Using Isotopic and Hydrochemical Indicators to Identify Sources of Sulfate in Karst Groundwater of the Niangziguan Spring Field, China

Karst groundwater in the Niangziguan spring fields is the main source to supply domestic and industrial water demands in Yangquan City, China. However, the safety of water supply in this region has recently suffered from deteriorating quality levels. Therefore, identifying pollution sources and causes is crucial for maintaining a reliable water supply. In this study, a systematic sample collection for the karst groundwater in the Niangziguan spring fields was implemented to identify hydrochemical characteristics of the karst groundwater through comprehensive analyses of hydrochemistry (piper diagram, and ion ratios,) and stable isotopes (S and H-O). The results show that the karst groundwater in the Niangziguan spring fields was categorized as SO4·HCO3-Ca·Mg, HCO3·SO4-Ca·Mg, and SO4-Ca types. K+, Cl-, and Na+ are mainly sourced from urban sewage and coal mine drainage. In addition, SO42− was mainly supplied by the dissolution of gypsum and the oxidation of FeS2 in coal-bearing strata. It is noteworthy that, based on H-O and S isotopes, 75% of the karst groundwater was contaminated by acidic water in coal mines at different degrees. In the groundwater of the Niangziguan spring field, the proportions of SO42− derived from FeS2 oxidation were 60.6% (N50, Chengxi spring), 30.3% (N51, Wulong spring), and 26.0% (N52, Four springs mixed with water). Acid mine drainage directly recharges and pollutes karst groundwater through faults or abandoned boreholes, or discharges to rivers, and indirectly pollutes karst groundwater through river infiltration in carbonate exposed areas. The main source of rapid increase of sulfate in karst groundwater is acid water from abandoned coal mines.


Introduction
Karst regions occupy approximately 25 percent of the land surface of the Earth [1][2][3]. Karst groundwater is an important water resource that is used as a drinking and industrial water source by a significant proportion of the world's population [4,5]. However, karst aquifers are much more vulnerable to pollution than other aquifers, and the restoration of contaminated groundwater is difficult [6][7][8]. Karst groundwater is a "hidden resource" in northern China and the prevention and monitoring of karst groundwater pollution are more difficult than that of surface waters due to its inaccessibility. Among numerous contaminants that contribute to groundwater pollution, Fe [9][10][11][12], nitrate [13][14][15][16][17], and sulfate [18][19][20] are cited most frequently. Fe in karst groundwater is mainly affected by Fe-rich acid mine drainage (AMD). Nitrate in karst groundwater includes synthetic and organic fertilizers, as well as municipal and industrial sewage. Sulfate minerals in karst groundwater originate from natural sources like the dissolution of sulfate, atmospheric deposition, and AMD caused by human activity.
Niangziguan spring is located in Yangquan City, Shanxi Province. The average annual flow of spring water is 9.8 m 3 ·s −1 . Spring water in the Niangziguan spring fields is the main source to supply domestic and industrial water demands in Yangquan City, China. Since 2010, the sulfate radical of karst spring water has been more than 250 mg·L −1 . Karst well water and springs are the primary method used for the discharge of groundwater from karst aquifers [21,22]. Therefore, the monitoring of karst well water and springs can provide more insights into the hydrogeological and hydrogeochemical processes that occur in underground environments.
The hydrogeochemical characteristics are crucial to reflect the circulation and hydrodynamic conditions of regional groundwater, and to understand the water quality and distribution characteristics. To date, methods commonly used in the karst water hydrogeochemistry are the hydrochemical type, saturation index, and ion ratio methods [23][24][25]. Advanced techniques have been also used such as isotopic analysis, statistical methods, and hydrogeochemical simulations. The interactions between precipitation and groundwater can be investigated using stable isotopes of hydrogen and oxygen [26][27][28][29]. The application of isotope tracers (δ 34 S-SO 4 2 , δ 18 O-SO 4 2 ) in identifying sources and fate of SO 4 in groundwater has been successfully achieved for several decades [30][31][32][33][34][35]. The geochemical evolution of karst water and its associations with urbanization impacts have not been fully addressed in the karst regions throughout the world [36][37][38][39][40][41][42]. This study characterizes spatial-temporal variations in hydrochemistry in the Niangziguan spring region, a typical karst area of fast urbanization in Shanxi, using hydrochemical reconnaissance and monitoring of S, D, and O isotopes [43][44][45][46][47][48][49]. The aim of our study is to evaluate the impacts of rapid urbanization on karst water quality. Identifying pollution sources and causes is crucial for reliable water supply.

