Identification of the hydrochemical characteristics of the groundwater system in the mining area and the controlling factors of the water's chemical components is necessary to protect groundwater resources. In this study, 80 sets of groundwater samples were collected from three aquifers of the Liuzhuang coal mine (northern Anhui Province, China), and a total of eight indicators were selected for quantitative analysis of the chemical components of water. Conventional mathematical and statistical methods and Piper trilinear diagrams show that the cations in the groundwater samples of the mine area are mainly K+ + Na+ (92.4%), while the anions in the Cenozoic and Carboniferous aquifers are mainly Cl, reaching 57.2% and 55.2%, respectively, and the anions in the Permian aquifer are mainly HCO3- (52.6%). Most of the water chemistry types are Cl-Na, HCO3-Na, and HCO3-Cl-Na. Analysis on the basis of Gibbs plots showed that the aquifer system in the mine area is primarily controlled by the water–rock interaction. The results of ion ratio analysis, principal component analysis, and cluster analysis showed that the dissolution of hydrochloric acid and alternating cation adsorption is more prominent in the Cenozoic and Carboniferous aquifers, while desulfurization is more significant in the Permian aquifer.

  • This paper is the first to analyze the hydrochemical characteristics and controlling factors of three aquifers in the Liuzhuang coal mine (Northern Anhui Province, China).

  • The study determines the major cations and anions as well as the types of water chemistry in each aquifer in the study area.

  • The study shows that the aquifer system in the study area is mainly controlled by water–rock action.

  • Correlation analysis, ion ratio analysis, principal component analysis and cluster analysis were used to show the hydrochemistry of the three aquifers in the study area and to reasonably infer their genesis.

Groundwater is an important component of water resources, and groundwater resources in China account for approximately one-third of the total water resources, whereas approximately 40% of groundwater in China is affected by coal mining activities (Cai et al. 2017; Liang et al. 2018). In recent years, water scarcity and pollution have been increasing, and the utilization rate of mine water is less than one-third; thus, it is practical to utilize groundwater resources (Liu et al. 2022). With increasing mining efforts, the contradiction between coal mining and groundwater resource management has become increasingly prominent (Xiao et al. 2018; Sun et al. 2021). Coal has been the basic source of energy in China (Shahbaz et al. 2015; Lu et al. 2022). However, coal mining activities have destroyed groundwater resources (Sun et al. 2021). Therefore, it is deemed important to determine methods to protect groundwater resources.

Various experts and scholars have conducted in-depth studies on water geochemistry, and the Piper trilinear diagram is the most widely used method for the graphical representation of water chemistry, reflecting water chemistry based on the distribution of major ions in the diagram. Adimalla (2019) combined the Piper trilinear diagram with Geographic Information System technology and found that groundwater in the rapidly urbanizing areas of Telangana, southern India, is heavily contaminated with nitrate, and Ahmad & Mazhar (2020) found that groundwater in parts of the Ramganga basin could be used for irrigation by various methods such as the Piper trilinear map, Gibbs plot, and correlation analysis. Sethy et al. (2016) applied ion ratio analysis to the southern Ganges plain to determine the suitability of groundwater for irrigation in the study area and the main controlling factors that control changes in water chemistry characteristics. Discriminant analysis establishes a discriminant function based on certain discriminant criteria, determines the coefficients to be determined in the discriminant function using the water chemistry information of the study area, and calculates the discriminant index, which can be divided into the distance discriminant method, the Fisher discriminant method (Fan et al. 2018; Tan et al. 2023), the Bayes discriminant method (Fritz et al. 2016; Yan et al. 2020; Fasaee et al. 2021), and so on. The organic combination of isotope and water chemistry methods can quantitatively or qualitatively evaluate groundwater flow and transport patterns, providing a powerful means to finely characterize groundwater genesis, recharge sources, and hydraulic connections (Zhang et al. 2009; Vreča & Kern 2020). Du et al. (2019) used stable isotopes to analyze the seasonal and spatial variability of tap water in urban Lanzhou City. In addition, Qiu et al. (2016) found that heavy isotope enrichment in soil moisture is associated with evapotranspiration in the high northwestern Tibetan Plateau, and there is a consistency between river water and groundwater recharge sources. However, Bowling et al. (2017) argued that isotope measurements are inadequate for plant water sources and that integration with other geochemical and biomechanical methods is essential. Poetra et al. (2020) coupled geomorphology with PHREEQC to invert the water chemistry of the study area, revealing that ion exchange, dissolution, and rainfall are the main hydrogeochemical processes. The study of water resources has gradually evolved into a combination of multiple analytical methods, providing new ideas for solving geochemical problems.

The Liuzhuang coal mine is an important producing mine in northern Anhui Province, where decades of coal mining activities have triggered a series of groundwater environmental problems. Thus, in this paper, the hydrogeochemical characteristics of the Cenozoic (Q), Permian (P), and Carboniferous (C) aquifers in the Liuzhuang coal mine were investigated by using Piper trilinear diagram analysis, correlation analysis, Gibbs plot analysis, ion ratio analysis, principal component analysis, and cluster analysis to analyze the main ion–chemical components of the aquifers and to explore the controlling factors of the hydrogeochemistry of the groundwater and to provide a reference for the mining area to solve the environmental problems of groundwater and the green and safe production of mines.

The Liuzhuang coal mine is located approximately 20 km north of Yingshang County (Fuyang City, Anhui Province), the southern flank of Chenqiao backslope (Figure 1). It is 16 km long in the east–west direction and 3.5–8 km wide in the north–south direction, covering an area of 82.21 km2. The mine area borders the Xieqiao mine from the F5 fault in the east and the Kouzidong mine from the F12 fault in the west. Based on the distribution of tectonic patterns and faults, it can be divided into the following zones: east, central, and west, and the overlapping tectonic zone of the F1 fault group. The historiography of the east zone is monoclinic development; the central zone is in the west-turning part of Xieqiao oblique, faults are more developed, and the west zone and the overlapping body of the broken sandwich blocks the coal stratigraphy complex. The Huaihe River alluvial plain is a largely flat landform and has few ditches and ponds, with the Ji River flowing southeastward, and a transitional climate between the northern temperate zone and subtropics, with evident seasonality.
Figure 1

Overview map of the study area: (a) land use coverage map of Fuyang City; (b) geographical topographic map of Yingshang County; (c) geographical topographic map of Anhui Province; and (d) geological structure sketch of Liuzhuang coal mine.

Figure 1

Overview map of the study area: (a) land use coverage map of Fuyang City; (b) geographical topographic map of Yingshang County; (c) geographical topographic map of Anhui Province; and (d) geological structure sketch of Liuzhuang coal mine.

Close modal

The Liuzhuang coal mine is a fully concealed well-field covered by thick loose layers, primarily mining Permian coal seams. It is distributed in the northern flank of the Xieqiao diagonal and the west-turning part, and all concealed under the huge thick loose layers, overlying the Taiyuan Formation and Ordovician.

