Spatial and seasonal variability in the water chemistry of Kabar Tal wetland (Ramsar site), Bihar, India: multivariate statistical techniques and GIS approach

Freshwater is one of the most important constituents of life. Human interactions with water resources involve fresh streams, marshes, lakes, and groundwater. The rapid population growth, agriculture, and industrialization have forced environmentalists to determine the chemical, physical and biological characteristics of natural water resources (Regina & Nabi 2003). Wetlands are most productive ecosystems and play a very important role in the biogeochemical cycling of nutrients (Sánchez-Carrillo et al. 2014). In wetlands, nutrient dynamics is generally influenced by many processes like transport, transformation, storage, release, and removal of the inflow regime of wetlands (Shardendu et al. 2012). Wetlands function as water purification systems, helps in flood control, acts as sinks for pollutants as well as for organic carbon and therefore performs a vital role as climate stabilizers (Suhani et al. 2020). Wetlands are under increasing threat due to both point sources such as municipal and industrial wastewater and non-point sources like urban and agricultural run-off. Nitrate and phosphate are the major nutrients in the aquatic ecosystem which promote the productivity in an aquatic ecosystem (Weigelhofer et al. 2018). Phosphorous is an essential promoter of aquatic plant growth and a limiting nutrient for eutrophication in the lake ecosystem (Conley et al. 2009).

The use of various statistical techniques viz., principal component analysis (PCA), and discriminant analysis (DA) has been used by several researchers (Dash et al. 2018; Kaushik et al. 2021) to identify those parameters which govern the water quality and determine the geochemical processes. It helps in sorting out samples having similar properties or a particular group or cluster (may be seasonal or restricted to a particular site) (Singh et al. 2004). Geographic information system (GIS) is one of the best intensive tools for processing, analyzing, storing, and displaying spatial distribution of data which improves the decision making ability of the conservation experts. Similar studies using the GIS tool has been conducted on Deepor Beel, by Dash & Kalamdhad (2021), for spatial mapping of water quality parameters.

Kabar Tal (KT) is one of the most important floodplain wetlands of Bihar (India) as it serves as a habitat for various migratory birds and plays important ecological functions for the regional hydrological balance. Recently, KT wetland has been designated as Ramsar sites (no. 2436) making it ‘Wetlands of international importance’ (Ramsar 2020). The wetland is shrinking at an alarming rate due to encroachment for agricultural practices (Ghosh et al. 2004; WISA 2016). Besides, it is also being used as a domestic wastewater drainage system by surrounding villages (Ambastha et al. 2007) resulting in the deterioration of water quality. Many researchers (Singh & Roy 1990; Siddiqui & Ramakrishna 2002; Ranjan & Kumari 2018) consider KT wetland to be at the initial stage of eutrophication.

The physicochemical characteristics of water alter seasonally; therefore it is very important to highlight its influence on the hydrochemistry of the wetland. Very few studies related to water quality have been carried on KT wetland related to water quality, which includes Siddiqui & Ramakrishna (2002), Singh & Jayakumar (2016), Ranjan et al. (2017), Ranjan & Kumari (2018) and Singh et al. (2020). However, all earlier studies are based on either one-time sampling during winter season (e.g., Ranjan et al. 2017) or a very small sample size. In addition, previous findings by Singh & Jayakumar 2016 and Singh et al. 2020, etc; have not focused on hydrochemical characterization and identification of water types (such as by using piper plot, etc.) and have also not evaluated the water quality for irrigation purpose.

Therefore, the present study was undertaken to assess the hydrochemistry of KT wetland to understand the influence of seasons (i.e., summer, monsoon, and winter) on water chemistry and also to delineate the sources of major ions and nutrients in the water. The literature survey shows that this study would be the first to address sources and processes along with influence of spatial and temporal variability in surface water chemistry of KT wetland.

KT wetland (Figure 1) located at 86°05′E to 86°09′E, 25°30′N to 25°32′N, also known as Kabar Lake Bird Sanctuary or Kanwar Jheel is situated in the Begusarai district of Bihar, India. This wetland lies in the middle Gangetic plains, with alluvial landscape and has an elevation of 37–39 m above mean sea level (MSL). It is the largest wetland among the various wetlands found in the Kosi-Ganga interfluve in the northern plains (Singh & Sinha 2019). The site is one among 18 wetlands within an extensive floodplain complex; it floods during the monsoon season to a depth of 1.5 meters. It experiences three climatic seasons, i.e., summer (March to mid-June), monsoon (mid-June to October), and winter (November to February), and characterized by tropical monsoon type climate with an average annual rainfall of 1,200 mm. The average temperature ranges from 25 °C to 38 °C in summer and 8 °C to 25 °C in winter.