Study Area
The Niangziguan spring field is located in the eastern part of Shanxi Province adjacent to Hebei Province in China, which includes: (1) the urban areas, suburbs, Pingding County, and Yu County of Yangquan City; (2) Xiyang, Heshun, Zuoquan, Shouyang, and Yuci of Jinzhong City; (3) Taiyuan City and Yangqu County. These areas are shown in Figure 1. The distribution area of coal measures strata in the spring area is 4728 km 2 , accounting for 63.5% of the whole spring area. Therefore, the industrial structure dominated by coal has become a feature of this area. Taking a comprehensive view of the main districts and counties in the spring area (the districts and counties of Yangquan City and Shouyang, Xiyang, Heshun, and Zuoquan of Jinzhong City), the output value of coal mining industry generally accounts for more than 50% of the total output value. The total area of the district is 7436 km 2 ranging from 36 • 55 N to 38 • 15 N and 112 • 20 E to 113 • 55 E. The average annual water evaporation in the study area was 1202 mm while the average annual precipitation of the whole region was 542.4 mm during the period from 1955 to 2019 at Yangquan Station within the China Meteorological Data Network. While the average annual temperature is 10.9 • C, the average temperature in January is −4.6 • C and the extreme minimum temperature is −28 • C.

Hydrogeochemistry
The chemical composition of karst water is controlled by factors such as geological conditions, hydrodynamic conditions, and human engineering activities in the region, and it can better preserve such influence information [50][51][52][53][54]. Therefore, hydrogeologists often use hydrogeochemical methods to study complex karst water systems [55][56][57][58].

Hydrogeochemistry of the Karst Groundwater
Based on a total of 52 sets of analytical results for the karst groundwater, it was found that the pH values of karst water range from 6.96 to 8.00, with an average of 7.60. The average concentrations of Ca 2+ , Mg 2+ , K + +Na + , Cl − , and SO 4 2− are 143.5 mg·L −1 , 36.2 mg·L −1 , 45.6 mg·L −1 , and 323.9 mg·L −1 , respectively. The ranges of each indicator are shown in Table 1 and Figure 2. Figure 3 shows a piper diagram of ionic concentration for the collected samples. The karst water (spring) generally shows cationic high concentrations of Ca 2+ and Mg 2+ and low concentrations of K + +Na + , and anionic high concentration of SO 4 2− , low concentrations of Cl -, and high concentration of HCO 3 − (Table 1).   Figure 5, the concentrations of Cl − and SO 4 2− ranged from 14.2 to 522.9 mg·L −1 (an average of 136.1 mg·L −1 ) and 183.2 to 4042 mg·L −1 (an average of 1240.7 mg·L −1 ), respectively. These results are closely related to the Shukarev classification types for the coal mine acidic water and surface water. In other words, the coal mine acidic water is usually grouped by SO 4 -Ca·Mg and SO 4 -Mg·Ca types while the surface water is mainly classified by SO 4 -Ca·Na and SO 4 -Ca types. Therefore, the coal mine acidic water generally exhibited high concentrations of Ca 2+ , Mg 2+ , and SO 4 2− while the surface water showed a higher concentration of Cl − than that of the acidic water from coal mines ( Figure 6).     Table 1 and Figure 7 show a statistical summary of stable isotopes in the karst groundwater samples in the study area, the δ 34 S-SO 4 2− values ranged from −6.    Figure 7 show the chemical properties and composition of AMD and river samples. The results showed that the δ 34 S-SO 4 2− values for three surface water samples varied from −5.40‰ to 4.79‰, with an average of 1.11‰ while seven samples from the acidic water at coal mines showed a range of δ 34 S-SO 4 2 from −6.6‰ to 0.3‰, with an average of −3.3‰. The δD values ranged from −66‰ to −57‰, with an average of −61.7‰. Further, the δ 18 O values were from −8.9‰ to −7.4‰, with an average of −8.2‰.

