Introduction

Groundwater resources, as an indispensable component of the Earth’s ecosystem, play a crucial role in maintaining biodiversity, supporting agricultural production, and ensuring the safety of human drinking water(Giordano 2009; Huang et al. 2023; Nadeem et al. 2023). However, this valuable natural resource is facing dual challenges from both natural and anthropogenic factors (Gorelick et al. 1983). Natural factors include the diversity of hydrogeological conditions, the migration characteristics of groundwater through different lithological media, the complexity of water–rock interactions, the impact of tidal forces on coastal groundwater systems, and the potential alterations in the chemical composition of groundwater due to seawater intrusion in coastal areas(Chen et al. 2010; Hsieh et al. 2015; Xu et al. 2023). The rapid growth of the population, accelerated urbanization, deepening industrialization processes, and the continuous expansion of agricultural activities have not only exacerbated the over-extraction of groundwater but have also led to widespread pollution issues (Gorelick 1982).

Groundwater pollution is primarily caused by industrial discharges, the use of agricultural fertilizers and pesticides, urban sewage infiltration, and leaks from landfill sites, leading to the accumulation of contaminants including heavy metals. These pollutants pose long-term threats to both the environment and human health due to their persistence and bioaccumulation potential. Such elements can result in neurological damage, kidney diseases, and even cancer (Raza Altaf et al. 2021; Gao et al. 2017), as well as disrupt ecosystems, leading to a decline in biodiversity (Teng and Altaf 2022). In China, this issue is particularly pronounced, especially in regions with dense industrial activity (Huang et al. 2018). Mining operations not only lead to a continuous increase in the concentration of heavy metals in groundwater but may also result in the depletion of groundwater resources and deterioration of water quality (Burri et al. 2019). Taking Zibo City in Shandong Province as an example, the region relies on groundwater as a primary source for drinking water and agricultural irrigation. However, mining extraction and closure activities have had profound impacts on the groundwater system, including alterations in hydrodynamics and hydrochemical properties. These changes have subsequently had adverse effects on the health of local residents and on agricultural production. Although prior research (Guo et al. 2004; Xu and Xu 2003; Zhang et al. 2016) has employed numerical simulation and sensitivity analysis to uncover the origins and migration mechanisms of groundwater pollution in mining areas, providing a scientific foundation for management, the accurate identification of pollution sources in the region and the assessment of potential risks to human health still necessitate further study.

Prompt and precise discernment of pollution origins is paramount for the mitigation and management of groundwater contamination. Historically, multivariate statistical models have been a preferred tool among scholars to decipher intricate hydrochemical datasets (Abd El-Wahed et al. 2022; Trabelsi and Zouari 2019). Yet, groundwater quality is shaped by a myriad of elements, ranging from lithological processes to both human-induced and agro-industrial activities (Singh et al. 2017). Such interrelated factors complicate the task of pinpointing pollution origins (Yu et al. 2022a, b, c; Zhao et al. 2022). Numerous scholars have employed a variety of techniques, including factor analysis (FA), principal component analysis (PCA), absolute principal component scores-multiple linear regression (APCS-MLR), and cluster analysis (CA), to analyze hydrochemical data of both groundwater and surface water, identifying both natural phenomena and anthropogenic impacts (Blake et al. 2016; Salamon et al. 2006; Sunkari et al. 2021). While APCS-MLR amalgamates the advantages of PCA and MLR, its reliance on extensive sample volumes escalates research costs (Yu et al. 2022a). Cluster analysis is sensitive to initial data values, with the determination of an ideal cluster count remaining problematic (Zhang et al. 2012). In recent years, the positive matrix factorization (PMF) model, which considers the uncertainty of variable concentrations, is able to provide optimized source variable resolution, and has gradually been introduced into the field of groundwater research (Haghnazar et al. 2022; Niu et al. 2022). Research employing both PMF and APCS-MLR has gauged their effectiveness in elucidating the natural and man-made determinants shaping groundwater's physicochemical attributes (Li et al. 2021). Such findings highlight PMF’s superior capacity for a more unbiased source attribution. Therefore, our study harnesses the PMF model to discern and scrutinize pollution antecedents.