The lake during dry season covers an area of 2600 hectares whereas; in monsoon, it expands up to an area of 7,400 ha. The Union Government of India has notified KT wetland as a wetland of national importance and also been declared as an IBA (Important Bird Area) site of Bihar and a Ramsar site. The wetland supports many waterfowl in the Gangetic plain along with a variety of species of migratory and resident avifauna; and an important stopover along the Central Asian Flyway, with 58 migratory water birds. Apart from this, the wetland is also a good source of fish varieties, paddy cultivation, and other resources like fodder (Irfan & Pasha 2013). Many varieties of aquatic flora is found in KT wetland, which includes Pistia stratiotes, Hydrilla verticillata, Eleocharis plantaginea, Eichhornia crassipes, Myriophyllum spicatum, Zostera marina, Egeria densa, Lemna obscura, Phragmites australis, Typha (Shardendu et al. 2012). KT inhabits three critically endangered species of vultures, namely the red-headed vulture (Sarcogyps calvus), white-rumped vulture (Gyps bengalensis) and Indian vulture (Gyps indicus), and two water birds, the sociable lapwing (Vanellus gregarius) and Baer's pochard (Aythya baeri) (Ramsar 2020).

Overall 83 water samples, 29 samples each in summer (May 2015), and monsoon (July 2015), and 25 samples in winter (February 2016) were collected. Sampling points were fixed by using the portable Global Positioning System (GPS, Garmin Etrex-20). Water samples were collected based on water availability and possible sites for sampling. Samples were stored in an icebox and brought to the laboratory for analysis. The pH, electrical conductivity (EC), total dissolved solids (TDS) and temperature were recorded on the field using a Thermo-Orion portable meter (Thermo Scientific Orion Star A329). Sodium and potassium were estimated by flame photometer; however, calcium and magnesium were analyzed by the EDTA complexometric titration method. Sulphate, nitrate, and phosphate were estimated by spectrophotometric method, chloride was analyzed by argentometric method and HCO₃⁻ by titrimetric methods (APHA 2012). The analytical data quality was controlled by standardization, procedure blank measurements. All observations were recorded in duplicate and average values were reported. The data used for doing all statistical analysis, as well as preparing contour map, were based on 29 samples each from summer and monsoon and 25 samples from winter season.

Xi and Yi represent two different parameters and N = the total number of observations.

PCA overall is a mathematical approach which does not depend upon assumptions (Mazlum et al. 1999) and is used to decrease the number of the sample without losing on the original sample information (Helena et al. 2000; Singh et al. 2017). Eigen value plays a significant role in importing the variations of these observed data (Praus 2007). In the present study, principal component (PC) having Eigen value >1 is considered. The factor axis was varimax-rotated to extract and minimize the variations among the variables for each factor.

• i represents the number of groups (G);

• ki = constant inherent to each group;

• n = number of parameters used for classification of every parameter; and

• wj = the weighting coefficient assigned by DA to a particular parameter (pj).

PCA, DA and correlation matrix for water samples was carried out by using ‘Statistical Package for Social Sciences (SPSS), version-10.0’. Microsoft Excel 2013 was used for hydrochemical plotting.

In this study, the contour map is prepared through interpolation, by Inverse Distance Weighted (IDW) method. IDW is the method, which implements the assumption that the things that are closer to one another are more alike than those farther apart. The data obtained through laboratory analysis were tabulated and interpolated using GIS tools. ArcGis 10.1 software was used for preparing contour map.

Classification of surface water quality (Table 1) for irrigation purpose is done on the basis of Na% [Sodium % = [(Na + K) * 100]/[Ca + Mg + Na + K]], Sodium adsorption Ratio (SAR) = Na/[(Ca + Mg)/2]^1/2), Residual Sodium Carbonate (RSC) [RSC = (CO₃ + HCO₃)−(Ca + Mg)], and Salinity Hazard (EC) (μS/cm) (Richards 1954).