The lithology of the Eocene Quaternary is mainly medium and fine sand and silt with sandy clay. By the end of the Paleocene, the southern part of the area is low hills, and it oversteps the pre-mountain plain to the north and gradually thickens from southeast to northwest. By the late Pliocene, the terrain of this area tended to be flat after the deposition of stripping and razing and gully depressions. The bottom gravel layer, commonly known as the ‘red layer’, consists of grayish white, brownish red, reddish red, and purplish red rocks of different sizes and gravels; gravelly sand; and gravelly clay, occasionally interspersed with thin layers of consolidated clay. The groundwater in the first aquifer of the Cenozoic loose layer is recharged by atmospheric precipitation and infiltration of surface water, and the thickness and distribution of the soil layer in the lower part of the third water barrier are, respectively, large and stable, which is the main water barrier in this area.

The thickness of the Permian System ranges from 749.32 to 910.18 m, with an average of 829.77 m. The bottom is bounded by the top of the first tuff of the Taiyuan Formation, which is a consolidated contact. In addition, the depositional environment of the Permian System is the lower deltaic plain deposition developed from the land surface sea bay, which experienced the bay-filling, dendritic, and reticulated river system, transferred to the estuarine bay environment, and then developed to the upper deltaic plain and land-phase alluvial plain deposition. The main lithologies are sandstone, siltstone, sandy mudstone, mudstone, coal lineage, and so on, which form several sedimentary complexes. At the bottom, the top interface of Taiyuan Group 1 ash is in contact with the underlying Taiyuan Group tuffs. However, Permian coal sandstone aquifer recharge conditions are poor, water-rich and heterogeneous, and locally strong, with a poor hydraulic connection between the aquifers and generally no hydraulic connection between them.

The tuff of the Taiyuan Group is 96.35–118.75 m thick, with an average of 105.58 m. This group is a set of sea–land-interfacing deposits developed between thin-bedded tuff and siltstone, fine sandstone, mudstone, and coal line. Moreover, it is dominated by bioclastic tuffs, cryptocrystalline to fine-crystalline tuffs, and white dolomitic tuffs. The tuffs are rich in fossils such as sea lily stems, brachiopods, and corals, particularly the fossils of flies (10th–12th layers of limestone) are very abundant. The water richness is relatively good near the backslope axis, the big fault, and the outcrop area. The farther the coal strata are from the tuff outcrop area, and the depth of burial increases, so the water richness decreases.

The lower middle Ordovician is the base of the coal-bearing construction of Carboniferous and Permian aquifers, and the thickness of the area is approximately 250 m. Nevertheless, the development of fissures, soles, and cavities is uneven, and the overall water richness is weak. The lithology is primarily gray-white to flesh-red, crimson thick-layered cryptocrystalline and fine-crystalline tuffs, dolomitic tuffs, locally brecciated tuffs, interspersed with light gray-green and light gray muddy bands or thin sandy mudstones, with solution holes and fissures filled with mud and calcite veins; calcite and pyrite clusters can be seen when semi-filled, and the underlying strata are in pseudo-integrated contact.

Sampling and testing

The water samples were collected including downhole discharge and outlet points. The pre-cleaned sample bottles were first rinsed three times with water samples and then filled with samples, sealed, and brought back to the laboratory for storage at a low temperature (0–4 °C) until analysis. In this study, 80 groups were randomly selected from the collected water samples, including 35 groups of Cenozoic aquifers, 35 groups of Permian aquifers, and ten groups of Carboniferous aquifers. Then, samples were processed by skimmed cotton filters and tested. The test items include K+, Na+, Ca2+, Mg2+, Cl, , , total dissolved solids (TDS), pH, and so on. The tests all used superior pure reagents, and the test vessels were treated with deionized water for rinsing and drying. Among them, the K+, Na+ test method was used for flame atomic absorption spectrophotometry (Corning410); the Ca2+, Mg2+ test method for EDTA (ethylene diamine tetraacetic acid) titration; the Cl, test method for ion chromatography (792Basic IC); the test method for acid–base titration; and the pH value with WpH3110 portable multi-parameter water quality monitor determination. TDS is equal to the total mass concentration of eight ions (, , Cl, , Ca2+, Mg2+, K+, and Na+) minus half of the content (Table 1). The ion balance error of the collected groundwater samples is within the allowable range; the sample data are valid, and the test formula is as follows:
formula
(1)
where E is the relative error, k denotes the milligram equivalent concentration of cations (mEq/L), and a is the milligram equivalent concentration of anions (mEq/L).
Table 1

Indicator testing of groundwater samples

Test subjectTest methods
K+, Na+ Flame atomic absorption spectrophotometry (Corning410) 
Ca2+, Mg2+ EDTA titration 
Cl,  Ion chromatography (792Basic IC) 
 Acid–base titration 
pH WpH3110 portable multi-parameter water quality monitor 
TDS Total mass concentration of eight ions (, , Cl, , Ca2+, Mg2+, K+, and Na+) minus half of the content 
Test subjectTest methods
K+, Na+ Flame atomic absorption spectrophotometry (Corning410) 
Ca2+, Mg2+ EDTA titration 
Cl,  Ion chromatography (792Basic IC) 
 Acid–base titration 
pH WpH3110 portable multi-parameter water quality monitor 
TDS Total mass concentration of eight ions (, , Cl, , Ca2+, Mg2+, K+, and Na+) minus half of the content 

Statistical analysis methods

Trilinear diagram analysis

The trilinear diagram was introduced by Piper in 1944 and is often applied to interpret some geochemically relevant problems (Ray & Mukherjee 2008; Karmegam et al. 2011). The percentages of major cations (K+ + Na+, Ca2+, and Mg2+) and anions (Cl, , and ) milligram equivalent per litre are visualized on the graph, where the diamond-shaped areas are divided into nine zones, each with its unique hydrochemical characteristics, allowing the study of the formation conditions of different aquifers.

Correlation analysis

Correlation analysis is often used to analyze the chemical characteristics of groundwater and its genesis analysis. The chemical reactions occurring in groundwater in different aquifers are different, and any reaction will follow a chemical equilibrium system; thus, there must be a certain correlation between ions (Noori et al. 2010). The absolute value of the correlation coefficient directly reflects the strength of the correlation; the closer the absolute value is to 1, the stronger the correlation, and the closer the absolute value is to 0, the weaker the correlation.

Gibbs plot analysis

Gibbs considered three types of mechanisms, namely, atmospheric precipitation, water–rock interaction, and evaporative crystallization, as the main factors controlling the dissolved salt composition of groundwater and showed a good visualization of the hydrochemical composition of groundwater and its genesis mechanism by the Gibbs diagram (Gibbs 1970). The water–rock interaction is the main factor for groundwater, and it often shows the trend of evaporation, because, with the evolution of groundwater and the influence of mineral saturation, groundwater will show the evolution trend from HCO3-type to SO4-type to Cl-type.

Ion ratio analysis

The ion ratio analysis is a common method of statistical analysis of water chemistry to obtain the ratio coefficients of the main characteristic ions, which can profile the groundwater chemical evolution process and determine the groundwater genesis and formation process (Gomo & Vermeulen 2014; Pazand & Javanshir 2014). The occurrence of chemical reactions in groundwater causes certain changes in ionic composition, and the ratio relationship between ions can directly reflect the genetic changes of groundwater chemical components.

