Introduction

Normally, the small amount of fluoride (F−) is essential for development of bones and formation of dental enamel (Handa 1975). However, low concentration of F− (< 0.60 mg/L) in drinking water results in dental decay, while an excess of it (> 1.50 mg/L) causes dental fluorosis (WHO 2011; BIS 2012). Drinking water with F− content in between 0.60 and 1.50 mg/L is, thus, stipulated as safe limit for human health.

In the world, more than 260 million people suffer from fluorosis (Amini et al. 2008) due to drinking of water with higher F−content (> 1.50 mg/L). Contamination of groundwater with F−is widespread especially in Africa, China, India, Iran, Kenya, Nigeria, Pakistan, South America, and Srilanka (Ayoob and Gupta 2006), posing high health risk to people (Wu and Sun 2016; Li et al. 2016a). In India, the first case of endemic fluorosis was reported due to higher concentration of F− in groundwater as long as 1937 in the Prakasam district (erstwhile Guntur District), Andhra Pradesh (Short et al. 1937). About 61 million people suffer from ill-health in India caused by drinking water with consumption of F− contamination in 20 states, including Andhra Pradesh, Assam, Bihar, Chhatisgarh, Delhi, Gujarat, Haryana, Jammu and Kashmir, Jharkhand, Karnataka, Kerala, Madhya Pradesh, Maharashtra, Orissa, Punjab, Rajasthan, Tamil Nadu, Telangana, Uttar Pradesh, and West Bengal (Susheella 1999; CGWB 2015).

A lot of research on occurrence of F− in groundwater has been carried out in different parts of the world: Harrison (2005) discussed the impact of F− content on human health in the UK. Ayoob and Gupta (2006) and Ozsvath (2009) carried out a literature review on health impacts, sources, and mobility of F−, which provides a comprehensive insight into the relation among F−, human health, and the environment. Chae et al. (2007) studied the hydrogeochemistry of bed rocks in order to evaluate the geological and geochemical controls on the concentration of F− in groundwater in South Korea. Fordyce et al. (2007) and Wen et al. (2013) assessed the areas of high F− waters and their health effects in Central Europe and China. Madhnure et al. (2007) stated that the long residence time of water in deeper aquifers than in shallow aquifers of Maharashtra (India) was responsible for higher concentration of F− in groundwater. Ayenew (2008) reported the high concentration of F− caused by low concentration of Ca2+ under alkaline environment in Ethiopian groundwater. Hudak (2009) observed the contamination of F− groundwater, especially in public areas as well as in irrigated areas in West Texas. Suman et al. (2009) studied the controlling factors of F− content in groundwater in a semi-arid region (Sirsa City, Haryana) of India, using statistical and multivariate techniques. Li et al. (2010) performed a geochemical model to explain an evolution mechanism of F− in groundwater in Pengyang County, China. Fluorite, apatite, biotite, hornblende, and clay are the minerals, which have the greatest effect on geochemistry of F− in groundwater (Dar et al. 2011), when they are contacted with groundwater and dissolved, leading to increase of F− in groundwater. Li et al. (2011, 2014, 2017) concluded in their study conducted in different parts of China that the high F− was controlled by hydrogeological conditions, but man-made activities and mixing of different recharge waters were also important reasons. Ghrefat et al. (2014) found that the geochemical environment is the main controlling factor of F− content in groundwater from the region of Midyan Basin, Saudi Arabia. Wu et al. (2015) assessed the hydrochemical characterization of drinking groundwater with reference to F− content in an arid area of China.

Handa (1975), Jacks et al. (1993, 2005), Shaji et al. (2007), Mamatha and Rao (2010), Reddy et al. (2010, 2015), Reddy (2014), Brindha et al. (2011), Chidambaram et al. (2013), Kumar et al. (2014), Rao et al. (2014), Singaraja et al. (2014), Dhanya Raj and Shaji (2017), and Satyanarayana et al. (2017) examined the factors responsible for higher concentration of F− in groundwater of different states (Andhra Pradesh, Karnataka, Kerala, Tamil Nadu, Telangana) of South India and found that the major source of F− is caused by dissolution of F−-bearing minerals (fluorite, apatite, hornblende, and biotite), ion exchange, and evaporation and also the usage of chemical fertilizers.

From the cited literature reports, it is clear that the geochemical environment of groundwater is the main controlling factor of F− migration, while the anthropogenic activities like application of chemical fertilizers are the adding factors for enrichment of F− in groundwater. However, the intensity and activity of the chemical reactions are generally different from place to place, depending upon the availability of the source material, movement of groundwater flow, and climatic conditions. These result in the varying of F− content. In view of the increased importance worldwide on the concentration of F− in relation to its effects on human health, the present research has been carried out to assess the controlling factors of F− in the groundwater, using Piper and Gibbs diagrams, correlations, saturation indices, ionic ratios, and multivariate analysis, in a part of Prakasam district, Andhra Pradesh, South India. Hence, the present study will be of great significant to the scientific community.

Study area

Location

The present study area (15o 24′−15o-39′30″ N and 80o 15′−80o 35″ E) is located in a part of Prakasam district, Andhra Pradesh, India (Fig. 1). It covers an area about 325 km2, falling in the Survey of India toposheets 66 A/2 and 66 A/3. Ongole is the main city. The population is about 5,00,000. The area experiences a semi-arid climate, with minimum and maximum temperatures of 28 and 39 °C, respectively. The annual rainfall is about 910 mm. Southwest monsoon (June to September) contributes 69%, northeast monsoon (October to December) 29%, and the remaining seasons (winter and summer) 12% of the total rainfall. The river Gundlakamma flows through the area, which is an ephemeral nature. The drainage pattern is dendritic to sub-dendritic. The drainage density is generally less than 0.4 km/km2 (CGWB 2013).