Sodium % <20; SAR <10; RSC < 1.25, indicates that the waters of the wetland is excellent for irrigation for all seasons, if RSC is >2.5 it indicates increased salt content which might lead to chocking of soil pores and thereby reduce air passage (Inayathulla & Paul 2013).

The change in temperature at different sampling points of KT wetland was analyzed during the summer, monsoon, and winter seasons are shown in Table 2. The results show that there is not much variation in temperature between summer (28.5 °C–34 °C) and monsoon (28 °C–35 °C) however, during winter temperature ranges between 18 °C and 23.5 °C. Temperature in the wetland varies at different segments with Northern part showing high temperature (∼33 °C, ∼23 °C) than southern part (∼29 °C, 18 °C) during summer and winter, respectively. During monsoon (∼30 °C) it is almost uniform showing slight variation on the periphery of the northern, southern, and eastern region. It is observed that the temperature during winter is higher in the northeastern and northwestern regions (∼23 °C) whereas the southern part has lower temperature (18 °C) (Figure 2(a)). The pH shows spatial variation during summer. It is highest at the southern side (∼7.2) and lowest in the northwestern (∼6.5) and northeastern region (∼6.8). In monsoon, pH is highest at the south eastern side (∼6.8) and lowest at the northern side (∼6.5) whereas during winters, the pH is highest at the northern side (∼7.97) as shown in Figure 2(b).

The seasonal variation of temperature in the water is due to the variability in water depth and its quantity (Ling et al. 2017). The temperature fluctuation of KT wetland at various seasons is less significant in areas where volume of water is greater.

The average value of pH during winter (7.9) is slightly more than in summer and monsoon, indicating the neutral to alkaline nature of water. pH value is affected by biological activities like photosynthetic and respiration rates in wetland (Weisse & Stadler 2006; Shah et al. 2019). Therefore, higher photosynthesis uptakes more CO₂ and thus increases pH in winter. The low value of pH in summer could be due to the decomposition of accumulated organic matter, and its biological oxidation, releasing CO₂ which in turn reduces the pH (Langmuir 1997). However, a low value in monsoon may be due to high turbidity of water and elevated temperatures which leads to less photosynthesis and accumulation of more free CO₂ (Adebisi 1981; Edoreh et al. 2019). A low rate of photosynthesis can also cause a decline in pH due to inefficiency of light-dependent reactions on a cloudy day during monsoon as temperatures hardly affect the light-dependent reactions of photosynthesis (Marra & Heinemann 1982; Wu et al. 2017). Additionally, transport of humic and fulvic materials in colloidal suspension may also contribute to lower pH values (Langmuir 1997; Weng et al. 2002).

The electrical conductivity (EC) of water is a measure of dissolved ions which originates from the weathering processes and decaying plant matter (Sarwar & Majid 1997; Ray et al. 2021), and also input of organic and inorganic waste (Wright 1982; Dey & Dey 2015). During the present analysis, EC was found to be significantly higher in the summer (371 μS/cm) than monsoon and winter. The higher value of EC during summer shows the ions getting concentrated due to lesser water availability and temperature induced rapid evaporation (Gupta & Paul 2013; Singh et al. 2020). However, it was noticed to be least in monsoon due to dilutions of the wetland water through rainfall and runoff from its adjoining areas like agricultural lands and water bodies (Ambastha et al. 2007). Electrical conductivity during summer is maximum in the northeastern region by up to 850 (μS/cm), reflecting the effect of an increase in temperature over northeastern part (Figure 2(c)); and lowest at the southern part (∼250 μS/cm). EC shows the variation (from 110 to 360 μS/cm) in winter as shown in Figure 2(c) whereas, during monsoon, EC is uniform and shows a slight increase in the northern region (∼290 μS/cm) and reduction at the southern region (∼60 μS/cm). This may be the case due to an increase in runoff at the northern region from the agricultural fields.