Principal component analysis

The principal component analysis is the most commonly used multivariate mathematical analysis method to linearly map variables for dimensionality reduction. Generally, the composite index generated by the transformation becomes a principal component, and each principal component has its significant characteristics, which can more easily reveal the regularity among the original variables. With fewer data dimensions, the correlations of the original data reflected the maximum extent, thereby effectively eliminating redundant information and making the analysis results more valuable to use (Liu et al. 2003; Lee et al. 2007; Zhu et al. 2017).

Cluster analysis

As a classical exploratory analysis method in multivariate statistics, cluster analysis is frequently cited in the fields of geochemistry, mathematics, computer science, and medicine, and the technique has been developed for different applications (Frades & Matthiesen 2010; Zou 2020). Water chemistry characterization is achieved by classifying the distance proximity of conventional ionic components based on the similarity and difference of the intrinsic structure of groundwater chemical data, with greater intrinsic structural similarity being closer and greater difference being farther apart.

Water chemical composition analysis

Conventional ion content analysis

Table 2 and Figures 2 and 3 interpret the ion–chemical concentration distribution rules and characteristics of different aquifers from different perspectives. As shown in Table 2, the order of conventional ion concentrations in the Cenozoic aquifer is Cl (530.0 mg/L) > K+ + Na+ (483.1 mg/L) > (328.8 mg/L) > (103.5 mg/L) > Ca2+ (27.4 mg/L) > Mg2+ (12.4 mg/L). Among the cations, K+ + Na+ dominates at 92.4%, and among the anions, Cl dominates at 57.2%. The regular ion concentration distribution pattern of the Permian aquifer is (684.3 mg/L) > K+ + Na+ (627.6 mg/L) > Cl (504.5 mg/L) > (113.3 mg/L) > Ca2+ (20.7 mg/L) > Mg2+ (20.0 mg/L). The percentages of anion and cation K+ + Na+ were 52.6% and 93.9%, respectively. The conventional ion concentrations in the Carboniferous aquifer from high to low were Cl (543.3 mg/L) > K+ + Na+ (508.2 mg/L) > (301.1 mg/L) > (139.8 mg/L) > Ca2+ (30.7 mg/L) > Mg2+ (15.4 mg/L). The predominant cations were cationic K+ + Na+ and anionic Cl, accounting for 91.7% and 55.2%, respectively. In terms of TDS, the order was Permian aquifer (1,974.7 mg/L) > Carboniferous aquifer (1,461.9 mg/L) > Cenozoic aquifer (1,349.3 mg/L). TDS is an important index reflecting the hydrogeological conditions of groundwater recharge and runoff; the higher the TDS value, the longer the contact time between the water and surrounding rocks, and the more the dissolved substances contained in the water. It is assumed that some areas of this aquifer have good recharge, runoff, or discharge conditions. The pH values of the three aquifers were not significantly different (Figure 3(h)), and the overall pH values in the study area ranged from 7.6 to 10.4, with alkaline water quality.
Table 2

Statistics of the main chemical indicators of the aquifer system

SourceIndexIonic component content (mg/L)
TDS (mg/L)pH
K+ + Na+Ca2+Mg2+Cl
Q (n = 35) Maximum 844.6 37.5 20.7 1,091.8 201.7 439.3 2,391.0 9.3 
Minimum 138.3 12.7 3.7 23.1 21.9 150.4 446.5 7.7 
Average 483.1 27.4 12.4 530.0 103.5 328.8 1,349.3 8.3 
Cv 64% 30% 30% 88% 66% 22% 59% 4% 
P (n = 35) Maximum 1,024.0 81.7 99.0 926.4 776.4 1,668.9 2,957.3 8.8 
Minimum 324.6 2.8 2.2 14.4 5.8 24.3 1,014.8 7.6 
Average 627.6 20.7 20.0 504.5 113.3 684.3 1,974.7 8.1 
Cv 27% 91% 110% 53% 173% 67% 21% 4% 
C (n = 10) Maximum 830.7 91.4 33.8 1,001.6 319.7 483.0 2,459.0 10.4 
Minimum 141.8 11.9 4.5 12.7 16.1 39.0 682.0 7.7 
Average 508.2 30.7 15.4 543.3 139.8 301.1 1,461.9 8.4 
Cv 47% 77% 56% 66% 83% 40% 41% 9% 
SourceIndexIonic component content (mg/L)
TDS (mg/L)pH
K+ + Na+Ca2+Mg2+Cl
Q (n = 35) Maximum 844.6 37.5 20.7 1,091.8 201.7 439.3 2,391.0 9.3 
Minimum 138.3 12.7 3.7 23.1 21.9 150.4 446.5 7.7 
Average 483.1 27.4 12.4 530.0 103.5 328.8 1,349.3 8.3 
Cv 64% 30% 30% 88% 66% 22% 59% 4% 
P (n = 35) Maximum 1,024.0 81.7 99.0 926.4 776.4 1,668.9 2,957.3 8.8 
Minimum 324.6 2.8 2.2 14.4 5.8 24.3 1,014.8 7.6 
Average 627.6 20.7 20.0 504.5 113.3 684.3 1,974.7 8.1 
Cv 27% 91% 110% 53% 173% 67% 21% 4% 
C (n = 10) Maximum 830.7 91.4 33.8 1,001.6 319.7 483.0 2,459.0 10.4 
Minimum 141.8 11.9 4.5 12.7 16.1 39.0 682.0 7.7 
Average 508.2 30.7 15.4 543.3 139.8 301.1 1,461.9 8.4 
Cv 47% 77% 56% 66% 83% 40% 41% 9% 

Note: Q, Cenozoic aquifer; P, Permian aquifer; C, Carboniferous aquifer; Cv, Coefficient of variation; TDS, total dissolved solids.

Figure 2

Distribution of coefficients of variation of the main indicators of the aquifer system. Q, P, and C represent Cenozoic aquifer, Permian aquifer, and Carboniferous aquifer, respectively.

Figure 2

Distribution of coefficients of variation of the main indicators of the aquifer system. Q, P, and C represent Cenozoic aquifer, Permian aquifer, and Carboniferous aquifer, respectively.

Close modal
Figure 3

Comparison of the distribution characteristics of the main indicators of the aquifer system. Q, P, and C represent Cenozoic aquifer, Permian aquifer, and Carboniferous aquifer, respectively.

Figure 3

Comparison of the distribution characteristics of the main indicators of the aquifer system. Q, P, and C represent Cenozoic aquifer, Permian aquifer, and Carboniferous aquifer, respectively.

Close modal

The coefficient of variation can eliminate the influence of the mean value and effectively reflect the degree of spatial variation in the content of groundwater chemical components. Combined with the analysis in Figure 2, the ions with higher coefficients of variation in the Cenozoic aquifer are all K+ + Na+, Cl, and . The coefficients of variation of Ca2+, Mg2+, and in the Permian aquifer are higher at 91%, 110%, and 173%, respectively. In addition, the variation coefficients of Ca2+, Cl, and in the Carboniferous aquifer are higher, and varies with environmentally sensitive factors. Moreover, it is speculated that the more complex formation role of water chemistry components in the Permian aquifer is due to the interference of coal mining activities.