Fig. 1
figure 1

Location of the study area, Prakasam district, Andhra Pradesh, South India. Numbers from 1 to 30 denote the locations of sampling points. Number in square denotes the contour line of spatial distribution of depth to groundwater level (m). Direction of groundwater flow is from northwest to southeast

Full size image

Geology

Topographically, the slope of the study area is towards southeast. The silty clay soils are the dominant type with the occurrence of calcareous concretions (CaCO3) locally known as kankar, which is a characteristic feature of the semi-arid climate. Geologically, the study area comprises of the Precambrian formations (Charnockites of the Eastern Ghats Super-group and unclassified Metamorphic Super-group of the Archaeans with younger intrusive bodies), Upper Gondwana sequence of Lower Cretaceous age, and Quaternary formations (Fig. 2). Pyroxene granulites come under the Charnockite group (CGWB 2013). Migmatized quartzo-felspathic-gneiss/schist associated with the amphibolites comes under the unclassified metamorphic rocks. Intrusive body comprising gabbro/norite occurs. The rocks are intensely migmatized at places. They constitute essentially of quartz, plagioclase and orthoclase feldspars, hornblende, hypersthene, sillimanite, garnet, apatite, and biotite. The rocks show a strike from NE–SW to NW–SE, with moderate to steep dip. Sandstones belong to the Upper Gondwana formations. Laterite, brown silty clay, black silty clay, brown silty sand, brown sand, and white to gray sand come under the Quaternary deposits, which occur over the basement rocks.

Fig. 2
figure 2

Geology of the study area (after CGWB 2013). Dotted lines denote the major fractures

Full size image

Hydrogeological conditions

The rocks of Precambrians and Gondwanas come under the consolidated formations, which occupy the large part of the study area (Fig. 2). They are not generally porous and permeable. But, they become aquifers due to weathering of the rocks and development of fractures in the rocks. However, the clay products, resulting from the highly weathering activity in the upper part of the bedrock, reduce the permeability to some extent. Thickness of the weathered rock portion varies from 3 to 6 m and the fractured rock portion from 5 to 35 m, depending upon the topography and rock type. The Quaternary deposits are the aquifers by virtue of their unconsolidated formations.

Groundwater is extracted through shallow wells and deep wells for domestic and irrigation purposes. The former wells are mainly confined to the weathered zone under unconfined conditions, while the latter wells extend into the fractured zone under semi-confined conditions. The depth to water levels is in between 3 and 16 m below ground level, fluctuating within the weathered and fractured rock portions. Spatial distribution of depth to groundwater level shows that the water levels are higher in northwestern side (> 15 m) and lower in southeastern side (< 5 m; Fig. 1), depending upon the porosity and permeability characters of lithology and slope of the area.

The present study area (325 km2) is a part of the regional hydrogeological set-up of the Prakasam district, spreading over an area of 17,626 km2. Depending upon the porosity and permeability characters of the geological formations present in the study area, the yield of the wells varies from place to place. The average yield of the wells located in the weathered and fractured rock portions is 50 and 100 m3/day, respectively, while it is 425 and 250 m3/day in the river alluvium and coastal zone, respectively (CGWB 2013). The irrigation wells are pumped at a discharge of about 4 L/s to meet the demand of high agricultural production (Reddy et al. 2015). The quality of groundwater is not generally potable, which is mostly brackish. Fluorosis is observed in this region during the field work.

Agriculture is the main occupation of the people. Irrigation is intensive and long-term practice in the study area. Soil amendments (gypsum) are used to alter the physical and chemical properties of soil permeability. The application of uncontrolled usage of chemical fertilizers (potash, phosphate, nitrogen and sulphate varieties) for higher crop yields is a common practice in the study area. Flood irrigation is very limited. Important crops are rice, groundnuts, chilies, tobacco, cotton, etc. Drainage conditions (dumping of wastes on the street sides and empty places, flowing of waste waters through leaked channels, leakage of septic tanks, etc) are poor.

Materials and methods

Sample collection and analysis

A total of 30 groundwater samples were monitored during summer 2012 and their locations are shown in Fig. 1. Most of them are located in the inland area. Very few are located along the coast, but they are far off from the influence of sea water. The groundwater samples were collected in half-liter polythene bottles, after the bottles were soaked in 1:1 HCl for 24 h and were rinsed with distilled water followed by deionized water. The bottles were tightly capped without air gap. Immediately, after collection of the groundwater, the samples are shifted to laboratory for the chemical analysis, following the standard water quality procedures of American Public Health Association (APHA 2012).

The groundwater samples were analyzed for pH, electrical conductivity (EC), total dissolved solids (TDS), total hardness (TH), calcium (Ca2+), magnesium (Mg2+), sodium (Na+), potassium (K+), bicarbonate (HCO\( {}_3^{-} \)), chloride (Cl−), sulphate (SO\( {}_4^{2-} \)), nitrate (NO\( {}_3^{-} \)), and fluoride (F−). The pH and EC were measured in the field, using portable pH and EC meters. The EC was used to calculate the concentration of TDS, following the procedure of Hem (1991). The TH (as CaCO3) and Ca2+ were analyzed volumetrically, using standard EDTA. The Mg2+ was computed, taking the difference between TH (as CaCO3) and Ca2+. A flame photometer was used for estimation of Na+ and K+ ions. The HCO\( {}_3^{-} \) was estimated by titrating with standard HCl. The Cl− was analyzed by titrating with standard AgNO3. The SO\( {}_4^{2-} \), NO\( {}_3^{-} \), and F− were determined, using turbidimetric procedure, colorimeteric method, and ion analyzer, respectively. The units of all chemical variables (except pH) are expressed in milligrams per liter (mg/L).

