Study on Natural Background Levels and Mechanisms of Groundwater Contamination in an Overexploited Aquifer Region: A Case Study of Xingtai City, North China Plain

Abstract

This study investigates the groundwater over-exploitation zone in Xingtai City, North China Plain, to address two critical gaps in the current understanding of groundwater chemistry: the lack of established natural background levels (NBLs) and the ambiguous mechanisms of groundwater contamination. Sixty shallow-groundwater samples were collected and analyzed using a combination of Piper diagrams, cumulative-probability statistics, contamination-index methods, and multivariate statistical techniques to determine NBLs and threshold values (TVs) for major chemical constituents and to clarify the contamination mechanisms. The results indicate that the groundwater is weakly alkaline, with the most prevalent water types being HCO₃–Na and SO₄·Cl–Na. The NBLs for Na⁺, Ca²⁺, Mg²⁺, Cl⁻, $\mathrm{S}{\mathrm{O}}_4^{2-}$ and $\mathrm{N}{\mathrm{O}}_3^{-}$ are 32.3 mg/L, 34.1 mg/L, 17.8 mg/L, 46.2 mg/L, 66.4 mg/L and 0.886 mg/L, respectively, and the corresponding TVs are 116 mg/L, 54.6 mg/L, 33.9 mg/L, 248 mg/L, 258 mg/L and 44.7 mg/L. Based on the TVs, 56.7% of the sampling sites are identified as anthropogenically contaminated. Principal component analysis reveals that groundwater over-extraction, industrial activities and water–rock interaction are the dominant drivers of groundwater contamination, whereas intensive abstraction, agricultural fertilization and domestic sewage discharge exert additional influence. The findings provide a scientific basis for pollution control and sustainable utilization of groundwater in over-exploited regions.

1. Introduction

Natural background levels (NBLs) of groundwater are fundamental for scientifically identifying contamination and establishing water-quality protection standards [1]. They not only reflect the intrinsic concentrations of chemical constituents under undisturbed hydrogeological conditions, but also serve as critical thresholds for distinguishing natural geochemical processes from anthropogenic impacts [2]. However, groundwater chemistry is influenced by a complex interplay of factors, including geological settings, climate, land use, and human activities [3]. Under the influence of intensive anthropogenic pressures, groundwater is currently threatened by diffuse agricultural pollution, as well as point sources such as industrial effluents and domestic sewage [4]. Accurate delineation of regional NBLs is essential for safeguarding drinking-water security and supporting the sustainable use of water resources.

In recent years, substantial advancements have been made globally in the determination of groundwater NBLs and the characterization of contamination. Urresti-Estala et al. [5] employed statistical, iterative and distribution-function methods to establish NBLs in Milan, Italy, and the Guadalhorce River basin, Spain, respectively. Their research revealed that industrial wastewater discharge and agricultural fertilizer application are primary factors causing concentrations to exceed NBLs and triggering contamination. Zhou et al. [6] demonstrated that agricultural activities are the primary driver of elevated NBLs and subsequent contamination in the Xi’an alluvial plain. Furthermore, numerous studies have investigated contamination mechanisms and controlling factors: in rapidly urbanizing areas, groundwater quality deterioration is mainly attributed to domestic sewage, industrial wastewater and chemical fertilizers [7,8], whereas in coastal zones, seawater intrusion, water–rock interactions, sewage discharge and agricultural pollution dominate [9,10]. Evidently, differing anthropogenic patterns and geological backgrounds among regions lead to distinct contamination mechanisms. With increasing groundwater abstraction, extensive cone-of-depression has formed in many areas, altering both hydraulic and hydrochemical regimes [11] and exacerbating groundwater quality degradation. Nevertheless, investigations into NBLs and contamination mechanisms in over-exploited zones remain scarce, constraining effective groundwater protection in such regions.