The human health risk model, formulated by the US Environmental Protection Agency, has traditionally been pivotal in gauging environmental quality by quantifying exposure rates across diverse populations (Lu et al. 2018). However, these conventional risk evaluations often hinge on static parameters, potentially skewing exposure implications for distinct demographic segments. Discrepancies in variables like adult body weight and water intake can introduce uncertainties, thereby clouding the resultant assessments (Gao et al. 2022). At present, few studies have integrated the uncertainties of health risks into a comprehensive assessment of health hazards. To counteract these shortcomings, we drew upon the technique of Monte Carlo simulation, which capitalizes on the range and distribution of pre-defined parameters to offer a more nuanced assessment of human health risks posed by pollutants (Jia et al. 2023). Monte Carlo simulation generates multiple random iterations, ascertaining the likelihood of risk occurrence and providing a more accurate estimate of the extent and distribution of human health risks(Wang and Li 2022). Though Monte Carlo’s prowess in analyzing health threats from soil contaminants is well-documented, its integration in groundwater quality research is nascent. Thus, we melded the positive matrix factorization model with the Monte Carlo simulation, aiming to pinpoint precise pollutant concentration benchmarks. This avant-garde approach sidesteps traditional risk evaluation constraints, presenting a more encompassing insight into environmental pollutants' health implications.

This study focuses on the Fengshui mining area in Zibo City, Shandong Province, aiming to thoroughly explore the impacts of mining, industrial, and agricultural activities on the chemical composition of groundwater. The main objectives of the study are (1) to systematically analyze the current status of groundwater quality in the Fengshui mining area; (2) to accurately identify and quantify the primary sources of groundwater pollution and their impacts on overall water quality; (3) to assess the potential health risks of specific pollutants to different population groups. Utilizing innovative methods, this research is committed to the meticulous evaluation of groundwater quality, tracking the sources of pollution, and analyzing their impacts on public health. With an emphasis on a pioneering approach, this study endeavors to meticulously assess groundwater quality, trace pollution origins, and discern the subsequent health repercussions. Through this endeavor, we seek to furnish a scientifically robust paradigm for groundwater pollution prevention and the assurance of drinking water safety.

Materials and methods

Study area

The study area is centrally located within the Zhangdian District of Zibo City, adjacent to Fengshui Town. It extends southward to Luocun Town, westward to Dadan Village, and eastward to Bingxu Village, as illustrated in Fig. 1. The study area is situated within the northern temperate zone and characterized by a continental monsoon climate. The study area is located in the north temperate zone and is characterized by a typical continental monsoon climate. The annual average precipitation in this region is 621.2 mm, with the majority of rainfall occurring between July and September. This area represents a well-developed industrial region, hosting a multitude of enterprises. These enterprises span across various sectors, notably including chemical engineering, construction materials, and coal industries.

Fig. 1
figure 1

Location of study area and sampling point distribution

Full size image

Geology and hydrogeology

The study area is positioned on the eastern flank of the Zibo syncline basin, primarily characterized by its geological structures dominated by faults and folds, including the Mansi River Fault and the Hongshan Syncline. It is delineated by the Chaomizhuang Graben to the east and north, bounded by the Manlihe Fault to the south, constrained by the Nanding Fault to the west, and demarcated to the southeast by the groundwater divide between the Zibo River and the Xiaofu River karst water system (Zhu et al. 2000). This region encompasses a wedge-shaped hydrogeological unit known as Fengshui-Yuedian. The terrain of the study area features higher elevations in the southeast, gradually transitioning to lower hills and gently sloping plains toward the northwest (Ying et al. 2004). The area’s groundwater resources are primarily constituted by sedimentary rock layers from the Ordovician to the Quaternary period, categorized into karst water, clastic rock fissure water, and loose rock pore water (Yu et al. 2022a, b, c). Groundwater levels are higher in the southeast and gradually decrease towards the northwest, with the main recharge mechanism being monsoonal rainfall, and discharge processes including artificial extraction, evaporation, and lateral outflow. Groundwater is principally utilized for agricultural irrigation and drinking purposes, with irrigation relying predominantly on tube wells and canal systems. A significant feature of the study area is the presence of numerous abandoned coal mines. The water accumulated in these deserted mine shafts often contains a variety of complex pollutants and has a water table that exceeds the aquifer of the Ouhui Formation, posing a potential risk of groundwater and soil contamination.

