assessment of Upper Ganga and Yamuna river systems during COVID-19 pandemic-induced lockdown: imprints of river rejuvenation

Abstract

Clean river water is an essential and life-sustaining asset for all living organisms. The upper Ganga and Yamuna river system has shown signs of rejuvenation and tremendous improvement in the water quality following the nationwide lockdown due to the coronavirus pandemic. All the industrial and commercial activity was shut down, and there was negligible wastewater discharge from the industries. This article addresses the water quality assessment from the study area, which is based on the original data of physical parameters, major and trace elements, and stable isotopes (hydrogen and oxygen) systematics during the nationwide lockdown. The impact of the lockdown could be seen in terms of an increase in dissolved oxygen (DO). Water samples were collected from the Upper Ganga and Yamuna river basins (Alaknanda, Bhagirathi, and Tons rivers) during an eight-week lockdown in Uttarakhand, India. We discussed the signs of rejuvenation of riverine based on physical parameters, major ions, trace elements, isotopic ratios, and water pollution index (WPI). Results reveal that the water quality of the entire upper basins of the Ganga has significantly improved by 93%, reflecting the signs of self-rejuvenation of the rivers. Multivariate analysis suggests a negative factor loading for an anthropogenic element (${NO}_{3}^{-}$), implying that they contribute little to the river water during the lockdown. Further, bicarbonate (${HCO}_{3}^{-}$) is a dominant element in both river basins. The geochemical facies are mainly characterized by the (${{Ca}^{2+} :{Mg}^{2+} : HCO}_{3}^{-}$) type of water, suggesting that silicate rock weathering dominates with little influence from carbonate weathering in the area. The positive factor loadings of some cations, like${HCO}_{3}^{-}$,${Ca}^{2+}$, and ${Mg}^{2+}$ reflect their strong association with the source of origin in the lockdown phases. Stable isotopic reveals that the glaciated region contributed the most to the river basin, as evidenced by the low d-excess in riverine water compared to anthropogenic contributions. Rivers can self-rejuvenate if issues of human influence and anthropogenic activities are adequately resolved and underline our responsibility for purifying the ecosystem. We observed that this improvement in the river water quality will take a shorter time, and quality will deteriorate again when commercial and industrial activity resumes.

Introduction

The COVID-19 pandemic originated in December 2019 in Wuhan, China, and spread quickly globally [18, 91, 128]. Nationwide lockdown (phase-1) was imposed in India on March 25th, 2020, and extended up to April 14th, 2020, to break the chain of the COVID-19 pandemic and extended until May 17th, 2020 (Phase-2). The World Health Organization apprised that there were five lakh confirmed cases, while more than three lakhs lost their lives across 216 countries due to COVID-19 [120]. Unfortunately, the deadly pandemic (COVID-19) continues in 2021, and from March 2021, more than 50,000 cases have been reported in India every day and found maximum in May. Presently, the number of COVID-19 deaths reported to WHO (cumulative total) worldwide to date is 7.1 million, and confirmed cases are 776 million and increasing (https://data.who.int/).

In India, the environmental conditions were temporarily improved because of the reduced pollution load during the nationwide lockdown. The aerosol concentrations were measured at a 20-year low based on the satellite data on optical depth measurements over the Indo-Gangetic Plains (IGP). It was possible due to the restrictions imposed on industries, air, rail, road transport, etc. [70]. The Ganga River Basin (GRB) is the largest source of fresh water for the most densely populated region globally [69, 79]. However, the effect of the COVID-19 lockdown on the water quality of the River Ganga was debatable. As stated by the Central Pollution Control Board (CPCB) of India, the Ganga River (2601 km length) has been exposed to high pollution over the last few decades [72]. Although the Ganga River is an essential life-supporting component of the rapidly growing population of India, human health is endangered due to its water pollution level. Earlier, the river water was polluted by various human activities such as industrialization, urbanization, agricultural practices, and overexploitation [8]. It has been stated that the pollution level of the riverine system has dropped during the lockdown period [39], as the primary pollution sources (e.g., industries, tourism activities, pilgrimage, hotels and lodges, shops) of the Ganga River were closed. However, the domestic discharges from the household continued during the lockdown period [34].

Garrels and Mackenzie [35] raised a fundamental question about the derivation of major ions in river water from its sources, which was later successfully addressed globally by several studies [6, 10, 13, 16, 25, 27, 30,31,32, 75, 89, 98, 103, 108]. Riverine water quality across the globe has degraded due to unplanned urbanization, human intervention, population growth, and anthropogenic activity [7, 39, 44, 100, 103]. Monitoring the water quality of major rivers is essential as these provide water and food security to about 3 billion people around the globe [7].

Water quality monitoring to quantify the impact of rapid population and industrial waste in the riverine system in India has gained momentum in the last two decades [29, 54, 103]. The Ganga–Brahmaputra basin spreads over one million km² and ranks among the most densely populated regions, with a population density of > 300 people/km², and is home to over 0.6 billion people in India [80]. India has minimal geochemical and isotopic data on rivers related to water quality assessment, including the Ganga River and its tributaries, which stated how the riverine water quality has deteriorated over time [11, 60, 93, 103, 115].

The print and electronic media stated that the overall pollution in the Ganga river is down by up to 50% due to the imposition of lockdown [41, 76, 77, 109, 110]. That has made the Ganga water potable by filtering without further treatment [5, 45]). Dissolved Oxygen (DO), Biochemical oxygen demand (BOD), and other physical parameters levels were improved during the lockdown, making the water drinking and bathing purposes [26]. These results were based on some observations (satellite data and physical parameters). Still, no geochemical and isotopic data is available to assess riverine water quality based on major ions, trace metals, and environmental isotopes. Based on isotope and geochemical studies, we report the scientific evidence for improved water quality in the Upper Ganga and Upper Yamuna River systems (UGRS and UYRS). If the concentrations of physical parameters, trace elements, and major ions surpass the permissible limit, they may harm humankind and livelihood [12, 19, 73]. Thus, measuring these parameters is essential to understand the water quality of any water masses.

