Introduction

In semi-arid and arid areas, where precipitation is rare and evaporation is very high, groundwater becomes the most imperious water resource. Groundwater is very imperative in maintaining ecosystems during socio-economic development (Li et al. 2018a; Subba Rao et al. 2019a; Wu et al. 2019a). The rigorous extraction of groundwater, decreased rainfall frequency and decline in its natural restoration capacity inevitably intensify the groundwater depression. Simultaneously, the groundwater pollution becomes noxious because of the migration of solutes by hydrodynamic dispersion from where the solute amalgamated in aquifers groundwater (Freeze and Cherry 1979). Various water quality crises subsist in surface water and groundwater in India, where 85% of the total rural population depend upon the groundwater resources (Nagaraju et al. 2016). The perturbation of toxic pollutants such as heavy metals, arsenic, nitrate and fluoride distresses the quality of groundwater in India (Marghade et al. 2012; Rina Kumar et al. 2018; Subba Rao et al. 2019b). Numerous studies have been carried out to determine the sources for pollutants such as fluoride, arsenic, heavy metal, landfill leachate, sea water intrusion and industrial effluents throughout India (Marghade et al. 2010; Brindha and Elango 2013; Karunanidhi et al. 2013; Selvam 2014; Singaraja et al. 2014; Brindha et al. 2016, Subba Rao et al. 2017; Duraisamy et al. 2018; Anand et al. 2019; Subba Rao and Chaudhary 2019; Aravinthasamy et al. 2019).

The water dearth distracts the attention of researchers towards sustainable water resources management instead of source appointment of pollutants. The most important step for the sustainable water resources management is understanding the mechanism of natural circulation processes including geochemical processes (composition of soil and rock through which the groundwater drifts, contact time and contact surface between groundwater, mixing, ion exchange), climatic factors (precipitation and evaporation) and unnatural sources (irrigation runoff and industrial and domestic sewage discharge) which govern the groundwater quality during its movement from the recharge to discharge areas (Li et al. 2017).

Now, researchers have concentrated on geochemical characterisation, using various graphical and mathematical modelling approaches. Multivariate statistical evaluation (cluster analysis and principal component analysis) along with geochemical modelling (graphical tools, Pearson’s correlation, chloro-alkaline indices, thermodynamic calculations, etc.) has been successfully applied in numerous hydrogeochemical studies (Thilagavathi et al. 2015: Karunanidhi et al. 2014; Marghade et al. 2015; Magesh et al. 2016; Subba Rao 2017a, b, c; Adimallaa et al. 2018; Rina Kumar et al. 2018; Sreedevi et al. 2018; Li et al. 2019a; Subba Rao et al. 2019a; Wu et al. 2014, 2019b). Principal component analysis (PCA) helps to compress and arrange large data sets to provide significant understanding. Calculated mass balance by simulation analysis epitomises the alterations of the components along the flow path (Li et al. 2007; Subba Rao et al. 2017). Marghade et al. (2019) carried out a comprehensive study in a semi-arid part of Yavatmal District, Maharashtra, India. The study shows that 55% of groundwater samples have fluoride concentration above the recommended limit which mainly comes from weathering and dissolution of fluorite-, apatite-, and fluoride-rich silicates under the influence of high \({\rm H}{\text{C}}{\rm O}_{3}^{ - }\) and Na+ contents.

The review of the literature suggests that the factors controlling groundwater chemistry are precipitation, evaporation, chemical composition of soils and rocks through which the groundwater flows, resident time of water with aquifer material, dissolution and ion exchange, domestic waste, irrigation return flow and industrial discharge. For assessing these factors, the multivariate statistical analysis, graphical tools, chloro-alkaline indices and thermodynamic calculations have been widely used.

The earlier study of adjoining region has mainly been focused on fluoride contamination in groundwater and its health impact (Marghade et al. 2019). The detailed geochemical characterisation is very essential for assessing the groundwater quality not only for its present use but also for sustainable development. In this perception, an endeavour has been made to evaluate the factors controlling geochemistry, using geochemical and multivariate approaches. The human health risk assessment has also been carried out which would be helpful for sustainable groundwater resources management.

Study area

The Wani region of Yavatmal district, Maharashtra, India, is located in between 19′53″ to 20′10″ N latitude and 78′4″ to 79′08″ E longitude (Fig. 1a). The area experiences hot summer and mild winter with an annual average temperature range of 15.5–42.0 °C. The annual rainfall in the region is about 850 mm between June and September (CGWB 2013). The drainage pattern is of sub-dendritic (Fig. 1c). The major land-use type in the region is cultivation land.

Fig. 1
figure 1

a Location, b geological and c drainage map of the study region

Full size image

Geology and hydrogeology

Geologically, the eastern part of the present study region contains Penganga limestones of Neoproterozoic age signified by fluorite-bearing veins at some places (CGWB 2013). The Penganga limestone is overlain by Barakar and Kamthi sandstone formations of Gondwana Supergroup of Permo-Carboniferous age (Fig. 1b). The formation of coal began in carboniferous period. They are most productive water-bearing formations. The Gondwana rocks are overlain by Lameta Formation represented by clay with cherty limestone and sandstones which are in turn overlain by Cretaceous to Paleogene basalts of Ajanta and Chikhli formations of Sahyadri Group of Deccan Volcanic Provence. Plagioclase feldspar and pyroxence constitute the primary minerals of basalt with secondary quartz, calcite and zeolites. Quaternary alluvium and black soils consisting of silt, clay and sand overlie basalts.

