Introduction

Groundwater plays a vital role in various ecosystems and water cycle. It is an indispensable water body unit on the earth1,2,3,4. According to the 2022 China Water Resources Bulletin, the total water supply and consumption in Hebei Province reached 18.24 billion cubic meters, with groundwater sources accounting for 39.6% of the total. Agricultural water constitutes the primary use of water resources, representing 55.0% of total water consumption5. Groundwater serves as the primary component of the water supply in arid and semi-arid regions of the north, playing a crucial role in ensuring domestic water supply as well as supporting industrial and agricultural production. Zhangjiakou City, often referred to as the northern gateway to Beijing, is situated in the upper reaches of the Guanting Reservoir. It carries the significant responsibility of serving as a water conservation area for the capital, as well as supporting the ecological environment. The dam area, recognized as an ecological protection zone, plays a crucial buffering role against pollution stemming from grassland agriculture, animal husbandry, and urban pollution. Additionally, the primary source of water supply for the region is groundwater4. Therefore, ensuring the safety of groundwater resources in Zhangjiakou is of paramount importance. With the development of the urban economy and the expansion of agricultural scale, the problem of nitrate pollution in groundwater in North China has become increasingly prominent, especially in rural areas where groundwater is used as a drinking water source7.Given the flow and circulation characteristics of groundwater, once polluted, recovery becomes exceedingly challenging, and manual treatment often fails to achieve satisfactory results8. Consequently, analyzing the chemical characteristics and water quality status of groundwater in the urban area of Zhangjiakou has emerged as a critical prerequisite for safeguarding the ecological environment of Zhangjiakou and its neighboring districts, as well as ensuring the safety of drinking water for residents.

Water quality evaluation is a fundamental approach for analyzing water quality. Numerous groundwater evaluation methods, such as single factor evaluation method, the Nemero pollution index method9, the fuzzy comprehensive evaluation method10, grey clustering analysis analysis11, the water quality comprehensive index method, the water pollution index method and the comprehensive water quality identification index method12. Various evaluation methods have been effectively implemented in practice. For example, Su et al.13evaluated the water quality data from three monitoring sections of Lugu Lake in 2019, 2020 and 2010 by single factor index method and traditional Nemero pollution index for water quality

Evaluation. Mohamed et al.14 employed hydrochemical analysis and a fuzzy logic method to evaluate the groundwater quality in the northern plain of Chengdu.Their findings suggest that traditional methods are fully applicable for evaluating of groundwater quality. Gai et al.15 successfully used an evaluation method based on improved grey relational analysis and particle swarm optimization multi-classification support vector machines to assess the water quality of national river sections in 2021, providing a scientific basis for environmental planning. The comprehensive index method and the comprehensive water quality identification index method have also been widely implemented in water quality assessments across North China5,16,17,18. Common analytical methods for pollutants in water include factor analysis20, principal component analysis21, and the positive matrix factorization algorithm (PMF) model and so on. PMF is a recently developed coordination technique22, which can reduce non-negative factor profiles and non-negative factor contributions. It is mainly used to identify air emission sources23 and has also been successfully applied in aquatic environments24,25.

This study analyzes the temporal variations in groundwater hydrochemical characteristics in the drinking water source area of Zhangjiakou City, utilizing monthly variation index data from 2021 to 2023. The research employs the comprehensive water quality identification index method and the fuzzy comprehensive evaluation method to assess the water quality of underground drinking water sources. Additionally, it incorporates the positive matrix factorization (PMF) model to conduct both qualitative and quantitative analyses of potential pollution sources. This integrated approach aims to provide theoretical support for the stability of the groundwater environment in Zhangjiakou.

Materials and methods

Research area overview

The study area is situated in Zhangjiakou City, located in the northwest of Hebei Province, China. This region experiences a temperate continental monsoon climate characterized by warm and rainy summers, and cold and dry winters. The area is minimally influenced by marine water vapor, with the majority of precipitation occurring during the flood season from June to September. Groundwater flow direction in the study area generally mirrors that of the surface water, predominantly flowing from the northeast to the southwest through the Qiaoxi and Qiaodong districts, originating from the southwest of Chongli district. The mineralization degree of the water body is low, indicating relatively fragile water quality, with calcium carbonate (Ca–CO3) as the primary water quality type. The magmatic rocks of the Zhangjiakou Formation are well developed, exhibiting a large distribution area and a complete range of rock types26. In the water mining area, the underground karst type consists of pore water from the alluvial layer in the intermountain basin and fissure water from the karst caves in the karst mountain.