Ionic-balance-error

Ionic-balance-error between the total concentrations of cations (Ca2+, Mg2+, Na+, and K+) and the total concentrations of anions (HCO\( {}_3^{-} \), Cl−, SO\( {}_4^{2-} \), NO\( {}_3^{-} \), and F−) was checked for accuracy of each groundwater sample, following the formula (Eq. 1), where all ions are expressed in milliequivalent per liter (meq/L). The calculated ionic-balance-error was observed to be within the acceptable limit of ± 5% (Domenico and Schwartz 1990).

$$ \mathrm{Ionic}\hbox{-} \mathrm{balance}\hbox{-} \mathrm{error}=\frac{\sum \mathrm{Cations}-\sum \mathrm{Anions}}{\sum \mathrm{Cations}+\sum \mathrm{Anions}}\times 100 $$
(1)

Chloro-alkaline indices

Chloro-alkaline (CA) indices viz. CA1 and CA2 proposed by Schoeller (1965) were calculated, following the formulas (2 and 3), where all concentrations of ions are expressed in meq/L. They are widely used to assess the possible mechanism of ion exchange between the groundwater and host rocks (Marghade et al. 2012; Li et al. 2013a; Wu and Sun 2016).

$$ \mathrm{CA}1\kern0.75em =\left[\frac{{\mathrm{Cl}}^{\hbox{-} }-{\mathrm{Na}}^{+}+{\mathrm{K}}^{+}}{{\mathrm{Cl}}^{\hbox{-} }}\right] $$
(2)
$$ \mathrm{CA}2\kern0.75em =\left[\frac{{\mathrm{Cl}}^{\hbox{-} }-\kern0.5em {\mathrm{Na}}^{+}+{\mathrm{K}}^{+}}{{\mathrm{SO}}_4^{2\hbox{-} }+{\mathrm{HCO}}_3^{\hbox{-} }+{\mathrm{CO}}_3^{2\hbox{-} }+{\mathrm{NO}}_3^{\hbox{-} }}\right] $$
(3)

Saturation index

The intensity of soluble minerals is expressed as saturation index (SI). Thus, it is used to evaluate the degree of equilibrium between water and respective mineral (Garrels and Christ 1965; Fetter 1990; Li et al. 2010), which is calculated as (4).

$$ \mathrm{SI}=\frac{{\mathrm{K}}_{\mathrm{IAP}}}{{\mathrm{K}}_{\mathrm{SP}}} $$
(4)

where IAP is the ion activity product and SP is the solubility product.

If SI is zero, it indicates saturated state (equilibrium) with respect to the particular mineral. If SI is less than zero, it shows under-saturated state (dissolution) with respect to the concerned mineral. If SI is more than zero, it suggests over-saturated state (precipitation) with respect to the particular mineral (Li et al. 2010).

Statistical analysis

The chemical data of the groundwater obtained from the chemical analysis are subjected to statistical analysis to (a) measure the distribution and central tendency of the data set, (b) know about the interrelationship of the chemical variables, and (c) assess the sources of chemical species. Multivariate statistical analysis was also carried out to find out the possible sources of chemical variables (Ca2+, Mg2+, Na+, K+, HCO\( {}_3^{-} \), Cl−, SO\( {}_4^{2-} \), NO\( {}_3^{-} \) and F−), using Statistic (version 6) software (StatSoft Inc. 2001).

In R-mode analysis, principal component analysis (PCA) was performed with varimax orthogonal rotation and Kaiser normalization to reduce the data set and also to explain the possible sources of controlling processes of chemical variables (Kaiser 1958; Davis 1986; Suman et al. 2009; Nosratic and Eechhaut 2012; Kim and Park 2016; Ravikumar and Somashekar 2017).

Results and discussion

Groundwater quality

The chemical composition of the groundwater is shown in Table 1 and Fig. 3. The pH varies from 7.2 to 8.3, with an average of 7.81, indicating an alkaline condition. This is within the recommended limit of 6.5 to 8.5 for drinking water in all groundwater samples (WHO 2011; BIS 2012). The TDS is in the range of 995 to 3290 mg/L and its average is 1884.27 mg/L, showing a wide variation in the degree of water quality. As per the TDS classification (Fetter 1990), about 97% of the total samples (1 to 6 and 8 to 30) come under the brackish water type (1000 to 10,000 mg/L) and the remaining 3% (sample 7) under the freshwater type (< 1000 mg/L). In all groundwater samples, the concentration of TDS is more than its permissible limit of 500 mg/L prescribed for drinking water. This causes a gastrointestinal irritation (WHO 2011; BIS 2012).

Table 1 Chemical composition of groundwater samples. All are expressed in mg/L except pH. Concentrations in bold denote the number exceeding the safe limits prescribed for drinking water
Full size table
Fig. 3
figure 3

Box plot for the relative variation of groundwater quality

Full size image

Among cations, the Ca2+ varies from 60 to 75 mg/L and the Mg2+ from 42 to 105 mg/L, being averages of 65.67 and 76.87 mg/L, respectively (Table 1). The concentration of Ca2+ is within the safe limit of 75 mg/L, except in two water samples ( 8 and 23), while that of Mg2+ is higher than that of its prescribed limit (30 mg/L) recommended for drinking water in all water samples (WHO 2011; BIS 2012). It leads to a scale formation in water distribution structures. The Na+ is in between 189 and 823 mg/L with its average of 444.73 mg/L. The concentration of Na+ is higher than its allowable limit of 200 mg/L for drinking water (BIS 2012) in 90% of the total water samples (1, 4 to 6 and 8 to 30). It effects on human health, who suffer from the hypertension. The concentration of K+ varies from 12 to 121 mg/L, being an average of 54.33 mg/L. Generally, it is less than 10 mg/L in groundwater (Davis and Dewiest 1966). The higher concentration of K+ than that of 10 mg/L is, thus, considered to be a man-made pollution.