Consequently, this study selects Xingtai City in the North China Plain as the study area. Xingtai hosts the largest groundwater depression cone in China and represents a critical agricultural and industrial base. Groundwater over-extraction and quality deterioration are severe in this region. The aquifer system is subject to both naturally elevated geochemical backgrounds and anthropogenic contamination. The NBL characteristics and contamination mechanisms in this region may thus differ markedly from those reported elsewhere. This study employs cumulative probability methods, contamination index methods, and multivariate statistical techniques to determine NBLs, elucidates hydrochemical signatures driven jointly by natural and anthropogenic factors, and clarifies the mechanisms of water-quality deterioration. The results provide a scientific basis for rational groundwater development and contamination control in over-exploited regions.

2. Materials and Methods

2.1. Study Area

The study area is situated in the central-eastern part of Xingtai City, Hebei Province, China, covering a total area of 12,414 km², 70.8% of which is the plain region. Xingtai has a north-temperate continental semi-arid monsoon climate with four distinct seasons. The multi-year mean air temperature is 13.5 °C. Precipitation is concentrated mainly in July and August, with a multi-year mean of 519.7 mm, whereas the multi-year mean evaporation is 1953 mm [12].

2.2. Geological and Hydrogeological Setting

Xingtai city is located in the North China Platform in terms of regional geology. The western part is the Taihang Mountains, while the eastern part is the North China Plain. The aquifer system comprises primarily porous and karstic aquifer. Porous aquifer is dominant in the central and eastern plains, while karstic aquifer is primarily found in the western mountainous area; the present study focuses on the porous aquifer in the central and eastern plains. Carbonate rocks are widely distributed in the area. The regional groundwater flow direction is from west to east. Recharge sources include atmospheric precipitation, surface-water infiltration and irrigation return flow; discharge is dominated by artificial abstraction.

2.3. Sampling and Analytical Procedures

In May 2024, 60 shallow-groundwater samples were collected (Figure 1). Polyethylene bottles (500 mL) were used for sampling. All bottles were first rinsed with distilled water and then conditioned by rinsing three times with groundwater at the sampling point to ensure sample integrity. A submersible pump was employed for purging and sampling. A 1-inch submersible pump was used for well purging and sampling at a pumping rate of 1–3 L/min, with a purge volume of 100 L. Each well was purged until field pH and EC stabilized before sample collection. Each sample was split into two aliquots: one aliquot for cation analysis (Na⁺, K⁺, Ca²⁺ and Mg²⁺) was acidified on site with ultrapure nitric acid to pH < 2; the second aliquot for anion analysis (Cl⁻, $\mathrm{S}{\mathrm{O}}_4^{2-}$, F⁻ and $\mathrm{N}{\mathrm{O}}_3^{-}$) was left unacidified. All bottles were filled without headspace, stored in the dark at 4 °C and analyzed within one week of collection.

Field pH was measured with a portable multi-parameter analyzer (Hach-HQ40D, Hach Company, Shanghai). Cations (K⁺, Na⁺, Ca²⁺ and Mg²⁺) were determined by atomic absorption spectrophotometry (AA240FS/GTA120, Agilent Technologies, Santa Clara, CA, USA). Anions (Cl⁻, $\mathrm{S}{\mathrm{O}}_4^{2-}$, F⁻ and $\mathrm{N}{\mathrm{O}}_3^{-}$) were analyzed by ion chromatography (CIC-D180, Qingdao Shenghan Chromatography Technology Co., Ltd., Qingdao, China). $\mathrm{H}\mathrm{C}{\mathrm{O}}_3^{-}$ was measured by hydrochloric-acid titration. Total hardness (TH) was determined by EDTA titration, and total dissolved solids (TDS) by the drying-gravimetric method.

Following analysis, charge-balance errors for cations and anions were calculated and constrained to <5% to ensure data reliability.

2.4. Methodology

2.4.1. Determination of Hydrochemical Facies

The Piper trilinear diagram is commonly employed to portray the overall hydrochemical type of individual water samples; however, it does not directly provide the relative proportions of the different facies encountered within a dataset [13]. Consequently, it is used in conjunction with the Kurlov formula, in which anions and cations are ranked according to their milliequivalent percentages, and only those ions exceeding 25% of the total anionic or cationic charge participate in the classification [14]. This combined approach enables both the identification of the prevailing hydrochemical facies and the quantification of their areal frequencies.