The quality of water from a particular source (e.g., riverine/lake/groundwater) can assessed by its physio-chemical and biological parameters. A method known as the Water Quality Index (WQI) is classified by giving unequal weightage based on its importance to riverine water quality, but there are certain limitations to this method [64, 85, 116]. Therefore, to avoid complications and reduce the errors and sensitivity of water indexing, we adopted a new, improved concept called Water Pollution Index for the present study [42].

Since the analysis of a dataset of variable magnitude and density is challenging, multivariate statistical techniques, namely Principal Component Analysis (PCA), can be applied to assess the water quality [3, 37, 43, 126]. This method requires significant sources of ions in river basins [58, 61, 114, 121]. The PCA provides a simple solution to the problems, whereas the conventional method fails to interpret the data. Hence, multi-dimensional matrices and measurements have been preferred over new approaches to conventional ones [56].

Multivariate analysis, such as the Principal Component Analysis (PCA) and Hierarchical Cluster Analysis (HCA), are essential statistical methods to assess the water quality that includes the major sources of ions in river basins [63, 78, 121]. Analyzing the resulting dataset of variable magnitude and density is difficult. Therefore, multivariate statistical techniques are applied worldwide [3, 37, 43, 59, 114, 126]. Further, the hydrogeochemical evolution of the river water basin can be identified through the interpretation of PC loadings [4, 68].

Hydro-geochemical processes operating in the river basin are characterized by negative loadings on specific variables in terms of principal components. For example, low-negative loading on pH for Ca and Mg corresponds to high positive loading. The inverse correlation is generally seen between two negative loadings on pH and derivatives of carbonate ions [62].

In extensive hydrological studies, the stable isotopes of oxygen and hydrogen have been used as conservative tracers [38]. Numerous interpretations of hydrological processes occurring both on the surface and within aquifers have been influenced by stable water isotopes and solutes [50, 127]. For instance, they have been applied to identify water sources, flow patterns, and water mixing in various bodies of water [2, 38, 52]. In extensive hydrological studies, the stable isotopes of oxygen and hydrogen have been used as conservative tracers [38]. Numerous interpretations of hydrological processes occurring both on the surface and within aquifers have been influenced by stable water isotopes and solutes [50, 127]. For instance, they have been applied to identify water sources, flow patterns, and water mixing in various bodies of water [2, 38, 52].

We have calculated the WPI and compared the results with those carried out earlier in the same basin [16, 97] to quantify the improvement in water quality throughout the lockdown time. The rivers carry the products produced by continental weathering to the oceans and play a vital role in the evolution of global sea water [31, 32]. The carbonate and silicate weathering processes are also carried out at alterable scales, resulting in heterogeneity in river water composition [16].

This article addresses the water quality assessment for the first time during the COVID-19 pandemic-induced lockdown based on the original data of physical parameters, major and trace elements, and stable isotopes (hydrogen and oxygen) systematics. We discussed the signs of rejuvenation of riverine water against the average geochemical composition, isotopic ratios, and water pollution index (WPI) of the earlier studies of this region.

The main objectives of this study are: (1) To analyze the impact of the lockdown on the water quality of the Ganga River and its major tributary, Yamuna, in Uttarakhand; (2) To discuss issues and challenges to understand the magnitude of contamination and source relations and potential ways to improve the water quality using PCA and HCA; (3) To provide the essential implications for future restoration strategies on river rejuvenation; and (4) To generate the baseline data for future study on the water quality assessment from the Ganga river system.

Study area

General description

Present work is carried out in the headwaters of the Ganga and Yamuna river systems (UGRS) in the Himalayan region of Uttarakhand. The study area in the UGRS lies between latitudes 78^º00ʹ N-78^º50ʹ N and longitudes 29^º50ʹ-30^º20ʹE, which extends from about 10 km upstream of the Alaknanda River at Mulya and 25 km upstream of the Bhagirathi river (at Koteshwar Dam) to up to Devprayag, where these two rivers join to become the Ganga river and further up to Haridwar city. Similarly, the area in the upper reaches of the Upper Yamuna River System (UYRS) lies between latitudes 77^º50ʹN-78^º10ʹN and longitudes 30^º10ʹ-30^º30ʹE, which extends from upstream of Yamuna River at Barkot till down to Dakpather, Vikas Nagar. The sample locations are given in (Fig. 1, Tables 1, and 2). The Upper Ganga River system (UGRS) comprises two glacier-fed rivers, namely the Bhagirathi River basin (catchment area: 7.8 × 103 km²) and the Alaknanda River basin (catchment area: 11.8 × 103 km²) [14, 105]. Bhagirathi River originates from the Gangotri glacier at an elevation of 3900 m a.m.s.l. [107] joins with the Alaknanda River, which has its source from Satopanth and Bhagirath Kharak glacier system, at an elevation of 3641 m a.m.s.l. [95]. At Devprayag, from where they flow as the mighty river Ganga.

Fig. 1

A Inset image of India, pictorial diagram. B and C showing the studied catchments and the sample locations in the Upper Ganga River System and Upper Yamuna River System

Table 1 Geographical distribution of the samples (Upper Ganga River System)

Table 2 Geographical distribution of the samples (Upper Yamuna River System)

The Yamuna River originates from Saptrishi Kund near Bander Punchh peak (latitude 31.01°N, longitude 78.46°E) in the Mussoorie range of the lower Himalaya [60]. The Yamuna and its major tributaries (the Tons, the Giri, the Aglar, the Bata, and the Asan) form the Himalaya Yamuna River System (YRS). The YRS is fed mainly by the Indian summer monsoon and has its maximum discharge during July–September, the monsoon season. Its catchment receives about 80% of its annual rainfall during these months [21]. Further, the Yamuna and the Tons receive water from glacier melt, a major contributor to the discharge from April to June during the study period. The Yamuna joins the Ganga at Allahabad in the plains.