The groundwater is the major source of drinking, while seasonal rainfall is the main source of groundwater recharge. The geographical and geological settings are the major factors for occurrence and distribution of groundwater in the region. Water retention capacity of karstified Penganga limestones, Gondwana sandstones and Quaternary alluvium is prominent. Groundwater occurs under unconfined condition in weathered zones and under semi-confined to confined conditions in fracture zones. The depth to water level varies according to lithology. In Penganga Group rocks, it is 4–16 m below ground level (bgl) during pre-monsoon and 2.3–7.6 m bgl during post-monsoon period. In Gondwana, it varies from 7.5 to 3.6 m bgl during pre- and post-monsoon seasons, respectively, while in Deccan Traps, it varies from 2.1 to 7.6 m bgl during pre- and 1.3–5.7 m bgl during post-monsoon season. The drainage map of the present study region is given in Fig. 1c. Evapotranspiration and artificial extraction are the main discharge pathways of groundwater.

Groundwater sampling and analytical procedure

Seventy-two bore wells of the study region which are used for drinking and irrigation purposes without any treatment are selected and sampled. The physico-chemical parameters of groundwater are analysed following the standard procedure of American Public Health Association (APHA 1999). The physical parameters pH and electrical conductivity (EC) are measured onsite by respective meters. Calcium (Ca2+) and magnesium (Mg2+) are determined volumetrically with standard ethylene diamine tetra acetic acid. Chloride (Cl−) is determined, using standard solution of silver nitrate. Carbonate (CO3 2−) and bicarbonate (HCO3 −) are determined by titrating with standard HCl solution. Sodium (Na+) and potassium (K+) are measured by flame photometric method. Fluoride (F−), sulphate (\({\text{SO}}_{4}^{2 - }\)) and nitrate (\({\text{NO}}_{3}^{ - }\)) are determined, using UV–visible spectrophotometer. The ionic data are used to calculate total dissolved solids (TDS). The precision of the chemical analyses is carefully checked by taking duplicate samples. The ionic balance errors for the analyses are within ± 5% level.

Modelling methods

The Microsoft Excel 2007 was used to perform statistical analysis of data obtained after analysis. The bivariate diagrams are used to recognise the reactions between groundwater and the aquifer matrix (Kshetrimayum 2015). To comprehend the data set obtained from the study region, Pearson’s correlation coefficient method and principal component analysis are used.

Geochemical modelling is used for studying hydrochemical evolution mechanisms. The PHREEQC software is used to compute saturation index (SI) values of minerals. Inverse geochemical modelling is also computed by using PHREEQC interactive software, which is used to compute the number of moles of minerals and gases transferred from dissolved state to precipitated/degassed from the solutions (Li et al. 2010; Subba Rao et al. 2017). For calculation of mass balance, the differences in compositions between the initial and final end-member solutions have been used. PCA was performed by using varimax rotation method with Kaiser normalisation (Hussin et al. 2016; Subba Rao et al. 2017).

Health risk assessment

Human health risk assessment is the technique to evaluate the nature and possibility of ill-health effects in humans who consumed contaminated groundwater regularly (Adimallaa and Qian 2019; Li et al. 2016a, 2019b). In this study, drinking water consumption was the major exposure pathway for nitrate and fluoride. The IRIS (Integrated Risk Information System) model introduced by US Environmental Protection Agency (USEPA 2004) is used to measure the exposure dose (E) through ingestion pathway and probable non-carcinogenic risk of hazard quotient or hazard index (HQ) (He and Wu 2019; He et al. 2019; Karunanidhi et al. 2019):

$$E = \frac{{{\text{CPW}} \times {\text{IR}} \times {\text{ED}} \times {\text{EF}}}}{{{\text{ABW}} \times {\text{AET}}}}$$
(1)
$${\text{HQ}} = \frac{E}{\text{RfD}}$$
(2)

where E is the chronic daily intake (mg/kg/day), CPW is the concentration of a particular contaminant in groundwater (mg/L), IR is the human ingestion rate (2.5 L/day for adults, 0.78 L/day for children), ED is the exposure duration (years: 64 and 12 for adults and children), EF is the exposure frequency(days/years: 365 days), ABW is the average body weight (Kg: 57.5 and 18.7 for adults and children, respectively), AET is the average time (days: 23,360 and 4380 for adults and children), HQ is the non-carcinogenic for hazard quotient and RfD is the reference dose of fluoride and nitrate 4.00E − 02 and 1.1E + 00 mg/kg-day, respectively (Integrated Risk Information System, USEPA 2004).

Hazard index interprets 10−6 risk level for carcinogens, and HQ is 1 for non-carcinogens. If a person is exposed daily for a lifetime to 1 µg of the nitrate/fluoride per litre of drinking water, the unit risk = 2 × 10−6 per µg/L, meaning two excess cancer cases are expected to develop per 1000,000 people.

Results and discussion

General groundwater chemistry

Groundwater of study region is neutral to slightly alkaline (pH 6.7–8.6). A total dissolved solid (TDS) diverges from 240.8 to 4139.8 mg/L with an average of 1056.3 mg/L (Table 1).

