The Gaocheng-Wuji region is situated in the central part of Shijiazhuang, North China Plain (NCP) (Figure 1). It extends between 37°51′21.29″N and 38°18′39″N in latitudes and from 114°8′33″E to 115°8′30″E in longitudes, with a covering area of ∼1,296 km². Geomorphologically, this area belongs to the piedmont plain of the NCP. This region belongs to the semiarid continental climate. The average annual rainfall approximates 500 mm and more than 50% occurs in July and August. There are three rivers, namely Hutuo River, Ci River, and Wangyang River, running through the study area from the west towards the east (Figure 1C). While rivers in this area usually dry up due to the reservoirs construction in the upstream, resulting in surface water scarcity. Groundwater plays an irreplaceable role in water supply for human society in this area. The land use type in the study area is dominantly separated into three types, i.e., agricultural land, urban land, and industrial land. The urban lands are mainly distributed in the central area with the name of Gaocheng and in the northeast with the name of Wuji (Figure 1C). Industrial lands are concentratedly distributed in the west, namely the economic development zone of Shijiazhuang. Besides, there are many sporadic towns and villages distributing in the whole study area. However, agricultural land is the most dominant land use type and with a long history of agriculture and fertilizer practices.

Geologically, the present study area mainly belongs to the Hutuo river alluvial fan and is comprised of Quaternary alluvial deposits with the depth of 150–300 m (Cheng et al., 2017). These deposits were formed during the Pleistocene and Holocene with the lithology of pebble, gravel, coarse sand, sand, and fine sand. Thus, aquifers in this area are mainly unconfined with relatively high hydraulic conductivity (K = 27–70 m/day) (Chen et al., 2006) and it is easy for water and contaminants to move horizontally and vertically. Groundwater is recharged vertically by precipitation, irrigation water, and surface water seepage and horizontally by lateral flow in, and discharged mainly through anthropogenic pumping and lateral flow out. Regionally, groundwater flows from the northwest to the southeast (Li et al., 2016).

A total of 49 samples were collected from the phreatic aquifers across the study area (Figure 1C) from July to August in 2018. The sampling sites were selected based on the availability and even spatial distribution of boreholes/wells. All groundwater samples were obtained from boreholes/wells with depths in the range of 60–120 m. Boreholes and wells were pumped for more than 15 min to remove the stagnant water before conducting sampling strategies. The pH and electrical conductivity (EC) of target groundwater were monitored in this process, and the sampling was performed until these two in-situ physicochemical parameters became stable. Clean high-density polyethylene bottles were prewashed using the target groundwater in the field and used for groundwater sampling. All samples were stored under the condition of 4°C with the aid of incubators. Samples were sent to the laboratory of groundwater sciences and engineering of the institute of hydrogeology and environmental geology, Chinese Academy of Geological Sciences for hydrochemical analysis.

In-situ parameters of groundwater including water temperature (T), pH, EC were measured in the field using a portable device (Multi 350i/SET, Munich, Germany). Trace elements (Zn, Fe, Mn) and major cations (K⁺, Na⁺, Ca²⁺, Mg²⁺) were measured by inductively coupled plasma-mass spectrometry (Agilent 7500ce ICP-MS, Tokyo, Japan). NH₄ ⁺ and anions of SO₄ ²⁻, Cl⁻, NO₃ ⁻, NO₂ ⁻, F⁻ were determined using ion chromatography (Shimadzu LC-10ADvp, Kyoto, Japan). Acid–base titration was performed for analyzing HCO₃ ⁻ of groundwater. The quality assurance of analyzed results was ensured by checking the consistency and accuracy of hydrochemical analysis. The consistency of analyzed parameters were checked by conducting triplicate analyses. The accuracy of laboratory analysis was checked using the ionic charge balance error (ICBE) percentage using Eq. 1. All samples were with the ICBE within ±5%, suggesting the analysis results of all groundwaters were reasonably good. % I C B E = ∑ i m c e i + − ∑ i m c e i − ∑ i m c e i + + ∑ i m c e i − (1) Where c e i + and c e i − signify the charge concentration of a specific cation and anion, respectively, with the unite of milliequivalent per liter.

The potential hazards of aqueous contaminants in groundwater can be quantitatively evaluated with the aid of human health risk assessment (HHRA) model established by the United States Environmental Protection Agency (USEPA, 1989). Although contaminants can threaten the health of consumers through many pathways such as inhalation, dermal contact, and oral intake, the oral pathway is evidenced to be the harmful one (Yin et al., 2020). Therefore, the non-carcinogenic risk of contaminants via the oral pathway was the concern of the present study. Four populations were divided and considered according to the age and gender, i.e., infants (0–6 months), children (7 months–17 years), adult females (≥18 years) and adult males (≥18 years).