The seasonal variations shows (Table 2) that anions follows the trend SO₄²⁻ > Cl⁻ > HCO₃⁻ > NO₃⁻ > PO₄³⁻ in monsoon and winter, whereas in summer its trend is Cl⁻ > SO₄²⁻ > HCO₃⁻ > NO₃⁻ > PO₄³. Chloride (Cl⁻) occurs naturally in all types of water due to its high solubility. The higher concentration of chloride during summer may be associated with a high rate of evaporation as well as frequent run-off loaded with contaminated water from the surrounding areas. Sources of chloride in surface water are both natural and anthropogenic, such as weathering, leaching from sedimentary rocks and soils, the use of inorganic fertilizers, animal feeds, industrial effluents, irrigation drainage, and organic wastes of animal origin. Chloride enters surface water through both natural (weathering, sedimentary rock and soil leaching) and anthropogenic (inorganic fertilizers, animal feed, agricultural waste, irrigation runoff) sources (Venkatasubramani & Meenambal 2007). The chloride distribution in the study area is found to be highest at the western (234 mg/L) and northern (139 mg/L) periphery of KT wetland during summer. During monsoon, a single patch of high concentration of chloride (149 mg/L) is observed at the northeastern region whereas the concentration of chloride is relatively high in patches during winter as shown in Supplementary Figure S1(a).

The concentration of HCO₃⁻ is higher (17.4 mg/L) in winter than summer and monsoon (11.2 mg/L). The source of HCO₃⁻ in the aquatic ecosystem is through the dissolution of gases, carbonate equilibrium, and weathering of carbonate rocks (Shah et al. 2019). The increase in HCO₃⁻ concentration in winter can be attributed to seasonal variability in the chemical weathering of carbonate and silicate minerals (Tipper et al. 2006). The higher presence of HCO₃⁻ pertains generally to the dissolution of minerals like calcite and dolomite (Cai et al. 2007). In the winter season, higher dissolution of CO₂, due to the inverse relationship of dissolved CO₂ and temperature may induce a higher rate of carbonate weathering (Langmuir 1997). HCO₃⁻ is found to be higher at two patches at the northeastern and southern boundary of Kabar Tal during summer and monsoon; it has been vividly represented on the contour map, shown in Supplementary Figure S1(b). During winter, the bicarbonate concentration ranges from 12 to 24 (mg/L) but at the northeastern side wavering towards the center, it is 5.9 (mg/L) at a single patch.

Phosphorus is one of the key nutrients which limit primary productivity in many freshwater ecosystems, whereas nitrogen is a common limiting nutrient in marine ecosystems (Cloern 2001). However, in some freshwater environments, particularly in the tropics and subtropics, N is the primary limiting nutrient for phytoplankton production, owing to excessive P load and long growing seasons (e.g., Yang et al. 2008). The nitrogen in water is generally present in forms like nitrate, nitrite, ammonia and organic form sources are urea, amino acids, etc. The most significant source of NO₃⁻ in an aquatic ecosystem such as wetland and lake is the biological oxidation of nitrogen-rich organic matter such as domestic sewage, agricultural runoff, and industrial effluents (Vrzel et al. 2016). The high concentration of nitrates is beneficial for irrigation, but it causes eutrophication which promotes the growth of algae and macrophytes (Trivedy & Goel 1984). NO₃⁻ concentration is greater in summer (0.5 ± 1.1 mg/L) than in monsoon (0.3 ± 0.6 mg/L) and winter (0.3 ± 0.8 mg/L. Similar trends in nitrate were also observed in many aquatic ecosystems (Garg et al. 2006; Sinha & Biswas 2011; Prabhahar et al. 2012; Singh & Jayakumar 2016; Singh & Deepika 2017). Nitrate concentration lies between 0.2 and 2.2 (mg/L), highest at the northern extreme, and two patches at the southeastern region during summers, and the distribution of nitrate is not found to be homogenous, especially during monsoons as nitrate cannot be attributed to a single source. Supplementary Figure S1(c) shows the heterogeneity of nitrate distribution in the wetland. Nitrate concentration gradually increases towards northern regions in winter.

The concentration of PO₄³⁻ in monsoon (63.3 ± 64.2 μg/L) is reported to be higher than in summer and winter. The source of PO₄³⁻ in the surface water comes mostly through anthropogenic sources. The high concentration of PO₄³⁻ and NO₃⁻ in monsoon may be due to the runoff from the agricultural area which contributes to both nitrate and phosphate (Desai et al. 1995; Bandela et al. 1999; Anshumali & Ramanathan 2007). In general, it is reported that 80% of lake and other reservoirs' eutrophication is controlled by phosphorus (Zhao 2004). The lowest values of phosphate were observed in summer and winter which may be due to low inflow of wastewater, biological utilization, and removal by absorption on to sediment and suspended particles (D'Sousa et al. 1981; Rajasegar 2003). PO₄³⁻ concentration is observed high in the northern region during monsoon (257.7 μ/L) and winter (69.1 μ/L), whereas during summers it is concentrated towards the southern part with maximum recorded value of 180 μ/L. Monsoon season also exhibits a very high concentration of PO₄³⁻ (∼255 (μg/L) towards the northern end of the KT wetland as shown in Supplementary Figure S1(d). The high concentration of phosphates in the northern part and some patches coupled with nitrate concentration depicts that the agricultural runoff (fertilizer) is contributing to the presence of excess ionic entities (Ambastha et al. 2007; Singh et al. 2020).