Geological setup

The Upper Ganga Rivers system and its tributaries flow mainly through the Higher Himalayan crystalline (HHC) and Lesser Himalayan sedimentary (LHS) lithology. However, some of their tributaries originate in the Tethyan sediments. HHC in the study area consists of high-grade gneisses, metabasite, quartzite, schist, and granite, with a trace amount of carbonate and calc-silicate rocks [81, 123]. After originating from the HHC, the Bhagirathi River drains through a deep gorge at Gangotri, crosses the Bhagirathi leucogranite, and travels to the village of Bhatwari where it crosses the Vaikrita Thrust at Gangnani [106]. According to [101], the Vaikrita Thrust marks the upper boundary of the Main Central Thrust (MCT), which is also defined as a zone of high ductile strain bounded by an upper thrust, MCT-II (Vaikrita Thrust) and a lower thrust, MCT-I (Munsiari Thrust).

Alaknanda River originates about 13 km upstream from the temple at Badrinath in the HHC rocks in the footwall of the South Tibetan Detachment (STD) Fault. After originating from the glacier, the river drains through a narrow and deep gorge of high-grade HHC rocks—grading from garnet-biotite-muscovite-schists from the base of the MCT to sillimanite-kyanite schists, psammitic gneiss, migmatites, and pervasive pegmatite veins and dykes of Malari leucogranite to the base of the STD Fault [47]. It crosses the MCT (8 km thick) near Helang, a small village about 12 km downstream from Joshimath, and suddenly, the gradient of the river decreases as it enters the carbonate formations of LHS rocks. The Yamuna originates near the Bandar Poonch peak in the Higher Himalaya, where it flows predominantly to the Higher Himalayan Crystallines of the Almora and the Ramgarh groups, having granodioritic to quartz-dioritic composition [33, 117]. The presence of calc-schists and marble with sulphide mineralization has been reported in the upstream area of Hanuman Chatti [21]. In the downstream, the river drains in the southwest direction, flowing through various litho-units in the outer and inner belts of the Lesser Himalayan [117]. The Tons River is the major tributary of the Yamuna in the Himalaya, which drains the western part of the Yamuna catchment. It originates in the glacier beyond Har-ki-dun and drains through crystallines and sedimentaries in its upstream and carbonates in the downstream reaches before joining the Yamuna at Kalsi, near Dehradun [21].

Climatic condition

The study area, Alaknanda, Bhagirathi, Yamuna, and Tons rivers (headwater) are situated in the Himalayan region of Uttarakhand State. The Alaknanda and Bhagirathi rivers confluence at Devprayag, where the Ganga river adopts its formal name. The Alaknanda and Bhagirathi river basins occupy an area of about 19.6 × 10³ km² up to Rishikesh, where the Ganga River enters the plain [17]. The Upper Ganga and Yamuna rivers and tributaries mainly depend on glacier melt and precipitation. The climate of the study area varies from alpine to subtropical. Precipitation is generally received through monsoon rainfall from June to September, and maximum snowfall occurs in December, January, and February. The average annual rainfall in the UGRS and UYRS varies between 1000 and 2500 mm [104]. The amount of rainfall is higher in the upper reaches than in the plains for both the valleys [99]. The annual temperature varies from 0 to 30 °C for the UGRS and UYRS [99].

Methodology

The total water samples (n = 34) were collected from the entire stretches of the UGRS and UYRS (Bhagirathi, Alaknanda, Ganga, Yamuna, and Tons) from the higher Himalaya up to the foothills (Fig. 1, Tables 1, and 2) followed by earlier set international protocols [21, 23, 111, 113] during the COVID-19 lockdown (May and June 2020). The sampling was done from the flowing river water near the bank of the river using a mug or bucket. After rinsing it twice with the sample water, the bottles (High-Density Polyethylene (HDPE) were filled to the top to avoid any headspace. After that, the bottle mouth was sealed with teflon tape. The samples were filtered onsite with a 0.22 µm nylon membrane filter in the field (Millipore®). and stored in bottles for analysis at 4–5 °C. The samples were transported to the Wadia Institute of Himalayan Geology (WIHG) for laboratory measurements and stored in bottles for analysis at 4–5 °C. Field photographs of some of the sample collection sites from the study area are given in (Fig. 2).

Fig. 2

Field photographs of some sample collection sites from the study area. A, B Bhagirathi River at Jakholi (BS-1). C Sample collected in High-density Poly Ethylene (HDPE) bottle. D Alaknanda at Devprayag (AS-2). E Ganga at Har Ki Pauri, Haridwar (G-5)

The in-situ measurement of physical parameters, namely pH, TDS (mg/l), and EC (μS/cm), were carried out as per the method described by [21, 111, 113] using multi-parameter electrodes and probes (Hach^®) with an average precision of ± 0.01 for EC and TDS, ± 0.05 for pH (Table S1 and S2 in the supplementary material). Bicarbonate (HCO₃⁻) was analyzed on pH based auto-titrator (Metrohm^®).

The major ions (Cl⁻, F⁻, SO₄²⁻, NO₃⁻, Na⁺, K⁺, Mg²⁺, Ca²⁺) were measured using Ion Chromatography (Dionex series ICS-5000). Primary standards (Dionex, seven anion standard-II, product no. 057590, and six cation-II standard product no. 046070) traceable to the National Institute of Standard Technology (NIST) were measured for calibration of ICS-5000 before the sample analysis followed the standard protocols [111,112,113]. Normalized Ions Charge Balance (NICB) is presented in Tables S1 and S2 (supplementary material). The data quality is checked by NICB within 5% [22, 111]. The trace elements and dissolved silica concentration were measured in unacidified filtered water samples using a Quadrupole Inductively Coupled Plasma Mass Spectrometer (ICP-MS). In water samples, measurement reproducibility was better than (± 5%) for trace elements dissolved in silica and major ions.