Sources of K + , Na + , and Cl −
The main natural sources of K + , Na + , and Cl − are atmospheric precipitation, human activity emissions, and dissolution of salt rocks. Figure 8 shows the concentrations of Na + , K + , and Cl − in precipitation in the spring field were in ranges of 1.4 to 23.4 mg·L −1 , 0.25 to 9.2 mg·L −1 , and 1.6 to 24.7 mg·L −1 , respectively. The concentrations of Na + , K + , and Cl − in the surface water of the spring field showed ranges of 40.2 to 523.4 mg·L −1 , 3.3 to 18.0 mg·L −1 , and 3.1 to 144.0 mg·L −1 , respectively. In contrast, the concentrations of Na + , K + , and Cl − for the acidic water at coal mines were 23.7 to 343.0 mg·L −1 , 0.1 to 3.8 mg·L −1 , and 3.1 to 144.0 mg·L −1 , respectively. This study also measured the concentrations of Na + , K + , and Cl − for the karst groundwater, which showed ranges of 6.26 to 193.6 mg·L −1 , 0.8 to 6.6 mg·L −1 , and 6.0 to 164.9 mg·L −1 , respectively. As shown in Figure 9, Cl − has a positive correlation with Na + and K + in karst groundwater. Na + , K + , and Cl − mainly originated from urban domestic sewage or coal mine drainage [59][60][61].

−
This study showed that the karst groundwater has the main ions of Ca 2+ , Mg 2+ , and HCO 3 − , which are mainly derived from dissolution of carbonate rocks and gypsum as well as dedolomitization. There are large differences in the mechanisms of groundwater runoff.

Dissolution of Carbonate Rocks and Gypsum
The dissolution of calcite and dolomite in Ordovician strata is the main source of Ca 2+ , Mg 2+ , and HCO 3 − . The dissolution equations are shown in Equation (1): The dissolution of gypsum may cause an increase of Ca 2+ and SO 4 2− in the karst groundwater. Figure 10a shows a strong positive correlation between Ca 2+ and Mg 2+ in the karst groundwater. Figure 10b also shows a strong positive correlation between Ca 2+ and SO 4 2− in the karst groundwater, indicating that the dissolution of gypsum plays an important role in the concentration of Ca 2+ and SO 4 2− .

Dedolomitization
As the solubility of CaSO 4 is greater than that of CaCO 3 and CaMg(CO 3 ) 2 due to the coion effect, the precipitation of calcium carbonate (CaCO 3 ) will occur in order to maintain the equilibrium of Ca 2+ concentration. When the precipitation of CaCO 3 takes away part of the HCO 3 − from the solution, dolomite will be dissoluted to maintain the mass balance of HCO 3 − concentration, which is called dedolomitization [62,63].
We assumed that the dedolomitization is real. Dedolomitization as described in Equation (1) leads to an increase of concentration in SO 4 2− and Mg 2+ in the karst groundwater and precipitation of calcite; subsequently, the mass concentrations of HCO 3 − and Ca 2+ remain unchanged. Figure 11 shows that SO 4 2− and Mg 2+ have a strong correlation (R 2 = 0.87), demonstrating that dedolomitization is the main source of Mg 2+ .

Water-Rock Interaction
The chemical composition of the karst groundwater in the Niangziguan spring field is mainly dependent on the dissolution of calcite, dolomite, and gypsum. Figure 12a shows the relationship between SO 4 2+ +Cl − and HCO 3 − .The points below the 1:1 line represent that the dissolution of evaporite is accompanied. The results imply that the karst groundwater in the Niangziguan spring field is mainly affected by the dissolution of carbonate rocks, along with the dissolution of evaporites (e.g., gypsum, Cl − in AMD or rivers, etc.). Figure 12b also shows the relationship between SO 4 2+ +HCO 3 − and Ca 2+ +Mg 2+ . Most analytical points are lying on the straight line or below the 1:1 line, indicating that karst groundwater is affected by acid drainage and surface water. The relationship between Na + +K + -Cl − and Ca 2+ +Mg 2+ -SO 4 2+ -CO 3 − is shown in Figure 12c. Most of the analytical points from the runoff area lie near the cation exchange line, which means that the hydrochemistry of groundwater is dependent on not only the dissolution of calcite, dolomite, and gypsum but also on cation exchange. Figure 12d shows the relationship between Ca 2+ /Na + and HCO 3 − /Na + to evaluate the effects of materials from different sources (evaporite, silicate, precipitation, and carbonate rocks) in the processes of the groundwater circulation. The result demonstrates that the main factors influencing the hydrochemical features of the karst groundwater in the Niangziguan spring field are precipitation and dissolution of carbonate rocks.
The TDS values, a comprehensive indicator of water quality, obtained from 52 water samples in the karst groundwater within the study area ranged from 280.3 to 1859 mg·L −1 , an average of 775.7 mg·L −1 . The TDS values in the karst groundwater are larger in order of the recharge area, runoff area, discharge area, and the detention area. Figure 13 show the relationship between TDS and SO 4 2− . It is prominent that TDS has a strong linear correlation with SO 4 2− contents (R 2 = 0.922). SO 4 2− in regional karst groundwater is mainly derived from the dissolution of gypsum in the Ordovician carbonate rocks while SO 4 2− in acidic water is from coal mines and meteoric precipitation. The average TDS values in the Niangziguan spring field is 688.6 mg·L −1 , lower than the average values analyzed in this study, because the TDS values in the karst water are affected by the runoff condition, the dissolution of gypsum, and the pollution caused by acidic water from coal mines.