Box plots can visually distinguish the characteristics of the raw data distribution (Chen et al. 2021), and Figure 3 shows the comparison of the same indicator on different aquifers. In most indicators, the distribution patterns of the Cenozoic and Carboniferous aquifers are closer, and in the Permian aquifer, the content is lower, while the content is evident in the other two aquifers, and it is assumed that the desulfurization reaction is more significant in the Permian aquifer.

Trilinear diagram

The groundwater flow causes shifts between ions of water chemistry types, and investigating the water chemistry types of aquifer systems in the study area can help understand groundwater chemistry characteristics and their evolutionary patterns (Ma et al. 2019). The trilinear diagram can visualize the anion and cation milligram-equivalent percentages, which, in turn, makes the study of water chemistry types simpler and clearer. As shown in Figure 4, the water samples of the Cenozoic and Carboniferous aquifers are more dispersed in overall distribution and rich in water sample types. In the water samples of the Cenozoic aquifer, the cations are more than 60% of K+ + Na+ and less than 40% of Ca2+ + Mg2+, and the anions are mainly Cl, followed by . The main water chemistry types are Cl-Na, HCO3-Na, and HCO3-Cl-Na types, of which the Cl-Na type accounts for 45%. The cations in the water samples of the Carboniferous aquifer show a trend toward K+ + Na+ enrichment, with the overall greater than 80% and Ca2+ + Mg2+ less than 20%; and the distribution of anions in the water samples is relatively scattered, with Cl dominating and the content of accounting for 10%–70% of the distribution, with the main hydrochemical types of Cl-Na, Cl-HCO3-Na, and SO4-HCO3-Na. The overall distribution of the Permian aquifer water samples in the trilinear map is concentrated, the water chemistry type changes less, and the main water chemistry type is HCO3-Cl-Na, Cl-Na, and HCO3-Na. There are 20 groups of HCO3-Cl-Na type in 26 groups of water samples, accounting for approximately 77%, reflecting the characteristics of deep-water quality.
Figure 4

Analysis of water chemistry type of the aquifer system.

Figure 4

Analysis of water chemistry type of the aquifer system.

Close modal

Correlation analysis

There is a certain connection between ions in the groundwater; thus, the correlation analysis of water chemistry data is necessary (Li et al. 2019). From the correlation matrix of the aquifer system (Tables 35), in the Cenozoic aquifer, the correlation coefficients of with K+ + Na+ and Cl are 0.821 and 0.819, respectively, and shows a strong negative correlation with K+ + Na+, Cl, and . In the Permian aquifer, the correlation coefficient between K+ + Na+ and is 0.637, thereby indicating that the dissolution of silicate (feldspar) may exist in the Permian aquifer. Meanwhile, in the Carboniferous aquifer, the correlation coefficient between and Cl is 0.637 (p< 0.05). The correlation coefficients between K+ + Na+ and Cl in the Cenozoic and Carboniferous aquifers are as high as 0.999 and 0.998 (p< 0.01). It is assumed that the dissolution effect of rock salt in the related aquifers has a greater influence on both of them. The correlation between Ca2+ and Mg2+ in all three aquifers is significant in different degrees, and it is assumed that the sources of the two are consistent to some extent. The correlation coefficients of Ca2+, Mg2+, and in the Permian aquifer are significant at the 0.05 level, while the correlation coefficients of Ca2+, Mg2+, and in the Carboniferous aquifer are 0.896 and 0.710, which are presumed to be related to the dissolution of sulfate. The correlation coefficients of K++Na+, Cl and TDS in the Cenozoic aquifer are 0.952 and 0.946, respectively, and the correlation coefficient between K+ + Na+ and TDS in the Permian aquifer is 0.848. The correlation coefficients between K+ + Na+ and Cl and TDS in the Carboniferous aquifer are 0.971 and 0.973, respectively, thereby indicating that K+ + Na+ is the key ion affecting the TDS variation in the aquifer system.

Table 3

Correlation coefficient matrix of the Cenozoic aquifer

ElementsK+ + Na+Ca2+Mg2+ClTDSpH
K+ + Na+        
Ca2+ 0.271       
Mg2+ −0.017 0.512*      
Cl 0.999** 0.296 −0.001     
 0.821** 0.432 0.044 0.819**    
 0.656** 0.168 0.459* 0.656** − 0.747**   
TDS 0.952** 0.215 0.002 0.946** 0.762** − 0.599**  
pH 0.195 −0.091 0.736** 0.179 0.164 0.619** 0.171 
ElementsK+ + Na+Ca2+Mg2+ClTDSpH
K+ + Na+        
Ca2+ 0.271       
Mg2+ −0.017 0.512*      
Cl 0.999** 0.296 −0.001     
 0.821** 0.432 0.044 0.819**    
 0.656** 0.168 0.459* 0.656** − 0.747**   
TDS 0.952** 0.215 0.002 0.946** 0.762** − 0.599**  
pH 0.195 −0.091 0.736** 0.179 0.164 0.619** 0.171 

Note: Significant correlation values are indicated in bold.

*Significant correlation at the 0.05 level (two-tailed).

**Significant correlation at the 0.01 level (two-tailed).

Table 4

Correlation coefficient matrix of the Permian aquifer

ElementsK+ + Na+Ca2+Mg2+ClTDSpH
K+ + Na+        
Ca2+ −0.139       
Mg2+ −0.239 0.480*      
Cl 0.189 0.420* 0.241     
 −0.125 0.406* 0.423* 0.36    
 0.637** − 0.472* − 0.442* 0.199 0.538**   
TDS 0.848** 0.064 −0.002 0.238 −0.018 0.575**  
pH −0.18 −0.13 0.142 −0.042 0.069 −0.303 −0.349 
ElementsK+ + Na+Ca2+Mg2+ClTDSpH
K+ + Na+        
Ca2+ −0.139       
Mg2+ −0.239 0.480*      
Cl 0.189 0.420* 0.241     
 −0.125 0.406* 0.423* 0.36    
 0.637** − 0.472* − 0.442* 0.199 0.538**   
TDS 0.848** 0.064 −0.002 0.238 −0.018 0.575**  
pH −0.18 −0.13 0.142 −0.042 0.069 −0.303 −0.349 

Note: Significant correlation values are indicated in bold.

*Significant correlation at the 0.05 level (two-tailed).

**Significant correlation at the 0.01 level (two-tailed).

Table 5

Correlation coefficient matrix of the Carboniferous aquifer

ElementsK+ + Na+Ca2+Mg2+ClTDSpH
K+ + Na+        
Ca2+ −0.218       
Mg2+ −0.038 0.896**      
Cl 0.998** −0.205 −0.028     
 −0.033 0.710* 0.584 −0.044    
 0.626 −0.213 0.021 0.637* −0.592   
TDS 0.971** −0.045 0.155 0.973** 0.133 0.556  
pH −0.545 −0.46 −0.569 −0.557 −0.1 −0.607 −0.567 
ElementsK+ + Na+Ca2+Mg2+ClTDSpH
K+ + Na+        
Ca2+ −0.218       
Mg2+ −0.038 0.896**      
Cl 0.998** −0.205 −0.028     
 −0.033 0.710* 0.584 −0.044    
 0.626 −0.213 0.021 0.637* −0.592   
TDS 0.971** −0.045 0.155 0.973** 0.133 0.556  
pH −0.545 −0.46 −0.569 −0.557 −0.1 −0.607 −0.567 

Note: Significant correlation values are indicated in bold.