Table 1 Analytical data of groundwater samples (n = 72)a
Full size table

About 21% of sampling points exceed maximum permissible limit of 1500 mg/L marking them not recommended for drinking purposes (Marghade et al. 2015; Subba Rao 2017b). The dominant Na+ level in the groundwater diverges from 7.4 to 1364 mg/L with a mean of 221 mg/L. Almost 40% of sampling points have Na+ concentration above the prescribed limit of 200 mg/L (Table 1). The amount of K+, Ca2+ and Mg2+ in the sampling points diverges in the ranges of 0–201, 10–364, 1.2–153.1 mg/L, respectively.

\({\rm H}{\text{C}}{\rm O}_{3}^{ - }\) is the dominant anion over other anions and diverges from 103.7 to 1159 mg/L (Table 1). High \({\rm H}{\text{C}}{\rm O}_{3}^{ - }\) concentration in groundwater reduces the available adsorption sites of the sorbents that lead to the release of fluoride from minerals (Subba Rao 2017a). The semi-arid climate-induced prolonged evaporative enrichment of water leads to high Cl− concentration (Subba Rao 2006). Chloride varies from 17.8 to 1136 mg/L. \({\text{SO}}_{4}^{2 - }\) in groundwater ranges from 35.1 to 925 mg/L. The high \({\text{SO}}_{4}^{2 - }\) concentration may be due to geogenic processes and high evaporation rate in the study region (Adimallaa and Qian 2019). The elevated nitrate (1.6–803 mg/L with an average of 140.2 mg/L) concentrations in drinking water cause fatal disease “blue baby syndrome” (methaemoglobinemia) in infants (Marghade et al. 2011; Subba Rao 2017b). About 66.6% of the groundwater samples were found to surpass the tolerance limit of 45 mg/L for nitrate (WHO 2011). It is observed that samples having high TDS have high concentration of Na+ and \({\text{NO}}_{3}^{ - }\). This strongly shows that the groundwater quality is influenced by human activities and mainly by irrigation return flows, drainage effluents and animal waste. The 0.5–1.5 mg/L of fluoride in drinking water is a very essential element for human health (WHO 2011), but its higher concentration becomes risky for human health. Only 12% of the groundwater samples contain the F− concentration above 1.5 mg/L. (Table 1).

The divergences in the standard deviation values (0.4–1483) (Table 1) due to various hydrogeochemical reactions indicate a varied distribution of salts in the groundwater (Subba Rao et al. 2017). The difference in the coefficient of variation values also shows a varied distribution of the chemical variables in groundwater. The highest coefficient of variation is of \({\text{NO}}_{3}^{ - }\) (120.4%) followed by Na+ (118.4%), \({\text{SO}}_{4}^{2 - }\) (115.1%) and Cl− (102.9%). pH has lowest coefficient of variation (5.3%). This observation suggests that Na+ and \({\text{NO}}_{3}^{ - }\) may be the main key factors in controlling groundwater chemistry.

The Pearson’s correlation coefficient method generates matrix with significant correlations between the parameters. The electrical conductivity shows high correlation coefficient with Ca2+, Na+, Cl− and \({\text{SO}}_{4}^{2 - }\), which elucidates large involvement of these elements to the groundwater chemistry. Significant positive correlation (Table 2) observed between TDS and Na+, Ca2+, Cl− and \({\text{SO}}_{4}^{2 - }\) indicates contribution of these ions towards hydrochemical characteristics of groundwater (Tay et al. 2017). Strong positive correlation of Na+ with Cl− (r = 0.81) and \({\text{SO}}_{4}^{2 - }\) (r = 0.80) indicates high pollution level of groundwater (Table 2). This is further supported by strong positive correlation of Ca2+ with Cl− (r = 0.75) and \({\text{NO}}_{3}^{ - }\) (r = 0.72). An alkaline environmental condition due to high bicarbonate concentration (positive correlation of F− with pH, r = 0.43) promotes F− leaching (Subba Rao 2003, 2011).

Table 2 Correlation coefficients matrix
Full size table

A significant low correlation between K+ and \({\text{NO}}_{3}^{ - }\) (r = 0.34) and \({\text{SO}}_{4}^{2 - }\) (r = 0.48) indicates the presence of other sources of these ions irrespective of fertiliser runoff (Table 2). Negative correlation between F− and Ca2+ (r = − 0.18) is endorsed for the influencing factors for fluorite (CaF2) dissolution in groundwater and the removal of Ca2+ from groundwater through the precipitation of CaF2 due to high load of Ca2+ in groundwater (Subba Rao et al. 2013). The Na+-enriched groundwater samples contain high F− concentration which is supported by positive correlation (r = 0.33) between F− and Na+. The high R 2 value than \({\rm H}{\text{C}}{\rm O}_{3}^{ - }\) confirmed the significant contribution of Na+ in controlling the groundwater F− geochemistry (Magesh et al. 2016; Subba Rao 2017a, c). The less significant positive correlation between F− and K+ (r = 0.14) and negative correlation between F− and \({\text{NO}}_{3}^{ - }\) (r = − 0.05) indicate less contribution of anthropogenic input such as fertilisers and domestic waste in enrichment of F− in the study region.

