Introduction

Groundwater, as an easily accessible resource, not only meets the domestic water needs of people but also supports agricultural and industrial activities (Jiang et al. 2022). This dependence on groundwater is expected to increase in the future owing to the water demands of a rapidly rising global population (Xiao et al. 2022a). Excessive abstraction of groundwater exceeding the natural recharge inevitably leads to declining groundwater levels, seawater intrusion, land subsidence, and pollution (Wilopo et al. 2021; Orhan 2021). Groundwater pollution is also caused by contaminants released by anthropogenic activities that percolate into subsurface aquifers (Goyal et al. 2021; Motlagh et al. 2020). Assessment of groundwater quality and human health risks related to groundwater pollutants are thus essential research themes for scientists and scholars worldwide (Varol et al. 2021; Snousy et al. 2022).

Major contaminants detected in groundwater are nitrate, fluoride, toxic metals, pesticides, pharmaceuticals, hydrocarbons, and radioactive substances (Bedi et al. 2020; Li et al. 2021a). Researchers around the globe have tested groundwater samples for the presence of such contaminants and evaluated their suitability for drinking and irrigation purposes using geospatial tools, multivariate statistics, and index-based approaches (Sarma and Singh 2021). For example, Rahman et al. (2018) assessed the groundwater quality of Gopalganj district in Bangladesh and reported that most hydrochemical parameters exceeded the limits for drinking water standards. For the Guanzhong Basin of north-western China, Li et al. (2021b) analysed 25 groundwater samples and reported that residents of their study area are at risk from high fluoride levels. Ram et al. (2021) applied water quality index and GIS methods to samples from Uttar Pradesh, India, to categorize the groundwater quality as excellent, good, poor, and unsuitable. For Malda district in Eastern India, water quality assessment showed that 14% of the groundwater samples fell in poor category (Sarkar et al. 2022).

The occurrence of high levels of nitrate and fluoride in groundwater has been reported in many studies (Rezaei et al. 2017; Rufino et al. 2019; Makubalo and Diamond 2020; He et al. 2021; Liu et al. 2022). Nitrate enters the groundwater system from excessive fertilizer use, agricultural runoff, sewage and septic tank leaks, manure systems and animal wastes (Duvva et al. 2022; Dhakate et al. 2023). Consumption of groundwater with nitrate levels greater than 45 mg/L can lead to methemoglobinemia. Also known as blue baby syndrome, this condition reduces the ability of blood to transport oxygen, causing breathlessness, cardiac arrest or death, especially in infants (Ceballos et al. 2021; Golaki et al. 2022; Gugulothu et al. 2022; Panneerselvam et al. 2022). Fluoride naturally occurs in groundwater from geogenic sources and weathering of fluoride-bearing minerals (Subba Rao 2017; Mukherjee and Singh 2018). Surplus application of phosphate fertilizers further enhances the fluoride pollution of groundwater (Karunanidhi et al. 2020; Subba Rao et al. 2021). The intake of fluoride within the permissible limits for drinking water prevents tooth decay and dental cavities and helps in bone formation (Adimalla and Venkatayogi 2017; Sathe et al. 2021). But the long-term intake of excessive fluoride (≥ 1.5 mg/l) may lead to neurological effects, and dental and skeletal fluorosis (Mukherjee and Singh 2018; Ambastha and Haritash 2021; Liu et al. 2022).

In India, high levels of nitrate and fluoride have been reported in many regions — Maharashtra (Nawale et al. 2021), Telangana (Adimalla and Li 2019; Subba Rao et al. 2021; Duvva et al. 2022), Punjab (Singh et al. 2020), Haryana (Kaur et al. 2020; Rishi et al. 2020), Assam (Sathe et al. 2021), Rajasthan (Rahman et al. 2021; Jandu et al. 2021), Tamil Nadu (Karunanidhi et al. 2020; Khan et al. 2021), Jharkhand (Giri et al. 2021) and Uttar Pradesh (Maurya et al. 2020). The United States Environmental Protection Agency has developed a framework to assess the health risks from the usage of fluoride and nitrate contaminated groundwater (USEPA 1989; 1997; 2004; 2014). Many researchers around the world have adopted this model in their studies to evaluate the hazard index (HI) for different categories of people — males, females, children and infants (Singh et al. 2020; Chen et al. 2021; Reddy et al. 2022). The acceptable limit of non-carcinogenic risk is when HI ≤ 1. If the HI is more than 1, then exposure to contaminated groundwater has serious adverse effects on health (Adimalla et al. 2019). Many studies have reported that infants and children are more at risk than adults (Gao et al. 2020; Adimalla and Qian 2021).