*Significant correlation at the 0.05 level (two-tailed).

**Significant correlation at the 0.01 level (two-tailed).

Analysis of control factors

Gibbs diagram analysis

In the Gibbs plot, if the horizontal and vertical coordinates are high, the ions in the groundwater are controlled by evaporative crystallization; if the TDS values are medium and the horizontal coordinates are low, the ions in groundwater are controlled by hydromorphism; if the vertical coordinates are low and the horizontal coordinates are high, the ions in groundwater are controlled by atmospheric rainfall (Jiang et al. 2020). From the Gibbs plot of water samples, the water chemical composition of the samples involved in this study is primarily controlled by hydroelastic action, but the performance is slightly different among the three aquifers, in which most of the water samples of the Cenozoic and Permian aquifers have low Cl/(Cl + HCO3) (Figure 5(b)). No significant evaporation control trend is found primarily because of the extractive thinning. The impact of water release, the great disturbance to the Permian aquifer water samples, fast water flow, and significant renewal resulted in an insufficient water–rock interaction between the Permian aquifer water samples and the surrounding rocks. Moreover, the excavation disturbs the water samples of the Cenozoic aquifer very much, thereby resulting in a faster percolation rate and lower overall Cl/(Cl + HCO3) values in the Cenozoic aquifer. Most of the Carboniferous aquifer water samples have high Cl/(Cl + HCO3) values because the carbonate rocks have high solubility and the aquifer water samples are not disturbed; hence, the Carboniferous aquifer water samples remain in the aquifer for a long time and can fully interact with the surrounding rocks.
Figure 5

Gibbs plot of groundwater chemistry in the study area.

Figure 5

Gibbs plot of groundwater chemistry in the study area.

Close modal

Given that Na/(Na + Ca) is a good indicator of ion exchange, the range of ρNa+/ρ(Na+ + Ca2+) in groundwater is 0.8–1.0, and the content of Ca2+ is much lower than that of Na+, thereby indicating that there may be dissolution of rock salt or alternating cation adsorption in the groundwater of the study area resulting in an increase of Na+ in the water. This is more obvious in Cenozoic and Carboniferous aquifers (Figure 5(a)). The more evident Na-Ca ion exchange may exist in the water samples of the Cenozoic boundary, which is related to the presence of a large number of clastic rocks in this aquifer.

Ion ratio analysis

Hydrochemical reactions in groundwater cause changes in ion concentrations and, based on the chemical equilibrium system, certain patterns in the ratio change between groundwater ion concentrations, which is an important basis for studying the role of groundwater chemistry formation (Zhang et al. 2022). There exists mutual influence and an equilibrium relationship between various ions in the groundwater, and the analysis of the ratio change between various ion concentrations in the groundwater can obtain important information about the chemical characteristics of the groundwater. To study the hydrochemical characteristics and control factors of groundwater in the three aquifer systems of Liuzhuang coal mine, this paper draws and analyzes the relationship between the ratio of different ions in each aquifer.

The dissolution of aquifer carbonates (dolomite, calcite, gypsum, etc.) and sulfate is the main source of Ca2+ and Mg2+ in groundwater, and the selected groundwater samples in Figure 6(a) are all located above 1:1, thereby indicating that the Ca2+ + Mg2+ content in groundwater is less than that of + , suggesting that the dissolution of carbonates and alternating cation adsorption exist simultaneously. With Equations (2) and (3), it can be seen that the dissolution rate of dolomite is smaller than that of calcite under the same conditions, so it is presumed that Ca2+ in the aquifer is involved in alternating adsorption as the main ion.
Figure 6

Ion ratio analysis graph.

Figure 6

Ion ratio analysis graph.

Close modal
According to the calcite dissolution process (Equation (2)), at this time ρ()/ρ(Ca2+) is 2:1; the dissolution process of dolomite is shown in Equation (3), and at this time ρ()/ρ(Ca2+) is 4:1. As shown in Figure 6(b), a few groundwater samples are distributed on both sides of 2:1 and 4:1, indicating the existence of dolomite and calcite dissolution or alternating cation adsorption in each of the three aquifers. Most water samples fall in the area above the 4:1 line, indicating that is enriched relative to Ca2+, and it is presumed that there is desulfurization (Equation (4)) or alternating cation adsorption.
formula
(2)
formula
(3)
formula
(4)
Ca2+ and originate from the process of gypsum (Equation (5)) when ρ(Ca2+)/ρ() is 1. In addition, pyrite oxidation (Equation (6)) and alternating cation adsorption are also sources of . The coal mining activities provide an oxidizing environment for pyrite. Combined with the distribution of groundwater samples in Figure 6(c) on both sides of 1:1, it indicates that the dissolution of gypsum and oxidation of pyrite occurred in all three aquifers.
formula
(5)
formula
(6)
Atmospheric rainfall, water–rock action, and evaporative crystallization are all possible sources of Na+ and Cl. Na+ in groundwater primarily comes from the dissolution of rock salt when ρ(Na+)/ρ(Cl) is 1. In Figure 6(d), only one Permian aquifer water sample is above the 1:1 line, and the rest are below the 1:1 line, thereby indicating that the dissolution of silicate is not the only source of Na+ in the three aquifers; there is also alternating cation adsorption (Equation (7)), and alternating cation sorption is strongly expressed in some areas of the Permian aquifer.
formula
(7)
The degree of alternating cation adsorption in groundwater can be reflected by whether or not the slope of ρ(Na+ − Cl)/ρ[(Ca2+ + Mg2+) − ( + )] is close to −1. The Na+ in the geotechnical soil reacts with Ca2+ and Mg2+ in the groundwater by replacement, which decreases Ca2+ and Mg2+ and increases Na+ in the groundwater, achieving the effect of alternating cation adsorption (Equation (8)). As shown in Figure 6(e), most of the water samples are distributed along the 1:1 line; thus, alternating cation adsorption occurs in all three aquifers, particularly in the Cenozoic and Carboniferous aquifers, which is the most evident.
formula
(8)
Desulfation can be reflected by (Equation (9)); the smaller its value, the stronger the desulfation, the more ions are produced, and the more H2S gas is volatilized, which is also the process of coal formation. As shown in Figure 6(f), the desulfation of the Permian aquifer is relatively strong, and the location is rich in carbon sources, which provides better conditions for the occurrence of desulfation.
formula
(9)