Groundwater types

The sequence of cation dominance is Na+ > K+ > Ca2+ > Mg2+ and anion dominance is \({\rm H}{\text{C}}{\rm O}_{3}^{ - }\) > Cl−> \({\text{SO}}_{4}^{2 - }\) > \({\text{NO}}_{3}^{ - }\) > F− in the groundwater of the study region. The groundwater types derived from the major ion chemistry are represented in trilinear plot (Fig. 2) (Piper 1944). About 38% of sampling points fall in Na–Cl and 32% in Ca–HCO3 water-type zones. Mixed water types Ca–Mg–Cl and Ca–Na–HCO3 are observed in 8 and 15% of the total samples. It is interpreted that the freshwater (Ca–HCO3) shifts towards the saline water (Na–Cl) through the mixed water (Ca–Mg–Cl and Ca–Na–HCO3), indicating the influence of geogenic and human activities on the groundwater chemistry (Fig. 2).

Fig. 2
figure 2

Piper plot

Full size image

Risk assessment for human health

The non-carcinogenic health risk for adults and children is due to the consumption of polluted drinking water. The risks are calculated by IRIS model for those samples which contain high \({\text{NO}}_{3}^{ - }\) and F− in the study region (Table 3).

Table 3 Calculated health risk from nitrate and fluoride pollution for adult and children
Full size table

As nitrate is a non-geogenic contaminant, it is considered as an indicative of anthropogenic activities such as manures, animal wastes and agricultural runoff (Marghade et al. 2011). The \({\text{NO}}_{3}^{ - }\) concentration varies from 1.2 to 803 mg/L in groundwater of the study region. About 67% of the total groundwater samples have \({\text{NO}}_{3}^{ - }\) concentration above the tolerance limit of 45 mg/L (WHO 2011). The health risk values for nitrate range from 2.00E − 02 to 1.01E + 01 with an average of 2.36E + 00 calculated for children (Table 3). For adults, HQ values for nitrate vary from 1.21E − 02 to 6.05E + 01 with an average of 1.42E + 00. In 22% and 7% of the total groundwater samples, HQ has very high (> 3) non-carcinogenic risk for nitrate for children and adults, respectively. The values, which are higher for children, are due to prolonged drinking of nitrate-contaminated water, causing blue baby syndrome in children of the study region. The health risks assessment also revealed that 58% and 44% of the total groundwater samples exceed the tolerance limit of 1.0E + 00 prescribed for non-carcinogenic risk in children and adults, respectively (Table 3).

Similarly, for fluoride, calculated HQ values are higher than the reference dose of 1.0E + 00 mg/kg-day in 19 (40%) and 11 (23%) of the total groundwater samples for children and adults, respectively, increasing the probability of dental and skeletal fluorosis (Table 3). The high-risk values for fluoride in 12% of the total groundwater samples are due to concentration of fluoride above the prescribed limit of 1.5 mg/L.

Processes regulating groundwater chemistry

Various factors such as geogenic, ion exchange, anthropogenic and climatic conditions are accountable for hydrogeochemical amendment of groundwater chemistry

Carbonates and gypsum dissolution

The value of Ca2+/Mg2+ molar ratio equal to one indicates the dissolution of dolomite and more than one indicates higher calcite contribution and release of Ca2+ ions into the groundwater during recharging of aquifer (Fig. 3a). The Ca2+/Mg2+ molar ratio greater than one specifies the dominance of calcite dissolution over dolomite dissolution in the groundwater. However, the dispersion of sampling points above and below the uniline of 1:1 verifies the existence of calcite and dolomite dissolution in the groundwater, but neglects the dominancy of any single process (Wu et al. 2019a). The contribution of carbonate (such as calcite CaCO3 and dolomite CaMg (CO3)2 to hydrochemistry is significantly indicated by (Ca2+ + Mg2+) and (\({\rm H}{\text{C}}{\rm O}_{3}^{ - }\) + \({\text{SO}}_{4}^{2 - }\)) bivariate plot (Fig. 3b). Most of the groundwater samples fall along the 1:1 line, indicating the dominance of carbonate component. The dispersal of groundwater sampling points from equiline indicates the contribution of cation exchange and/or the dissolution of silicate for high Ca2+ and Mg2+ percentage in the groundwater. However, all groundwater samples fall above the 1:1 line in the (Ca2+ + Mg2+) versus \({\rm H}{\text{C}}{\rm O}_{3}^{ - }\) diagram (Fig. 3c), indicating less dissolution of carbonates (Wu and Sun 2016).

Fig. 3
figure 3

Bivariate plots of a Ca2+ and Mg2+, b Ca2+ + Mg2+ and \({\rm H}{\text{C}}{\rm O}_{3}^{ - }\) + \({\text{SO}}_{4}^{2 - }\), c Ca2+ + Mg2+ and \({\rm H}{\text{C}}{\rm O}_{3}^{ - }\), d Ca2+ + Mg2+ − 0.5 \({\rm H}{\text{C}}{\rm O}_{3}^{ - }\) and \({\text{SO}}_{4}^{2 - }\) and e Ca2+ and \({\text{SO}}_{4}^{2 - }\) f Na+ and Cl− g Na+ versus TC

Full size image

When the aquifer contains both dolomite and calcite, it is possible that the gypsum dissolution accelerates the dedolomitisation process due to simultaneous precipitation of calcite and dissolution of dolomite (Fijani et al. 2017). The dissolution of gypsum tends to increase calcium concentration, causing precipitation of calcite (Li et al. 2016b, c). The maximum number of groundwater sampling points falls above the zero line of binary plot of [(Ca2+ + Mg2+) − 0.5 \({\rm H}{\text{C}}{\rm O}_{3}^{ - }\)] versus \({\text{SO}}_{4}^{2 - }\) (Fig. 3d), which proves that gypsum is also a source of Ca2+ and \({\text{SO}}_{4}^{2 - }\) in the groundwater system (Subba Rao et al. 2019a, b). However, the binary plot between Ca2+ and \({\text{SO}}_{4}^{2 - }\) (Fig. 3e) resolved that gypsum dissolution is highly masked by other factors such as recharge water chemistry, precipitation, evaporation and anthropogenic sources. This is also supported by the deviation of sampling points from 1:1 line (Li et al. 2013, Wu and Sun 2016).