Sulphate is a naturally occurring substance constituting sulphur and oxygen and is present in various mineral salts that are found in soil. Sulphate may leach from the soil, fertilizers, decaying plant and animal matter commonly released into the water. The highest SO₄²⁻ concentrations were observed in monsoon (59 ± 21.8 mg/L). This could be attributed to the dissolution of SO₄²⁻ rich minerals, and also by human interventions through the application of fertilizers (Grasby et al. 1997). The similar trends of seasonal variations in physicochemical characteristics of River Yamuna reported a higher sulphate concentration of 80 mg/L during the wet season than the in dry season (Ravindra et al. 2003). Sulphate concentration during monsoon reaches up to 118 (mg/L) in the southwestern region whereas during summer and winter it is confined 80 (mg/L) with variation at different sampling points as depicted in Figure 2(d). Ranjan et al. (2017) has analyzed sulphate concentration in KT area and reported concentration of 10.89 mg/L, which is lower than present study (47.7 mg/L). As this area is surrounded by agriculture field and the cropping patterns are variable, we can conclude that the application of fertilizers has increased in this area in a heterogeneous manner. Whereas, the elevation in SO₄²⁻ concentration might have occurred due to dissolution of SO₄²⁻ rich minerals with time (Grasby et al. 1997). Simultaneously Mg²⁺ and Cl⁻ is moderately increasing; by observing the pattern of increment in Mg²⁺ we can say that it is associated with SO₄²⁻ and Cl⁻ coupling when entering the system (Jhingran 1975).

The seasonal variations show (Table 2) that cations follow the trend Ca²⁺ > Mg²⁺ > Na⁺ > K⁺ in summer and winter, whereas in monsoon its trend is Ca²⁺ > Mg²⁺ > K⁺ > Na⁺. Calcium is one of the most important nutrients for the aquatic organisms as it an essential element for the formation of the cell walls and is one of the important factors for physiological function (Yadav et al. 2013). The calcium ions are the dominant cations in the surface water. The mean value of calcium in KT wetland was found to be higher in the winter season (22.7 ± 11.5 mg/L) and lower during summer (13.7 ± 7.7 mg/L). Similar results were noted by (Munawar 1970; Singh & Jayakumar 2016). The high concentration of calcium ions may be due to the weathering of rocks such as limestone, marble, calcite, dolomite, gypsum, fluorite and apatite, etc. (Singh et al. 2011). Although calcium is found to be abundant in water naturally, as well as by induced weathering of rocks, still the addition of sewage waste from nearby residential areas of KT wetland may be accountable for its dominance (Angadi et al. 2005; Udhaya Kumar et al. 2006). Ca²⁺ during monsoon and winter reaches up to 40 (mg/L) and during summers they are confined to a maximum of 28 (mg/L). The distribution is shown in Supplementary Figure S1(e).

Magnesium is generally found in minerals and rocks linked up with calcium and iron compounds, but its concentration is always less than calcium (Tulsankar et al. 2020). It is a very essential macronutrient for the chlorophyll bearing autotrophs like algae and plants which manufacture their food (White & Brown 2010). For phytoplankton, the limiting factor for its growth is magnesium (Dijkstra et al. 2019). The mean concentration of magnesium recorded seasonally ranged between (12.9 ± 6 mg/L) to (5.5 ± 4.8 mg/L), maximum concentration was observed during winter (12.9 ± 6 mg/L) and minimum concentration during monsoon (5.5 ± 4.8 mg/L). A similar result was observed by Kumar et al. (2014) also. Magnesium also enters the system in association with anions like chloride and sulfate (Jhingran 1975). Mg²⁺concentration show 1.2–15.8 (mg/L) variation during summers, maximum concentration was observed at the edges of the KT on the northeastern side. During monsoon, 2.4–19.4 (mg/L) concentration is observed with two patches recording high concentration in the northeastern and southeastern region. In winter season, a vivid patch of higher concentration is observed as shown in Supplementary Figure S1(f) ranging from 2.5 to 25 (mg/L).