2.4.2. Calculation of NBLs

Cumulative-probability plots, widely accepted for partitioning hydrochemical datasets into distinct populations [15], allow an intuitive delineation of the NBLs for single or multiple constituents.

Let the cumulative distribution function be F(X), where F(X) = P_r (X ≤ x). For a discrete random variable X, F(x) is expressed as

$$\mathrm{F}(\mathrm{X})={\displaystyle \sum _{\mathrm{k}=1}^{\mathrm{n}}{\mathrm{P}}_{\mathrm{r}}}(\mathrm{X}={\mathrm{x}}_{\mathrm{k}})$$

(1)

in which x_k denotes the largest value in X that does not exceed x, thereby defining the upper bound for cumulative-probability evaluation.

The first inflection point on the cumulative-probability curve separates the first, geogenic population from subsequent populations [16]; the abscissa of this inflection is adopted as the NBLs for the constituent under consideration.

2.4.3. Calculation of TVs

Given that hydrogeological settings and the intensity of anthropogenic pressures vary across regions, the groundwater-quality criteria required for each area also differ accordingly. Consequently, the natural background level (NBL) of each hydrochemical constituent is compared with the Grade-III groundwater quality standard of China or the corresponding guideline value recommended by the World Health Organization, which are adopted as the reference value (REF). The TVs is then calculated using the BRIDGE methodology [17], as expressed by the following equations:

TV = (REF + NBL)/2    If   NBL < REF

(2)

TV = NBL    If    NBL > REF

(3)

Establishing the TVs for groundwater hydrochemistry enables a more accurate assessment of the contamination status. Any constituent whose concentration exceeds its respective TV can be defined as water-quality pollution resulting from anthropogenic activities.

2.4.4. Determination of Pollution Load

The contamination percentage (%) for each well was calculated using the pollution index proposed by Pacheco and van der Weijden [18] and subsequently refined by Soumya et al. [19].

$$\mathrm{P}\mathrm{e}\mathrm{r}\mathrm{c}\mathrm{e}\mathrm{n}\mathrm{t}\mathrm{a}\mathrm{g}\mathrm{e} \mathrm{o}\mathrm{f} \mathrm{P}\mathrm{o}\mathrm{l}\mathrm{l}\mathrm{u}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n} \mathrm{I}\mathrm{n}\mathrm{d}\mathrm{e}\mathrm{x} (\mathrm{P}\mathrm{P}\mathrm{I})={\displaystyle \frac{[\mathrm{C}{\mathrm{l}}^{-}]+[\mathrm{S}{\mathrm{O}}_4^{2-}]+[\mathrm{N}{\mathrm{O}}_3^{-}]}{[\mathrm{C}{\mathrm{l}}^{-}]+[\mathrm{S}{\mathrm{O}}_4^{2-}]+[\mathrm{N}{\mathrm{O}}_3^{-}]+[\mathrm{H}\mathrm{C}{\mathrm{O}}_3^{-}]}}\times 100$$

(4)

In the equation, all ionic concentrations are expressed in meq·L⁻¹. By examining the relationship between the pollution load and the relevant ionic ratios, the provenance of contaminants can be inferred.

2.4.5. Identification of Controlling Factors Using Principal Component Analysis (PCA)

PCA is a multivariate statistical technique that employs an orthogonal linear transformation to reduce the dimensionality of a complex, correlated dataset into a smaller set of uncorrelated principal components. These components capture the dominant structures and patterns within the data and thereby facilitate the identification of the principal factors governing groundwater contamination [20]. In recent years, PCA has been widely applied to the identification of pollution sources in soil and water environmental pollution [21].

3. Results and Discussion

3.1. Hydrochemical Characteristics of Groundwater

Table 1. Statistical table of groundwater chemical indicators.