Among anions, the HCO\( {}_3^{-} \)ranges from 480 to 1025 mg/L and its average is 731.33 mg/L (Table 1). The threshold limit of HCO\( {}_3^{-} \)is 300 mg/L in groundwater (BIS 2012). In all groundwater samples, the concentration of HCO\( {}_3^{-} \)exceeds its recommended limit. The concentration of Cl− is from 215 to 1060 with an average of 568.83 mg/L. This is more than its safe limit of 250 mg/L prescribed for drinking water (BIS 2012) in 90% of the total groundwater samples (1, 2, 4 to 6, 8, 9, and 11 to 30). It develops salt water, which is a common phenomenon in the study area. The SO\( {}_4^{2-} \)content varies from 22 to 125 and its average is 71.47 mg/L. In all groundwater samples, the concentration of SO\( {}_4^{2-} \)is within its safe limit of 200 mg/L allowed for drinking water (BIS 2012). The NO\( {}_3^{-} \) is in between 31 and 97 me/L, being an average of 56.20 mg/L. The NO\( {}_3^{-} \) is higher than its recommended limit of 45 mg/L permitted for drinking water (BIS 2012) in 73% of the total groundwater samples (1, 5, 6, 8 to 11, 13, 14, 16 to 18, and 20 to 30). This causes a blue baby disease. However, it is not observed from the study area. The higher concentration of NO\( {}_3^{-} \), which is more than 10 mg/L, indicates an anthropogenic pollution (Cushing et al. 1973). The concentration of F− is from 0.70 to 2.80 mg/L with an average of 1.34 mg/L. Spatial distribution of F− is shown in Fig. 4. In 47% of the total groundwater samples (4, 8, 9, 11, 12, 21 to 28 and 30), the F− content is higher than its safe limit of 1.50 mg/L assigned for drinking water (WHO 2011; BIS, 2012). It causes a fluorosis, which is observed from the study region during the field work.

Fig. 4
figure 4

Spatial distribution of F− (mg/L). Forty seven percent of the area exceeds the safe limit of F− content (1.50 mg/L) prescribed prescribed for drinking water

Full size image

From Table 1, it is noted that the chemical variables show a wide range of standard deviations (SD). The TDS, Cl−, HCO\( {}_3^{-} \), and Na+ have the higher values of SD (781.93, 303.48, 200.97, and 228.72, respectively) compared to the other chemical variables (pH, Ca2+, Mg2+, K+, SO\( {}_4^{2-} \), NO\( {}_3^{-} \), and F−). The values of SD in these chemical parameters vary from 0.40 to 28.49. It is also significant to note that the K+ (28.49) and SO\( {}_4^{2-} \)(28.12) as well as Mg2+ (18.87) and NO\( {}_3^{-} \)(18.91) have almost the similar values of SD, followed by Ca2+ (8.88), while the pH (0.40) and F− (0.56) have the lowest values of SD. The difference in the values of SD among chemical variables clearly suggests that there is a wide dispersion of salts in the groundwater system. This is a result of differences in (a) the availability of sources of salts and (b) the processes of mineral dissolution, mineral solubility, ion exchange, and evaporation over the control of water chemistry (Ghrefat et al. 2014). Further, the values of coefficient of variation (CV) vary from 5.12 to 67.81%. The difference in the CV among chemical variables also clearly supports the spatial variation of chemical composition of groundwater.

Hydrogeochemical evolution

A trilinear diagram (Piper 1953) helps for characterization of hydrogeochemical evolution (Fig. 5). This diagram consists of three distinct fields, including two triangular fields on the lower side to plot cations (left side) and anions (right side) and a diamond-shaped field on upper side of these two triangular fields to plot an overall chemical quality of groundwater for characterization of hydrogeochemical evolution through assessment of various water types viz. zone-5 (carbonate hardness or temporary harness or freshwater), zone-6 (non-carbonate hardness or permanent hardness), zone-7 (non-carbonate alkali or saline water), zone-8 (carbonate alkali or excess alkalinity), and zone-9 (mixed or transition water).

Fig. 5
figure 5

Hydrogeochemical evolution (after Piper 1953). Zone-5 denotes the carbonate hardness (temporary harness or freshwater), zone-6 non-carbonate hardness (permanent hardness), zone-7 non-carbonate alkali (saline water), zone-8 carbonate alkali (excess alkalinity) and zone-9 mixed (transition) water. Arrow mark denotes the flow path of groundwater

Full size image

As shown in Fig. 5, the chemical data of the groundwater sampling points fall in the right lower corner of the cation triangle, which reveals that the Na+ is a dominant ion in the groundwater. Similarly, the groundwater samples are observed from the central lower part of the anion triangle, indicating the higher concentrations of both HCO\( {}_3^{-} \)and Cl− ions in the groundwater. Most groundwater sampling points fall in the central part of the diamond-shaped field and their trend direction is towards the right side from the left side, following the flow path. It suggests that the freshwater (zone-5) moves towards the saline water (zone-7) through the mixed water (zone-9). The changing of Na+-HCO\( {}_3^{-} \)type gradually towards the Na+-Cl− type is a result of influence of anthropogenic activity along the flow path.