The study area, which serves as the collection zone for drinking water sources in the urban region of Zhangjiakou, is categorized into the ecological conservation area encompassing Beijing, Tianjin, and Hebei. This area functions as an ecological protection barrier. Based on the standards of the water supply plant’s water source collection, the region can be divided into three primary collection zones, as illustrated in Fig. 1. These zones include the Taobeiying Water Source Area (TWS) located near E114.92, N40.89; the Gushi Water Source Area (GWS) near E114.89, N40.87; and the Yaozhanbu Water Source Area (YWS) near E114.85, N40.74, all situated within the Qiaoxi District

Fig. 1
figure 1

The distribution of elevation, sampling points and surface river system in the study area.

Full size image

Data sources and research methods

Data source

The groundwater index data comes from the non-public database of the Zhangjiakou Municipal Environmental Protection Bureau, and the authenticity and reliability of the data are guaranteed.

Comprehensive water quality index method

This method is an improvement and optimization based on the water quality classification of China’s ‘Groundwater Quality Standard ’ (GB / T 14848-2017). It can fully represent the water quality category and water quality target compliance and is an intuitive and clear water quality evaluation method12. The composition structure is as follows:

$${\text{I}}_{\text{wq}}={X}_{1}.{X}_{2}{X}_{3}{X}_{4}$$
(1)

In the equation: Iwq is the comprehensive water quality identification evaluation index, X1 is the groundwater quality level, X2 is the degree of the index between the water quality category and the previous category, X3 is the number of index water quality categories in the detection points that do not meet the functional area target, and X4 is the comparison result of water quality level and water quality functional area target. When the water quality level is better than or equal to the water environment functional area target, it is 0.

Fuzzy comprehensive evaluation method under the principle of weighted evaluation

Fuzzy comprehensive evaluation is a type of fuzzy logic based on fuzzy logic that expresses the influence relationship of multiple boundary-uncertain and quantitative complex factors as fuzzy logic relationships and then conducts a comprehensive evaluation using a fuzzy mathematical model. The calculation steps are as follows:

The groundwater evaluation factors included 22 basic water quality indicators. According to the standard, groundwater is divided into I–V. The membership degree calculation formula for different water quality indicators is as follows:

Type I water ( j = 1 ) :

$${r}_{i1}\left\{\begin{array}{c}1, {m}_{i}\le {l}_{i1}\\ \frac{{l}_{i2}-{m}_{i}}{{l}_{i2}-{l}_{i1}} , {l}_{i1}\le {m}_{i}\le {l}_{i2} \\ 0, {m}_{i}>{l}_{i2}\end{array}\right.$$
(2)

Type II ~ IV water (\(\text{j}\)=2,3,4) :

$${r}_{ij}\left\{\begin{array}{c}0, {m}_{i}<{l}_{i\left(j-1\right)}, {m}_{i}\ge {l}_{i\left(j+1\right)}\\ \frac{{m}_{i}-{l}_{i\left(j-1\right)}}{{l}_{i\text{j}}-{l}_{i\left(j-1\right)}} , {l}_{i\left(j-1\right)}\le {m}_{i}\le {l}_{ij} \\ \frac{{l}_{i\left(j+1\right)}-{m}_{i}}{{l}_{i\left(j+1\right)}-{l}_{ij}} , {l}_{ij}\le {m}_{i}\le {l}_{i\left(j+1\right)} \end{array}\right.$$
(3)

Type V water (\(\text{j}\)=5):

$${r}_{i5}\left\{\begin{array}{c}0, {m}_{i}<{l}_{i4}\\ \frac{{m}_{i}-{l}_{i4}}{{l}_{i5}-{l}_{i4}} , {l}_{i4}\le {m}_{i}\le {l}_{i5} \\ 1, {m}_{i}>{l}_{i5}\end{array}\right.$$
(4)

In the equation: \({m}_{i}\) is the detection value of the i-th evaluation factor, and the unit is mg / L ; \({l}_{ij}\) is the j-level standard value of the i-th evaluation factor, in mg / L ( i × j = 22 × 5) ; \({r}_{ij}\) is the membership degree of the i-th evaluation factor to the j-th level water.