The non-carcinogenic risk posed by a specific aqueous contaminant can be expressed by the hazard quotient (HQ) and calculated as follows: C D I = C × I R × E D × E F A T × B W (2) A T = E D × 365 (3) H Q = C D I R f D (4) Where, CDI represents the chronic daily ingestion dose of the concerned contaminant via oral pathway; C refers to the aqueous contaminant concentration; IR is the water oral ingestion rate; ED and EF express the duration and exposure frequency, respectively; AT denotes the average exposure time; BW signifies the average body weight of concerned population; RfD is the reference dosage via oral pathway. Parameters that used in the HHRA were presented in Table 1 as below:

The quality of groundwater can be comprehensively understood with the aid of water quality index (WQI). It has been widely used over the world and proved to be robust and effective. However, its accuracy heavily relies on the weights assigned to each physicochemical parameter involved in the WQI appraisal. The entropy-weighted water quality index (EWQI) method introduces the information entropy to objectively assign the weights and can greatly improve the accuracy of assessment results (Xiao et al., 2020).

For the EWQI procedure, firstly, the eigenvalue matrix “X” should be constructed using “m” groundwater samples and “n” physicochemical parameters (Eq. 5). Secondly, the “X” is standardized to “Y” (Eq. 6) using the normalizing function (Eq. 7). Thirdly, the entropy weight “w_j” is calculated (Eqs 8–10). Fourthly, the rating scale is quantitatively determined (Eq. 11). Finally, the EWQI value is computed (Eq. 12). X = [ x 11 x 12 x 21 x 22 ⋯ … x 1 n x 2 n ⋮ ⋱ ⋮ x m 1 x m 2 ⋯ x m n ] (5) Y = [ y 11 y 12 y 21 y 22 ⋯ … y 1 n y 2 n ⋮ ⋱ ⋮ y m 1 y m 2 ⋯ y m n ] (6) y i j = x i j − ( x i j ) min ( x i j ) max − ( x i j ) min (7) w j = 1 − e j ∑ i = 1 n ( 1 − e j ) (8) e j = − 1 ln ⁡ m ∑ i = 1 m ( P i j × ln ⁡ P i j ) (9) P i j = y i j ∑ i = 1 m y i j (10) q j = 100 × C j S j (11) E W Q I = ∑ j = 1 m ( w j × q j ) (12) Where, “i” and “j” denote the sample no. and index no., respectively; “e_j ” represnts the information entropy; “C_j ” and “S_j ” are the measured physicochemical parameter value and the drinking water standard of the physicochemical parameter, respectively.

The quality of groundwater can be classified into five categories, i.e. Rank 1: Excellent quality (EWQI < 50), Rank 2: Good quality (EWQI, 50–100), Rank 3: Medium quality (EWQI, 100–150), Rank 4: Poor quality (EWQI, 150–200), Rank 5: Extremely poor quality (EWQI, >200) (Amiri et al., 2014; Luo et al., 2021).

The physicochemical parameters of groundwater can provide preliminary insights into the overall geochemical characteristics of sampled groundwaters and have been statistically presented in Table 2. The related desirable limits recommended by the Chinese guidelines (GAQS, 2017) and the World Health Organization (WHO, 2017) were also listed in the table for comparison.

The sampled phreatic groundwaters in the study were with the temperature ranging from 14.1°C to 22.5°C, averaging at 16.7°C, which was very close to the annual average air temperature of 14.2°C in the study area. The pH of groundwater ranged from 7.11 to 8.07 with an average of 7.52 (Table 2), demonstrating a slightly alkaline nature. All sampled groundwaters had the pH value within the desirable limit range of 6.5–8.5 recommended by Chinese Guidelines for drinking purposes. The EC values were in the range of 329–2,224 μS/cm with a mean of 851 μS/cm. The TDS varied from 172–1,875 mg/l and averages at 510 mg/l. Both of EC and TDS suggested the phreatic groundwater in the study area had a relatively large variation of salinity. The total hardness of sampled groundwater ranged from 100 to 1,166 mg/l with an average of 356 mg/l. As shown in Figure 2A, the sampled groundwaters in the study area were predominantly moderately-hard fresh water, with a small portion of soft fresh, hard fresh, very-hard fresh and very-hard brackish water.