Evaporite dissolution

A large percentage of Na+ and Cl− to the groundwater come from the dissolution of evaporite (such as halite). Many groundwater sampling points fall along the 1:1 line of Na–Cl plot (Fig. 3f), signifying the dissolution of halite for high concentration of Na+ and Cl− (Hussin et al. 2016). A number of groundwater samples have Na+/Cl− ratio > 1, which inferred that Na+ is released from weathering of feldspars such as plagioclase instead of halite dissolution (Hussin et al. 2016). In about 55% of the total groundwater samples, the value of Na+/Cl− ratio is more than one, indicating high contribution of Na+ in the groundwater by hydrolysis of silicates (Subba Rao and Surya Rao 2010; Subba Rao et al. 2017).

Silicate dissolution

Sodic plagioclase mineral may be the probable source of high Na+ over Cl− in the study region (Eq. 3). Most groundwater sampling points plot above the equiline of bivariate graph between \({\rm H}{\text{C}}{\rm O}_{3}^{ - }\) + \({\text{SO}}_{4}^{2 - }\) and Ca2+ + Mg2+ (Fig. 3b), indicating the less significant role of silicate weathering than carbonate and sulphate mineral dissolution in the groundwater chemistry (Hussin et al. 2016; Adimalla and Rajitha 2018):

$$2\,{\text{Na}}^{ + } \left[ {{\text{AlSi}}_{3} {\text{O}}_{8} + 2{\text{CO}}_{2} } \right]^{ - } + 11{\text{H}}_{2} {\text{O}} \to 2\,{\text{Na}}^{ + } + 2{\text{HCO}}_{3}^{ - } + {\text{Al}}_{2} {\text{Si}}_{2} {\text{O}}_{5} \left( {\text{OH}} \right)_{4} + {\text{Mg}}^{2 + } + 4{\text{H}}_{4} {\text{SiO}}_{4} \,{\text{Albite}}$$
(3)

Ion exchange factor

As discussed earlier, the control of anions (\({\rm H}{\text{C}}{\rm O}_{3}^{ - }\) + \({\text{SO}}_{4}^{2 - }\)) over cations (Ca2+ + Mg2+) and the Na+/Ca2+ ratio higher than unity in 55% of the total groundwater samples also verify the existence of ion exchange processes (Marghade et al. 2015; Li et al. 2018b). It is further verified by chloro-alkaline indices CAI-1 and CAI-2 proposed by Schoeller (1977).

The negative CAI index values (base ion exchange process) indicate a softening process in which exchange of Ca2+ and Mg2+ of the water with Na+ and K+ from the aquifer material takes place. However, in reverse ion exchange process, positive values indicate hardening process in which Na+ and K+ get adsorbed on the aquifer material with simultaneous release of Ca2+ and Mg2+ (Hussin et al. 2016).

The values of CAI-1 (0.44–47.05) and CAI-2 (0.26–47.06) are positive in about 85% and 99% of the total groundwater samples (Table 4), specifying that reverse ion exchange may be due to anthropogenic inputs and rainfall. In this case, cation exchange is expressed as

Table 4 Chloroalkaline indices
Full size table
$$2{\text{Na}}^{ + } + {\text{CaX}}_{2} = {\text{Ca}}^{2 + } + 2{\text{NaX}}$$
(4)

The excess of Ca2++ Mg2+ over \({\rm H}{\text{C}}{\rm O}_{3}^{ - }\) + \({\text{SO}}_{4}^{2 - }\) shifts the groundwater sampling points towards the left on bivariate plot of (Ca2+ +Mg2+) and (\({\rm H}{\text{C}}{\rm O}_{3}^{ - }\) + \({\text{SO}}_{4}^{2 - }\)), which also supports the existence of reverse ion exchange (Fig. 3b) (Marghade et al. 2015).

Anthropogenic factors

The irrigation return flow may be the source of Ca2+, Mg2+, Na+, \({\rm H}{\text{C}}{\rm O}_{3}^{ - }\), Cl− and \({\text{SO}}_{4}^{2 - }\) in groundwater of the agricultural rural region (Subba Rao and Surya Rao 2010; Subba Rao et al. 2017). The sanitary facilities in the present region are in general poor. The strong evidence of anthropogenic perturbation is the strong positive correlation between \({\text{NO}}_{3}^{ - }\) and Cl− (r = 0.67; Table 2), which suggests that both the ions have same source, i.e. mainly human activities (Marghade et al. 2011). However, the negative correlation between \({\text{NO}}_{3}^{ - }\) and \({\rm H}{\text{C}}{\rm O}_{3}^{ - }\) (r = − 0.18) shows different sources for both the ions. \({\text{NO}}_{3}^{ - }\) is mainly derived from non-lithological inputs (human activities), while \({\rm H}{\text{C}}{\rm O}_{3}^{ - }\) is mainly derived from lithological inputs (soil CO2).