The North-West region of Delhi has industrial, residential and agricultural areas. In the last two decades, this region has experienced major changes in land use. Much of the agricultural lands and rural built-up area have been converted to urban areas. Identifying the mechanisms of groundwater pollution arising from the rapid urbanization in this region is important. This study was carried out in the North-West region of Delhi, India, to (a) evaluate the hydrogeochemistry of groundwater and its suitability for drinking and irrigation, (b) assess the spatial extent of fluoride and nitrate contamination, and (c) estimate the corresponding non-carcinogenic health risks for men, women and children using the USEPA method. The results of this study will be helpful in understanding how increasing urbanization influences groundwater quality and affects human health.

Materials and methods

Study area

The investigated area lies in the North-West region of the National Capital Territory (NCT) of Delhi. The NCT covers a geographical area of 1483 km2 and falls in the Yamuna River sub-basin, which controls its drainage system. The NCT has adjoining smaller cities — Faridabad, Gurugram, Ghaziabad and Noida which contribute to a total of 3000 km2 of urban area (Chaudhuri and Sharma 2020). The region is characterized by hot summers and cold winters. The average rainfall is 581 mm. July, August and September are the main monsoon months that receive 81% of the total rainfall. There are planned residential and industrial areas in the North-West region of NCT with some agricultural lands near the adjoining state of Haryana (CGWB n.d). Thus, this area has both urban as well as rural populations. Land use maps from Bhuvan (2021) show that agricultural regions have decreased in the last 15 years while urban areas have increased in this region.

In Delhi, the aquifer geology is complex, varying from Quartzite to Older and Younger Alluvium (CGWB 2021; Sarma and Singh 2022). North West District is characterized by unconsolidated Quaternary alluvium deposits from the Middle to Late Pleistocene Age (CGWB n.d). Sand, silt, and clay are the major soil types in the region in varying proportions. In most of the district, water levels are 5–10 m below ground level, with deeper water levels (> 15 mbgl) observed in the northern part. The district is bordered by the Yamuna River in the northeast which controls the drainage system. The total annual groundwater recharge has been estimated as 8630.7 ham and total annual ground water draft for all uses has been estimated as 9015.2 ham as on 2011 (CGWB n.d). Groundwater exploration studies by the Central Ground Water Board, India, showed that discharge in exploratory wells and piezometers ranged from 150 to 2816 lpm and drawdown ranged from 0.72 to 17.23 m (CGWB n.d). The overall stage of ground water development of the area is 112.36%. The Central Ground Water Board has classified the sub-regions of the district as semi-critical or over-exploited.

Sample collection and analysis

Groundwater samples from 58 locations in the study area were collected from handpumps and bore wells with a depth range of 15–35 mbgl in January 2021. The coordinates of the sampling locations were recorded using a portable GPS device. Location map of the study area and sampling points were prepared by GIS software ArcMap 10.7.1 (Fig. 1). The wells were pumped for 5–10 min to remove the interference from any stagnant water. The water samples were collected in distilled water rinsed polyethylene bottles of 1 l capacity. The sample bottles were sealed, labelled and stored at 4 °C. The analytical procedures for estimating the groundwater parameters were carried out according to the standard methods given by the American Public Health Association (APHA 2017).

Fig. 1
figure 1

Groundwater sampling locations in the study area

Full size image

The physical parameters — pH, electrical conductivity (EC) and total dissolved solids (TDS) were measured on site using a portable multi-parameter meter (Orion Star A320). Prior to use, the pH meter was calibrated using buffer solutions of pH 4.0, 7.0 and 10.0 and the EC meter was calibrated using standard solutions with EC = 1413 μS/cm and 12.9 mS/cm. Total hardness (TH as CaCO3), total alkalinity (TA as CaCO3), chloride (Cl−), carbonate (CO32−), bicarbonate (HCO3−), calcium (Ca2+) and magnesium (Mg2+) ions were determined by titrimetric methods. Sulphate (SO42−), nitrate (NO3−), and phosphate (PO43−) were measured using UV–Visible Spectrophotometers (Labtronics 290 and LabIndia Analytical UV 3092). Fluoride (F−) was measured using an electrode meter. Sodium (Na+) and potassium (K+) ions were determined using a flame photometer (Systronics 128). The analytical test methods, their corresponding reagents and detection limits are presented in Table S1. The accuracy of the chemical analysis was validated by charge balance errors (CBE), and samples with ± 10% error were considered only (Domenico and Schwartz 1990; Adimalla et al. 2019; Rahman et al. 2021; Panneerselvam et al. 2022). Eliminating samples above this error, 52 samples were considered for further analysis (Fig. S1). The CBE was calculated as \(\mathrm{CBE}= \frac{\sum \mathrm{cations}- \sum \mathrm{anions}}{\sum \mathrm{cations}+ \sum \mathrm{anions}} \times 100\).