Sampling and testing

During the high-water period in October 2020, a total of 47 samples were collected. The geographical coordinates of each sampling site were recorded using a Global Positioning System (GPS) to analyze the groundwater quality in the study area at that time. The sampling process was conducted within a relatively compact time frame. Groundwater samples were preserved for at least seven days. The entire process of water sample collection adhered to the “Technical Specifications for Groundwater Environmental Monitoring” (HJ/T164-2004). Samples were collected and stored in 1 mL polyethylene bottles. Each polyethylene bottle was rinsed with pure water 2–3 times before sampling and then with the original water sample 2–3 times. After collection, the water samples were immediately sealed, labeled, and accompanied by submission forms, then promptly transported to the laboratory.

In the laboratory, the pH values were precisely measured using a PHS-3C pH meter, while Na + concentrations were analyzed through flame photometry. Sulfate (SO42−), chloride (Cl−), and fluoride (F−) concentrations were measured using inductively coupled plasma-optical emission spectrometry (ICP-OES), and nitrate (NO3−) levels were quantified using a visible spectrophotometer. The analysis of Mn, Ba, Al, and Fe was conducted through inductively coupled plasma-optical emission spectrometry (ICP). Total dissolved solid (TDS) measurements were performed with a portable multi-probe instrument under controlled laboratory temperatures ranging from 21 to 25 °C. To ensure the accuracy and reliability of the groundwater quality assessment, this study implemented post-analytical standards including instrument calibration, precision, linearity, detection limits, and the use of blank reagents for validating the analytical procedures. These quality control measures encompassed rigorous calibration of all instruments and the inclusion of quality control samples during the analysis to monitor analytical stability. Replication of samples and the control of the relative error within 5% for the cation–anion balance further validated the consistency and accuracy of the data.

Water quality indexing

The water quality index (WQI) is primarily utilized to evaluate individual water parameters, while also gauging the collective impact these parameters have on the overall water quality (Patel et al. 2023). The computational procedure is delineated below:

$$W_{i} = \frac{{w_{i} }}{{\sum\limits_{i = 1}^{n} {w_{i} } }}$$
(1)
$$Q_{i} = \frac{{C_{i} }}{{S_{i} }} \times 100$$
(2)
$$SI_{i} = W_{i} \times Q_{i}$$
(3)
$$WQI = \sum\limits_{i = 1}^{n} {SI_{i} }$$
(4)

where Wi is the weight of each chemical parameter and wi is the relative weight of each chemical parameter. Ci is the value of the concentration of different chemical parameters in a single sample; Si is the value of different chemical parameters in a single sample in groundwater quality standard (GB14848-2017) Category III water standard. Qi is the quality evaluation level of different chemical parameters in a single sample. Si is the value of the ith chemical parameter assessment level. Groundwater quality is categorized into five classes based on WQI values: excellent (< 50), good (50–100), medium (100–150), poor (150–200), and very poor (> 200) (Ravindra et al. 2023).