The characterization of hydrogeochemical evolution obviously explains the change in degree and type of mineralization of groundwater due to different stages of interaction of recharged water through soil cover and weathered/fractured rock portions with the contaminated water caused by human activities. During its initial stage of movement, the water exchanges Ca2+ for Na+ by ion exchange process (Hem 1991). When the groundwater is dominated by Na+, it should have HCO\( {}_3^{-} \) as a predominant ion due to CO2, which releases from the decay of organic and root respiration in soil zone (Rogers 1989). This combines with H2O to form HCO\( {}_3^{-} \) (Eqs. 5 and 6). As a result, the groundwater shows Na+-HCO\( {}_3^{-} \)type water. It causes the higher concentration of F− in the groundwater (Eq. 7), which is in consistency with the results of Chae et al. (2007). In the latter stage, the groundwater shows Na+-Cl− type water due to influence of man-made pollution on the fresh groundwater system. These two ions increase the TDS, which includes F− content in the groundwater.

$$ {\mathrm{CO}}_2\left(\mathrm{g}\right)+{\mathrm{H}}_2\mathrm{O}(1)\kern0.5em \rightleftarrows \kern0.5em {\mathrm{H}}_2{\mathrm{CO}}_3\left(\mathrm{aq}\right) $$
(5)
$$ {\mathrm{H}}_2{\mathrm{CO}}_3\left(\mathrm{aq}\right)\kern0.5em \rightleftarrows \kern0.5em {\mathrm{H}}^{+}\left(\mathrm{aq}\right)+{\mathrm{H}\mathrm{CO}}_3^{-}\left(\mathrm{aq}\right) $$
(6)
$$ {\mathrm{CaF}}_2\left(\mathrm{s}\right)+2{\mathrm{H}\mathrm{CO}}_3^{-}\left(\mathrm{aq}\right)\kern0.5em \rightleftarrows \kern0.5em {\mathrm{CaCO}}_3\left(\mathrm{s}\right)+{\mathrm{H}}_2\mathrm{O}\ \left(\mathrm{l}\right)+{\mathrm{CO}}_2\left(\mathrm{g}\right)+2{\mathrm{F}}^{-}\left(\mathrm{aq}\right) $$
(7)

Mechanism controlling groundwater chemistry

The hypothesis explained in the hydrogeochemical characterization is further verified by Na+ + K+:Na+ + K+ + Ca2+ and Cl−:Cl− + HCO\( {}_3^{-} \) ratios, where they are plotted against TDS (Fig. 6), following the procedure of Gibbs (1970). The diagrams of Gibbs explain the three mechanisms that control the water chemistry with respect to the influences of precipitation (rainwater), rock (soil and/or lithology), and evaporation (climate).

Fig. 6
figure 6

Mechanisms (precipitation, rock, and evaporation) controlling the groundwater chemistry (after Gibbs 1970)

Full size image

In the present study, the chemical data of the groundwater sampling points fall in rock and evaporation domains (Fig. 6). No single sample is observed from the domain of precipitation (TDS < 100 mg/L). Falling of a few groundwater samples in the domain of rock indicates that the geogenic origin is the main responsible factor for the increase of ions from the domain of precipitation towards the domain of rock (TDS 100 to 1000 mg/L). This is a result of interaction of water with soils and/or weathering of rock forming minerals, before reaching the water table.

A general reaction for water-rock interaction is shown in Eq. (8). According to this reaction, the groundwater should have the Na+ and HCO\( {}_3^{-} \)as the dominant ions. However, the influence of sources of evaporation and/or anthropogenic origin increases the concentrations of Na+ and Cl− ions (Todd 1980; Hem 1991; Drever 1997), which make the higher mineralization, i.e., TDS (> 1000 mg/L) in the groundwater. As a result, the chemical data of the groundwater samples moves towards the domain of evaporation from the domain of rock in the study area (Fig. 6).

$$ \left({\mathrm{Na}}^{+},{\mathrm{Mg}}^{2+},{\mathrm{Ca}}^{2+},{\mathrm{K}}^{+}\right)\ \mathrm{silicates}\ \left(\mathrm{s}\right)+{\mathrm{H}}_2{\mathrm{CO}}_3\left(\mathrm{aq}\right)\kern0.36em \to \kern0.5em {\mathrm{H}\mathrm{CO}}_{\overline{3}}\ \left(\mathrm{aq}\right)+{\mathrm{Na}}^{+}\left(\mathrm{aq}\right)+{\mathrm{Mg}}^{2+}\left(\mathrm{aq}\right)+{\mathrm{Ca}}^{2+}\left(\mathrm{aq}\right)+{\mathrm{K}}^{+}\left(\mathrm{aq}\right)+{\mathrm{H}}_4{\mathrm{SiO}}_4\left(\mathrm{aq}\right)+\mathrm{Clays}\ \left(\mathrm{s}\right) $$
(8)

Perel’man (1977) reported that the higher concentration of TDS enhances the ionic strength. This leads to an increase of the solubility of CaF2. This causes the higher concentration of F− in groundwater, as observed in the present study area (Table 1). Kumar et al. (2014) and Li et al. (2014) also observed the similar results from the regions of Vellore district (India) and Weihe River (China), respectively.

Controlling factors of F− content

Correlation coefficient analysis

Pearson’s correlation coefficient (r) is widely used tool to get information on associations between individual chemical variables (Jiang et al. 2009; Suman et al. 2009; Vasanthavigar et al. 2013). It demonstrates the overall data set that indicate their controlling factors (Li et al. 2011, 2012, 2014). The calculated correlation coefficient matrix is presented in Table 2. The binary plots were plotted in Fig. 7 to assess the possible controlling factors of F− content in the groundwater. If the geogenic origin is the only source for dissolved ions, their spreading points are close to the linear trend. If there is any deviation in the spreading points from the linear trend, it indicates the involved of other sources (like anthropogenic origin) for the dissolved contents.