The membership fuzzy relation matrix R of each evaluation factor is obtained as follows:

$$R=\left[\begin{array}{ccc}{r}_{11}& \cdots & {r}_{1j}\\ \vdots & \ddots & \vdots \\ {r}_{i1}& \cdots & {r}_{ij}\end{array}\right]$$
(5)

Establish the weight subset of the index evaluation \(W=\left\{{w}_{1},{w}_{2},\cdots \left.{w}_{i}\right\}\right.\):

$${w}_{i}=\frac{{m}_{i}}{\sum_{j=1}^{j}{l}_{ij}}/\sum_{i=1}^{i}\frac{{m}_{i}}{\sum_{j=1}^{j}{l}_{ij}}$$
(6)

In the equation : \({w}_{i}\) is the evaluation weight of the i-th evaluation factor.

The evaluation matrix (B) is obtained by multiplying the weight matrix and membership fuzzy relation matrix:

$$B=WR$$
(7)

The general common fuzzy evaluation principle is the maximum membership degree evaluation principle, that is, the water quality grade corresponding to the maximum value is selected as the water quality category of the evaluation object in the fuzzy subset, and the judgment of the water quality category lacks continuity. In some cases, failure problems may occur27. Therefore, this study uses the weighted rank principle to evaluate:

$${B}_{WA}=\sum_{j=1}^{j}j{b}_{j}/\sum_{j=1}^{j}{b}_{j}$$
(8)

In the equation, BWA is a fuzzy comprehensive evaluation value, and bj is the membership degree of the evaluation point to the j-level standard. The water quality evaluation classification is presented in Table 1.

Table 1 Classification of water quality evaluation results.
Full size table

Positive matrix factorization algorithm (PMF)

PMF is a multi-factor source analysis model first proposed by Paatero et al. in 1994 and approved and improved by the United States Environmental Protection Agency ( EPA)28. The mathematical principle of the model is as follows:

$${X}_{ij}=\sum_{k=1}^{P}{G}_{ik}\times {F}_{ki}+{E}_{ij}$$
(9)

In the equation: X is the sample concentration matrix, G is the factor contribution matrix, F is the factor concentration matrix, E is the residual matrix, P is the number of factors, i is the number of samples, j is the number of chemical substances. The operation of the PMF model requires groundwater pollution concentration index and uncertainty as input files, and EPA PMF5.0 software was used to calculate the pollution source and its contribution rate.

After testing, when six factors were set, the convergence result of 20 basic runs was ‘yes’, and Qtrue / Qrobust < 1.5. At this point, the model fits the input data and meets the analysis requirements29. The Q-values are listed in Table 2.

Table 2 Results of 20 basic run goodness-of-fit parameters ( Q ).
Full size table

Results

Present situation of groundwater source water quality

Table 3 shows that the pH values of GWS, TWS, and YWS ranged from 7.5 to 8.5, 7.5 to 8.4, and 7.4 to 8.2, respectively. The total dissolved solids (TDS) content was 340–724 mg/L, 364–672 mg/L and 380–693 mg/L, respectively. The oxygen consumption levels measured by the potassium permanganate index (CODMn) were 0.43–1.38, 0.56–1.34, and 0.2–1.62 mg/L, respectively. The concentration ranges of nitrate (NO3-) were 1.22–15.7 mg/L, 2.21–17.8 mg/L and 6.49–12.2 mg/L, respectively, and the average values were 10.97 mg/L, 12.13 mg/L and 7.55 mg/L,respectively. The differences in pH, TDS, and CODMn content in the three water sources were very small, indicating that the local groundwater was almost maintained in a water environment with alkaline, low salinity, and almost no organic pollution, and the water quality was in good condition, which is in line with Chen30. The analysis of the chemical types of groundwater near Gushi and Taobeiying villages showed that the flow direction of groundwater and surface water was consistent and that the underground runoff conditions were good. The hydrochemical cycle is strong and exhibits a certain cleaning ability. There were differences in the content of chloride (\({\text{Cl}}^{-}\)), fluoride (\({\text{F}}^{-}\)), sodium ion (Na+), sulfate (\({\text{SO}}_{4}^{2-}\)), ammonia nitrogen (\({\text{NH}}_{4}^{+}\)), nitrite (\({\text{NO}}_{2}^{-}\)) and other indicators in the groundwater of the water extraction area, but the concentration was small, which did not exceed the Class III water quality of the Sanitary Standard for Drinking Water (GB 5749-2022) and the Groundwater Quality Standard (GB / T 14,848-2017), while the concentration of nitrate (\({\text{NO}}_{3}^{-}\)) met the Class III standard of groundwater. However, in GWS and TWS, there was a long-term exceedance of drinking water limits.