Positive matrix factorization model

The positive matrix factorization (PMF) model stands as an influential multivariate analytical instrument, predominantly harnessed in probing soil contamination origins. This statistical approach functions by deconstructing a given sample data matrix (X) into two separate matrices: the factor contribution matrix (G) and the factor profile matrix (F) (Zhang et al. 2020). Central to the model’s computations is the aspect of uncertainty, which unfolds as delineated below:

$${x}_{ij}=\sum\limits_{k=1}^{p}{g}_{ik}{f}_{kj}+{e}_{ij}$$
(5)
$$Q=\sum\limits_{i=1}^n\sum\limits_{j=1}^m\left[\frac{x_{ij}-\sum\limits_{k=1}^pg_{ik}f_{kj}}{u_{ij}}\right]^2$$
(6)
$$\mathrm{For }{{\text{x}}}_{{\text{ij}}}\le {\text{MDL}},{u}_{ij}=\frac{5}{6}\times MDL$$
(7)
$${\text{Else}}, {u}_{ij}=\sqrt{(\sigma \times {\text{concentration}}{)}^{2}+(0.5\times MDL{)}^{2}}$$
(8)

where “i” and “j” are the indexes of the samples and parameters respectively, and xij is the concentration of each parameter in each sample, gik is the concentration of each parameter in each source, fkj is the contribution of each source in each sample, eij is the model error, and “p” is the number of factors, where uij is the uncertainly matrix and n and m are the numbers of samples and parameters, respectively. MDL is the method detection limit value, and \(\sigma\) is the relative standard deviation.

Spatial interpolation

The inverse distance weighting (IDW) methodology represents an advanced interpolation technique grounded in the extrapolation of proximities and relational weights between the subject interpolated nodes and their surrounding recognized discrete coordinates (Adhikary and Dash 2017). Proximity plays a pivotal role; discrete coordinates nearer to the interpolated outcome exert more significant influence, while those more distanced possess reduced weightage. The underlying mathematical representation is articulated below:

$$z_0=\sum_{i=1}^n\frac1{(D_i)^p}\times z_i\times\left[\sum\limits_{i=1}^n\frac1{(D_i)^p}\right]^{-1}$$
(9)
$${D}_{i}=\sqrt{({x}_{0}-{x}_{i}{)}^{2}+({y}_{0}-{y}_{i}{)}^{2}}$$
(10)

where z0 represents the estimated value; zi refers to the attribute value of the ith sample; p refers to the power of distance, which significantly affects the result of interpolation; Di represents the distance; and (x0, y0) and (xi, yi) represent interpolation and discrete point coordinates, respectively.

Health risk assessment

The health risk assessment (HRA) serves as a sophisticated quantitative technique that elucidates the nexus between noxious substances in groundwater and the ensuing implications on human health, founded on the likelihood of deleterious health impacts stemming from these agents(Fan et al. 2014; Zhang et al. 2018). To assess the health risks related to groundwater, this study adopted a health risk assessment model that is endorsed by the US Environmental Protection Agency (US EPA). It evaluated the potential health risks associated with heavy metals by investigating two pathways: dermal contact and ingestion of drinking water. To bolster the veracity and consistency of derived assessments, we meticulously fine-tuned model parameters to mirror the distinct attributes of the Chinese populace. The human health risk assessment in this study adhered to established methodologies documented in prior research reports. The calculation process is as follows:

$$ADD_{i} = \frac{CW \times IR \times EF \times ED}{{BW \times AT}}$$
(11)
$$ADD_{d} = \frac{CW \times SA \times PC \times ET \times EF \times ED \times CF}{{BW \times AT}}$$
(12)
$$HQ = \frac{ADD}{{RfD}}$$
(13)
$$HI = HQ_{i} + HQ_{d}$$
(14)

where ADDi is the average daily exposure dose by drinking water intake route, mg/(kd·d); ADDd is the average daily exposure dose by dermal route of ingestion, mg/(kd·d). CW denotes the measured concentration of contaminants in water, mg/L. IR represents the average water consumption rate, L/day. EF represents the average exposure frequency, d/a. ED represents the average exposure duration, a. BW indicates the average body weight of residents, kg. AT denotes the average time of non-carcinogenic effect, d. SA represents the surface area of skin exposed to pollutants, cm2. PC denotes the skin permeability coefficient of pollutants, cm/h. ET is the average exposure time, h/d. CF is a conversion factor. RfD is the standard dose of a non-carcinogenic substance in water, mg/(kd·d). HQi is the hazard quotient by drinking water intake route. HQd is the hazard quotient by dermal route of ingestion. HI is the hazard quotient of the non-carcinogenic heavy metals under the two exposure routes.