Stable isotopes of oxygen and hydrogen (δ¹⁸O and δD) were measured in the Laser Water Isotope Analyzer (LWIA, Picarro L2140-i wavelength-scanned cavity ring-down spectroscopy from Picarro Inc., CA, USA). An aliquot of an unacidified filtered water sample (a 0.22 micron-nylon membrane filter, Millipore^®) was filled in 2 ml glass vials and sealed with rubber caps. The vials with the sample were placed in a tray of a PAL auto-sampler connected to the Picarro L2140-ί. The (2 µl) samples were injected six times into the vaporization module of the analyzer by an auto-sampler at 110^○C before being sent into the laser cavity with ultra-pure nitrogen (99.999%) as the carrier gas. The primary standards of (GISP and VSMOW-2) provided by the Atomic Energy Agency (IAEA) were analyzed for the calibration of (LWIA). The reproducibility of measurements was 0.09‰ for δ¹⁸O and 0.5 ‰ for δD.

The Principal component analysis (PCA) is performed in the data sets for data transforming where the structure and the matrix of the data are often exposed once the boundaries of the PCA technique and detection of the scores-scores illustrations. PCA is a technique of Even though a statistical approach will process a collection of figures, whether analytically expressive or not, following this procedure, the new axes, termed principal components (PCs: F1, F2, F3, etc., using the scree plot), are selected based on a linear model Eq. (1).

$${F}_{jk}= {a}_{j1 }{k}_{2}+\dots +{a}_{jn}{k}_{n}$$

(1)

where F_jk = PC value is j for object k (the score for object j on component k), a_j1 = the loading of element one on component j, x_k1 = the length of the score for a variable one on item k, and n is the entire amount of variables observed.

The benefit of this technique is that the variance in the data set is mainly confined to the first few PCs, causing a reduction in the size of the multivariate matrix [126]. The physicochemical parameters (e.g., pH, EC, TDS, HCO₃⁻, Na⁺, K⁺, Mg²⁺, Ca²⁺, F⁻, Cl⁻, NO₃⁻, SO₄²⁻) were taken to evaluate the water pollution index (WPI) values of water samples based on recommended or standard permissible limits as suggested by the Bureau of Indian Standards [12, 119]. WPI is an integrated, weightage-free, and conventional indexing method that converts all input parameters into a one-value index to assess water quality [9, 42]. To compare the present results with earlier studies [16, 97], these parameters of the same basin were analyzed to estimate their WPI values. In addition, the WPI value of snow and ice from the Dokriani glacier was used to compare the results of different sources and assess the accuracy of this method. First, the pollution load (PL_i) of the ith parameters was calculated by adopting the following equation [42]:

$${PL}_{i}=1+\frac{({C}_{i}- {S}_{i})}{{S}_{i}}$$

(2)

where Ci = observed concentration of i^th parameter, Si = standard permissible limit of the i^th parameters. In the case of pH, the following equations were used to calculate the PL_i:

$${PL}_{i }= \frac{{C}_{i}-7}{{S}_{ia}-7}; if\, pH<7$$

(3)

here, the S_ia value is suggested to be the minimum acceptable pH value, i.e., 6.5.

$${PL}_{i }= \frac{{C}_{i}-7}{{S}_{ib}-7};If\, pH>7$$

(4)

where the S_ib value would be the maximum acceptable pH value, i.e., 8.5

Finally, the WPI was calculated using the following equation, as suggested by [42]:

$$WPI= \frac{1}{n} \sum_{i=1}^{n}{PL}_{i}$$

(5)

The relative WPI values were used to scale the water quality. The lower WPI value represents improved water quality, and the reverse is for relatively low water quality.

Results

Geochemical characteristics

A detailed description of the geographical distribution and altitude (m asl) of the collected samples is given in Tables 1 and 2. The major ions, trace elements concentrations, and isotopic compositions of samples are presented in Tables S1 and S2. The UGRS samples were slightly alkaline, with a pH between 7.6 and 8.2, whereas samples from the UYRS were more alkaline, with pH varying from 7.5 to 8.6 during the lockdown Phases 1 and 2. Correlation analysis also supported these results (Tables S5 and S6). The TDS concentration ranged from 47 to 63 mg/L. The dominance order of major ions from the UGRS and UYRS during the lockdown Phases 1 and 2 are as follows: UGRS (Alaknanda, Bhagirathi, and Ganga up to Haridwar)- HCO₃ (45%) > Ca (28%) > SO₄ (10.8%) > Mg (9.19%) > Na (3.67%) > K (1.66%) > Cl (0.96%) > NO₃ (0.43%) > F (0.17%), and UYRS (Yamuna, Tons up to Dakpathar, Vikas Nagar)- HCO₃ (54.1%) > Ca (22.1%) > Mg (8.3%) > SO₄ (6.5%) > Na (5.0%) > Cl (1.7%) > K (1.5%) > NO₃ (0.45%) > F (0.11%). The major ions data were plotted in a Piper diagram (Fig. 3), in which the cations and anions were shown in the bottom left and right triangles, respectively, and further presented in the central diamond to demonstrate the geochemical facies [86]. Most of the samples in the cation triangle were highlighted in the left corner, where Ca²⁺ concentrations were higher (28% for UGRS and 22.2% for UYRS, respectively). In contrast, major anions samples on the right side of the triangle showed elevated concentrations of HCO₃ (45% in the UGRS and 54% in the UYRS, respectively). The geochemical facies of both the UGRS and UYRS were characterized by Ca (25%)-Mg (9%)-HCO₃ (49%) type. The NO₃⁻ concentration was varied in the UGRS, with a maximum in Alaknanda (~ 23 μM) followed by Bhagirathi rivers (~ 20 on average). This concentration is lesser in the UYRS found near Kalsi before its confluence with the Tons River.