The mean concentration of sodium ranges from 4.75 mg/L (±1.8) to 2.40 mg/L (±1.6), maximum during winter 4.75 mg/L (±1.8) and minimum during monsoon 2.40 mg/L (±1.6). However, mean concentration of potassium ranges from 4.1 mg/L (±1.9) to 0.5 mg/L (±0.3), maximum during winter 4.1 mg/L (±1.9) and minimum during summer 0.5 mg/L (±0.3). Sodium and potassium are naturally present in water, although its concentration may increase due to silicate weathering, application of potash fertilizers, precipitation runoff, and from detergents and soap (Kumar et al. 2014). The K⁺ distribution in the study area is not homogenous as during summer and ranges from 0.2 to 1.49 (mg/L) with maximum value observed at the northern side of the KT wetland region (Figure 2(e)). During monsoon K⁺ ranges from 0.22 to 9.25(mg/L), maximum concentration observed towards the northern side again but slightly moving towards the west. In winters, it is concentrated at two patches in the north and southeastern region.

Na⁺ concentration ranges from 0.8 to 7.4 (mg/L) during summer. High concentration of Na⁺ is observed in north side of KT wetland whereas southern region exhibits lower Na⁺ concentration, 7.64 mg/L and 0.7 mg/L, respectively. However, during monsoon, Na⁺ concentrations ranges from 0.52 to 7.2 (mg/L) and the highest value was noted to be at two patches at the center (7.2 mg/L) and northwestern (6.57 mg/L) part as shown in Supplementary Figure S1(g). Winter shows the highest concentration of Na at northern extremes of KT wetland. Most of the dynamic ionic activity is thus seem to occur at the north and northwestern patch of wetland. The presence of higher ionic concentrations can change the hydrogeochemical dynamics which leads to alteration in functionality of the wetlands. The occurrence of higher ionic presence could be due to the depression hydrology where all entities meet and form a dynamic system of alternating storage and release of ions.

The correlation matrix (Supplementary Table S1(a)) plotted for the water samples collected in summer season shows good correlation between Ca²⁺ and HCO₃ (0.671), Mg²⁺ and HCO₃⁻ (0.678). Its main source may occur due to natural death and the decaying process occurring in the wetland. During summer, the temperature and humidity are relatively high; hence the decaying process is also high in releasing gases like CO₂, which dissolves in water. This helps in the formation of carbonic acid which leads to further breakdown resulting in an increased composition of bicarbonates.

The correlation matrix (Supplementary Table S1(b)) plotted for the monsoon water samples shows a good correlation between Na⁺ and EC (0.717) as Na⁺ ions are not generally up taken by the plants, thus, its concentration may have increased causing higher EC. Ca²⁺ and SO₄²⁻ show positive correlation (0.503). Similarly, Na⁺ and Cl⁻ (0.537) and Ca²⁺ and HCO₃ (0.692) pairs show high positive correlation. This may be present in the water samples due to weathering of sulphate minerals (e.g. gypsum) and transfer in the lake through agricultural runoffs.

The correlation matrix (Supplementary Table S1(c)) plotted for the winter NO₃⁻ and temperature (0.589), Na⁺ and temperature (0.501), NO₃⁻ and pH (0.707), Mg²⁺ and EC (0.542), shows good correlation. These correlation values suggest that during winter, temperature governs the physicochemical character of the surface water of the wetland.

The surface water of the study area may have some similarities. To understand this, the study area has been classified hydrochemically using major cations and anions data by plotting Piper trilinear diagram (Piper 1944). Piper plot (Figure 3) shows that surface water during summer is dominated by Ca²⁺ and Cl⁻ but during the monsoon and winter period, the water shows the dominance of Ca²⁺-Mg²⁺-SO₄²⁻ and Cl⁻ types of ions, indicating possible runoff input of sulphur into the system. Excess input of sulphur into the system might be due to the agricultural runoff from the nearby agricultural fields. The presence of chloride type water during summer is mainly due to the evaporation of wetlands surface water. Overall, the hydrochemical facies in all the seasons indicates that strong acids exceed weak acids in the wetland and alkaline substances also dominate over alkalis.