The concentration of major cations and anions was also in relatively wide range (Table 2). Ca²⁺ is the dominant cation in the phreatic groundwater (Figure 2B) with its concentration in the range of 30.29–280.70 mg/l and averaging at 85.7 mg/l. Na⁺ and Mg²⁺ ranked the second and third in the major cations, with the concentration varying from 12.63 to 185.80 mg/l (a mean of 40.05 mg/l) and between 5.98 and 113.00 mg/l (a mean of 34.54 mg/l), respectively. While the K⁺ was the least abundant ion in major cations and ranged from 1.14 to 12.24 mg/l with an average of 2.52 mg/l. For major anions, HCO₃ ⁻ ranked the first, and followed by SO₄ ²⁻ and Cl⁻ (Figure 2B). The concentration of major anions was in the range of 158.00–693.10 mg/l with an average of 288.52 mg/l for HCO₃ ⁻, 4.81–459.00 mg/l with a mean of 81.98 mg/l for SO₄ ²⁻, 5.25–570.40 mg/l with an average of 67.86 mg/l for Cl⁻. The hydrogeochemical facies of phreatic groundwater were dominated by HCO₃-Ca type, and only a small portion belonged to Mixed Cl-Mg·Ca, Mixed HCO₃-Na·Ca and Cl-Ca type (Figure 2B), demonstrating an overall fresh nature for phreatic groundwater in the study area. The relatively salty groundwaters are mainly distributed in the middle area of the study area along the Hutuo River, extremely in the urban Gaocheng (Figure 3A).

Elements (including nitrogen, fluoride and heavy metals) that may threat the quality of water supply were detected. Most of these elements were found with very low concentrations and even below the detection limits. The detected elements were summarized in Table 2.

Nitrogen is the common and widely distributed contaminant in groundwater across the world, and also found with relative high concentration for its three forms, i.e., NO₃ ⁻, NO₂ ⁻, NH₄ ⁺, in the study area. The concentration of NO₃ ⁻ was in a significant large range, and varied from 2.00 mg/l to 313.40 mg/l with an average of 30.23 mg/l. Although the mean concentration of NO₃ ⁻ in the phreatic aquifer was far below the desirable limit of 50 mg/l recommended by WHO (2017) and most groundwaters (83.67%) were within this limit, some groundwaters were detected with extremely high NO₃ ⁻ concentration (up to 313.40 mg/l, more than 6 times of the recommended desirable limit) yet. The concentration of NO₂ ⁻ and NH₄ ⁺ ranged from 0.001 to 0.320 mg/l and from 0.02 to 0.30 mg/l, respectively, averaging at 0.033 and 0.06 mg/l. About 28.57 and 10.20% of groundwaters were found with the NO₂ ⁻ and NH₄ ⁺ concentration exceeding the drinking water desirable limits (Table 2). The high NO₃ ⁻ groundwaters were dominantly distributed in the urban Gaocheng (Figure 3B), and the high NO₂ ⁻ groundwaters distributed in the area from the urban Gaocheng to the urban Wuji (Figure 3C). The exceeding NH₄ ⁺ in groundwater is sporadic in spatial (Figure 3D).

Fluoride is a common contaminant in groundwater. Its concentration varied from 0.22 to 0.99 mg/l with an average of 0.54 mg/l in phreatic groundwater in the study area. All groundwaters are with the F⁻ concentration within the desirable limit of 1 mg/l recommended by WHO (2017). But it should be noted that the maximum concentration of F⁻ in sampled groundwater approximated the desirable limit. As shown in Figure 3E, the phreatic groundwater was with the F⁻ contaminant in a relatively low level, only very limited sites were observed with the F⁻ concentration approximating the desirable limit. For heavy metals, only Fe was found beyond the drinking water desirable limit, while others were all within the desirable limits (Table 2). The concentration of Fe in phreatic groundwater varied between 0.005 and 4.980 mg/l with an average of 0.230 mg/l, and about 14.29% of sampled groundwater exceeded the desirable limit of 0.3 mg/l.