Fig. 3

Piper diagram showing the hydrochemical facies of the Upper Ganga and Yamuna River system

The concentrations of measured trace elements (in ppb), namely barium (Ba), strontium (Sr), lithium (Li), lead (Pb), nickel (Ni), molybdenum (Mo), copper (Cu), and cobalt (Co) are presented in Table S1 and S2. The present study observed no health hazard-causing element exceeding their permissible upper limit [12, 119]. In addition, the lowest concentrations of trace elements were observed in the UGRS and UYRS during the lockdown.

Gibbs plot and mixing models

Gibbs plot characterizes the controlling mechanisms in the riverine system, which explain the three end-members, i.e., precipitation, rock weathering, and role of evaporation from bottom to top based on the rations of [Na⁺/(Na⁺ + Ca²⁺) and Cl⁻/(Cl⁻ + HCO ₃)] (Fig. 4) [36]. The samples with low TDS and a high ratio of [Na⁺/(Na⁺ + Ca²⁺) and Cl⁻/(CI⁻ + HCO₃⁻] reflect the influence of precipitation. Samples with a medium concentration of TDS along with Na⁺/(Na⁺ + Ca²⁺) and Cl⁻/(Cl⁻ + HCO₃) ratios of < 0.5 showed the dominance of rock weathering that is presented on the middle left side of the plot. The concentration of TDS ≥ 300 mg/L and the ratio of Na⁺/(Na⁺ + Ca²⁺ or Cl⁻/(Cl⁻ + HCO₃) up to 1.0 implied evaporation or evapo-crystallization as a dominant source of ions. The mixing model prepared using the current data set (Fig. 5) revealed that different groups of major ions could result in different rock weathering rather than other sources.

Fig. 4

Gibbs plot showing the Ratios of a (Na⁺/(Na⁺  + Ca²⁺) and b (Cl⁻/(Cl⁻ + HCO₃⁻) as a function of TDS in the Upper Ganga and Yamuna river system

Fig. 5

a Mixing diagram of [Ca/Na Vs. HCO3/Na (background data is taken from [71], b mixing chart of Ca/Na Vs. Mg/Na (background data obtained from Gaillardet et al. [31]

Assessment of water pollution index (WPI)

The WPI index based on the standard permissible limits in surface water is recommended by WHO [119] and BIS [12]. In surface waters, we categorized the WPI into four types to infer the water quality as follows: excellent (WPI < 0.5), good (0.5 < WPI < 0.75), moderate (0.75 < WPI < 1), and highly polluted (WPI > 1). The WPI of water samples collected during the lockdown from the UGRS and UYRS considerably varied from 0.10 to 0.61. The WPI in the studied riverine systems was lower than that reported in a previous study by Chakrapani [16]. Before the confluence with the Yamuna, the Tons River at Khadar had the highest WPI value (0.28–0.32). In contrast, the WPI values for other major river basins (Alaknanda, Bhagirathi, and Ganga) were nearly the same (average ~ 0.14) (Table 3). Notably, the WPI values were the lowest for the snow (~ 0.002) and ice (~ 0.04) of the Dokriani glacier, indicating the freshwater resources feeding the UGRS and UYRS. Furthermore, the change in the percentage of water quality between 2005 and the lockdown period in 2020 was maximum for the Ganga River (Table 3).

Table 3 Changes in water quality of the Alaknanda, Bhagirathi, Ganga, and Tons during the Pre-lockdown and lockdown phase

4.4. Stable isotopes (δ¹⁸O and δD) systematic

The UGRS and UYRS are mainly fed annually by fresh snow/glacier-melt waters [22, 49, 58, 67, 92, 118] (Fig. 1). We analyzed stable hydrogen (δD) and oxygen (δ¹⁸O) isotope ratios of studied riverine systems during the COVID-19 pandemic lockdown Phase 1 and 2. The δD and δ¹⁸O data are presented in Tables S1 and S2, whereas the slope, intercept, and deuterium excess data are shown in Table 4.

Table 4 Slope, Intercept, and deuterium excess calculated based on stable isotopes of oxygen (δ¹⁸O_vsmow) and hydrogen (δD _vsmow) of the upper Ganga and Yamuna River systems and other riverine systems of northwest Himalaya, India

Samples from the UGRS had δD values ranging from − 76.16 to − 59.53 with an average of − 65.91 ± 0.1 and δ¹⁸O values from − 10.16 to − 8.55 with an average of − 9.40 ± 0.02; whereas, samples from the UYRS had δD values varying from − 59.16 to − 43.15 with an average of − 52.9 ± 0.1 and δ¹⁸0 values from − 8.76 to − 6.20 with an average of − 7.65 ± 0.02 during the lockdown Phase 1 and 2.

The correlation plot between δD and δ¹⁸O isotope ratios is shown in Fig. 6. The statistical summary, including slope, intercept, p-value, and r² of the stable isotope data, along with previously published data, is given in Table 4. The Global Meteoric Water Line (GMWL) [96] and Indian Summer Monsoon Line (ISML) [118] are also plotted for comparison with their best-fit lines (Fig. 6 and Table 4).

Fig. 6

XY-Plot of stable Isotopic (δD_vsmow Vs. δ¹⁸O_vsmow) systematics of Upper Ganga River and Upper Yamuna River System

Isotopic values in the UGRS were nearly close (slope:7.45 ± 0.23, intercept: − 8.15 ± 2.0) to the observations reported by Ramesh and Sarin [92]. Each tributary/stream of the UGRS and UYRS reflects a distinct isotopic signal due to changes in elevations in its catchment [49]. The samples collected during the lockdown Phase 1 and 2 displayed the maximum contribution of meltwater in the UGRS and UYRS.