The groundwater samples were evaluated for drinking purpose by comparing the observed value against the recommended limits given by the Bureau of Indian Standards (BIS) and World Health Organization (WHO). The spatial distribution maps of the groundwater parameters were created using the inverse-distance weighted (IDW) interpolation technique in ArcMap 10.7.1 software. The hydrogeochemical characteristics of the groundwater samples were studied by plotting Piper trilinear diagram (Piper 1944) in AquaChem software. Chloro-alkaline indices CAI-1 and CAI-2 (Schoeller 1965) were calculated to understand the mechanisms of ion-exchange and rock-water interactions (Subba Rao 2017). CAI-1 and CAI-2 were calculated as per the following equations (all ions in meq/l).

$$\mathrm{CAI}-1= \frac{{\mathrm{Cl}}^{-}- {\mathrm{Na}}^{+}+ {\mathrm{K}}^{+}}{{\mathrm{Cl}}^{-}}$$
(1)
$$\mathrm{CAI}-2= \frac{{\mathrm{Cl}}^{-}- {\mathrm{Na}}^{+}+ {\mathrm{K}}^{+}}{{\mathrm{SO}}_{4}^{2-}+ {\mathrm{HCO}}_{3}^{-}+ {\mathrm{CO}}_{3}^{2-}+ {\mathrm{NO}}_{3}^{-}}$$
(2)

Water quality index values (WQI) were calculated to determine the suitability of the groundwater samples for drinking. The WQI was based on the values of pH, TDS, TH, TA, Cl−, F−, SO2−4, NO−3, Ca2+, and Mg2+. The following equation was used to calculate the WQI.

$$\mathrm{WQI}= \frac{\sum {W}_{i}{Q}_{i}}{\sum {W}_{i}}$$
(3)

where Qi is the quality rating for each parameter given by Qi = 100 *[(Vi − Vo)/(Si − Vo)], Vi is the observed value of ith parameter, Vo is the ideal value of parameter in pure water (0 for all parameters; 7.0 for pH), Si is the recommended standard value of ith parameter and Wi is the unit weight of each parameter (Wi = K/Si). For calculation of Wi, K is proportionality constant given by K = 1/∑(1/Si)).

In order to determine the suitability of the samples for irrigation, the parameters such as soluble sodium percentage (SSP), residual sodium carbonate (RSC), sodium absorption ratio (SAR), permeability index (PI), Kelley’s ratio (KR) and magnesium hazard (MH) were calculated as per their respective formula (Aravinthasamy et al. 2021) (Table 1). The suitability of samples for irrigation was also determined using the US Salinity Laboratory classification (USSL 1954). IBM SPSS Statistics software version 26 was used for multivariate statistical techniques — principal component analysis (PCA) and Pearson correlation matrix.

Table 1 Equations to calculate suitability of water for irrigation
Full size table

Health risk assessment

The USEPA considers high nitrate and fluoride in drinking water as non-carcinogenic risks to human health (USEPA 1989). The exposure routes to such contaminated water may be either through oral ingestion (drinking) and/or dermal contact (bathing). Considering these exposure pathways, the chronic daily intake (CDI in mg/kg/day) through oral ingestion, and dermally absorbed dose (DAD in mg/kg/day) through bathing were calculated. The non-carcinogenic risk through drinking water exposure route in terms of CDI was calculated by Eq. (4).

$$\mathrm{CDI}= \frac{\mathrm{CPW}*\mathrm{IR}*\mathrm{ED}*\mathrm{EF}}{\mathrm{ABW}*\mathrm{AET}}$$
(4)

where CDI 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 (L/day: 2.5 L/day for adults and 0.78 L/day for children), ED is the exposure duration (years: 64, 67 and 12 for men, women and children respectively), EF is the exposure frequency (days/years: 365 days for children and adults), ABW is the average body weight (Kg: 65, 55 and 15 for males, women and children respectively), and AET is the average time (days: 23,360, 24,455 and 4380 for males, women and children, respectively) (USEPA 2014). The health risk due to dermal exposure was calculated by using the following equation.

$$\mathrm{DAD}= \frac{\mathrm{CPW}*\mathrm{TC}*\mathrm{Ki}*\mathrm{EV}*\mathrm{SSA}*\mathrm{CF}*\mathrm{ED}*\mathrm{EF}}{\mathrm{ABW}*\mathrm{AET}}$$
(5)

where DAD is the dermally absorbed dose (mg/kg/day), TC indicates the contact duration (h/d: 0.4 h per day for adults and children), Ki is the dermal adsorption parameters (cm/h: 0.001 cm/h), EV is the bathing frequency (times/day: considered as 1 time in a day), SSA is the skin surface area available for contact (cm2: 16,600 and 12,000 cm2 for adults and children, respectively), CF is the unit conversion factors (0.001), ED is the exposure duration (years: 64, 67 and 12 for males, women and children, respectively), EF is the exposure frequency (days/years: 365 days for children and adults), ABW is the average body weight (Kg: 65, 55 and 15 for men, women and children respectively), and AET is the average time (days: 23,360, 24,455 and 4380 for males, women and children, respectively) (USEPA 1997; Adimalla and Qian 2021).