Supply Source of Karst Groundwater
The δ 34 S-SO 4 2− isotope has been widely used to trace the source of sulfate in water. In addition, H-O isotopes are ideal tracers to detect the origin and evolution of groundwater. Thus, these stable isotopes in groundwater were used to identify the recharge sources of groundwater in the study area.

Recharge Sources
The δD values for the karst groundwater samples in the study area ranged from −74‰ to −49.8‰, an average of −68.62‰ while the δ 18 O values ranged from −10.1‰ to −5.8‰, an average of −9.25‰. Figure 14 show the relationship between δD and δ 18 O. The result shows that the intercept and slope of the δD-δ 18 O relation line are both lower than the global precipitation line (δD = 8δ 18 O + 10) due to an inland arid climate in the study area. The results also show that meteoric precipitation is the main recharge source of karst water in the Niangziguan spring field. The H-O isotopic composition of surface water was similar to that of the karst groundwater in the runoff area, implying that the karst groundwater is mostly recharged from the seepage of surface water. While the surface water showed SO 4 2− concentrations ranging from 183.2 to 4042 mg·L −1 , an average of 1089.6 mg·L −1 , SO 4 2− concentrations for the acidic water in coal mines ranged from 1138 to 16218 mg·L −1 , an average of 5079 mg·L −1 . In contrast, the karst groundwater showed SO 4 2− concentrations ranging from 21.7 to 1043 mg·L −1 , an average of 323.9 mg·L −1 (Figure 15). Precipitation and dissolution of gypsum, oxidation of FeS 2 in coal-bearing strata are the sources of SO 4 2− in the karst groundwater. However, SO 4 2− in precipitation (23.95 mg·L −1 on average) has little influence on the karst groundwater. Therefore, it was assumed that most SO 4 2− is sourced from both oxidation of FeS 2 and dissolution of gypsum. The δ 34 S values of SO 4 2− formed dissolution of gypsum and formed oxidation of FeS 2 are 26.2‰ (the mean value of gypsum samples) and −3.3‰ (the mean value of AMD in Niangziguan spring catchment), respectively ( Figure 16). A 34 S isotopic analysis was conducted to determine the source of SO 4 2− by calculating the proportion of SO 4 2− sourced from FeS 2 oxidation in the karst groundwater by mass balance. As shown in Table 3, the proportions for the main discharge points (N50, N51, and N52) were 60.6%, 30.3%, and 26.0%, respectively.

Conclusions
The karst groundwater in the Niangziguan spring field was mainly classified into SO 4 ·HCO 3 -Ca·Mg, HCO 3 ·SO 4 -Ca·Mg, and SO 4 -Ca types. The SO 4 2− exceeding ratio of karst groundwater samples in this study was 51.9%. K + , Cl − , and Na + are mainly sourced from urban sewage and coal mine drainage. Dedolomitization gradually led to an increase of Mg 2+ , especially in the processes of the groundwater runoff. While the main sources of SO 4 2− supply were the dissolution of gypsum and the oxidation of FeS 2 in coal-bearing strata, the hydrochemical characteristics of the study area were mainly affected by dedolomitization, the dissolution of dolomite, salt, rock and gypsum.
The isotopic analyses found that acidic water in coal mines contributed to the pollution of the karst groundwater at different degrees. In the groundwater of the Niangziguan spring field, the proportions of SO 4 2− derived from FeS 2 oxidation were 60.6% (N50), 30.3% (N51), and 26.0% (N52), respectively.
The acid water of coal mine directly recharges and pollutes karst groundwater through faults or abandoned boreholes, or discharges to rivers and indirectly pollutes karst groundwater through river infiltration in carbonate exposed areas. The main cause for the rapid growth of sulfate in karst groundwater is acid water from abandoned coal mines. Institutional Review Board Statement: "Not applicable" for studies not involving humans or animals.
Informed Consent Statement: "Not applicable" for studies not involving humans.
Data Availability Statement: All data, models, and code generated or used during the study appear in the submitted article.

Conflicts of Interest:
The authors declare no conflict of interest.