Table 3 Statistics of representative chemical parameters of groundwater in water source area from 2021 to 2023.
Full size table

According to the box line-normal distribution Fig. 2, it can be seen that the nitrate index of TWS is more dispersed than that of other water sources, and the concentration fluctuates greatly. The nitrate detection data of YWS are relatively concentrated, and the monthly concentration changes little and the deviation tendency concentration is low. The potassium permanganate index reflects the pollution of organic and inorganic oxides in groundwater. The concentration of the water source index was low and was less affected by organic matter. The nitrate concentration is expressed by the line chart in the form of monthly variation: on the whole, the groundwater situation in the upstream water supply collection area is inferior to that in the downstream of the urban area; the nitrate concentration in GWS and TWS gradually decreased in 2021 and then increased in 2022. The nitrate concentration in the two places remained beyond the limit of drinking water health standards, except that the nitrate concentration in October of GWS in 2021 and March of TWS in 2022 suddenly decreased to 1.22 mg/L and 2.21 mg/L, respectively. In recent years, the nitrate concentration of YWS has been relatively stable, fluctuating between 6.49 and 8.32 mg/L. Only in March and April 2022 and November 2023, the nitrate concentration exceeded the standard. The concentration of nitrate in groundwater is increasing annually in GWS and TWS, and the problem of drinking water pollution is increasing annually, but there is no significant law in seasonal variation. The nitrate concentration level of YWS was low and stable, and the three times of excessive concentration may be due to sudden environmental pollution.

Fig. 2
figure 2

Nitrate index and oxygen consumption index box line-normal distribution diagram (a) and nitrate concentration time change line chart (b) in water source collection area.

Full size image

Groundwater quality assessment

According to the rainfall situation, the monthly test results were divided into flood (June to September) and non-flood (January to May, October to December) season index datasets. The 22 indicators that will be evaluated for water quality are counted according to groundwater standards. As shown in Figs. 3 and 4, all the indices are in class III and above the groundwater quality standard, which meets the requirements of the groundwater quality target in this area. Most of the indices are in classes I and II, and only a few indices, such as arsenic (As), total radioactive α, and organic carbides, are in class III. Almost all the detection values of \({\text{NO}}_{3}^{-}\) and total hardness (calculated by CaCO3) are in class III. By comparing the results of index categories in the flood and non-flood seasons, it can be seen that the results of index categories in the non-flood season are relatively good, and the water quality has improved. However, compared with the flood season, this change was not significant. Even in the non-flood season, the ammonia nitrogen index in some months increases to Class III, and \({\text{NO}}_{3}^{-}\) decreases to Class II in a short time. With changes in rainfall, the transformation of ammonia and nitrate nitrogen occurs in groundwater quality.

Fig. 3
figure 3

Index classification statistics undergroundwater standard in flood season.

Full size image
Fig. 4
figure 4

Index classification statistics undergroundwater standard in non-flood season.

Full size image

According to the comprehensive water quality identification index method (Figure 5), the Iwq of different water sources in 2021–2023 was in grade I or II, and the groundwater quality was excellent, even in 2022, the overall water quality was better. In 2021, the Iwq results for TWS and YWS were more stable, but there was a sudden increase in groundwater quality in December of that year and March of the following year. In 2022, the evaluation results of groundwater source water quality were all in the range of grade I, but the evaluation value changed greatly, and the Iwq results showed a downward bias in the second half of the year. The evaluation results for different water sources had different time biases in 2023. The water quality of GWS was better in cold season, and the Iwq value of TWS increases in spring and autumn. The evaluation results of YWS are that the water quality in spring and winter is better than that in summer and autumn.

Fig. 5
figure 5

The cobweb diagram of the comprehensive water quality identification index method in different years (a) 2021 (b) 2022 (c) 2023.