Oral and dermal hazard quotient for the nitrate and fluoride were computed by the following equations:

$${\mathrm{HQ}}_{\mathrm{oral}}= \frac{\mathrm{CDI}}{\mathrm{RfD}}$$
(6)
$${\mathrm{HQ}}_{\mathrm{dermal}}= \frac{\mathrm{DAD}}{\mathrm{RfD}}$$
(7)

where HQoral and HQdermal are the non-carcinogenic oral and dermal hazard quotient, respectively. CDI and DAD are chronic daily intake (mg/kg/day) and the dermally absorbed dose (mg/kg/day), respectively, and RfD represents the reference dose of a specific contaminant (USEPA 1989). The oral reference dose of nitrate is 1.6 mg/kg/day and that of fluoride is 0.06 mg/kg/day, obtained from the database of IRIS (Integrated Risk Information System) (USEPA 1989). The HQ values can be used to evaluate the health risk alone where adverse health effects are seen if HQ > 1. However, the hazard index (HI) gives the total hazard presented by exposure to multiple contaminants through multiple pathways. In this study, it is calculated as the sum of the hazard quotients calculated for oral and dermal risk exposure (HQoral and HQdermal) to nitrate and fluoride given by:

$${\mathrm{HI}}_{i}= {\mathrm{HQ}}_{\mathrm{oral}}+ {\mathrm{HQ}}_{\mathrm{dermal}}$$
(8)
$${\mathrm{HI}}_{\mathrm{total}}={\textstyle\sum_{i=1}^n}{\mathrm{HI}}_i$$
(9)

Based on the HItotal values, no significant non-carcinogenic risk occurs if HItotal ≤ 1. However, if HItotal > 1, then there is significant non-carcinogenic risk (USEPA 1991; 2004).

Results and discussion

Groundwater chemistry

The statistics of the physicochemical parameters of the groundwater samples — minimum, maximum, mean and standard deviation are summarized in Table 2. The pH of the samples ranges between 7.5 and 8.4 with a mean value of 8.0, indicating slightly alkaline conditions. The pH values of the water samples fall within the acceptable limits set by BIS (2012). The EC values vary significantly, with a range of 254–15,440 μS/cm and a mean value of 3699 μS/cm. The elevated values of EC indicate high ionic strength, mineral content and dissolved solids. TDS values range from 128 to 7770 mg/l, with a mean value of 1854 mg/l. Only 17% of the samples are within the BIS acceptable limit of 500 mg/l. According to the classification of TDS given by Freeze and Cherry (1979), TDS < 1000 mg/l indicates fresh water while TDS between 1000 to 10,000 mg/l indicates brackish water. Based on this classification, 44% and 56% of the samples fall in the fresh and brackish water categories respectively. Davis and DeWiest (1966) classified groundwater as desirable for drinking if TDS < 500 mg/l, permissible for drinking if TDS is between 500 to 1000 mg/l, useful for irrigation if TDS is between 1000 to 3000 mg/l and unsuitable for drinking and irrigation if TDS > 3000 mg/l. Based on this classification, 17% of the samples were desirable for drinking, 27% were permissible for drinking, 37% were suitable for irrigation and 19% were unfit for both drinking and irrigation. The spatial distribution maps of pH, EC and TDS are given in Fig. S2.

Table 2 Statistics of groundwater quality parameters (n = 52) and comparison with drinking water standards (BIS and WHO)
Full size table

The concentrations of the cations Ca2+, Mg2+, Na+ and K+ range from 20 to 872 mg/l, 20–1580, 4–1006 and 1.4–71.4 mg/l respectively with mean values of 170, 356, 296 and 12.2 mg/l respectively. The concentrations of dissolved anions such as HCO3−, Cl−, PO43− and SO42− vary from 268 to 1696, 20 to 4700, 0.00 to 0.61 and 35 to 2840 mg/l respectively with the mean concentrations of 949, 704, 0.13 and 443 mg/l, respectively. The TH values range from 180 to 7108 mg/l as CaCO3 with mean of 1883 mg/l as CaCO3. According to the classification for total hardness by Sawyer and McCarty (1967), water is termed “very hard” if TH > 300 mg/l as CaCO3 and “hard” if TH is between 150 and 300 mg/l as CaCO3. Based on this classification, 92% of the samples have “very hard” water, and 8% of the samples fall in “hard water” categories (Table S2). This is evident from the high levels of bicarbonate ions present in the samples. The standard deviation of SO42− is higher than its mean which indicates that sulphate levels in the water samples fluctuate randomly. The dominant major cations in the groundwater samples are in the order of Mg2+  > Na+  > Ca2+  > K+, while the dominant anions are HCO3− > Cl−  > SO42−  > NO3−  > CO32−  > F−. The elevated concentrations of HCO3− along with Mg2+ and Ca2+ ions in some samples indicate that the study area might be affected by dissolution of carbonate minerals (like calcite and dolomite) and/or silicate minerals by carbonic acid (CGWB 2016; Snousy et al. 2022). Excess Na+ over Cl− indicates rock weathering (or cation exchange) while the vice versa indicates reverse ion exchange (Subba Rao et al. 2017; Gugulothu et al. 2022). For the studied samples, about 85% had excess Cl− over Na+ indicating that reverse ion exchange was the primary source of these ions. High sodium intake (> 200 mg/l) leads to problems of hypertension, kidney and nerves (Rishi et al. 2020). Na+, Mg2+ and K+ arise from anthropogenic sources such as wastewater, return flows from irrigation and potassium fertilizers (Subba Rao et al. 2021). The high Cl− concentration may be due to the release of untreated sewage and industrial effluents in the region. Chloride imparts a salty taste to the water and may have laxative effects. The industrial activities in the study region may also be the reason for the high SO42− levels found in the water samples. High sulphate concentrations along with high Mg2+ are known to cause gastro-intestinal problems (CGWB 2016).