Statistical analysis

Utilizing the SPSS software (version 27.0), descriptive statistical procedures and Pearson correlation analyses were conducted. The ArcGIS 10.7 software was employed for the creation of detailed maps, facilitating the geospatial elucidation of pollution trends. Source analysis was performed via the PMF EPA (version 5.0) application. The crafting of graphical representations was accomplished using both the Origin 2022 and R software tools. To generate more accurate risk values, Monte Carlo simulations alongside sensitivity analyses were executed utilizing Oracle Crystal Ball (version 11.1.2.4.400).

Results and discussion

Groundwater physico-chemical characteristics

Descriptive statistics of the samples from the study area are shown in Table 1. The pH in the water samples ranged from 6.84 to 7.98, with a mean value of 7.43, indicating an overall slightly alkaline condition. The range of TDS was 913.73–3664.16 mg/L, with an average of 2119.48 mg/L. According to the Groundwater quality standard (GB14848-2017, in Chinese), the maximum value observed was 3.66 times the limit for Category III. In terms of the average concentrations of various parameters in groundwater, heavy metals show a trend of Mn > Fe > Al > Ba, while anions and cations exhibit a trend of SO42−  > Ca2+  > Cl−  > Na+  > Mg2+  > NO3−  > K+  > F−. Among the heavy metals in groundwater, only Mn exceeded the standard (25.53%), while the exceeding rates for other metals were 0. Although Mn is an essential element for human survival and development, high concentrations of Mn in drinking water can lead to neurological diseases, which are particularly evident in the elderly (Hu et al. 2022; Liu and Han 2021). The coefficients of variation (CV) for K+, F−, and Mn in the study area all exceed 1, indicating significant variations in these elements’ concentrations among the sampling points, with some points having extremely high concentrations, suggesting the potential presence of point source pollution (Liu et al. 2022). Significantly, SO42− concentrations surpassed the prescribed limit in 93.62% of groundwater samples, accompanied by a coefficient of variation (CV) of 0.55 and a skewness of 0.50, highlighting a pervasive sulfate contamination across the study area. The average F− concentration was 1.84 mg/L, and approximately 40.43% of the groundwater samples exceeded the standard limit, confirming the presence of fluoride pollution in the groundwater. Similar to Mn, the intake of high concentrations of fluoride can also cause damage to the nervous system. In addition, long-term consumption of water with high fluoride levels can lead to dental fluorosis and skeletal fluorosis, manifested as osteoporosis and osteosclerosis (Lauria et al. 2022). Therefore, the pollution in groundwater deserves special attention.

Table 1 Statistical summary of groundwater quality indicator concentrations(mg/L) in the study area
Full size table

Assessment of water quality index (WQI)

Within the designated research zone, the employment of the water quality index (WQI) offers a holistic assessment of the groundwater’s quality (Naik et al. 2022). Figure 2 a graphically represents the WQI values corresponding to each sample point and element within the study parameters. Remarkably, approximately 68.09% of the sites rendered WQI readings below 50, while around 21.28% exhibited values ranging between 50 and 100, suggesting a predominantly satisfactory water quality across the majority of locations within the research domain. Sites manifesting elevated WQI are predominantly characterized by augmented concentrations of Mn and F−. The spatial distribution of WQI, delineated in Fig. 2b, indicates that areas with elevated WQI values are predominantly situated in the northern sector of the research region, whereas those with lower WQI readings are chiefly localized within the central and southern portions. This region is home to chemical and ceramic manufacturing units. Deviations from standard safety practices and potential releases of hazardous materials might be contributing to groundwater contamination within this vicinity.