The Deuterium excess (d-excess) has been practiced for a long time as one of the diagnostic tools to estimate the contribution of water vapor from different sources for a particular location on the globe [48, 67, 84, 118]. d-excess is described by an equation, d = δD-8* δ¹⁸O [24], which indicates the deviation in a set of data points from a line with slope 8 in δD vs. δ¹⁸O through a simple regression equation (Table 4). The UGRS and UYRS had a d-excess ranging from 6.1 to 9.4 with a mean of 8.0 ± 1.9.

Multivariate metrics

To validate the geochemical analysis, we performed a multivariate analysis that included the PCA in the riverine waters. The PCA is generally calculated in the surface water to evaluate the interrelationship within the available geochemical data set and simplify the complex data [62]. The PCA values of water samples from the UGRS and UYRS, including the Tons river basin, during lockdown Phase 1 and 2 (May and June 2020) are shown in Table 5. From scree plots, the inflection point (eigenvalue > 1) for the UGRS starts at the principal component (PC) sequence number four. In contrast, it occurs at the PC sequence number two for the UYRS (Fig. 7). Any factor with an Eigenvalue > 1 is considered more significant [63]. However, to maintain the differences in inflection points between the UGRS and UYRS, we have extracted three PCs (F1, F2, and F3: Eigenvalues > 1), which are more significant as compared to others and can be utilized to assess the dominant hydro-geochemical processes and their variability (Table 5). Cumulative variance in the UGRS was almost the same in May and June but varied substantially in the UYRS during the same months. A higher eigenvalue indicates higher PC variability, which is evident in the UYRS from May to June. Factor loadings measure the closeness between the input variables and the PC. The first three PCs (F1, F2, F3) account for 83.06% in UGRS, 90.28% in the UYRS of the total variance during May, and 80.80% in the UGRS, and 95.56 in the UYRS of the total variance during June.

Table 5 Loadings of experimental variables on principal components analysis (PCA) for the Upper Ganga River System (UGRS) and Upper Yamuna River System (UYRS) in May and June 2020

Fig. 7

Scree plots of principal components showing Eigenvalues and cumulative (%) for a Upper Ganga River Basin in May; b Upper Ganga River Basin in June; c Upper Yamuna River Basin in May; d Upper Yamuna River Basin in June

In the UGRS (May), the concentrations of pH, F⁻, and NO₃⁻ showed negative loadings for the first principal component (F1), while concentrations of all other variables had positive factor loadings, with higher loadings for EC, TDS, Mg²⁺, and Ca²⁺. A higher value of EC is attributed to a strong linkage with major cations like Ca²⁺, Na⁺, and Mg²⁺ [9]. For the second principal component (F2), the concentration of many variables such as pH, EC, TDS, HCO³⁻ and Ca²⁺ had negative loadings, whereas Cl⁻, SO₄²⁺, Na⁺, and K⁺ showed high positive loadings, contributing more to the water sample. In the third principal component (F3), no higher positive loadings (> 0.8) were found.

In the UGRS (June), the scenario of factor loadings of PCs was quite different compared to May. For F1, the concentration of Na⁺ (~ 0.89) had high positive loadings. The concentration of all physical variables and HCO₃⁻ had negative loadings. Other remaining variables had lower (< 0.5) or moderate positive loadings (0.5 > F1 > 0.8) (Table 5). This indicates that the water quality of the riverine system may be attributed to less anthropogenic and industrial waste input in the river basin during the COVID-19 lockdown. For F2, no variables had high positive loadings (> 0.8), while the concentration of some physio-chemical variables such as temperature, TDS, EC, Mg²⁺, and SO₄²⁻ showed moderate positive loadings. The pH, F-, and NO₃⁻ concentrations were characterized by negative loadings for F2. The PC (F3) is not significant for assessing concentrations of variables due to low and moderate loadings, except HCO₃⁻.

In the UYRS (May), the first three principal components together account for 90.28% and 95.55% of the total variance of the data set. The concentration of variables such as pH, F⁻, and NO₃⁻ had negative loadings similar to the UGRS (May) load for F1. The HCO₃⁻, K⁺, Mg²⁺, and Ca²⁺ showed high positive loadings, whereas the concentrations of SO₄²⁻, Cl⁻, F⁻, and TDS had moderate positive loadings. For F2, no variables showed high positive loadings. However, most of the physio-chemical variables indicated moderate positive loadings. For F3, concentrations of Na⁺ had high positive loadings.

Discussion

Geochemical source identification and implications against water quality improvement

The COVID-19 pandemic enforced partial and total lockdowns worldwide and provided an unprecedented opportunity to test how lockdown has improved the water quality by reducing the contamination in the water resources. The rock-water interactions resulted in the dissolved ionic compositions in the riverine system. Furthermore, the major ions are added from many sources like solid/liquid precipitation, dust/aerosol, and anthropogenic activities. The congruent and incongruent dissolutions of rock weathering dominate the water composition [16, 57]. In contrast to the pre-lockdown period, we observed increased pH (~ 8 on average) in the post-lockdown, showing a general improvement in the water quality of UGRS and UYRS during the lockdown. Relative increases in pH, i.e., a more alkaline nature of the riverine water in a specific time, may result from reduced anthropogenic activities [28, 129]. The water quality of water bodies, including the Ganga, Yamuna, Mandakini, Alaknanda, Bhagirathi, and Gaula, and the Naini and Bhimtal lakes, were examined based on physical parameters such as pH, TDS, Hardness, etc. [74] during Covid-19 lockdown and observed that all improved due to decreased human activity (tourist, religious activities, rafting, and other sports), as well as a decrease in the flow of industrial effluents. The TDS exhibits the lowest concentrations during the lockdown and is equivalent to the glacier melt [112], which suggests the dominance of major cations from natural weathering sources [28]. Similar variation was observed for EC in both river systems. The correlation matrix showed a strong positive relation between EC and TDS at a 0.05 level of significance (Table S4 and S5 in the supplementary material). EC is directly proportional to the TDS content in the water [15]. No study examined the water quality based on geochemical and isotopic chemistry during the COVID-19 lockdown in this study area. However, based on the physical and other parameters such as dissolved oxygen (DO), biochemical oxygen demand (BOD), total coliforms (TC), and pH, the Central Pollution Control Board of India and the Indian Institute of Technology Roorkee suggested that the water quality of the Ganga and Yamuna River has improved by 40–50% [20]. Khan et al. [53] also reported improved water quality during the COVID-19 lockdown, and industrial wastewater runoff was significantly stopped. A similar study was carried out by Chakraborty et al. [15] in the Damodar River during the COVID-19 lockdown and pre-lockdown phases and found that water samples were substantially contaminated during pre-lockdown. As a result of the halting of the heavy metal industries over three months, 90.90% of water samples were upgraded to good quality, whereas 9.10 percent of samples were moderately polluted.