The Wilcox plot, Figure 4(a), which uses sodium percentage and EC values helps in a better understanding of water usage for agricultural purposes pictorially. The study shows that wetland water in all seasons shows excellent nature for agricultural purposes except for few sampling locations in summer (S16, S17, S24, and S29) and monsoon (S11 and S21) where water quality falls in the group of unsuitable category. The unsuitable nature resulting from high sodium% during these seasons might be due to the influence of agricultural activity.

The SAR vs EC plot (Figure 4(b)) shows the usage of water for both drinking and agricultural purposes. The low EC (i.e. <750 μS/cm) indicates its suitability for drinking purposes and moderate EC value (750–1,250 μS/cm) indicate the acceptable nature of water. Low SAR values (<10) indicate its suitability for agricultural activity. The SAR vs EC plot shows that water in KT wetland in all seasons falls under the low-risk category (C1S1) indicating suitability for both drinking and agricultural purposes.

The Gibbs plot (Gibbs 1970) helps to understand the dominant source of ions to the hydrological system. The low values of total dissolved solids (TDS) and high Na/Cl content indicates precipitation source whereas, high TDS value indicates evaporation dominated source. The present study (Figure 5(a) and 5(b)) shows that all data lies within the rock dominance region indicating weathering as the main source of ions regulating the water chemistry. The Mg/Na and HCO₃/Na vs Ca/Na plot helps to find the dominance of weathering in the region (Figure 6(a) and 6(b)). It shows the dominance of carbonate weathering along with a minor amount of silicate weathering in the Kabar Tal wetland's water.

As depicted from the SO₄²⁻ vs Ca²⁺ plot (Figure 7(a)); HCO₃⁻ vs Ca²⁺ plot (Figure 7(b)); Na/Cl vs Cl⁻ plot (Figure 7(c)); Ca²⁺ + Mg²⁺ vs HCO₃⁻ plot (Figure 7(d)); it can inferred that the dominance of ions in the study area is mainly due to weathering of both carbonate and silicate minerals. To understand the presence of dominant mineral, the biplots for various cations and anions were studied (Figure 7(a) and 7(d)). The Na⁺/Cl⁻ vs Cl⁻ plot (Figure 7(c)) helps us decipher the dominance of evaporation over silicate weathering for Na⁺ and Cl⁻ concentration in the water. The plot (Figure 7(c)) depicts that during summer, the evaporation dominates whereas during winter and monsoon silicate weathering dominates. The plots (Figure 7(b) and 7(d)) show that all the data points fall near to the Y-axis (Ca²⁺ and Ca²⁺ + Mg²⁺) indicating Ca and Mg chemistry is dominated by the predominance of calcite and dolomite. The sulphate concentration in the region is derived from both natural (dissolution of sulphate minerals like gypsum) and anthropogenic (fertilizers) activities as the data plots scatter to large extent in the biplot (Figure 7(a)).

The first four PCs (Eigen value >1) are the most significant principal components, which represent 70.81% of the variance in water quality of KT wetland (Table 3). PC 1 explains 23.52% of the variation and is influenced highly by HCO₃⁻, Ca^2+, and Mg²⁺; whereas the loading of Na⁺ is very weak. This PC 1 suggest that carbonate weathering plays a major role in governing water chemistry in summer. PC 2 is responsible for 18.01% of the variation and is strongly influenced by temperature. The temperature in summer is relatively higher than winter and monsoon seasons and it is evident from the PC 2 that temperature is one of the determining factors which controls water chemistry. PC 3 contributes to 16.16% of the variation and is influenced highly by Cl⁻ and Na⁺ ions. Since high evaporation rate in summer due to high temperature leads to an increase in concentration of Cl⁻ and Na⁺ in water and controls the hydro geochemistry. However, PC 4 contributes to 13.13% of the variation and is highly influenced by SO₄²⁻ and is least influential in summer than other anions.