Fig. 2
figure 2

a Bar graph of calculated WQI values of every sampling point and b spatial distribution of WQI values of groundwater

Full size image

Source identification and apportionment

In this study, the PMF model was employed with an integration of 12 distinct parameters to quantitatively assess the impact of both natural and anthropogenic influences on groundwater quality (Peng et al. 2022). The fluctuation of the signal-to-noise (S/N) ratio, ranging from 2.7 to 9.0, unequivocally accentuates the exemplary integrity and robustness of the amassed dataset. As shown in the Fig. 3, the four factors present the most credible scenarios to depict the significant origins of the primary elements in the groundwater of the research zone. Figure 3a displays the varying contribution percentages of the four factors: 36.76% for factor 1, 11.97% for factor 2, 25.96% for factor 3, and 25.31% for factor 4. Figure 3c illustrates the variability in percentage contributions of distinct factors across a spectrum of chemical elements. To accurately discern between natural and anthropogenic influences on water chemistry, Pearson correlation analysis was employed to clarify the interrelationships among water chemistry constituents.

Fig. 3
figure 3

Analysis of elemental sources in the study area. a The contribution rate of each factor by the PMF model. b The relationship between the elements was determined by combining Pearson correlation analysis and PMF modeling, where the width of the curve in b represents the contribution factor. c Elemental profiles obtained from the PMF model

Full size image

As elucidated in Fig. 3a, the principal component, denoted as factor 1, exerts a predominant influence on the region’s groundwater characteristics. The constituents predominantly associated with factor 1 include K+, Ca2+, Mg2+, SO42−, and Mn, contributing respectively at rates of 40.64%, 59.27%, 68.35%, 63.53%, and 71.40%. As noted earlier, SO42− concentrations exceed the accepted groundwater limit in 93.62% of the sampled groundwater instances. The detected levels of K+, Ca2+, and Mg2+ might reflect the weathering and dissolution dynamics of silicate minerals. This phenomenon is markedly evident within the mining district under study, where historical mining activities have not only expedited the physical breakdown of rock structures but also enhanced chemical weathering processes by exposing fresh mineral surfaces (Ding et al. 2022). The heightened concentrations of SO42− could be attributed to the oxidation of sulfide minerals, whereas the presence of Mn is probably a consequence of manganese ore extraction followed by its weathering (Jiang et al. 2020). Figure 3b discloses a positive correlation amongst K+, Ca2+, Mg2+, SO42−, and Mn, suggesting these ions exhibit consistent hydrogeochemical behaviors in the groundwater milieu (Liu and Han 2021). Hence, factor 1 is designated as the primary geological and industrial activity factor.

Factor 2 is primarily influenced by chloride ions, contributing 57.97%, with the substantial contribution of sodium ions (Na+) amounting to 35.1% also meriting attention. The enrichment of Cl− and Na+ in groundwater is a direct consequence of the interaction between the groundwater and the surrounding rock matrix of the shallow aquifer in the Zibo Fengshui area. Specifically, this process is closely associated with the weathering of silicate minerals within the study area, which include, but are not limited to, feldspar, mica, and clay minerals. This weathering facilitates the release of Cl− and Na+ from the rock matrix, followed by their dissolution and transport in groundwater. Ravindra et al. (2019) further corroborated this phenomenon, highlighting the significance of silicate mineral weathering on the chemical composition of groundwater. This weathering and dissolution process is a key factor in shaping the chemical characteristics of groundwater, revealing the complexity of the interaction between groundwater and rock matrix and emphasizing the role of silicate minerals in regulating the chemical composition of groundwater. Factor 2 predominantly encapsulates the impact of silicate mineral weathering and dissolution inherent to the geological setting, thereby justifying its denomination as the leaching-enrichment factor.