Full size image

Following a fuzzy comprehensive evaluation, the results are illustrated in Figure 6. From 2021 to 2023, the groundwater evaluation results for GWS, TWS, and YWS consistently fall within Grade I and Grade II categories. However, the annual variations depicted in the cobweb diagram are still pronounced. In 2022, the water quality evaluation score was lower, indicating relatively better water quality, while 2023 showed a slight decline. Notably, the groundwater quality was at its lowest in 2021. The changes and discrepancies in the water quality evaluation results align closely with the findings derived from the comprehensive water quality identification index method.

Fig. 6
figure 6

The cobweb diagram of fuzzy comprehensive evaluation method in different years (a) 2021 (b) 2022 (c) 2023.

Full size image

According to the water quality evaluation results classified by the Groundwater Quality Standard (GB / T 14848-2017), it is evident that the overall groundwater quality in 2022 was better, while the water quality in 2021 remained stable. In 2023, there was a seasonal bias among different water sources, with evaluation results indicated better water quality during the cooler seasons. In contrast, summer and autumn, characterized by increased agricultural activities, higher rainfall, and elevated temperatures, showed a negative trend in water quality evaluation results. These two evaluations effectively illustrate the current state of local groundwater quality. However, the weighted fuzzy comprehensive evaluation method can more distinctly highlight the overall changes in water quality across different years and months. The water quality assessment meets the groundwater Class III standard for the local functional area target. Nevertheless, due to the use of groundwater as one of the urban drinking water supply sources in Zhangjiakou City, the nitrate index exceeds the drinking water standard. After adjusting the target of the water functional area to meet the drinking water standard, the results of the comprehensive water quality identification index method did not change significantly, with variations limited to three or four decimal points, which minimally impacted the evaluation results. Relatively speaking, the weighted fuzzy comprehensive evaluation method provides a more accurate reflection of water quality. Due to the decentralization of evaluation impacts by other indicators, the standard range has been redefined. This redefinition reduces the fluctuation of evaluation results, leading to more objective water quality assessments.

Traceability of groundwater pollutants based on PMF operation

According to the requirements of the parameter unit, the pH, the number of coliform bacteria, and the total bacterial count were excluded from all indexes. Additionally, certain interference samples and those with a poor signal-to-noise ratio were excluded. Consequently, pollution data that exhibited a high degree of reflection were selected. The monitoring data from different water sources were run 20 times in random mode, demonstrating a remarkable fitting effect of the matrix, from which the optimal solution was chosen. GWS, TWS, and YWS all identified six concentration source factors, as illustrated in Figure 7.

Fig. 7
figure 7

Based on the PMF model, the factor contribution ratio (a) GWS (b) TWS (c) YWS is calculated.

Full size image

Analysis of groundwater material composition based on PMF operation in water source area

GWS factor 1 was identified as having an average contribution rate of 14.68% from livestock and poultry breeding activities, with Copper (Cu) and CODMn demonstrating particularly high contribution rates of 42.01% and 38.46%, respectively. Cu is classified as an artificial metal, closely associated with human activities. Previous studies indicate that to mitigate livestock and poultry diseases, heavy metals such as Cu, Zn, and Pb are commonly added to animal feed. Notably, the addition of CuSO4 to piglet feed can reach levels as high as 100–250 mg/kg, while concentrated feed may contain Cu concentrations ranging from 1000 to 1500 mg/kg. A significant portion of these heavy metal elements present in feed is subsequently released into the natural environment through livestock and poultry manure31,32,33,34. Concurrently, the organic matter and partially reduced inorganic matter found in feces contribute to increased the CODMn level. \({\text{SO}}_{4}^{2-}\), \({\text{F}}^{-}\), and \({\text{Cl}}^{-}\) were identified as significant contributors to factor 2, which has an average contribution rate of 16.51%. These substances contributed 42.57%, 39.43% and 38.48% to factor 2. The high salinity and elevated sulfate ion characteristics of these substances were identified as indicative of human domestic sewage35. Factor 3 was interpreted as pesticide application, and its representative indicators were chloroform (CHCl3, 56.42%), carbon tetrachloride (CCl4,58.69%), benzene (C6H6, 56.41%), and toluene (C7H8, 65.57%). This interpretation arises from the extensive coverage of economic crops by the water source. To prevent pests and promote healthy plant growth, pesticides are applied during the growth stages of crops. Common pesticides types include organochlorine and organophosphorus pesticides36. The \({\text{SO}}_{4}^{2-}\), \({\text{NO}}_{3}^{-}\), and total hardness (calculated by CaCO3) of the factor 4 contribute more to the output. According to field research, the main local cash crops are pepper and corn. During crop planting and management, the land undergeos turning and fertilization. The application of agricultural fertilizers such as calcium, compound fertilizers, potassium sulfate, zinc sulfate, and urea, leads to increased concentrations of \({\text{SO}}_{4}^{2-}\) and \({\text{NO}}_{3}^{-}\) in the soil. Consequently, factor 4 was identified as the loss of agricultural fertilizers in the soil. Factor 5 was interpreted as physical effects such as leaching and weathering from underground rock minerals, with an average contribution rate of 15.78%. Arsenic (As) is its representative source. Factor 6 was explained as the average contribution rate of 15.04% of soil leaching, 34.57% of CODMn, a small part of CHCl3, C6H6, C7H8 contributed to factor 6. Pesticide residues in the environment are difficult to decompose and accumulate in the soil. They are mixed with domestic sewage produced by human activities with surface water and shallow groundwater to form organic matter that is prone to reduction reactions in the anaerobic environment37.