Table 2 Pearson’s correlation coefficients. Values in bold denote the high positive correlations
Full size table
Fig. 7
figure 7

Correlation between a pH and F−, b Ca2+ and F−, c Mg2+ and F−, d Na+ and F−, e HCO\( {}_3^{-} \)and F−, f SO\( {}_4^{2-} \) and F−, g K+ and F−, and h NO\( {}_3^{-} \) and F−

Full size image

The positive correlation of pH with F− (r = 0.64; Fig. 7a and Table 2) suggests that the F− tends to be higher in the higher level of pH due to precipitation of Ca2+ and dissolution and/or replacement of F− with OH− (Li et al. 2014). Further, the Ca2+ and Mg2+ ions have positive correlations with F−. However, the Ca2+ shows low correlation with F− (r = 0.30; Fig. 7b), while the Mg2+ has good correlation with F− (r = 0.72; Fig. 7c). This difference can be explained in a general perspective with respect to CaF2 and MgF2 minerals. For instance, the SI of CaF2 is 3.11 and of MgF2 is 1.56 (Table 3), indicating that these salts are in the precipitated state. The higher precipitation of CaF2 than that of MgF2 may be caused by differences in their solubility (Hem 1991). On the other hand, high positive correlation is observed between HCO\( {}_3^{-} \) and Mg2+ (r = 0.76) and low correlation between HCO\( {}_3^{-} \) and Ca2+ (r = 0.20). This suggests that MgCO3 is the main source of these ions rather than the CaCO3, which is further supported by positive SI of CaCO3 (0.93) and negative SI of MgCO3 (−0.20). The natural occurrence of calcareous concretions in the soils is the direct evidence of precipitation of CaCO3 due to semi-arid climate of the study area, supporting the above finding.

Table 3 Average values of saturation index (SI), chloro-alkaline (CA) index, and geochemical ratio
Full size table

Significantly, it is noted that the concentration of Ca2+ (average 65.67 mg/L) is lower than that of Na+ (average 444.73 m/L). The present study area (325 km2) is a discharge part of the regional hydrogeological set-up of the Prakasam district (17,627 km2). Therefore, the difference in the values of Ca2+ and Na+ is a result of the discharge area. Similar observation has been made by Wu et al. (2015) in an arid area of China. Further, the higher Na+ than the Ca2+ in the groundwater is a result of two geochemical processes—(a) ion exchange, which leads to an increase of Na+ by a decrease of Ca2+, and (b) precipitation of CaCO3, which reduces the concentration of Ca2+ due to semi-arid climate (Li et al. 2016b, c). Jack et al. (1993) stated that the rate of evaporation increases the concentration of HCO\( {}_3^{-} \), while the precipitation of carbonates occurs.

The Mg2+ salts do not precipitate readily as the Ca2+ salts due to differences in their solubility (Hem 1991). As a result, very low correlation exists between Na+ and Ca2+ (r = 0.10) compared to that between Na+ and Mg2+ (r = 0.79; Table 2). On the other hand, the solubility of F− is low (Nordstrom and Jenne 1977). However, it increases with the increase of Na+ in groundwater (Apambire et al. 1997). Since the Na+ is the dominant ion in the groundwater (Table 1) due to rock weathering (Stumm and Morgan 1996), it favors the release of F− (Li et al. 2014). Consequently, good correlation is observed between Na+ and F− (r = 0.85; Fig. 7d). Because of the rock weathering, the HCO\( {}_3^{-} \) can also release along the Na+ into the groundwater (Eq. 8). This accelerates the dissolution of F−-bearing minerals more effectively (Shyam and Kalwania 2012), which favors the release of F−. This pollutes the groundwater. A high positive correlation is, thus, observed between HCO\( {}_3^{-} \)and F− (r = 0.94; Fig. 7e).

Sodium has good positive correlation with Cl− (r = 0.98; Table 2). This indicates that these two ions are derived from the same source and also they have the higher solubility conditions (Hem 1991; Li et al. 2016b). Therefore, the SI of NaCl shows the negative value (−2.06; Table 3), which indicates the dissolution state of Na+ and Cl− ions in the groundwater. However, the calculated ratio of Cl− to Na+ is less than unity (0.83; Table 3). If the groundwater is contaminated with sea water, the ratio between Cl− and Na+ should be more than one, because the sea water contains more concentration of Cl− (19,000 mg/L) than that of Na+ (10,500 mg/L; Hem 1991). But, the influence of sea water on the groundwater system is not so happened in the present study area so that the sources of Na+ and Cl− ions should be related to other origins. Thus, the dissolution of sodium feldspars is the primary source of Na+. Further, the clays are also the sources of both Na+ and Cl− ions in addition to the drainage wastes (Todd 1980; Hem 1991).

Since the clays occur in the present study area, it is important to assess the role of ion exchange between Ca2++Mg2+ and Na++K+, which control the geochemistry of groundwater. Chloro-alkaline (CA) indices viz. CA1 and CA2 proposed by Schoeller (1965) are widely used to evaluate the mechanism of ion exchange in groundwater (Marghade et al. 2012; Li et al. 2016c). If CA is positive, it indicates an exchange of Na+ and K+ from the water with Ca+ and Mg2+ of the rocks. It is called base-exchange reaction. If CA is negative, it specifies an exchange of Ca2+ and Mg2+ of the water with Na+ and K+ of the rocks. It is called cation-anion exchange reaction. In the present study area, the CA shows negative values (CA1 = −0.29 and CA2 = −0.31; Table 3). This indicates that the Ca2++Mg2+ in the groundwater exchanged for Na++K+ from aquifer materials, which is also supported by the ratio Na+:Ca2+ (3.04; Table 3). Further, the negative trend of Ca+ + Mg2+ − HCO\( {}_3^{-} \) + SO\( {}_4^{2-} \) vs Na+ + K+ − Cl− also indicates the existence of reverse cation-anion exchange phenomenon in the groundwater (Fig. 8). Marghade et al. (2012), Ahmed et al. (2013), Li et al. (2013b), and Wu et al. (2015) also reported the similar observations in India (Nagpur city), Egypt (Eastern Nile Delta), and China (Ordos Basin and Shizuishan city).