Nitrate levels in the samples range from 0.24 to 380.19 mg/L, with a mean of 65.29 mg/L (Fig. S3(a)). According to WHO (2011), there is no health risk for humans if nitrate levels are below 45 mg/l. However, nitrate between 45 and 100 mg/L causes health effects on children and adults and > 100 mg/L have very high health risk. As per this classification of nitrate, 60% of groundwater samples fall under the “no health risk” category, while 19% and 21% of groundwater samples fall under the “high health risk” and “very high health risk” categories. The spatial distribution map of nitrate is presented in Fig. 2a. Nitrate is predominant in shallow aquifers and easily reaches the groundwater from the surface owing to its high solubility in water (Adimalla and Qian 2021). Nitrate is thus largely anthropogenic in nature and majorly sourced from agrochemicals, open land dumping, domestic, animal and manufacturing wastes (Duvva et al. 2022; Panneerselvam et al. 2022). The high levels of nitrate in the study region may be due to fertilizers such as diammonium phosphate and urea which are commonly utilized in North India. Because of the widespread use of such fertilizers, nitrate can drain away from soils and percolate into the groundwater. Rahman et al. (2021) lists landfill leachate as one of the contributors to nitrate contamination of groundwater. The Bhalaswa landfill in the study region has been operational since 1993 (Sidhu et al. 2015), and its leachate percolating into the groundwater may also contribute to high nitrate levels.

Fig. 2
figure 2

Spatial distribution maps of a nitrate and b fluoride in the study area

Full size image

The fluoride concentration ranged from 0.05 to 7.90 mg/l with a mean of 2.23 mg/l (Fig. S3b). Fluoride concentration less than 0.6 mg/l may cause dental caries while greater than 1.5 mg/l may cause severe problems of fluorosis. The concentration of fluoride was below 0.6 mg/l in 21% of the samples and exceeded the permissible limit (1.5 mg/l) in 58% of the groundwater samples. The spatial distribution map of fluoride is presented in Fig. 2b. The high fluoride distribution is identified in northern, southern, central and western parts of the region. Fluoride-rich minerals and usage of phosphate fertilizers are the chief sources of elevated fluoride levels. The anionic exchange controlling the fluoride content in the study region is enhanced by the alkaline nature of water (Duvva et al. 2022; Xiao et al. 2022b). Several studies have reported high concentrations of \({\mathrm{NO}}_{3}^{-}\) and \({\mathrm{F}}^{-}\) in north-west Delhi and the neighbouring state of Haryana (Singh et al. 2017; Kaur et al. 2020; Ambastha and Haritash 2021; Masood et al. 2022).

Hydrochemical facies

The Piper trilinear diagram (Piper 1944) suggests the dominance of groundwater chemistry. For the collected groundwater samples, Piper diagram was plotted using AquaChem software, version 10 (Fig. 3). In the cation plot, maximum samples fall in magnesium type (67%) while in the anion plot, maximum samples fall in the bicarbonate type (56%). Mg2+ is the dominant ion as a result of weathering of silicate rocks (Adimalla 2019). In the diamond shape, maximum samples (48%) fall in CaHCO3 type followed by mixed CaMgCl type (30%). CaHCO3 type water indicates that River Yamuna and irrigation canals are primarily responsible for the aquifer recharge in the absence of adequate rainfall. The Piper classification indicates that major processes regulating groundwater chemistry in the study region are ion exchanges, rock-water interactions, mineral weathering and anthropogenic influences (Snousy et al. 2022; Panneerselvam et al. 2022).