HCO₃⁻ was the principal constituent compared to the other ions in both river systems. The sources of various dissolved elements in the riverine system [57] are given in Table S3. Chemical constituents (major and trace elements) in the riverine systems have various sources, including sea salts with chemical, physical, and biological processes carried by atmospheric circulation, which are finally deposited through solid and liquid precipitation. These components undergo different rock weathering, evaporation, and anthropogenic activities [16, 21, 82, 125]. Geochemical facies have described the dominance of calcium, influenced by local sources [51, 102], and HCO₃⁻ dominated as a major anion, followed by the SO₄²⁻ and Cl⁻. SO₄²⁻ and Cl⁻ had lower concentrations in both riverine systems, indicating a negligible contribution from the anthropogenic sources [122]. The concentration of Na⁺ and K⁺ is reduced during lockdown, particularly in the UGRS, compared to the UYRS. However, their concentration is greater than the limiting value in some places, as described by [12]. These major ions showed the presence of urban wastewater, which is increased during lockdown phases near populated areas. Cl⁻ and NO₃⁻ enrichments are attributed to anthropogenic activity caused by agricultural waste, rural land uses, and substantial population growth [1, 94].

Six sub-categories are available in the diamond-shaped piper diagram [66]. A detailed description of major ions chemistry of riverine systems along with surface and its controlling factors was described by Gibbs [36]. Few samples from the UYRS fall in the evaporation category, representing the effect of local environmental conditions [87]. In contrast, samples from the UGRS fall in the rock-weathering dominant category (Fig. 4).

The mixing model (Fig. 5) describes the rock-water interaction as silicate weathering that produces mainly sodium, potassium, calcium, magnesium, silica, and bicarbonate in the riverine system. Whereas carbonate weathering produces calcium, magnesium, and bicarbonate, and the dissolution of evaporates, chloride, and nitrate [16, 21, 22, 31, 40, 51, 71, 125]. The traces of SO₄²⁻, Cl⁻, and Na⁺ could also be found in the riverine water, obtained from the dissolution of halite, pyrite, gypsum, and anhydrite [83, 87]. Data collected during the lockdown period from the UGRS and UYRS displayed the dominance of silicate weathering with little influence of carbonate weathering.

The WPI exhibits that the water quality started to be contaminated from the beginning of the twentieth century and continued to increase until recent years due to anthropogenic activities, including untreated industrial effluents and urban sewage near the river basins. The WPI values of water samples collected by Sarin et al. [97] are comparable to those of the lockdown phase for the Bhagirathi and the Ganga rivers (Table 3), suggesting that these river systems rejuvenated the water quality to its initial stage, i.e., around 1990. The water of Alaknanda could not be recycled because the WPI value for Alaknanda River was still higher (during lockdown 2020) than that of the water sample taken around 1990. COVID-19 lockdown provides clear information on environmental deterioration triggered by several anthropogenic actions in the past two decades. The lockdown Phases 1 and 2 has significantly reduced pollution due to the temporary closing of the industries and other pollution-deriving agents. However, industrial activities are essential for public livelihood, so the local governing bodies should practice several awareness programs to improve river water quality and the environment.

The isotopic values of both the basins fall on the Global Meteoric water line (GMWL) and the Glacial meltwater line of the Garhwal, northwest Himalaya [118]. However, in this study, the samples collected during the pre-monsoon season showed their trend away from the Indian Summer Monsoon Line (ISML) (Fig. 6), which may result from the westerlies' influence. The stable isotopes of δ18O and δD were depleted in the UGRS, showing that isotopic ratios of δ18O and δD in the basin are more glacierized compared to enriched values of isotopic ratios of δ18O and δD in the UYRS. The Yamuna River (slope: 5.73 ± 0.59 and intercept: − 8.15 ± 4) originates from the Yamunotri Glacier near Bandarpunch Peak [90], which is smaller than the Gangotri Glacier. The Bhagirathi River emerges from the second-largest glacier in India, and the Alakananda River contributes to this, which originates from the Satopanth glacier. Hence, the contribution of glacial melt/ fresh snow in the UYRS is also far less than that of the UGRS.

In May and June 2020, we noticed a lower d-excess in the UGRS and UYRS, less comparable to prior research in the same region [21, 92] (Table 4). The observed low d-excess indicated the maximum contribution of glacial melt/fresh snowmelt in the UGRS and UYRS riverine water during the sample collection time.

Statistical substantiation to water quality inferences

In the PCA, the positive loading displays the increased contribution of the variables with the increasing load in the dimensions, while negative loadings show less contribution of the parameters [63]. In our study, the first principal component had the negative factor loading for an anthropogenic element like NO₃ during the lockdown, suggesting its less contribution to UGRS and UYRS river waters [65]. At the same time, positive factor loadings of some cations like HCO₃⁻, Ca²⁺, and Mg²⁺ reflect their strong association with the sources of origin. Positive loadings of the Na⁺ indicated that the ionic enrichment factor was dominant due to urban waste [65].