The dominant contributors to factor 3 are sodium ions (Na+) and fluoride ions (F−), accounting for 54.26% and 99.74%, respectively. Within the study area, fluoride contamination primarily originates from historical coal mining activities. The presence of conduits and fractures between the abandoned mine shafts and the karst aquifer within the study area facilitates the diffusion of fluoride pollutants from the abandoned mine shafts into the karst water system, thereby inducing interstratal contamination in the groundwater system (Liu et al. 2013). This process has led to a significant increase in the fluoride concentration in the groundwater, with an exceedance rate reaching 40.43% (He et al. 2019). Figure 3b delineates a positive correlation between Na+ and F−, suggesting a similar hydrogeochemical behavior among these ions. Consequently, it is reasonable to label factor 3 as the mine wastewater factor.

Factor 4 distinctly manifests the influence of Fe, Al, NO3−, and Ba. Existing literature identifies NO3− as a prevalent component of fertilizers, thereby representing a significant marker of agricultural activities. The concentrations of Fe, Al, and Ba are predominantly influenced by the geological context, potentially attributable to the weathering and dissolution processes of select crustal minerals (Rajkumar et al. 2020). Diverse agricultural practices, including fertilization and irrigation strategies, might indirectly elevate the abundance of these elements within the soil matrix, subsequently influencing their distribution in the groundwater. Therefore, it is appropriate to label factor 4 as the primary geological and agricultural activity factor.

Health risk assessment

Deterministic health risk assessment

The risks of F− and NO3− in groundwater were evaluated using the deterministic Human Health Risk Assessment (HHRA) model recommended by the US Environmental Protection Agency (EPA) (USEPA 2019). This model takes into account two exposure pathways: direct ingestion of the water and dermal exposure. Figure 4 reveals the spatial distribution characteristics of the potential health risks brought about by increased concentrations of F− and NO3− in groundwater. As illustrated in Fig. 4a and b, in the groundwater samples covered by the study, 95.7% and 61.7% exceeded the threshold of the non-carcinogenic health risk index (HI = 1) for adults and children respectively. Contrarily, the situation with NO3− is somewhat more optimistic, as shown in Fig. 4c and d, with only 17% and 2.1% of the groundwater samples exceeding the acceptable health risk boundaries for children and adults, respectively. These findings reveal that both F− and NO3− in groundwater pose certain levels of non-carcinogenic health risks, with children being at a higher risk than adults. Similar results that children are more at non-carcinogenic risk in groundwater than adults were reported by scholars (Gao et al. 2022). However, due to the generally higher health risk index of F− and their wider impact, the potential health risks they pose to the study population warrant significant attention.

Fig. 4
figure 4

Spatial distribution map. a The HI for fluoride in children; b the HI for fluoride in adults; c the HI for Nitrate in children; c the HI for Nitrate in adults

Full size image

Uncertainty of health risk assessment

The non-carcinogenic risks to adults and children from the main groundwater pollutants (NO3− and F−) through ingestion and dermal exposure pathways were simulated using the Monte Carlo method. This simulation, involving 10,000 random experiments, assessed the non-carcinogenic risks at 5% and 95% confidence levels for both adults and children. In the cumulative probability distribution graph (Fig. 5b), the average total non-carcinogenic risk for adults and children, as represented by the hazard index (HI), was found to be 0.38 and 1.04, respectively. The non-carcinogenic risk probability for children was found to be 2.74 times higher than that for adults, and 31.89% of children were found to be in an unacceptable non-carcinogenic health risk category. This can primarily be attributed to their increased average daily exposure dose, a consequence of their lower average body weight. As per Fig. 5a, the proportions of adults and children at unacceptable non-carcinogenic risk (HI > 1) from NO3− were 4.41% and 15.92%, respectively. In contrast, as observed from Fig. 5c, the proportions of adults and children at unacceptable non-carcinogenic risk (HI > 1) from F− were 30.66% and 90.12%, respectively. Overall, the burden of non-carcinogenic risk is significantly higher in children than in adults, with F− serving as the primary driving factor of this elevated risk.