Fig. 3
figure 3

Piper trilinear classification of groundwater samples

Full size image

The chloro-alkaline indices CAI-1 and CAI-2 help in understanding the mechanism of ion exchange. If the index is positive, it implies an exchange of sodium and potassium ions from the water with calcium and magnesium ions of the rocks (base–exchange reaction). If it is negative, it indicates vice versa, i.e., calcium and magnesium of water exchanging with sodium and potassium from rocks (cation–anion exchange reaction) (Subba Rao 2017). For the studied groundwater samples, most of the samples demonstrated positive CAI values (Fig. 4a) indicating the cation–anion exchange reaction where Na+ and K+ from the water continuously exchanges with Ca2+ and Mg2+ from aquifer materials due to rock-water interactions (Rashid et al. 2022). Moreover, plot of (Na+ + K+)–Cl− against (Ca2+ + Mg2+)–(HCO3− + SO42−) can be expressed as y =  − 1.1249x + 5.5638 with a correlation coefficient of 0.9226 (Fig. 4b). The negative slope of − 1.1249 confirms that the relationship between Na+, K+, Ca2+ and Mg2+ is influenced by reverse ion exchange process (Kumar and James 2016):

Fig. 4
figure 4

Plot of a CAI-1 against CAI-2 and b (Na+  + K+)–Cl− against (Ca2+  + Mg2+)–(HCO3−  + SO42.−)

Full size image

2Na+  + Ca2+(Mg)–Clay → Na+–Clay + Ca2+  + Mg2+

Multivariate statistical analysis

Principal component analysis

The application of PCA was first done by checking the Bartlett’s test of sphericity and Kaiser–Meyer–Olkin (KMO) sampling adequacy. PCA requires KMO sampling adequacy to be > 0.50 for the dataset (Snousy et al. 2022). The Bartlett test of sphericity was in accordance with p value < 0.0001, and KMO sampling adequacy was 0.671 for the groundwater samples. These values confirm that the dataset is suitable for PCA. The PCA was performed in SPSS software using varimax rotation method with Kaiser normalization. Factors loading values are classified as weak (0.30–0.50), moderate (0.50–0.75), and strong (> 0.75) (Wu et al. 2020). Table 3 reveals that five significant components are calculated (with eigenvalues > 1), which represent 83% of the total variance. The eigenvalues represent how much variance there is in the dataset, and the variance represents the amount of variation in the dataset that can be attributed to each principal component. Component 1 explains 46.6% of the total variance and has positive loading of EC, TDS, TH, Cl−, Mg2+, Ca2+, Na+ and SO42− implying that EC and TDS are primary governed by the major cations and anions through mineral dissolution, rock-water interaction, ion-exchange and anthropogenic factors (Elemile et al. 2021). Component 2 explains 13.5% of the total variance and has positive loading of carbonate, bicarbonate and total alkalinity. This implies that TA is driven by dissolution of carbonate and bicarbonate minerals in the study area. Component 3 explains 8.6% of the total variance and has positive loading of K+ and PO43− indicating the use of potash and phosphate fertilizers. Component 4 explains 7.5% of the total variance and has positive loading of F−, with moderate loading of CO32− and pH and negative loading of Ca. This implies that the concentration of fluoride is due to weathering of fluorite minerals (CaF2) enhanced by carbonate weathering and alkaline conditions (Barzegar et al. 2017; Xiao et al. 2022b). Finally, component 5 explains 6.8% of the total variance and has strong positive loading of nitrate indicating that origins of nitrate in the water samples may be purely anthropogenic — fertilizer use, sewage and animal wastes. Figure S4 represents the PCA plot of the components in rotated space.

Table 3 The main five principal components extracted form groundwater samples
Full size table

Pearson correlation matrix analysis

The relationships between the physicochemical parameters were analysed by PCMA. The correlation matrix is presented in Table 4. pH has negative correlation with EC, TDS, TH, Cl−, SO42−, Ca2+, Mg2+, and Na+, consistent with studies by Swain et al. (2022) and Panneerselvam et al. (2021). EC shows identical liner correlation with TDS (r = 1.000) with a 99% confidence level and significant positive correlation with Na+ (r = 0.825), Ca2+ (r = 0.835), Mg2+ (r = 0.923), Cl− (r = 0.935) and SO42− (r = 0.769). This is consistent with the results of PCA. The TDS has a strong positive correlation with Na+ (r = 0.825) and Cl− (r = 0.939) indicating that rock weathering and sewage seepage have caused the salinity to increase. Ca2+ shows significant positive correlation with Mg2+ (r = 0.756), Cl− (r = 0.828) and SO42− (r = 0.693). Mg2+ also shows significant positive correlation with Cl− (r = 0.867) and SO42− (r = 0.823). These correlations indicate that major ion chemistry in the groundwater samples is influenced by the dissolution of aquifer materials, rock-water interactions and domestic wastewater infiltration (Snousy et al. 2022). NO3− shows negative correlation with pH which is also reported in Stylianoudaki et al. (2022) and Glass and Silverstein (1998).