Fig. 8
figure 8

Relationship between Ca2+ + Mg2+ – SO\( {}_4^{2-} \)+ HCO\( {}_3^{-} \) and Na+ + K+ – Cl−. Negative trend denotes the reverse cation-anion exchange phenomenon

Full size image

It is significant to note that the F− shows a good positive correlation with SO\( {}_4^{2-} \)(r = 0.74; Fig. 7f), K+ (r = 0.71; Fig. 7g), and NO\( {}_3^{-} \)(r = 0.69; Fig. 7h) and their correlation values are more or less the same. This indicates a common source of these ions. The chemical fertilizers (potash, phosphate, nitrogen, and sulphate varieties) are the main sources of K+, Cl−, SO\( {}_4^{2-} \), NO\( {}_3^{-} \), and F− ions, as reported by Jiang et al. (2009), Chidambaram et al. (2013), and Li et al. (2016c). Further, F− also shows a good positive correlation with TDS (r = 0.86; Table 2). Perel’man (1977) and Rogers (1989) stated that the TDS increases the ionic strength, which increases the solubility of CaF2, causing the higher concentration of F− in the groundwater (Table 1). Furthermore, the TDS has a good correlation with Mg2+ (r = 0.84), Na+ (r = 0.99), K+ (r = 81), Cl− (r = 0.98), SO\( {}_4^{2-} \)(r = 0.90), and NO\( {}_3^{-} \)(r = 0.77). It represents the influences of geogenic and anthropogenic sources on the groundwater system. Otherwise, such relations should not have among these chemical parameters.

Multivariate analysis

Hierarchical cluster analysis

Hierarchical cluster analysis (HCA) classifies the chemical variables into clusters or groups, which provide a preface of water quality, depending upon the intensity of chemical variables (Suman et al. 2009). The groups are similar to each other group, but distinct from the other groups (Li et al. 2013b; Wu et al. 2014).

From the cluster analysis, all chemical variables (Ca2+, Mg2+, Na+, K+, HCO\( {}_3^{-} \), Cl−, SO\( {}_4^{2-} \), NO\( {}_3^{-} \), and F−) can be divided into two major groups. Group I is related to Ca2+, Mg2+, K+, SO\( {}_4^{2-} \), NO\( {}_3^{-} \), and F− ions, while group II is associated with Na+, HCO\( {}_3^{-} \), and Cl− ions (Fig. 9). The former group implies the weathering of plagioclase feldspar (Eq. 9), hornblende (Eq. 10), biotite (Eq. 11), and apatite (Eq. 12) minerals (Ca2+, Mg2+, Na+, K+, SO\( {}_4^{2-} \), and F−) and the influences of chemical fertilizers and drainage wastes (Mg2+, Na+, K+, SO\( {}_4^{2-} \), NO\( {}_3^{-} \), and F−) on the groundwater body. Further, the occurrence of sandstone is also responsible for F− content in the groundwater, as it contains 350 mg/L of F− (Rao et al. 2014). The latter group represents the weathering of plagioclase feldspar mineral (Na+ and HCO\( {}_3^{-} \); Eq. 9) and also the influence of soil salts developed by evaporation (Na+, HCO\( {}_3^{-} \), and Cl−) on the groundwater system. Similar observations have been reported from the other regions (Raju et al. 2009; Zhang et al. 2013; Christou et al. 2014; Rao et al. 2014; Reddy et al. 2010, 2015; Reddy 2014; Singaraja et al. 2014; Satyanarayana et al. 2017).

Fig. 9
figure 9

Dendrogram. Classification of chemical variables into groups depending upon the intensity of chemical variables, which are similar to each other group, but distinct from other groups

Full size image
$$ {\displaystyle \begin{array}{l}2{\mathrm{Na}}^{+}{\left[{\mathrm{Al}\mathrm{Si}}_3{\mathrm{O}}_8+2{\mathrm{CO}}_2\right]}^{\hbox{-} }+11{\mathrm{H}}_2\mathrm{O}\kern0.36em \to 2{\mathrm{Na}}^{+}+2{\mathrm{H}\mathrm{CO}}_{\overline{3}}+{\mathrm{Al}}_2{\mathrm{Si}}_2{\mathrm{O}}_5{\left(\mathrm{OH}\right)}_4+{\mathrm{Mg}}^{2+}+4{\mathrm{H}}_4{\mathrm{Si}\mathrm{O}}_4\\ {}\mathrm{Plagioclase}\ \mathrm{feldspar}\end{array}} $$
(9)
$$ {\displaystyle \begin{array}{l}{\left(\mathrm{Ca},\mathrm{Na}\right)}_{2\hbox{--} 3}{\left(\mathrm{Mg},\mathrm{F}\mathrm{e},\mathrm{Al}\right)}_5{\left(\mathrm{Al},\mathrm{Si}\right)}_8{\mathrm{O}}_{22}\left(\mathrm{OH},\mathrm{F}\right)\to {\left(\mathrm{Ca},\mathrm{Na}\right)}_{2\hbox{--} 3}{\left(\mathrm{Mg},\mathrm{F}\mathrm{e},\mathrm{Al}\right)}_5{\left(\mathrm{Al},\mathrm{Si}\right)}_8{\mathrm{O}}_{22}{\left(\mathrm{OH}\right)}_2+2{\mathrm{F}}^{-}\ \\ {}\mathrm{Hornblende}\end{array}} $$
(10)
$$ {\displaystyle \begin{array}{l}{\mathrm{KMg}}_3\left({\mathrm{AlSi}}_3{\mathrm{O}}_{10}\right){\mathrm{F}}_2+2{\mathrm{O}\mathrm{H}}^{-}\to {\mathrm{KMg}}_3\left({\mathrm{AlSi}}_3{\mathrm{O}}_{10}\right){\left({\mathrm{O}\mathrm{H}}^{-}\right)}_2+2{\mathrm{F}}^{-}\\ {}\mathrm{Biotite}\end{array}} $$
(11)
$$ {\displaystyle \begin{array}{l}{\mathrm{Ca}}_5{\left({\mathrm{PO}}_4\right)}_3\left(\mathrm{F},\mathrm{Cl},\mathrm{OH}\right)\to {\mathrm{Ca}}_5{\left({\mathrm{PO}}_4\right)}_3\left(\mathrm{Cl},\mathrm{OH}\right)+{\mathrm{F}}^{-}\\ {}\mathrm{Apatite}\end{array}} $$
(12)

As explained earlier, the HCO\( {}_3^{-} \) favors the solubility of F−-bearing minerals more effectively. As a result, these two ions are interdependent with each other. However, such relation is not observed from Fig. 9. For example, the F− is associated with group I and HCO\( {}_3^{-} \) with group II. This difference can be explained by HCA, which pools multiple (> 1) variables into groups based on distance measures (StatSoft Inc. 2001). Each group is sub-divided in a hierarchical manner. It is a pseudo-dimension reduction technique. The advantage is looking into small number of groups, each containing several variables. Thus, F− and HCO\( {}_3^{-} \) fall into two different groups in the present study area.