The PC loadings (F1 and F2) relating to the source and water quality are presented in Fig. 8 using the bi-plot for both the UGRS and UYRS during the lockdown. The size of the loadings on the individual PC(s) is represented by a vector drawn from the basis of a set of loadings coordinates, which characterize the sources from which they were derived or related [3, 55]. In the UGRS (May), cluster-1 containing six parameters (temperature, EC, TDS, HCO₃⁻, Ca²⁺, and Mg²⁺) lies at the rightmost position of the bi-plot (Fig. 8). Their position inferred that they contributed maximum variance and were strongly associated with F1, i.e., their response in the river water sample was strong for May. In the UGRS (June sampling), their relative position from F1 was shifted mostly along with F2, resulting in their relatively weak contribution to the water sample.

Fig. 8

Bi-plots of PCs 1 and 2 for pattern identification between physicochemical parameters for a Upper Ganga River System in May; b Upper Ganga River System in June; c Upper Yamuna River System in May; d Upper Yamuna River System in June 2020

The HCO₃⁻ position is near the origin of the bi-plot, showing the most negligible contribution in the water sample in May. However, chemically, they are consistent with the typical components of rock minerals. The NO₃⁻, which can be derived from organic matter and oxidation of ammonia, largely deviated from cluster 1. Hence, the geological processes consistent with cluster-1 are likely through rock weathering [78, 124]. The UGRS (June) had less variance for F1, contributing weakly to the river basin. A similar contribution was found for Fe²⁺ in June. Cluster-2 containing Na⁺, Cl⁻, SO₄²⁻, and K⁺ are characterized by positive scores on both F1 and F2. Since two sets of loading vectors (i.e., cluster-1 and 2) are almost 90°, they are uncorrelated from their source of origin. They provide a substantial contribution with F2 in May, which is reduced in June because they are shifted in the middle portion of F1 and F2. Cluster-3 is not well-defined in the bi-plot. pH shifted significantly from the origin in June, showing a significant variance along with F1 compared to the UGRS (May). In the UGRS (June), these clusters no longer remain the same, and their pattern is still somewhat unclear, as it is still hard to separate any trends within each group.

In the UYRS (May), cluster-1 containing Mg²⁺, Ca²⁺, K⁺, HCO_3-, SO₄²⁻, and Cl⁻ had positive scores for both F1 and F2. These parameters are strongly associated and positively correlated with their source of origin. They had positive loadings, indicating increased contribution to the river water in June. Cluster-2 containing pH, NO₃⁻, TDS, EC, and temperature showed a relatively low correlation with F1, and their clustering is not compact (Fig. 8). In the UYRS (June), the concentration of all variables was inconsistent because their factor scores for F1 and F2 were substantially changed. Only one well-defined cluster with positive scores on F1 and F2 could be found. The NO₃⁻ and Fe²⁺ showed a different source of origin, similar to the UGRS (May). In the UYRS (June), factor loading for cluster-1 was almost the same (Fig. 8).

Results from the correlation matrix demonstrate that TDS strongly correlates with EC during the study period (Table S4 and S5 in the supplementary material). Most cations (K⁺, Ca²⁺, Mg²⁺) strongly correlated with HCO₃, exhibiting their origin from a similar source [59, 88]. In the UGRS, Cl⁻ showed a good correlation with SO₄²⁻, Na⁺, and K⁺, while this relationship was no more substantial for UYRS in June. The NO₃ concentration in both riverine systems had no significant correlation with chemical constituents, reflecting the different sources of its origin. A similar correlation was found with F⁻ in both river systems. The strongest correlation among the chemical components of river waters was observed in UYRS during June (Table S5). Analysis of the correlation matrix is an effective method to describe the results obtained from chemical analysis [46, 63]. This analysis substantially supported our results derived from physio-chemical analysis.

Overall, multivariate analysis suggests that the source of physio-chemical parameters, such as anthropogenic activity, may be substantially reduced in the river basin in June. These findings illustrate that improvements in water quality during lockdown phases reflect the remarkable environmental impacts of anthropogenic activities. The present study offers a wake-up call to make development pathways environment-friendly.

Conclusions

Our findings draw the following conclusions and recommendations:

• Samples from the UGRS are slightly alkaline, with a pH ranging from 7.6 to 8.2, whereas the UYRS samples are more alkaline, ranging from 7.5 to 8.6.

• The TDS concentration in the UGRS and UYRS was the lowest-ever equivalent to the glacier melt, indicating a little mixing of effluent material from external sources.

• Bicarbonate (HCO3-) is one of the significant elements in both the river basins, followed by other cations and anions, indicating that the geo-genic process was dominant during the lockdown.

• Data obtained during the lockdown period from the Upper Ganga and Yamuna basins clearly show the dominance of silicate weathering with little influence of carbonate weathering.

• The Lockdown phase has significantly reduced the pollution level in water due to the temporary closing of the industries and other pollution-causing agents. Further, we found negligible hazardous trace elements (As, Pb, Ni, etc.) in the UGRS and UYRS.

• PCA evinces the negative factor loading for an anthropogenic element during the lockdown, signifying a reduction in the intake of pollution-causing elements to both UGRS and UYRS.

• The decreased isotope ratios of δ18O and δD suggest a more significant contribution from the glaciated areas in the basin.

• The stable isotopic values (δ18O and δD) are depleted at the headwaters of streams and tributaries and substantially enriched throughout the river basin at lower altitudes.

• The UGRS and UYRS were cleaner during the lockdown than during previous cleaning campaigns, which had cost significant money but never yielded satisfactory results.

• This lockdown shows that we are responsible for preserving nature's purity via sustainable development and resource protection.