Table 4 Pearson correlation coefficient matrix among physicochemical parameters in groundwater samples
Full size table

Water quality index for drinking

Based on the classification given in the study by Masood et al. (2022), the WQI obtained for the groundwater samples were evaluated (Table S3). The WQI < 50 is beneficial for health (“excellent” category) which is calculated for 12% of the samples, located in some isolated pockets in the study region. WQI between 50 and 100 is acceptable for drinking use (“good” category) which is calculated for 8% of the samples. Forty percent of the samples were impure with WQI 100–200 (“poor” category), and 25% of the samples needed treatment prior to use (“very poor” category) with WQI 200–300. The WQI > 300 were found in 15% of the samples which were completely unsuitable for drinking. The spatial distribution map of WQI is presented in Fig. 5. Poor, very poor and unsuitable water quality can be observed in most parts of the study region — central, northern, western, eastern and southern. Only a small area in the north eastern region has good water quality.

Fig. 5
figure 5

Spatial distribution map of water quality index

Full size image

Irrigation water quality

Agricultural areas in the study region are situated in the extreme northwest and western regions, where groundwater is the primary source of irrigation. Evaluating the suitability of groundwater for irrigation purposes was done by comparing the irrigation quality parameters with the recommended values (Table 5). The quality of water for irrigation is dependent on its mineral constituents which affect both plants and soil (Wilcox 1955; Alam et al. 2012). The EC is an indicator of the salinity of the groundwater which can influence crop growth. High levels of salinity can negatively affect crop development (Subba Rao 2017; Gugulothu et al. 2022). The salinity is low if EC < 250 μS/cm and very high if EC > 2250 μS/cm. For the study region, 54% of the samples have very high salinity. The sodium absorption ratio (SAR) values indicate the cation–exchange reaction in the soil. High values of SAR specify a situation where the absorbed calcium and magnesium have been replaced by sodium, posing a risk to soil structure (Saha et al. 2019). All the studied samples present a low sodic hazard in terms of the SAR (SAR values < 10). The USSL classification (USSL 1954) plots EC values against the SAR values (Fig. 6). The USSL diagram shows that majority of the samples fall in S1C2, S1C3, S1C4 and S2C4 classes, indicating low to medium sodium hazard and medium to very high salinity hazard in the study region.

Table 5 Irrigation quality of groundwater samples
Full size table
Fig. 6
figure 6

Groundwater suitability for irrigation according to USSL classification

Full size image

The residual sodium carbonate (RSC) is an indicator of the hazardous effects of carbonate and bicarbonate ions for irrigation purposes (Saha et al. 2019; Rishi et al. 2020). RSC values < 1.25 meq/l are fit for irrigation while RSC > 2.5 meq/l are unsuitable. Based on this classification, 83% of the samples are suitable for irrigation while only 15% are unfit. The soluble sodium percentage (SSP) indicates the sodium content in terms of %Na. The sodium-laden water reacts with soil and accumulates in the air spaces (or voids) in the soil. This leads to clogging of the soil particles and reduction in soil permeability which can affect the growth of plants (Todd 1980). The permissible limit of SSP is 60% for irrigation water. Based on this classification, 98% of the samples were permissible for irrigation. Kelley’s ratio (KR) measures sodium against calcium and magnesium (Kelley 1940). Water with KR > 1 indicates high sodium content and is unsuitable for irrigation. For the present study, KR values ranged from 0.04 to 1.77, with only 4 samples above KR value 1. Ninety-two percent of the samples were within the acceptable limit of KR < 1. Magnesium is important for soil productivity and maintaining soil structure. High levels of magnesium may result due to exchanges with Na+. This in turn renders the soil alkaline which causes loss of phosphorus (Paliwal 1972; Saha et al. 2019). The magnesium hazard index classifies water for irrigation as suitable if it is < 50 and unsuitable if it is > 50. In the present study, groundwater samples had high levels of Mg2+. Thus 98% of the samples were unsuitable for irrigation (Mg hazard > 50). Permeability of soil is affected by the continuous and long-term use of irrigation water and is regulated by soil Na+, Mg2+, Ca2+ and HCO3− (Snousy et al. 2022). The permeability index given by Doneen (1964) classifies water into three classes. Based on this classification, 17% of the samples were unsuitable for irrigation (class III), and 77% and 6% of the samples fall in class II and class I categories, respectively, which are suitable for irrigation.

Health risk assessment for nitrate and fluoride contamination

The groundwater in the study region is used by the local people for irrigation, industrial, and domestic purposes. Many residents in the area use the groundwater for drinking and showering. Since the samples collected had high nitrate and fluoride levels, the estimated concentrations of these pollutants were used for calculating the non-carcinogenic hazard quotient through oral and dermal exposure routes and the total hazard index according to Eqs. (4)–(9). The results obtained for hazard quotients for males, females and children are presented in Table 6.