TWS exhibits similarities to GWS, with comparable results from the factor analysis. The average contribution rate of concentration factor 1 was 17.60%, indicating that the dissolution of arsenic-rich rocks is the primary contributor to this source factor’s composition. Factor 2 identified radioactive α and β as the main contributing indicators, which may originated from wastewater discharged during the combustion of chemical fuels in industrial facilities or from geothermal energy produced by underground volcanic activities. When combined with the analysis of GWS and the distribution of local industrial enterprises, factor 2 was attributed to industrial wastewater. Factor 3 accounted for a contribution rate of 13.83% due to the loss of agricultural fertilizers, with representative indicators including Cu (24.43%), CaCO3 (21.24%), and SO42 (21.91%). The average contribution rate for factor 4 was 25.35%, which was explained by the representative indices of CHCl3, CCl4, C6H6, and C7H8, indicating pesticide application. Factor 5 reflected an average contribution rate of 16.11% from livestock and poultry breeding activities, with Cu and CODMn showing higher contribution rates. Finally, factor 6, with Cl− contributing 49.56% of its concentration, indicates that the contributions from CaCO3 and Na+ were negligible, thus identifying factor 6 as originating from rural domestic sewage.

The high contribution substances of YWS concentration factor 1 were \({\text{NH}}_{4}^{+}\) (32.44%), \({\text{NO}}_{2}^{-}\) (30.93%), \({\text{NO}}_{3}^{-}\) (29.95%), CODMn (28.13%). According to the high contribution rate of nitrogen, especially \({\text{NH}}_{4}^{+}\), Factor 1 can be identified as the increase in soil nitrogen resulting from human fecal pollution and the accumulation of livestock and poultry manure, which contributes to the rising nitrogen concentration in groundwater. Factor 2 was attributed to the discharge of industrial wastewater containing heavy metals, which has an average contribution rate of 14.80%. Factor 2 was mainly affected by the single element Cu (40.70%). Factor 3 is also affected by the mixed contribution of CaCO3 (30.31%), CHCl3 (27.53%), Na+ (27.50%), CCl4 (27.19%) and C6H6 (27.33%). CaCO3 and Na+ are one of the seven common groundwater ions, and the local groundwater type is HCO3-Ca. The simultaneous presence of alkanes and benzene ring organics are from the accumulation of agricultural pesticide residues on the ground into the groundwater. The integration of the two sources resulted in the concentration effect of factor 2. The average contribution rate of factor 4 was 12.67%. The presence of the single element manganese (Mn) at 65.02% leads to a significant abundance of iron and manganese minerals in Xuanhua County, particularly near the water source26. The analysis of factor 4 focused on the diffusion of manganese ore from point source pollution after leaching into groundwater. Factors 5 and 1 exhibit higher contributions to NH4+, NO2-, and CODMn simultaneously. However, unlike Factor 1, Factor 5 shows a higher contribution of Cl⁻ and As, which is similar to the composition found in municipal sewage38. The average contribution rate of Factor 6 was 26.65%, with four alkyl and benzene ring organics being the primary contributors, interpreted as indicative of pesticide use.