Fig. 5
figure 5

The probability distribution for non-carcinogenic risk: a non-carcinogenic risk of NO3− to adults and children; b total non-carcinogenic risk for adults and children; c non-carcinogenic risk of F− to adults and children

Full size image

Sensitivity analysis

Sensitivity analysis was deployed to delineate the magnitude of impact that exposure vectors and adjustable parameters have on non-carcinogenic health adversities (Ganyaglo et al. 2019). During the computation of the hazard index (HI) encompassing both adult and pediatric populations, variables like water concentration (CW), ingestion rate (IR), and body mass (BW) were meticulously considered for their sensitivity.

As depicted in Fig. 6, the influence of CW on HI is particularly significant for NO3− and F−, while IR also exhibits a positive correlation with HI. However, the body weights of adults and children show a negative correlation with HI. Consequently, within the examined demographic, attenuating the levels of NO3− and F− in the water emerges as the paramount approach to curtail non-carcinogenic health hazards. In Mahmood et al.’s (2019) research, an assessment of non-carcinogenic health risks associated with Cr was conducted. CW and IR emerged as the most significant variables influencing the results, with CW having the most pronounced impact. This study underscores the necessity of methodical risk evaluations through two primary channels of pollutant exposure: drinking water consumption and dermal contact. Yet, due to data limitations, we were unable to incorporate additional exposure sources (e.g., food intake and inhalation). This oversight could potentially underestimate the true risks for our examined population. For forthcoming inquiries, it is advisable to harness more inclusive datasets to amplify the precision of risk quantification pertaining to F− and NO3− exposure.

Fig. 6
figure 6

Sensitivity analysis of non-carcinogen exposure NO3− for adults (a) and children (b) groups and F− for adults (c) and children (d) groups

Full size image

Implication for the groundwater fluoride and nitrate management

Groundwater resources are indispensable for urban potable supply and agricultural irrigation systems. Yet, anthropogenic activities, encompassing domestic wastewater discharge, application of fertilizers, over-irrigation, and industrial emissions, coupled with inherent geological processes, have imperiled its quality. Specifically, the prevalence of fluoride and nitrate compounds poses intricate challenges to groundwater stewardship. In order to achieve sustainable utilization of water resources, effective measures must be taken to restrict wastewater discharge and to carry out rational irrigation activities (Miller et al. 2022). Furthermore, instituting an exhaustive groundwater surveillance system, facilitating real-time monitoring of aquifer levels and qualitative indices, emerges as a cardinal instrument to obviate further qualitative degradation (Tracy et al. 2022). At a societal echelon, augmenting the cognizance of the populace regarding sagacious irrigation protocols and the quintessence of groundwater resources is paramount for the holistic preservation and sustainable harnessing of subterranean water reserves.

Conclusions

This study integrates health risk assessment, groundwater chemical characterization, and comprehensive water quality evaluation to reveal the current state and primary influencing factors of groundwater quality within the study area. The findings indicate that the groundwater generally exhibits a slightly alkaline pH, with significant exceedances of sulfate, fluoride, and manganese concentrations, notably pointing to point source pollution issues for potassium, fluoride, and manganese. Water quality index (WQI) analysis suggests that while a majority of the monitoring sites exhibit good water quality, attention is warranted for water quality management in the northern and southern fringe areas of the study zone. Through positive matrix factorization (PMF) analysis, four principal pollution contributing factors were identified: primary geological and industrial activity factor, leaching-enrichment factor, mine wastewater factor, and primary geological and agricultural activity factor. These reflect that geological background and human activities predominantly influence the groundwater quality in the area. Health risk assessments indicate that fluoride poses a greater health risk than nitrate, with children being more vulnerable to these health risks than adults. Reducing the concentration of pollutants in groundwater appears to be the most effective method to mitigate non-carcinogenic health hazards, aligning with the outcomes of the sensitivity analysis. Consequently, it is imperative for relevant authorities to take immediate and effective actions to control groundwater pollution to safeguard the local population’s health.