Parameters	Range (mg/L)	Average (mg/L)	CV (%)	Standard *	Over Standard Rate (%)
pH	6.78–8.47	7.84	5.46	6.5–8.5	0
K+	0.420–15.4	2.21	137	-	3.3
Na+	6.26–1275	285	112	200	38.3
Ca2+	4.01–297	86.7	78.8	-	46.7
Mg2+	2.92–402	74.7	127	-	33.3
$\mathrm{H}\mathrm{C}{\mathrm{O}}_3^{-}$	40.2–1008	405	58.2	-	-
Cl−	15.4–2017	301	132	250	33.3
$\mathrm{S}{\mathrm{O}}_4^{2-}$	20.3–2355	344	149	250	31.7
$\mathrm{N}{\mathrm{O}}_3^{-}$	0–106	16.0	183	88.6	8.3
F−	0.11–3.91	0.84	83.1	1	23.3
TH	24.0–2202	533	97.1	450	40.0
TDS	169–5659	1271	99.6	1000	33.3

Note: * refers to the Class III water quality standard in the “Groundwater Quality Standard” [22]. The dash (-) indicates the absence of a standard. CV = coefficient of variation.

summarizes the descriptive statistics for twelve parameters: pH, K⁺, Na⁺, Ca²⁺, Mg²⁺, $\mathrm{H}\mathrm{C}{\mathrm{O}}_3^{-}$, Cl⁻, $\mathrm{S}{\mathrm{O}}_4^{2-}$, $\mathrm{N}{\mathrm{O}}_3^{-}$, F⁻, TH and TDS. The pH values range from 6.78 to 8.47, with a mean of 7.84, indicating weakly alkaline conditions; none of the samples exceed the permissible limit. The mean TH and TDS are 533 mg/L and 1271 mg/L, respectively. Notably, 40.0% of the samples exceed the standard for TH, and 33.3% exceed the standard for TDS.

Among cations, the mean concentrations decrease in the order Na⁺ (285 mg/L) > Ca²⁺ (86.7 mg/L) > Mg²⁺ (74.7 mg/L) > K⁺ (2.21 mg/L); with exceedance rate for Na⁺ is 38.3%. For anions, the order is $\mathrm{H}\mathrm{C}{\mathrm{O}}_3^{-}$ (405 mg/L) > $\mathrm{S}{\mathrm{O}}_4^{2-}$ (344 mg/L) > Cl⁻ (301 mg/L) > $\mathrm{N}{\mathrm{O}}_3^{-}$ (16.0 mg/L), with exceedance rates of 33.3% for Cl⁻, 31.7% for $\mathrm{S}{\mathrm{O}}_4^{2-}$ and 8.33% for $\mathrm{N}{\mathrm{O}}_3^{-}$. The elevated concentrations and high coefficients of variation for both anions and cations reflect strong anthropogenic impacts [23] and pronounced spatial heterogeneity in geological background and human activities.

3.2. Hydrochemical Facies

A Piper diagram was constructed to classify the hydrochemical types of the study area. The results show that Na⁺ dominates among cations (field D), and $\mathrm{H}\mathrm{C}{\mathrm{O}}_3^{-}$ dominates among anions (field F) (Figure 2).

Application of the Kurlov classification further indicates that the 60 samples represent 20 distinct hydrochemical facies, attesting to the complexity of the groundwater chemistry. The predominant types—HCO₃–Na, SO₄·Cl–Na·Ca, SO₄·Cl–Na, HCO₃·Cl–Na and HCO₃·Cl–Ca·Mg—account for 15.0%, 10.0%, 8.3%, 8.3% and 8.3% of the samples, respectively. The predominance of Na⁺ and $\mathrm{H}\mathrm{C}{\mathrm{O}}_3^{-}$ is generally associated with aquifers rich in carbonate and silicate minerals [24]. Conversely, the occurrence of the SO₄·Cl–Na·Ca facies suggests that the influence of industrial effluents or agricultural activities [25].