In addition, the variation in TDS with other ions in groundwater indicates land use and extent of pollution (Marghade et al. 2011). Correlation of Cl−, \({\text{SO}}_{4}^{2 - }\), \({\text{NO}}_{3}^{ - }\) and Na+ ions with TDS is used to show the impact of anthropogenic load on the aquifer chemistry (Li et al. 2016d). Pearson’s correlation matrix (Table 2) showed a significant strong positive correlation between TDS and Na+ (r = 0.90), \({\text{NO}}_{3}^{ - }\) (r = 0.67), Cl− (r = 0.95) and \({\text{SO}}_{4}^{2 - }\) (r = 0.80). The strong positive correlation between TDS and Na+ (r = 0.90) specifies the contribution of anthropogenic input along with silicate weathering as a cause of high level of Na+ in groundwater (Subba Rao et al. 2017).

Geochemical modelling

Geochemical modelling of groundwater is the best method to verify the hydrochemical findings. Saturation index (SI) is a thermodynamic parameter which specifies the tendency of minerals to dissolve or precipitate during the course of water. SI is computed from activity product and equilibrium solubility of minerals (Li et al. 2010; Subba Rao 2017b),

$${\text{Where}},\quad {\text{SI}} = \, K_{\text{IAP}} /K_{\text{SP}}$$
(5)

K IAP is ion activity product of the dissociated minerals and K SP is equilibrium solubility at mineral temperature.

The saturation index (SI) for halite (NaCl), gypsum (CaSO4.2H2O), dolomite (CaMg(CO3)2), calcite (CaCO3), fluorite (CaF2), anhydrite (CaSO4) and aragonite (CaCO3) minerals is calculated using PHREEQC Interactive 2.13.2 software and listed in Table 5. The SI values (− 2.76 to − 0.54) are in dissolution (undersaturation) state in all the groundwater samples with respect to fluorite, whereas 64% of the total groundwater samples show calcite oversaturation (SI values: − 0.86–1.60). The SI results verified the contribution of carbonate dissolution in rising F− concentrations in the groundwater by generating bulk \({\rm H}{\text{C}}{\rm O}_{3}^{ - }\) and F−. SI values of calcite are negative in 36% of the groundwater samples containing high fluoride concentration indicating equilibrium with respect to calcite (Magesh et al. 2016; Singh and Mukherjee 2014). The negative SI (− 8.46 to − 4.65) for halite (NaCl) is observed in the groundwater samples, indicating dissolution state resulting from the enrichment of groundwater by Na+. About 51.4 and 64% of the total groundwater samples show oversaturation of aragonite (SI: − 1.01–1.45) and calcite (SI: − 0.86–1.60) minerals, along with oversaturation of dolomite (SI: − 1.54–2.30) in 65.3% of the total groundwater samples (Li et al. 2016d).

Table 5 Saturation indices for mineral phases
Full size table

The groundwater shows the contribution of anhydrite in high Ca2+-bearing groundwater, as it is in unsaturated state along with carbonates (SI: − 2.98 to − 0.74; Table 5). The groundwater shows a low concentration of Ca2+ and a high concentration of \({\text{SO}}_{4}^{2 - }\) (Table 1). This is explained by the common ion effect in which the CaSO4 dissolution decreases the activity coefficient product γCa2+.γCO3 2−. However, the product (Ca2+)(\({\text{CO}}_{3}^{2 - }\)) increases due to the dissolution of calcite, gypsum, aragonite and anhydrite by releasing the higher amount of Ca2+ in the groundwater system (Subba Rao et al. 2017). Thus, precipitation of CaCO3 could occur to keep the solution in equilibrium with solid CaCO3 (Subba Rao et al. 2017). Further, CaCO3 precipitation becomes the crucial factor for dedolomitisation process, as discussed earlier.

Mass balance simulation analysis

Mass balance modelling is simply a hydrogeochemical simulation to calculate the extent of precipitation and mineral phase dissolving capacity, which is widely used to uphold different weathering reactions in water–rock systems (Galego et al. 2009). It is executed with PHREEQC interactive software. In this chemical/mathematical approach, various models are generated by comparing the geochemical components at upstream with geochemical components at downstream. Simulation exemplifies the difference in concentration of the components along the flow path due to the weathering reactions (Li et al. 2007; Subba Rao et al. 2017).

On the basis of geochemical findings, calcite, dolomite, fluorite, halite, gypsum, K-feldspar, albite and anhydrite minerals are considered as the possible mineral phases for geochemical modelling simulation analysis. The existence of ion exchange reactions in the groundwater system is identified geochemically. So, CaX2 and NaX are also incorporated in the simulation.

The mole transfer of phases is computed during simulation analysis carried out on the four flow paths listed in Table 6. Similar types of patterns are observed along all flow paths.