The hydrogeochemical components of groundwater can originate from various sources and be regulated by many processes (Rajmohan and Prathapar, 2016; Gao et al., 2020; Satheeskumar et al., 2020; Luo et al., 2021). Natural processes, usually in three categories of precipitation, rock, and evaporation dominance, lay the foundation of groundwater hydrochemical composition. Gibbs diagrams, i.e. the cation weight ratio of Na⁺/(Na⁺+Ca²⁺) and Cl⁻/(Cl⁻ + HCO₃ ⁻) versus TDS and anion weight ratio of Cl⁻/(Cl⁻ + HCO₃ ⁻) versus TDS (Gibbs, 1970), are effective tools to visibly reveal the main processes governing groundwater chemistry in nature and have been widely used worldwide (Xiao et al., 2018; Luo et al., 2021; Xiao et al., 2021). As demonstrated in Figure 4, all sampled groundwaters were plotted in the rock dominance rather than precipitation and evaporation dominance, suggesting that the hydrochemistry of phreatic groundwater in the study area is significantly and dominantly controlled by water-rock interaction and out of the influence of evaporation.

For further illustrate the water-rock interaction processes, the end-member diagrams were employed to constrain the specific rocks involved in these natural mechanisms. Three main categories of rocks, namely evaporates, silicates and carbonates, can be identified from the end-member diagrams of Mg²⁺/Na⁺ versus Ca²⁺/Na⁺ and HCO₃ ⁻/Na⁺ versus HCO₃ ⁻/Na⁺ (Venkatramanan et al., 2017; Adeyeye et al., 2021; Xiao et al., 2021). As demonstrated in Figure 5, the phreatic groundwaters were dominantly situated in the silicates dominance and the transition between silicates and carbonates, implying silicates weathering and carbonates dissolution were the main sources of ionic substances for phreatic groundwater in the study area. In addition, a few groundwaters were observed plotting in the transition between silicates and evaporites, demonstrating that evaporites dissolution also contributed to the hydrochemical components of phreatic groundwater at some sites to some degree.

The correlation matrix and saturation status modeling were employed to further illustrate the contribution of specific minerals to the natural chemical components of groundwater. As shown in Table 3, a positive relation coefficient was observed between Na⁺ and Cl⁻, indicating halite dissolution (NaCl) was the potential source of Na⁺ and Cl⁻ in groundwater. This was confirmed by the halite saturation index computing results (SI < 0, implying unsaturation) presented in Figure 6. But it should be noted that its relation coefficient was below the significant standard of 0.5, suggesting limited contribution from the halite dissolution. This was coincided with the conclusion implied by end-member diagrams (Figure 5). While SO₄ ²⁻ was observed having significant positive relation with Na⁺ (relation coefficient = 0.89), Ca²⁺ (relation coefficient = 0.71) and Mg²⁺ (relation coefficient = 0.69), indicating the dissolution of sulfate minerals (Na₂SO₄, CaSO₄, CaSO₄·2H₂O, MgSO₄, etc.) were the dominant process contributing to the chemical components of groundwater. These processes were confirmed by the saturation index (SI) results that all sampled groundwater were under unsaturated states (SI < 0) of sulfate minerals (such as anhydrite, gypsum, etc.) (Figure 6). Additionally, significant positive relations were observed between HCO₃ ⁻ and Ca²⁺ (relation coefficient = 0.70), Mg²⁺ (relation coefficient = 0.76), suggesting carbonate minerals including calcite, aragonite, dolomite were potential natural sources of groundwater chemical components (Table 3). While the saturation index results showed that only part of groundwaters were under unsaturated status of carbonate minerals, thus, the dissolution of calcite, aragonite, dolomite was the dominant processes contributing hydrochemical components of part groundwaters rather than all groundwaters. The contribution of carbonate dissolution is especially significant for groundwaters that with low TDS (Figure 6).

NO₃ ⁻ is a common contaminant for human society. Its natural limit is 10 mg/l. Any groundwater with NO₃ ⁻ concentration beyond this limit implies being impacted by anthropogenic contamination. The molar ratio diagram of NO₃ ⁻/Na⁺ versus SO₄ ²⁻/Ca²⁺ can demonstrate a detailed picture of the source of anthropogenic contaminants (Gao et al., 2021). As shown in Figure 7A, groundwaters at G1, G2, G5, G26, G30, and G45 were presented being contaminated by nitrate-containing fertilizers or domestic wastewater. It can be clearly seen that except G45 situated in the agricultural land, the other 5 groundwaters are located in the human settlements, suggesting that the NO₃ ⁻ in groundwater at G1, G2, G5, G26, G30 originated from domestic wastewater, while that at G45 was from nitrate-containing fertilizers. The correlation matrix suggested that NO₃ ⁻ in groundwater had significant positive relation with EC, TH and TDS (Table 3), indicating the inputs of anthropogenic NO₃ ⁻ also brought some other chemical components (such as Ca²⁺, Mg²⁺, SO₄ ²⁻, etc.) and increased the salinity of groundwater. This was evidenced by the similar distribution of high TDS groundwater (Figure 3A) and high NO₃ ⁻ groundwater (Figure 3B). Besides, a significant positive relationship was also observed between NO₃ ⁻ and NO₂ ⁻ and between NO₂ ⁻ and NH₄ ⁺, implying the existence of nitrosification and nitrification processes that transformed the NH₄ ⁺ into NO₂ ⁻ and finally into NO₃ ⁻. This conforms to the real condition of unsaturated zone and phreatic aquifers that with relatively good permeability and rich oxygen. Figure 7A also suggested groundwater at G29 and G43 were impacted by sulfur-containing fertilizers or industrial wastewater. Considering their agricultural land use type (Figure 1C), groundwater at these two sites were contaminated by the sulfur-containing fertilizers rather than industrial wastewater. Groundwaters at G29, G43 were found with higher TDS than background value, thus, the inputs of sulfur-containing fertilizers contaminants also increased the salinity of groundwater.