3.3. Controlling Factors and Sources of Major Ions

3.3.1. Controlling Factors

The evolution of groundwater chemistry is primarily governed by rock weathering, precipitation, and evaporation processes [26]. Gibbs diagrams were employed to identify the dominant processes in the study area. Figure 3a,b shows that most samples cluster in the rock-weathering and evaporation–concentration fields, indicating that these two factors are the primary controls on regional groundwater chemistry. A subset of samples plots outside the Gibbs envelope, suggesting additional influence from anthropogenic activities such as agriculture or industry [27].

3.3.2. Ion Sources

The ion ratio method is commonly used to identify the sources of ionic components in the aqueous environment. Na-normalized molar ratios are used to constrain the lithological end-members governing groundwater composition [28]. Published values indicate that silicate weathering yields Ca²⁺/Na⁺ and Mg²⁺/Na⁺ molar ratios of 0.35 ± 0.15 and 0.24 ± 0.12, respectively; carbonate weathering yields 50 ± 20 and 20 ± 8; and evaporite dissolution yields 0.17 ± 0.09 and 0.02 ± 0.01 [29]. Figure 4a shows that the majority of samples fall within the silicate and carbonate end-member ranges, with only a few approaching the evaporite field, implying that groundwater chemistry is mainly controlled by silicate and carbonate weathering, whereas evaporite dissolution plays a minor role.

The Na⁺/Cl⁻ molar ratio was calculated to further constrain the origin of Na⁺ and Cl⁻. In Figure 4b, Na⁺ and Cl⁻ exhibit a strong positive correlation (R² = 0.803), and 86.7% of the samples plot above the 1:1 line. These findings suggest a common source for Na⁺ and Cl⁻, most likely halite dissolution, while the excess Na⁺ (points above the 1:1 line) may also originate from silicate weathering or domestic sewage [30].

In natural waters, Ca²⁺ and Mg²⁺ are derived mainly from the dissolution of carbonate rocks, silicate minerals and evaporites, whereas $\mathrm{H}\mathrm{C}{\mathrm{O}}_3^{-}$ predominantly originates from carbonate dissolution and is less affected by anthropogenic inputs [31]. If carbonate dissolution is dominant, the molar ratio of [Ca²⁺ + Mg²⁺]/[$\mathrm{H}\mathrm{C}{\mathrm{O}}_3^{-}$] should approximate 1:1; when evaporites participate, [Ca²⁺]/[$\mathrm{S}{\mathrm{O}}_4^{2-}$] tends toward 1:1 [32]. Figure 4c indicates that 80.0% of the samples plot above the 1:1 line for [Ca²⁺ + Mg²⁺]/[$\mathrm{H}\mathrm{C}{\mathrm{O}}_3^{-}$], implying that silicate weathering also contributes to Ca²⁺ and Mg²⁺ in addition to carbonate dissolution [33]. Figure 4d shows that most samples cluster around the 1:1 line for [Ca²⁺]/[$\mathrm{S}{\mathrm{O}}_4^{2-}$], confirming the involvement of evaporite (e.g., gypsum) dissolution; however, 78.3% of the samples plot below the 1:1 line, suggesting that the excess $\mathrm{S}{\mathrm{O}}_4^{2-}$ is derived from other sources, such as industrial effluent discharge [34].

3.4. NBLs and TVs of Groundwater Chemical Components in Overexploited Areas

3.4.1. Analysis of NBLs and TVs

Cumulative-probability diagrams were constructed for Na⁺, Ca²⁺, Mg²⁺, Cl⁻, $\mathrm{S}{\mathrm{O}}_4^{2-}$, F⁻ and $\mathrm{N}{\mathrm{O}}_3^{-}$ using the regional hydrochemical dataset. As shown in Figure 5, the abscissa of the first inflection point on each curve was adopted as the NBLs. The derived NBLs are 32.3 mg/L for Na⁺, 34.1 mg/L for Ca²⁺, 17.8 mg/L for Mg²⁺, 46.2 mg/L for Cl⁻, 66.4 mg/L for $\mathrm{S}{\mathrm{O}}_4^{2-}$, 0.89 mg/L for $\mathrm{N}{\mathrm{O}}_3^{-}$ and 0.33 mg/L for F⁻. Based on these, NBLs and TVs were subsequently calculated using Equations (2) and (3) in Section 2.4.3.