Source of groundwater material composition

All results are summarized in Table 4. The groundwater quality of the three water sources in the urban area of Zhangjiakou was predominantly influenced by pesticide usage. Pesticides exhibit stability in soil environments, demonstrat low solubility in water, possess extended residual times, and are classified as persistent organic pollutants, leading to their accumulation in the environment36. Additionally, the application of agricultural fertilizers and the large-scale breeding of livestock and poultry imposed a significant burden on the self-purification capacity of the water environment. Furthermore, the impact of water cycling from primary strata on groundwater composition should not be overlooked. Although the results of factor analysis vary, they were generally consistent with the local economic structure and urban functional characteristics. It is noteworthy that, except for nitrate, the indicators assessed under drinking water standards did not exceed permissible limits. The concentration factor analysis results can only be considered potential pollution sources, and further investigation is required to understand the causes of elevated nitrate levels.

Table 4 The material source and factor contribution rate of different water sources based on PMF operation.
Full size table

Most studies have demonstrated that synthetic fertilizers, livestock and poultry manure, domestic and industrial sewage, atmospheric deposition, natural organic matter in soil, septic tanks, and landfills are the primary sources of nitrate pollution39,40,41. According to research conducted by Yang et al.42, the nitrate pollutants in the Xuanhua area, located downstream of the YWS, predominantly originate from soil nitrogen, feces, and sewage. These pollutants are significantly influenced by domestic sewage, industrial production activities, and agricultural irrigation in the region, with pollution from construction land being more severe than that from cultivated land. As a typical city centered on a primary industrial economy, Zhangjiakou experiences frequent agricultural activities in the study area. The water supply area of the North Water Supply Plant is extensively covered with economic crop greenhouses, and agricultural activities have a substantial impact on groundwater quality.

Discussion

The seasonal variation in water quality assessment results and the corresponding potential pollution factors of water sources reveal the influence mechanisms of multiple factors. Agricultural activities and meteorological conditions jointly affect the groundwater environment. During the summer and autumn, when agricultural activities were frequent, the excessive application of pesticides and fertilizers leaded to adverse changes in groundwater quality in the study area. Meanwhile, the overall quality of groundwater during the flood season declined, indicating that the leaching effect of precipitation on pollutants significantly outweighed the dilution effect brought about by the increase in water volume. However, the changes in nitrate nitrogen indicators were contrary to the overall situation, which might have been related to precipitation promoting the nitrogen cycling process. Furthermore, the issue of mixed flow of rainwater and sewage in aging urban pipeline systems was particularly prominent during heavy rainfall, where the overload of pipelines could significantly increase the risk of groundwater pollution.

The pollution source composition pattern of groundwater drinking water sources in Zhangjiakou was predominantly influenced by pesticides, followed by rural domestic wastewater, livestock and poultry farming, industrial wastewater, and municipal wastewater. This pattern is widely representative in northern China. For instance, multiple studies conducted in the nearby North China Plain have demonstrated that the detection rate of pesticide residues in groundwater within agricultural areas of this plain reaches as high as 78%, with pesticide pollution contributing 38–53% to overall groundwater pollution, thereby establishing it as the primary pollution source42,43. The impact of industrial and municipal wastewater is relatively minor, primarily concentrated in areas downstream of cities and around industrial zones44. The groundwater pollution characteristics of the North China Plain closely resemble those of Zhangjiakou. In the Northeast region, pesticide pollution remains the primary factor; however, due to a higher degree of livestock farming industrialization in the Northeast45, along with a cold climate, lower population density, and historical legacy issues stemming from old industrial bases, the proportional structure of agricultural and industrial pollution differs slightly from that of the North China region44. Furthermore, groundwater pollution patterns in the arid regions of Northwest China exhibit unique characteristics, with geological background pollution being the most prominent issue. For example, in areas such as the Hetao Basin in Inner Mongolia and the Junggar Basin in Xinjiang, severe natural high fluoride and high arsenic levels in groundwater pose significant concerns46.

A comparative analysis of groundwater pollution characteristics between Zhangjiakou City, the North China Plain, the Northeast Region, and the Northwest Arid Region reveals that agricultural intensification and pesticide pollution are prevalent issues and primary contributors to groundwater pollution in northern China. However, the contribution ratios of livestock manure, rural wastewater, and industrial and municipal wastewater to groundwater quality differ across spatial regions. Therefore, when formulating regional groundwater pollution prevention and control strategies, the primary focus should be on agricultural non-point source pollution as the main challenge. Differentiated prevention and control measures should be developed based on regional characteristics to achieve precise and scientifically informed pollution control.