Table 6 Mineral phase transfer amount
Full size table

Simulation analysis exposes the dissolution of dolomite, gypsum, halite, K-feldspar and CO2 along all simulated paths. The dissolution of dolomite is previously foreseen by Ca2+/Mg2+ molar ratio (Fig. 3a). Similarly, dissolution of gypsum is foreseen by binary plot of [(Ca2++Mg2+) − 0.5 \({\rm H}{\text{C}}{\rm O}_{3}^{ - }\)] versus \({\text{SO}}_{4}^{2 - }\) and plot of Ca2+ and versus \({\text{SO}}_{4}^{2 - }\) verified by positive values (Figs. 3d) calculated for gypsum phase by simulation analysis. The negative values (− 0.00,601, − 0.000045, − 0.00226) for calcite confirm the precipitation of calcite predicted by the plot of SI of gypsum against dolomite and ion exchange process (Subba Rao et al. 2017). The gypsum dissolution elevates the concentration of Ca2+, while the anhydrite decreases the concentration of Ca2+ due to its precipitation. The binary plot of Na–Cl (Fig. 3f) and Na+ versus TC forecasted the dissolution of halite (Fig. 3g), which is verified by the positive values (Table 6) of mass transfer for halite phase in the groundwater body. The possible sources for K+ are weathering of K-feldspars present and application of potash fertilisers. A positive value (Table 6) of mineral mass balance for K-feldspars indicates that the feldspars are in dissolution state.

The positive values (Table 6) of simulation for CO2 (g) signify the contribution of dissolution of soil CO2 in increasing the \({\rm H}{\text{C}}{\rm O}_{3}^{ - }\) concentration in the groundwater. The precipitation of calcite (along first three flow paths) and anhydrite is predicted by their negative simulation values (Table 6). Inclusively, the dolomite and gypsum dissolution increases the concentration of Ca2+, while the calcite and anhydrite precipitation and cation exchange reduce the concentration of Ca2+ simultaneously with the increase of Na+ in the groundwater. The dissolution of calcite in Model IV is supported by the positive value (0.0022) of CaX2 phase, indicating the role of ion exchange in chemical load of groundwater system. The negative values (− 0.000313, − 0.000075, − 0.0032) for CaX2 phase and positive values (0.000626, 0.000015, 0.0064) for NaX phase in model II, III and IV authenticated the presence of ion exchange foreseen by chloro-alkaline indices and (Ca2+ + Mg2+) and (\({\rm H}{\text{C}}{\rm O}_{3}^{ - }\) + \({\text{SO}}_{4}^{2 - }\)) bivariate plot (Fig. 3b). During this process, Na+ released by the dissolution of halite underwent exchange with Ca2+ absorbed on the rock surface.

It is found that the concentration of Na+ in the groundwater has increased by cation exchange, while precipitation of albite (negative values of simulation) lowers its concentration (Table 6). A very small amount of CO2 dissolved in flow path III (21–26) is due to coal deposits present in Barakar Formation of Gondwana Supergroup of Permo-Carboniferous age and continuous recharging responsible for the precipitation of fluorite.

Therefore, the simulation analysis of groundwater (Models I to IV) substantiates the water–rock interactions as the main parameters responsible for chemical characterisation of groundwater. In addition, the dry climate and man-made activities affect the rate of reaction and intensity (Li et al. 2016b).

Principal component analysis

The principal component analysis (PCA) with varimax-rotated R-mode was computed on the 72 set of 12 water parameters, such as pH, EC, TDS, Na+, K+, Ca2+, Mg2+, \({\rm H}{\text{C}}{\rm O}_{3}^{ - }\), Cl−, F−, \({\text{NO}}_{3}^{ - }\) and \({\text{SO}}_{4}^{2 - }\). According to Kaiser criterion (Wu and Sun 2016), only those principal components (PCs) are extracted with rotation of loadings for the maximum variance, which have eigenvalues larger than one.

The high positive loading of EC, TDS, Na+, K+, Ca2+, Cl−, \({\text{NO}}_{3}^{ - }\) and \({\text{SO}}_{4}^{2 - }\) in the first PC accounts for 49.37% of the total variance (Table 7). As discussed, significant positive correlation (r = 0.9, 0.78, 0.95 and 0.67) is observed between TDS and Na+ (Table 2), which signifies the contribution of these ions in governing the hydrochemical characteristics of groundwater. A very high positive PC loading of TDS (0.99) and Cl− (0.95) signifies the impact of anthropogenic loading on hydrochemistry of groundwater. Strong positive correlation between Na+ with Cl (r = 0.81) and \({\text{SO}}_{4}^{2 - }\) (r = 0.80) indicates high pollution level of groundwater (Barzegar et al. 2017; Subba Rao and Chaudhary 2019). The high content of Na+ (0.84) signifies the role of geogenic and anthropogenic sources for high concentration of Na+ in the groundwater (Marghade et al. 2015, Isa et al. 2017). Na+/Cl− ratio greater than one signifies silicate dissolution. The secondary salt precipitation due to irrigation return flows, elevated rate of evaporation and anthropogenic inputs are mainly governing the concentration of Cl− (Subba Rao 2017a). The significant positive correlation between Na+ and Cl− (r = 0.81) and \({\text{SO}}_{4}^{2 - }\) (r = 0.80), between Ca2+ and Cl− (r = 0.75) and \({\text{NO}}_{3}^{ - }\) (r = 0.72) supports the role of anthropogenic input over rock mineral interactions (Aksever and Buyuksahin 2017). The main sources of K+ in the study area are the weathering of feldspars and potash fertilisers. The plot of (Ca2+ + Mg2+) versus (\({\rm H}{\text{C}}{\rm O}_{3}^{ - }\) + \({\text{SO}}_{4}^{2 - }\)) (Fig. 3b) strongly supports the presence of reverse ion exchange in the study area (Marghade et al. 2015). The less significant positive correlation between Na+ and Ca2+ which makes Na+/Cl− ratio greater than one further strongly supports the existence of reverse ion exchange, which is further strengthened by first principal loading (Sudheer Kumar et al. 2017). A low dissolution of Mg2+ and Ca2+ ions due to the precipitation of calcite is indicated by negative PC scores in the first PC of 46 groundwater samples. Therefore, PC-1 reveals the silicate weathering and reverse ion exchange followed by human activities as the contamination sources.