Table 2. Statistics of natural background values and threshold characteristics of groundwater hydrochemical components.

Parameters	Standard (mg/L)	NBLs	TVs	Exceeding the TVs Points	Percentage Occupied (%)
Na+	200	32.3	116	40	66.7
Ca2+	75	34.1	54.6	33	55.0
Mg2+	50	17.8	33.9	30	50.0
Cl−	450	46.2	248	20	33.3
$\mathrm{S}{\mathrm{O}}_4^{2-}$	450	66.4	258	19	31.7
$\mathrm{N}{\mathrm{O}}_3^{-}$	88.6	0.89	44.7	7	11.7
F−	1	0.33	0.67	23	38.3

indicates that all constituents exhibit concentrations above their respective TVs. Over 50% of sampling sites exceed the TVs for Na⁺, Ca²⁺ and Mg²⁺, demonstrating pronounced anthropogenic impacts on groundwater quality.

The cumulative-probability curves further reveal that Mg²⁺, Cl⁻, $\mathrm{S}{\mathrm{O}}_4^{2-}$ and F⁻ display two inflection points, whereas Na⁺ and Ca²⁺ display three. The varying numbers of inflection points suggest that multiple, distinct anthropogenic sources, including groundwater over-extraction, agricultural fertilization, and industrial or domestic sewage discharge, are superimposed on the natural background [35].

3.4.2. Comparative Analysis of Groundwater Chemical NBLs in Different Regions

The NBLs of Ca²⁺ and Mg²⁺ in the study area (34.1 mg/L and 17.8 mg/L, respectively) are obviously lower than those reported in the Xi’an alluvial plain (76–80 mg/L and 43–184 mg/L) and Slovenia (110 mg/L and 25.9 mg/L) (

Table 3. Statistical Analysis of NBLs in Different Regions (mg/L).

Region	Na+	Ca2+	Mg2+	Cl−	$\mathbf{S}{\mathbf{O}}_4^{2-}$	$\mathbf{N}{\mathbf{O}}_3^{-}$	F−
Study area	32.3	34.1	17.8	46.2	66.4	0.89	0.33
Xi’an alluvial plain [6]	101–121	76.1–80.0	43–184	40.0–73.6	340–696	-	-
Milan area, Italy [2]	25.2	-	-	29.3	44	12	-
Rajasthan, India [1]	158	-	-	242	28	26	0.35
Slovenia [36]	12	110	25.9	30.8	26	6.3	-

). This difference can be attributed to the predominant carbonate rock composition of the unconfined aquifers in the latter two regions, which typically results in elevated concentrations of Ca²⁺ and Mg²⁺. Additionally, the NBLs of $\mathrm{N}{\mathrm{O}}_3^{-}$ in the study area (0.89 mg/L) is substantially lower than that observed in Milan, Italy (12 mg/L) and Rajasthan, India (26 mg/L). This lower NBLs of $\mathrm{N}{\mathrm{O}}_3^{-}$ is likely due to the limited application of fertilizers in the study area compared to the extensive agricultural activities in the latter regions, which have been shown to increase $\mathrm{N}{\mathrm{O}}_3^{-}$ background levels. Furthermore, the NBLs of $\mathrm{S}{\mathrm{O}}_4^{2-}$ in the study area and the Xi’an alluvial plain are higher than those in other regions, likely due to the high degree of industrialization in these areas.