Principal component analysis

After correlation analysis, strong correlations existed between most of the variables; hence, it was necessary to downscale the variables to study the controlling factors of the groundwater. K+ + Na+, Ca2+, Mg2+, Cl, , and with six ions were selected as the variables for analysis to be standardized, and the Kaiser–Meyer–Olkin (KMO) sampling fitness measure and Bartlett's sphericity were tested for significance. The groundwater samples in the study area were suitable for principal component analysis, and the correlation matrix of the original data was sought from the standardized data. The characteristic roots are calculated with the corresponding standard orthogonal eigenvectors. The first three principal components with large variance contribution were obtained based on the Kaiser criterion selection calculation, with eigenvalues of 2.252, 1.791, and 0.789 (Figure 7) and contribution rates of 37.534%, 29.846%, and 13.158% (Figure 8), respectively, with a cumulative contribution rate of 80.538% (Table 6). Thus, the load maps of the first three principal components can explain most of the information from the water sample data.
Table 6

Eigenvalue variance interpretation

CompositionsTotalVariance (%)Cumulative (%)
PC1 2.252 37.534 37.534 
PC2 1.791 29.846 67.380 
PC3 0.789 13.158 80.538 
PC4 0.658 10.961 91.499 
PC5 0.500 8.332 99.831 
PC6 0.010 0.169 100.000 
CompositionsTotalVariance (%)Cumulative (%)
PC1 2.252 37.534 37.534 
PC2 1.791 29.846 67.380 
PC3 0.789 13.158 80.538 
PC4 0.658 10.961 91.499 
PC5 0.500 8.332 99.831 
PC6 0.010 0.169 100.000 

Note: Bold formatting highlights that the three principal components of the load map can explain most of the information of the water sample data.

Figure 7

Eigenvalue gravel diagram.

Figure 7

Eigenvalue gravel diagram.

Close modal
Figure 8

Principal component diagram.

Figure 8

Principal component diagram.

Close modal
The maximum variance rotation method (varimax rotation) was used to do a proper rotation of the principal component axes, so that each principal component has the lowest number of variables with the highest loadings to highlight the variability of each principal component, thus simplifying the interpretation of the principal components and better revealing the intrinsic information of the water chemistry data. As shown in Table 7 and Figure 9(a–c), the loadings of Ca2+, Mg2+, and in principal component 1 (PC1) are 0.788, 0.673, and 0.760, respectively, and their elevated contents are primarily due to the dissolution of carbonate and sulfate and the oxidation of pyrite in the aquifer. Principal component 2 (PC2) accounts for a higher positive loading with K+ + Na+ and Cl, which may be the result of rock salt dissolution, and the alternating adsorption of cations in an environment with good groundwater runoff conditions may also lead to an increase in Na+. Loads of Mg2+ and in principal component 3 (PC3) are 0.586 and 0.562, respectively, and Mg2+ and can coexist in general, which can be regarded as the dissolution of Mg-containing carbonates and being dominated by the desulfurization effect of PC3.
Table 7

Principal component load matrix

Ordinal numberCompositionsPC1PC2PC3
K+ + Na+ −0.138 0.960 0.111 
Ca2+ 0.788 0.048 −0.050 
Mg2+ 0.673 0.060 0.586 
Cl 0.168 0.903 −0.297 
 0.760 0.074 0.168 
 −0.744 0.207 0.562 
Ordinal numberCompositionsPC1PC2PC3
K+ + Na+ −0.138 0.960 0.111 
Ca2+ 0.788 0.048 −0.050 
Mg2+ 0.673 0.060 0.586 
Cl 0.168 0.903 −0.297 
 0.760 0.074 0.168 
 −0.744 0.207 0.562 

Note: The main components of the load matrix in the strong correlation value are highlighted in bold.

Figure 9

Principal component load diagrams.

Figure 9

Principal component load diagrams.

Close modal
From the matrix of component score coefficients (Table 8), an integrated decision model with three principal components (Y1, Y2, and Y3) can be obtained, whose expressions (Equations (10)–(12)) are as follows:
formula
(10)
formula
(11)
formula
(12)
Table 8

Component score coefficient matrix

Ordinal numberCompositionsPC1PC2PC3
K+ + Na+ −0.061 0.536 0.140 
Ca2+ 0.350 0.027 −0.063 
Mg2+ 0.299 0.033 0.742 
Cl 0.075 0.504 −0.376 
 0.338 0.041 0.213 
 −0.330 0.116 0.711 
Ordinal numberCompositionsPC1PC2PC3
K+ + Na+ −0.061 0.536 0.140 
Ca2+ 0.350 0.027 −0.063 
Mg2+ 0.299 0.033 0.742 
Cl 0.075 0.504 −0.376 
 0.338 0.041 0.213 
 −0.330 0.116 0.711 
The principal component loading scores of the groundwater samples were plotted as scatter plots, and Figure 10(a)–10(c) show the PC1 and PC2, PC1 and PC3, and PC2 and PC3 scatter plots of the groundwater samples, respectively. In Figure 10(a), the water samples of the three aquifers are distributed in the first and second quadrants, indicating that the effect of hydrochloric acid dissolution and alternating cation adsorption has a greater influence on the water samples of the three aquifers. Meanwhile, the dissolution of carbonate and sulfate and the oxidation of pyrite have little effect on the water samples of some sampling points, and the Permian aquifer shows a relatively evident performance. The selected aquifer water samples are distributed in the first, second, and fourth quadrants of Figure 10(b), indicating that the chemical reactions represented by PC1 and PC3 all occur to different degrees with no apparent characteristics. In Figure 10(c), all three aquifer water samples are distributed in the first and fourth quadrants, indicating that the water samples of the three aquifers are significantly affected by the dissolution of hydrochloric acid and alternating cation adsorption, which corroborates the performance of Figure 10(a). The water samples of the Permian aquifer are strongly affected by the dissolution and desulfation of magnesium-containing carbonates.
Figure 10

Principal component load scores of groundwater samples.

Figure 10

Principal component load scores of groundwater samples.

Close modal
Figure 11

Cluster analysis tree diagram.

Figure 11

Cluster analysis tree diagram.

Close modal

Cluster analysis

Cluster analysis aims to collect data to group based on similarity, and in groundwater studies, variables from the same source are classified into the same group to facilitate the study of groundwater chemistry control factors, thus making the analysis simplified (Giridharan et al. 2009; Wu et al. 2009). This study used K+ + Na+, Ca2+, Mg2+, Cl, , and as input variables for the cluster analysis and applied them in SPSS software to calculate the distances by selecting intergroup connections and squared Euclidean methods to draw a cluster analysis dendrogram. As shown in Figure 11, the groundwater chemical control factors of the Cenozoic and Carboniferous aquifers have approximately the same pattern, and K+ + Na+ and Cl are classified into one group, which can be explained as rock salt dissolution or alternating cation adsorption. The variables are divided into two groups in the Permian aquifer. Group 1 includes K+ + Na+, Cl, and , which indicates the weathering of silicate and the dissolution of rock salt in this aquifer, and Group 2 includes Ca2+, Mg2+, and , which can be interpreted as the dissolution of carbonate and sulfate and the oxidation of pyrite. The results of this analysis can corroborate the results of the principal component analysis.