Table 7 Principal component loadings
Full size table

The second PC endorses the 17.59% of the total variance with high positive loading of \({\rm H}{\text{C}}{\rm O}_{3}^{ - }\) and \({\text{SO}}_{4}^{2 - }\) as well as high negative loading of \({\text{NO}}_{3}^{ - }\). The significant correlation between \({\rm H}{\text{C}}{\rm O}_{3}^{ - }\) and \({\text{SO}}_{4}^{2 - }\) (r = 0.62; Table 2) indicates the common source of these ions. The negative correlation between \({\rm H}{\text{C}}{\rm O}_{3}^{ - }\) and \({\text{NO}}_{3}^{ - }\) (r = − 0.18) signifies the different sources of these ions (Marghade et al. 2012). This again supports the fact that calcite dissolution is the main process governing \({\rm H}{\text{C}}{\rm O}_{3}^{ - }\) content. Non-geogenic sources are responsible for high nitrate content (Wu and Sun 2016). Thus, the second PC is assumed to be an indicative of influence of soil CO2 and soil amendment of gypsum (Wagh et al. 2017). The PC1 and PC2 (Table 7) shows high positive loadings (3.37, 5.80) for groundwater sampling point 31, which has the highest concentration of Ca2+ (364 mg/L), \({\rm H}{\text{C}}{\rm O}_{3}^{ - }\) (1159 mg/L) and \({\text{SO}}_{4}^{2 - }\) (925 mg/L). This suggests that the geogenic sources are masked by anthropogenic sources and reverse ion exchange in the groundwater.

The third PC accounts for 10.3% of the total variance, where pH (-0.80) and F− (0.87) show strong negative weights (Table 7). The pH has a low positive correlation with F− (r = 0.43; Table 2). The high \({\rm H}{\text{C}}{\rm O}_{3}^{ - }\) content in alkaline water (pH: 6.7–8.6) accelerates the dissolution of F− in groundwater (Subba Rao 2013, 2018). Simultaneously, negative correlation of F− with Ca2+ and Mg2+ indicates the precipitation of calcite that is reduction in Ca2+ content, which supports the dissolution of F− in the groundwater. Only 12% of the total groundwater sampling points have F− content above 1.5 mg/L due to the reverse ion exchange process. Thus, the third PC is assumed to be an indicative of F− loading in the groundwater (Table 7).

Conclusions

The application of statistical, geochemical and multivariate approaches gives an insight of hydrogeochemistry of groundwater to assess the groundwater quality from semi-arid Wani region of eastern Maharashtra, India. The main conclusions drawn are as follows:

  • Groundwater shows high concentrations of Na+ and \({\rm H}{\text{C}}{\rm O}_{3}^{ - }\) ions. A strong correlation of Na+ with Cl−, \({\text{SO}}_{4}^{2 - }\) and \({\text{NO}}_{3}^{ - }\) indicates the high pollution level of groundwater. Negative ion exchange between F− and Ca2+ and positive correlation between F− and Na+ support the acceleration of F− content in the groundwater. The precipitation of calcite and the dissolution of dolomite cause the dedolomitisation process in the study region. The positive CA1 and CA2, bivariate plots [(Ca2+ +Mg2+) and (\({\rm H}{\text{C}}{\rm O}_{3}^{ - }\) + \({\text{SO}}_{4}^{2 - }\)) indicate the reverse ion exchange, depending upon the groundwater flow path and residence time conditions. The strong relation of TDS with Cl−, \({\text{SO}}_{4}^{2 - }\), \({\text{NO}}_{3}^{ - }\) and Na+ ions is strongly reinforced the impact of human activities on the groundwater chemistry.

  • The health risks assessment showed the higher percentage of health risk among children (40%) on consumption of F−-polluted groundwater than that for adults (23%) in the study region.

  • Geochemical modelling of data supports the rock weathering reactions (calcite dissolution, dolomite dissolution, gypsum dissolution, oversaturation of aragonite, CaCO3 precipitation and dedolomitisation) taking place in the present study region.

  • Principal component analysis extracted a three-factor model. The high positive loading of PC1 (EC, TDS, Na+, K+, Ca2+, Cl−, \({\text{NO}}_{3}^{ - }\) and \({\text{SO}}_{4}^{2 - }\)) represents the carbonate weathering, silicate dissolution and ion exchange process by anthropogenic activities; PC2 (\({\rm H}{\text{C}}{\rm O}_{3}^{ - }\) and SO4) indicates the contribution of soil CO2 and human activities, and PC3 (F−) reveals the dissolution of fluorite by the precipitation of CaCO3.

  • The present study suggests the control of anthropogenic activities (wastes disposal and excessive use of agrochemicals) to prevent the contamination of groundwater resources. Regular groundwater quality monitoring stations are necessary. Creation of public awareness is essential for effective groundwater management. The study recommends that the rainwater harvesting techniques like surface water storage, recharge pits, check dams, etc., should be followed to improve the groundwater quality as per drinking water quality standards as well as to augment the groundwater levels which are the best methods for sustainable development and management of groundwater resources.