3.5. Contamination Characteristics and Controlling Factors in the Over-Exploited Zone

3.5.1. Contamination Characteristics

The contamination percentage at each sampling site was calculated using Equation (4) and interpreted in conjunction with the Na⁺/Cl⁻ molar ratio to distinguish between chemical weathering and anthropogenic impacts. Sites with contamination percentages < 50% are considered to be controlled primarily by chemical weathering, whereas those >50% are deemed anthropogenically contaminated. Figure 6 shows that 56.7% of the samples exhibit contamination percentages above 50%, signifying substantial anthropogenic influence.

Combined with the evidence presented in Section 3.3.2, it is concluded that groundwater in the study area is severely affected by intensive abstraction and other human activities. Owing to persistent over-extraction, the regional flow field has been altered, lengthening flow paths and residence times and enhancing mineral dissolution. Consequently, carbonates and silicates within the aquifer are extensively leached, leading to concentrations of Na⁺, Ca²⁺, Mg²⁺, Cl⁻and F⁻ that exceed the calculated TVs [37]. As demonstrated in Section 3.3.2, $\mathrm{S}{\mathrm{O}}_4^{2-}$ originates from both evaporite dissolution and industrial effluent discharge; the latter is the principal anthropogenic contributor to $\mathrm{S}{\mathrm{O}}_4^{2-}$ concentrations above the threshold. Furthermore, because abstracted groundwater is predominantly used for agricultural irrigation and domestic supply [38], irrigation return flows enriched in nitrogen fertilizers [39] and direct discharge of domestic sewage are responsible for $\mathrm{N}{\mathrm{O}}_3^{-}$ contamination, with 11.7% of the sampling sites exceeding the $\mathrm{N}{\mathrm{O}}_3^{-}$ threshold.

3.5.2. Controlling Factors of Groundwater Contamination

To elucidate the principal drivers governing groundwater quality deterioration, eleven hydrochemical variables (K⁺, Na⁺, Ca²⁺, Mg²⁺, Cl⁻, SO₄²⁻, $\mathrm{H}\mathrm{C}{\mathrm{O}}_3^{-}$, $\mathrm{N}{\mathrm{O}}_3^{-}$, TDS, TH and pH) were subjected to PCA. Prior to modeling, the Kaiser–Meyer–Olkin (KMO) measure and Bartlett’s test of sphericity were applied to verify the suitability of the dataset. The KMO value was 0.632 and Bartlett’s test value was 1 110 (p < 0.001), confirming the adequacy of the data for PCA. Three principal components with eigenvalues > 1 were extracted, together accounting for 51.8%, 18.9% and 14.6% of the total variance (

Table 4. Loadings of 11 selected variables on VARIMAX rotated factors in this area.

Parameters	PC1	PC2	PC3
$\mathrm{S}{\mathrm{O}}_4^{2-}$	0.941	0.205	0.108
Mg2+	0.912	0.308	0.091
TDS	0.860	0.396	−0.006
Cl−	0.850	0.073	0.496
K+	0.836	−0.306	−0.026
Ca2+	0.828	0.395	−0.165
TH	0.817	0.369	0.261
Na+	0.669	0.291	0.579
$\mathrm{H}\mathrm{C}{\mathrm{O}}_3^{-}$	0.241	0.855	−0.133
pH	−0.223	−0.712	−0.467
$\mathrm{N}{\mathrm{O}}_3^{-}$	−0.165	−0.084	0.897
Eigenvalue	6.65	1.61	1.14
Contribution rate (%)	51.8	18.9	14.6
Cumulative contribution rate (%)	51.8	70.7	85.3

).

Principal component 1 (PC1) exhibits an eigenvalue of 6.65 and explains 51.8% of the variance. It displays strong positive loadings on $\mathrm{S}{\mathrm{O}}_4^{2-}$, Mg²⁺, Cl⁻, TDS, TH, K⁺, Na⁺ and Ca²⁺. As discussed in Section 3.3.2, $\mathrm{S}{\mathrm{O}}_4^{2-}$ is derived primarily from evaporite dissolution and industrial effluents, whereas Na⁺, Ca²⁺, Mg²⁺, Cl⁻, TH and TDS are closely linked to rock weathering and water–rock interaction. Consequently, PC1 is interpreted as the combined influence of water–rock interaction and industrial wastewater discharge.