Based on the aforementioned analysis, the specific measures proposed include:

First, reducing the use of chemical pesticides and promoting water-saving, pesticide-reducing, and efficiency-enhancing cultivation techniques. This can be achieved by utilizing disease-resistant varieties, implementing crop rotation, and employing biological pesticides and physical control methods to minimize reliance on chemical pesticides. Additionally, aligning crop water requirements with efficient water-saving methods, such as drip irrigation and sprinkler irrigation, will facilitate the precise supply of water and nutrients47. To mitigate nutrient loss, it is recommended to replace chemical fertilizers with organic fertilizers and utilize slow-release fertilizers.

Additionally, attention should be directed towards the centralized treatment and resource utilization of livestock manure and rural sewage. It is essential to establish biogas and organic fertilizer production facilities in livestock breeding areas. Furthermore, promoting low-cost, easy-to-maintain decentralized sewage treatment facilities, such as constructed wetlands and biological oxidation ponds in villages located near water sources.

For YWS, which is impacted by urban mixed pollution, it is crucial to conduct clean production audits of industrial enterprises and investigate and regulate sources of industrial pollution. There is an urgent need to accelerate the improvement of the urban pipeline system and implement leakage prevention measures. Completing the rainwater and sewage separation renovation project promptly is vital to reduce sewage overflow during the rainy season. Advanced technologies, such as pipeline endoscopes and acoustic detection, should be employed to regularly inspect and swiftly repair points of leakage in the pipeline.

Moreover, it is imperative to establish a layered monitoring well network system within the water source protection zone of Zhangjiakou City and its upstream recharge area. By deploying multiple monitoring points, the vertical migration patterns of pollutants can be systematically tracked. Based on continuous multi-year groundwater quality dynamic monitoring data, a regional groundwater quality prediction model should be developed to inform scientific management and sustainable use of water resources in the Zhangjiakou region.

Conclusions

The purpose of this study is to explore the characteristics of drinking water quality and the sources of pollution in Zhangjiakou, using the concentration of groundwater index detection as the starting point. The following conclusions are drawn:

  • The nitrate index in the water source collection area of the upper reaches of the urban area significantly exceeded the drinking water standard. Since 2022, the concentration of nitrate has fluctuated and increased. In contrast, the groundwater quality in the lower reaches of the urban area was superior to that in the upper reaches, with a more stable nitrate concentration level, making it a more reliable source for drinking water.

  • All groundwater quality indexes met Class III standards or higher, showing minimal variation between flood and non-flood seasons. The evaluation results obtained from the comprehensive water quality identification index method and the fuzzy comprehensive evaluation method were consistent. Overall, the groundwater quality assessment indicated grades I or II, signifying excellent water quality. The evaluation results were most favorable in 2022, while water quality remained relatively stable in 2021. In 2023, groundwater quality from various sources showed improvement under conditions of lower temperatures and reduced rainfall, the evaluation results declined under high temperatures and increased rainfall.

  • The detection results of various water sources were analyzed using the PMF model to identify the concentration source factors. The composition of GWS groundwater material sources were determined as follows: pesticide use (19.25%) > fertilizer loss (18.73%) > domestic sewage (16.51%) > geological action (15.78%) > soil leaching (15.04%) > livestock and poultry breeding (14.68%). The composition of TWS groundwater material sources were: pesticide use (25.35%) > geological action (17.60%) > domestic sewage (16.47%) > livestock and poultry breeding (16.11%) > agricultural fertilizer loss (13.83%) > industrial wastewater (10.65%). The sources of groundwater material composition concentration in YWS were: pesticide use (26.65%) > mixed sources (18.21%) > industrial wastewater (14.80%) > municipal sewage (13.84%) > fecal pollution (13.82%) > manganese ore leaching (12.67%). Frequent agricultural activities and the chemical properties of pesticides positioned them as the largest potential source of drinking water pollution in the Zhangjiakou urban area. In addition to the direct manifestation of nitrate pollution, the control and selection of pesticides also warranted significant attention.