Combining the collected and tested groundwater chemical data of the Liuzhuang coal mine, the groundwater chemical composition of the Liuzhuang coal mine aquifer system and its controlling factors were studied by conventional water chemistry methods, correlation analysis, ion ratio analysis, principal component analysis, and cluster analysis. The following conclusions were obtained:

  • (1)

    The average cation content relationship of the aquifer system in the study area is all K+ + Na+ > Ca2+ > Mg2+, with K+ + Na+ dominating, accounting for more than 90% of the total. Conventional mathematical and statistical analyses have identified differences in the hydrochemical characteristics of different aquifers, and the average content relationship of anions in the Cenozoic and Carboniferous aquifers is Cl > > , with Cl accounting for 57.2% and 55.2%, respectively; the average content relationship of ions in the Permian aquifer is > Cl > , with accounting for 52.6%. The largest average content of TDS is in the Permian. The Permian aquifer has the highest average TDS content, followed by the Carboniferous aquifer, and the Cenozoic aquifer has the lowest. The main hydrochemical types of the Cenozoic aquifer are Cl-Na, HCO3-Na, and HCO3-Cl-Na; the main hydrochemical types of the Permian aquifer are HCO3-Cl-Na, Cl-Na, and HCO3-Na; and the main hydrochemical types of the Carboniferous aquifer are Cl-Na, Cl-HCO3-Na, and SO4-HCO3-Na.

  • (2)

    The Gibbs plot of the groundwater chemistry data indicates that the study area is mainly controlled by the water–rock interaction. Correlation analysis, ion ratio analysis, principal component analysis, and cluster analysis quantitatively analyzed the controlling factors of groundwater chemistry in the three aquifers, and found that dissolution of carbonate and sulfate, oxidation of pyrite, desulfation, and alternating cation adsorption occur to varying extents in all the three aquifers. Among them, the dissolution of rock salt has a greater influence on the Cenozoic and Carboniferous aquifers; the Permian aquifer is rich in carbon sources, which provides favorable conditions for the occurrence of desulfation, and the interference caused by coal mining activities on the Cenozoic and Permian aquifers also has a certain influence on the controlling factors of the chemical composition of the groundwater. In addition, ion ratio analysis and principal component analysis corroborate each other, confirming that cation alternating adsorption is more significant in the Cenozoic and Carboniferous aquifers and relatively weak in some water samples from the Permian aquifer. This study demonstrates the feasibility and effectiveness of conventional water chemistry analysis methods and multivariate statistical methods in analyzing and comprehensively interpreting the chemical composition of groundwater and its controlling factors, and provides useful guidance for solving the geochemical problems of the Liuzhuang coal mine.

Q.J. performed the experiments, analyzed and interpreted the data, and wrote the paper. Q.L. conceived and designed the experiments and contributed reagents, materials, analysis tools, and data. H.C. and Y.L. also contributed reagents, materials, analysis tools, and data. K.C. performed the experiments, and J.Z. supervised and revised the paper.

This work was financially supported by the Natural Science Foundation of Anhui Province under grant number 1908085ME145, the Postgraduate Research & Practice Innovation Program of Jiangsu Province (KYCX23_2760), the Fundamental Research Funds for the Central Universities (2023XSCX003), and the Graduate Innovation Program of China University of Mining and Technology (2023WLKXJ003). We sincerely thank the editors and reviewers for their valuable comments that greatly improved this paper.

All relevant data are included in the paper or its Supplementary Information.

The authors declare there is no conflict.