The first five PCs (Eigen value >1) are the most significant principals, which represent 75.57% of the variance in water quality of KT wetland (Table 3). PC 1 contributes to 20.19% of the variation and is highly loaded with Na⁺, and moderately loaded with nitrate and temperature. It is also observed that during a dry season Kabar Tal wetlands is used for agriculture purpose nitrogenous and phosphate-containing fertilizers are widely used (Ranjan et al. 2016). However, surface runoff from agricultural in monsoon increases concentration of nitrate which governs the water chemistry. PC 2 contributes to 18.76% of the variation with a high loading of Mg²⁺ and SO₄²⁻. PC 3 contributes to 13.44% of the variation with a high loading of HCO₃⁻. PC 4 contributes to 12.67% of the variation with high loading pH and EC. PC 5 contributes to 10.51% of the variation with a high loading of K⁺. The combined interpretation of PC2, PC3 and PC4 suggests that sulphate weathering plays the prominent role, followed by bicarbonate in monsoon.

The first five PCs (Eigen value >1) are the most significant principals, which represent 80.35% of the variance in water quality of KT wetland (Table 3). PC 1 contributes to 20.69% of the variation with a high loading of Mg²⁺ and temperature. PC 2 contributes to 20.37% of the variation with a high loading of HCO₃⁻, Ca²⁺, and EC. PC 3 contributes to 15.05% of the variation with a high loading of Na⁺. PC 4 contributes to 13.85% of the variation with a high loading of NO₃⁻ and Cl⁻. PC 5 contributes to 10.39% of the variation with a high loading of PO₄³⁻. The overall analysis of PC suggest that carbonate/dolomite weathering plays a significant role in increasing HCO₃⁻, Ca²⁺ and Mg²⁺ concentration in winter, which subsequently influences EC in winter season.

Seasonal (summer, monsoon, and winter) water quality variations were evaluated using DA. Discriminant functions (DFs) and classification matrices (CMs) were obtained from standard, forward, and backward stepwise modes of DA (Table 4). In forward stepwise mode, parameters having more significant to no significant change are included by the step-by-step process. However, in backward stepwise mode, parameters with less significant to no significant change are obtained through removing the step-by-step process. The standard DA mode, constructed DFs including 12 parameters are shown in Table 4. All the standard, forward, and backward stepwise mode DFs using 12, 5, and 8 discriminant variables, respectively, rendered the corresponding CMs, assigning 100% cases correctly (Table 5). Forward stepwise and backward stepwise DA showed that temperature, pH, EC, SO₄²⁻, and K⁺ are the major significant parameters followed by NO₃⁻, Cl⁻, Na⁺ (Table 4). Further, a less significant third group of the remaining four parameters, i.e., PO₄³⁻, HCO₃⁻, Ca²⁺, and Mg²⁺, is marked from the standard mode. Thus, the temporal DA results explain that temperature, pH, EC, SO₄²⁻, and K⁺ are the most significant parameters to discriminate among the three different seasons, these five parameters justify most of the expected temporal variations in the water quality (Table 5).

In this study, SAR (sodium adsorption ratio) value of the study area is less than 10, favorable, and suitable for irrigation purposes. Based on RSC values, all the samples of the three seasons had values less than 1.25 which also supports its suitability criteria. Gibb's plot reveals that weathering processes are one of the major sources of ions in water. The statistical technique reveals that temperature, pH, EC, SO₄²⁻, and K⁺ are the most significant parameters which are governing the water quality. In summer season, the dry region of KT is utilized for agricultural purpose. Thus, application of fertilizers might occur to get good yield. Therefore, during monsoon these agricultural fields get flooded and change the water chemistry and deteriorate the water quality of the wetlands. The agricultural areas and villages around the wetland also contribute to the nutrient enrichment in the wetland. Overall results illustrate that seasonal variation is playing a major role in nutrients dynamics and hydrochemistry of Kabar Tal wetland.

The authors would like to thank Department of Environmental science, Central University of South Bihar, Gaya. Dr R. K. Ranjan would like to thank the Principal Chief Conservator of Forests, Govt. of Bihar and Divisional Forest Officer, Begusarai, Bihar for granting permission to conduct sampling. The authors would also like to thank Dr Das Ambika Bharti, Department of Psychological Sciences, CUSB, Gaya and Dr Ram Sagar, Department of Botany, BHU, Varanasi; for assisting in multivariate statistical analysis. The authors would also like to thank two anonymous reviewers for their critical comments which have greatly improved the manuscript. The study is funded by SERB, Department of Science and technology, Government of India (Grant no. SR/FTP/ES-01/2014).

All relevant data are included in the paper or its Supplementary Information.