Table 6 Hazard quotient for (a) nitrate and (b) fluoride
Full size table

The risk through dermal contact for nitrate was very low for all 3 categories of people, and the values were less than 1 for all samples. This result was also observed in studies by Zhang et al. (2018) and Gao et al. (2020). The total hazard quotient for nitrate ranged from 0.006 to 9.163 (mean = 1.574) for males, 0.007 to 10.829 (mean = 1.860) for females and 0.010 to 15.917 (mean = 2.734) for children. The HQnitrate was greater than 1 for 44%, 46% and 52% of the samples for males, females and children respectively. Similar to nitrate, the risk through dermal exposure of fluoride was very low, demonstrating that the main health risk is through direct consumption. The total HQfluoride ranged from 0.031 to 5.078 (mean = 1.433) for males, 0.037 to 6.001 (mean = 1.693) for females and 0.055 to 8.820 (mean = 2.489) for children. HQfluoride > 1 was observed for 58%, 58% and 69% of the samples for males, females, and children respectively. The HQ for nitrate was found to be greater than the HQfluoride values across all demographics, indicating that nitrate poses a higher health risk to the residents of the study region. However, the percentage of samples with HQ > 1 was more for fluoride indicating that the spatial extent of risk was more for fluoride. The spatial distribution of zones with high health risks is presented in Fig. 7a and b.

Fig. 7
figure 7

Spatial distribution of a HInitrate, b HIfluoride, and c HItotal in the study area

Full size image

The total hazard index (HItotal) is a summary of the total risks posed by high levels of nitrate and fluoride (Table 7). For the studied groundwater samples, the HItotal was found to be greater than 1 for 75%, 79% and 85% for males, females and children respectively. This indicates that the majority of the population in the study area are at some health risk, primarily from consumption of contaminated groundwater. The spatial distribution map presenting the risk zones are given in Fig. 7c. The values of HItotal also indicate that the risk is of the order of children > females > males. Owing to their weak resilience and higher consumption per unit of body weight, children are at a greater risk from drinking contaminated water in the study region than adults (Chen et al. 2016; Adimalla 2020; Guo et al. 2022; Xiao et al. 2022a).

Table 7 Total hazard index for nitrate and fluoride through oral and dermal pathways
Full size table

Conclusions

This study analysed the groundwater quality and associated health risks in North-West Delhi, India, which is a rapidly urbanizing region. The hydrogeochemical mechanisms influencing the major ion chemistry were explored, and the characteristic pollutants were identified. The dominant cations in the groundwater samples were Mg2+  > Na+  > Ca2+  > K+, while the dominant anions were HCO3−  > Cl−  > SO42−  > NO3−  > CO32−  > F−. The groundwater is slightly alkaline and TDS, TA, TH, Cl−, SO42−, Ca2+ and Mg2+ exceeded the prescribed drinking water limits in 83%, 100%, 98%, 58%, 58%, 77% and 96% of the analysed samples, respectively. The groundwater in the study region is mostly unsuitable for human consumption.

Piper trilinear diagram showed that maximum samples fell in CaHCO3 type and CaMgCl type categories. The positive value obtained from chloro-alkaline indices showed that Na+ and K+ from water exchanged with Ca2+ and Mg2+ from the aquifer. Multivariate analysis using principal component analysis revealed five significant components which account for 83% of the total variance. Pearson correlation matrix indicated that major ion chemistry is influenced by several factors such as mineral dissolution, rock-water interactions and anthropogenic interferences. The water quality index for drinking was calculated for the collected groundwater samples based on the pH, TDS, TH, TA, Cl−, F−, SO2−4, NO−3, Ca2+, and Mg2+ values, and 15% of the samples were found to be unfit for drinking (WQI > 300). The water samples were analysed for irrigation quality, and results showed that all samples had low sodic hazard. However, 54% of the samples had high salinity, which adversely affects crop production.

Nitrate and fluoride were above the recommended limits of 45 mg/l and 1.5 mg/l in 40% and 58% of the samples, respectively. Wastewater infiltration and fertilizer use are the primary sources of NO3− and F−. High fluoride concentrations in the study region may also be due to geogenic sources. The hazard quotients for nitrate and fluoride suggested that non-carcinogenic health risk is higher for nitrate contamination. However, the spatial extent of HQ > 1 was more for fluoride, implying that more people are affected by fluoride pollution in the study region. Further, it was observed that the total hazard index was in the order of children > females > males. Due to differences in body weight, children are at a greater health risk than adults. Therefore, groundwater in the study region needs to be continuously monitored and should not be used for direct consumption to avoid adverse health effects. This study is helpful in understanding the chemistry of major contaminants in aquifers of regions that are transitioning from rural to urban areas.