Principal component analysis

Principal component analysis (PCA) is widely used for identification of influence of sources of chemical variables on variance in data matrix (Dalton and Upchurch 1978; Davis 1986; Jiang et al. 2009; Kim and Park 2016; Ravikumar and Somashekar 2017). In the present study, two principal components (PCs) are observed from the transformation of the chemical data of Ca2+, Mg2+, Na+, K+, HCO\( {}_3^{-} \), Cl−, SO\( {}_4^{2-} \), NO\( {}_3^{-} \), and F−, using varimax orthogonal rotation and Kaiser normalization, where the eigenvalue is more than one (Kaiser 1958; Davis 1986). The PC1 is related to the largest eigenvalues and is able to explain the greatest amount of variance in the data set, while the PC2 (orthogonal and uncorrelated with the PC1) explains the most of the remaining variance. These two PCs have a significant effect on variability of the groundwater chemistry.

Table 4 presents the results of principal component analysis. The total variation is 85.108%. The PC1 accounts for 72.824% of the total variation with high positive loadings of SO\( {}_4^{2-} \) (0.965), Na+ (0.943), Cl− (0.920), K+ (0.908), NO\( {}_3^{-} \) (0.879), HCO\( {}_3^{-} \) (0.862), Mg2+ (0.854), and F− (0.853). The PC2, which accounts for 12.284% of the total variation, is dominated by high positive loading of Ca2+ (0.961).

Table 4 Principal component analysis (PCA). Values in bold denote the high loadings of chemical variables
Full size table

In PC1, the Mg2+, Na+, K+, and F− ions represent the contribution of weathering of plagioclase and orthoclase feldspars, hornblende, biotite, and apatite; the Mg2+, Na+, K+, Cl−, SO\( {}_4^{2-} \), NO\( {}_3^{-} \), and F− ions reflect the influences of chemical fertilizers and drainage wastes; the Na+ and Cl− ions indicate the impact of evaporation; and the Na+, HCO\( {}_3^{-} \), and Cl− ions are the result of soil salts (Todd 1980; Hem 1991; Drever 1997; Manish et al. 2007; Suman et al. 2009; Chidambaram et al. 2013;Singaraja et al. 2014). In the PC2, the Ca2+ is a result of weathering of calcium feldspars (Hem 1991).

Since the Ca2+, Mg2+, Na+, K+, HCO\( {}_3^{-} \), SO\( {}_4^{2-} \), and F− are both the sources of lithological (geogenic) and non-lithological (anthropogenic) origins and the Cl− and NO\( {}_3^{-} \) are the sources of non-lithological origin, the PC1 can be explained as a result of both the geogenic and anthropogenic sources, while the PC2 is caused by the geogenic origin.

Conceptual model of geochemistry of fluoride

Based on the field conditions and hydrogeochemical data, it is clear that the source of F− in the groundwater is due to weathering of F−-bearing minerals (hornblende, biotite, and apatite) and sandstones. The processes of ion exchange (between Na+ and Ca2+) and evaporation (precipitation of CaCO3) enhance the concentration of Na+ by reduction of Ca2+, while the process of evaporation increases the concentration of HCO\( {}_3^{-} \). The Na+ and HCO\( {}_3^{-} \)ions, thus, favor the mineral weathering and dissolution and increase the solubility of F−-bearing minerals more effectively from the source. As a result, the F− shows a strong positive correlation with Na+ and HCO\( {}_3^{-} \). Further, the chemical fertilizers are the sources of K+, SO\( {}_4^{2-} \), NO\( {}_3^{-} \), and F− ions so that the F− shows a good positive correlation with K+, SO\( {}_4^{2-} \), and NO\( {}_3^{-} \). Therefore, the conceptual model explains the geogenic origin as the primary source for the occurrence of F− content and the anthropogenic input as the secondary source for further enrichment of F− in the groundwater.

Conclusions

Piper and Gibbs diagrams, correlations, saturation indices, ionic ratios, and multivariate analysis have been used to assess the controlling factors of F− in the groundwater in a part of South India. The following conclusions can be summarized.

  • The concentration of F− varies from 0.70 to 2.80 mg/L. It is higher (1.60 to 2.80 mg/L) than its safe limit of 1.50 mg/L recommended for drinking water in 47% of the total groundwater samples.

  • Piper and Gibbs diagrams suggest that the groundwater quality is characterized by Na+-HCO\( {}_3^{-} \)- and Na+-Cl−-type waters caused by water-rock interactions and anthropogenic sources, respectively.

  • A positive correlation of F− with pH, TDS, Mg2+, Na+, K+, HCO\( {}_3^{-} \), Cl−, SO\( {}_4^{2-} \), and NO\( {}_3^{-} \) and low with Ca+2, saturation indices, chloro-alkaline indices, and multivariate statistical analysis suggest that the mineral dissolution, ion exchange, and evaporation play a significant role as primary processes for the concentration of F− in the groundwater, while the influence of chemical fertilizers is the secondary factor for further enrichment of F− in the groundwater.