Principal component 2 (PC2) has an eigenvalue of 1.61 and contributes 18.9% of the variance. It exhibits a strong association with pH and $\mathrm{H}\mathrm{C}{\mathrm{O}}_3^{-}$, and a moderate correlation with Mg²⁺, Ca²⁺, TDS, and TH. Previous studies have shown that carbonate dissolution tends to depress pH [40]. In the present dataset, pH decreases with increasing concentrations of $\mathrm{H}\mathrm{C}{\mathrm{O}}_3^{-}$, Mg²⁺, Ca²⁺, TDS and TH, a pattern attributable to intensive groundwater abstraction. Over the past decades, persistent over-exploitation has created a pronounced cone of depression in the study area [41]. The consequent thickening of the unsaturated zone has shifted the hydrochemical regime toward more oxidizing conditions, promoting the oxidation of sulfide minerals, thereby elevating sulfate and TH concentrations while lowering pH [42]. Thus, PC2 is ascribed to the impacts of severe groundwater over-extraction.

Principal component 3 (PC3) possesses an eigenvalue of 1.14 and explains 14.6% of the variance. It is strongly positively loaded on NO₃⁻ and moderately correlated with Na⁺ and Cl⁻. Nitrate in groundwater typically originates from agricultural fertilizers, domestic sewage, industrial effluents, manure and atmospheric deposition [43]. Given that the majority of samples were collected from farmland and villages, agricultural fertilization and domestic wastewater are identified as the principal sources of nitrate. Moreover, domestic sewage is characterized by elevated Na⁺ and Cl⁻ concentrations. Accordingly, PC3 represents the influence of agricultural fertilization and domestic sewage.

4. Limitations and Future Directions

Although this study determined the NBLs and TVs of hydrochemical constituents in an over-exploited aquifer and identified the dominant contamination processes, several limitations remain. First, the temporal variability of groundwater contamination was not addressed, and a systematic analysis of the dynamic evolution of pollution is lacking. Second, source apportionment of dissolved ions relied solely on hydrochemical approaches, yielding only qualitative insights; the specific contributions of individual sources remain unquantified. Furthermore, while PCA elucidated the controlling factors, the relative importance of each factor was not quantitatively assessed.

In view of the above research deficiencies, future research could be carried out in the following directions: (1) implement multi-temporal, high-density sampling campaigns, employ time-series analyses to capture pollution dynamics, and integrate geostatistical interpolation with GIS to refine spatial sampling networks and delineate contamination patterns at finer scales; (2) combine stable-isotope tracers with multivariate statistical techniques to quantitatively apportion ion provenance and to discriminate the contributions of natural versus anthropogenic sources, thereby providing a dynamic and comprehensive scientific basis for groundwater protection.

5. Conclusions

Based on 60 shallow-groundwater samples and the combined application of hydrochemical and multivariate statistical techniques, this study determined the NBLs and TVs of major chemical constituents in a groundwater over-exploitation zone and identified the principal drivers of contamination. The dominant hydrochemical facies are HCO₃–Na and SO₄·Cl–Na, indicating pronounced anthropogenic influence. Using the established NBLs and TVs, 56.7% of the samples were identified as anthropogenically contaminated, underscoring the necessity of defining NBLs prior to contamination assessment.

The study demonstrates that groundwater over-extraction, industrial activities, water–rock interaction, agricultural fertilization and domestic sewage are the key factors controlling groundwater quality in the over-exploited region. These findings provide a scientific basis for pollution control in such areas. Nevertheless, the temporal dynamics of contamination and the quantitative contributions of individual sources remain inadequately resolved. Future investigations should adopt multi-temporal, high-resolution sampling strategies and integrate isotope tracers to elucidate spatio-temporal pollution evolution and to quantify source-specific contributions, thereby furnishing more comprehensive support for groundwater management and protection.