Ion exchange is usually one of the important natural processes regulating groundwater chemistry. It is especially significant for aquifers that with fine mediums. These kinds of hydrochemical processes can be clearly revealed by the chloro-alkaline indices including CAI-1 (Eq. 12) and CAI-2 (Eq. 13). C A I − 1 = C l − − ( N a + + K + ) C l − (13) C A I − 2 = C l − − ( N a + + K + ) S O 4 2 − + C O 3 2 − + H C O 3 − + N O 3 − (14)

If the aqueous Ca²⁺ or Mg²⁺ in groundwater exchanges the absorbed Na⁺ or K⁺ on sediments, i.e., cation-exchange reaction, the aforementioned two chloro-alkaline indices would be negative. While if the aqueous Na⁺ or K⁺ in groundwater exchanges the absorbed Ca²⁺ or Mg²⁺ on sediments, i.e., reverse cation-exchange reaction, both CAI-1 and CAI-2 are expected to be positive. Generally, cation-exchange reaction occurs in the groundwater with relatively low TDS and benefits groundwater mineralization. On the contrary, reverse cation-exchange reaction is highly possible to occur in the high TDS groundwater, in which the rich aqueous Na⁺ in groundwater tends to exchange Ca²⁺ or Mg²⁺ on sediments. As demonstrated in Figure 7B, most sampled groundwaters were with the two chloro-alkaline indices below 0, indicating that cation-exchange reaction was the dominant hydrochemical process and favored the enrichment of Na⁺ and K⁺ in groundwater. Additionally, a part of sampled groundwater was observed having the positive CAI-1 and CAI-2 value, implying occurred reverse cation-exchange reaction for these groundwaters. According to the statistics, groundwaters with negative chloro-alkaline indices had the TDS of 447 mg/l on average, and that with positive chloro-alkaline indices were with the TDS of 566 mg/l in average, suggesting the elevation of TDS had led to the ion exchange process transformed from cation-exchange to reverse cation-exchange for phreatic groundwater in the study area.

For the minor elements of F⁻, Zn, Fe, Mn, no significant relations were observed between them and the nitrogen contaminants (Table 3), indicating that they were not originated from the nitrogen-containing anthropogenic contamination. As discussed before, Zn and Mn were all in relatively low concentration and far below the guideline limits, thus, it is highly possible they originated from the natural dissolution of Zn-bearing and Mn-bearing minerals. This was confirmed by the simulation results of the saturation index that all Zn-bearing and Mn-bearing minerals were under unsaturated states for sampled phreatic groundwaters. For F⁻ contaminant, its concentration in all sampled groundwaters were within the guideline limit of 1 mg/l (Table 2). Although it presented a relatively high concentration (approaching 1 mg/l) sporadically in the study area (Figure 3E), no relation was observed between it and high salinity/other contaminants in terms to the spatial distribution and statistics. Consequently, it was highly possible that the F⁻ contaminant in groundwater may have come from the natural release of fluoride-bearing minerals. This was also evidenced by the saturation indexes simulation presented in Figure 6. For Fe in groundwater, it has no significant relationship with the salinity and nitrogen contaminants (Table 3), indicating it was not from the anthropogenic source. The saturation index simulation showed all Fe-bearing minerals were under unsaturated status and can potentially contribute aqueous Fe to groundwater if there are existing Fe-bearing minerals. Thus, the sporadic high Fe in groundwater of the study area were from the release of natural Fe-bearing minerals.