Ahmad
S.
&
Mazhar
S. N.
2020
Hydrochemical characteristics, groundwater quality and sources of solute in the Ramganga Aquifer, Central Ganga Plain, Bareilly District, Uttar Pradesh
.
Journal of the Geological Society of India
95
(
6
),
616
625
.
https://doi.org/10.1007/s12594-020-1488-y
.
Cai
J.
,
Varis
O.
&
Yin
H.
2017
China's water resources vulnerability: a spatio-temporal analysis during 2003–2013
.
Journal of Cleaner Production
142
,
2901
2910
.
https://doi.org/10.1016/j.jclepro.2016.10.180.
Du
M.
,
Zhang
M.
,
Wang
S.
,
Chen
F.
,
Zhao
P.
,
Zhou
S.
&
Zhang
Y.
2019
Stable isotope ratios in tap water of a riverside city in a semi-arid climate: an application to water source determination
.
Water
11
(
7
),
1441
.
https://doi.org/10.3390/w11071441.
Fasaee
M. A. K.
,
Berglund
E.
,
Pieper
K. J.
,
Ling
E.
,
Benham
B.
&
Edwards
M.
2021
Developing a framework for classifying water lead levels at private drinking water systems: a Bayesian belief network approach
.
Water Research
189
,
116641
.
https://doi.org/10.1016/j.watres.2020.116641
.
Frades
I.
&
Matthiesen
R.
2010
Overview on techniques in cluster analysis
. In:
Bioinformatics Methods in Clinical Research
(Matthiesen, R., ed.).
Humana Press
,
New York, USA
, pp.
81
107
.
https://doi.org/10.1007/978-1-60327-194-3_5.
Fritz
M.
,
Fernández-Serra
M.
&
Soler
J. M.
2016
Optimization of an exchange-correlation density functional for water
.
The Journal of Chemical Physics
144
(
22
),
224101
.
https://doi.org/10.1063/1.4953081.
Gibbs
R. J.
1970
Mechanisms controlling world water chemistry
.
Science
170
(
3962
),
1088
1090
.
https://doi.org/10.1126/science.170.3962.1088
.
Giridharan
L.
,
Venugopal
T.
&
Jayaprakash
M.
2009
Assessment of water quality using chemometric tools: a case study of River Cooum, South India
.
Archives of Environmental Contamination and Toxicology
56
(
4
),
654
669
.
https://doi.org/10.1007/s00244-009-9310-2.
Gomo
M.
&
Vermeulen
D.
2014
Hydrogeochemical characteristics of a flooded underground coal mine groundwater system
.
Journal of African Earth Sciences
92
,
68
75
.
https://doi.org/10.1016/j.jafrearsci.2014.01.014.
Jiang
Y.
,
Gui
H.
,
Yu
H.
,
Wang
M.
,
Fang
H.
,
Wang
C.
,
Chen
C.
,
Zhang
Y.
&
Huang
Y.
2020
Hydrochemical characteristics and water quality evaluation of rivers in different regions of cities: a case study of Suzhou City in Northern Anhui Province, China
.
Water
12
(
4
),
950
.
https://doi.org/10.3390/w12040950
.
Karmegam
U.
,
Chidambaram
S.
,
Prasanna
M. V.
,
Sasidhar
P.
,
Manikandan
S.
,
Johnsonbabu
G.
,
Dheivanayaki
V.
,
Paramaguru
P.
,
Manivannan
R.
,
Srinivasamoorthy
K.
&
Anandhan
P.
2011
A study on the mixing proportion in groundwater samples by using Piper diagram and PHREEQC model
.
Chinese Journal of Geochemistry
30
(
4
),
490
495
.
https://doi.org/10.1007/s11631-011-0533-3.
Lee
J.
,
Kim
J.-H.
,
Kim
H.-M.
&
Chang
H.-W.
2007
Statistical approach to determine the salinized ground water flow path and hydrogeochemical features around the underground LPG cavern, Korea
.
Hydrological Processes
21
(
26
),
3615
3626
.
https://doi.org/10.1002/hyp.6589.
Li
P.
,
Tian
R.
&
Liu
R.
2019
Solute geochemistry and multivariate analysis of water quality in the Guohua Phosphorite Mine, Guizhou Province, China
.
Exposure and Health
11
(
2
),
81
94
.
https://doi.org/10.1007/s12403-018-0277-y
.
Liang
Y.
,
Gao
X.
,
Zhao
C.
,
Tang
C.
,
Shen
H.
,
Wang
Z.
&
Wang
Y.
2018
Review: Characterization, evolution, and environmental issues of karst water systems in Northern China
.
Hydrogeology Journal
26
(
5
),
1371
1385
.
https://doi.org/10.1007/s10040-018-1792-4
.
Liu
R. X.
,
Kuang
J.
,
Gong
Q.
&
Hou
X. L.
2003
Principal component regression analysis with SPSS
.
Computer Methods and Programs in Biomedicine
71
(
2
),
141
147
.
https://doi.org/10.1016/S0169-2607(02)00058-5.
Liu
Y.
,
Liu
P.
,
Bo
L.
,
Li
Q.
,
Quan
G.
,
Zhuo
Y.
,
Han
Y.
&
Wang
Y.
2022
Research on mine water dispatching mode based on maximization of reuse rate
.
Sustainability
14
(
15
),
9289
.
https://doi.org/10.3390/su14159289.
Ma
L.
,
Guo
J.
,
Liu
W.
,
Zhang
D.
&
Yu
Y.
2019
Water conservation when mining multiple, thick, closely-spaced coal seams: a case study of mining under Weishan Lake
.
Mine Water and the Environment
38
(
3
),
643
657
.
https://doi.org/10.1007/s10230-019-00610-8.
Noori
R.
,
Sabahi
M. S.
,
Karbassi
A. R.
,
Baghvand
A.
&
Taati Zadeh
H.
2010
Multivariate statistical analysis of surface water quality based on correlations and variations in the data set
.
Desalination
260
(
1–3
),
129
136
.
https://doi.org/10.1016/j.desal.2010.04.053.
Pazand
K.
&
Javanshir
A. R.
2014
Geochemistry and water quality assessment of groundwater around Mohammad Abad area, Bam District, SE Iran
.
Water Quality, Exposure and Health
6
(
4
),
225
231
.
https://doi.org/10.1007/s12403-014-0131-9.
Poetra
R. P.
,
Adji
T. N.
,
Santosa
L. W.
&
Khakhim
N.
2020
Hydrogeochemical conditions in groundwater systems with various geomorphological units in Kulonprogo Regency, Java Island, Indonesia
.
Aquatic Geochemistry
26
(
4
),
421
454
.
https://doi.org/10.1007/s10498-020-09384-w.
Ray
R. K.
&
Mukherjee
R.
2008
Reproducing the Piper trilinear diagram in rectangular coordinates
.
Groundwater
46
(
6
),
893
896
.
https://doi.org/10.1111/j.1745-6584.2008.00471.x.
Sethy
S. N.
,
Syed
T. H.
,
Kumar
A.
&
Sinha
D.
2016
Hydrogeochemical characterization and quality assessment of groundwater in parts of Southern Gangetic Plain
.
Environmental Earth Sciences
75
(
3
),
232
.
https://doi.org/10.1007/s12665-015-5049-4.
Shahbaz
M.
,
Farhani
S.
&
Ozturk
I.
2015
Do coal consumption and industrial development increase environmental degradation in China and India?
Environmental Science and Pollution Research
22
(
5
),
3895
3907
.
https://doi.org/10.1007/s11356-014-3613-1.
Sun
K.
,
Fan
L.
,
Xia
Y.
,
Li
C.
,
Chen
J.
,
Gao
S.
,
Wu
B.
,
Peng
J.
&
Ji
Y.
2021
Impact of coal mining on groundwater of Luohe Formation in Binchang mining area
.
International Journal of Coal Science & Technology
8
(
1
),
88
102
.
https://doi.org/10.1007/s40789-020-00366-5.
Tan
R.-C.
,
Shao
Y.-J.
,
Wei
H.-T.
,
Zhang
J.-K.
,
Yu
M.-D.
&
Xiong
Y.-Q.
2023
Hydrothermal fluid characteristics of the Huayuan Pb-Zn orefield: constraints from apatite major and trace element compositions
.
Frontiers in Earth Science
10
,
1083799
.
https://doi.org/10.3389/feart.2022.1083799.
Vreča
P.
&
Kern
Z.
2020
Use of water isotopes in hydrological processes
.
Water
12
(
8
),
2227
.
https://doi.org/10.3390/w12082227.
Wu
M.-L.
,
Wang
Y.-S.
,
Sun
C.-C.
,
Wang
H.
,
Dong
J.-D.
&
Han
S.-H.
2009
Identification of anthropogenic effects and seasonality on water quality in Daya Bay, South China Sea
.
Journal of Environmental Management
90
(
10
),
3082
3090
.
https://doi.org/10.1016/j.jenvman.2009.04.017
.
Yan
B.
,
Ren
F.
,
Cai
M.
&
Qiao
C.
2020
Bayesian model based on Markov chain Monte Carlo for identifying mine water sources in submarine gold mining
.
Journal of Cleaner Production
253
,
120008
.
https://doi.org/10.1016/j.jclepro.2020.120008.
Zhang
C.
,
Zhang
J.
,
Zhao
B.
,
Zhang
H.
&
Huang
P.
2009
Stable isotope studies of crop carbon and water relations: a review
.
Agricultural Sciences in China
8
(
5
),
578
590
.
https://doi.org/10.1016/S1671-2927(08)60249-7.
Zhang
M.
,
Chen
L.
,
Yao
D.
,
Hou
X.
,
Zhang
J.
,
Qin
H.
,
Ren
X.
&
Zheng
X.
2022
Hydrogeochemical processes and inverse modeling for a multilayer aquifer system in the Yuaner Coal Mine, Huaibei Coalfield, China
.
Mine Water and the Environment
41
(
3
),
775
789
.
https://doi.org/10.1007/s10230-022-00851-0.
Zhu
B.
,
Wang
X.
&
Rioual
P.
2017
Multivariate indications between environment and ground water recharge in a sedimentary drainage basin in northwestern China
.
Journal of Hydrology
549
,
92
113
.
https://doi.org/10.1016/j.jhydrol.2017.03.058.
Zou
H.
2020
Clustering algorithm and its application in data mining
.
Wireless Personal Communications
110
(
1
),
21
30
.
https://doi.org/10.1007/s11277-019-06709-z.
This is an Open Access article distributed under the terms of the Creative Commons Attribution Licence (CC BY 4.0), which permits copying, adaptation and redistribution, provided the original work is properly cited (http://creativecommons.org/licenses/by/4.0/).