Temporal Variations of Water Chemistry in the Wet Season in a Typical Urban Karst Groundwater System in Southwest China

Abstract

It is important to investigate temporal variations of water chemistry for the purpose of improving water quality in karst groundwater systems. Groundwater samples were collected daily under various land uses of Guiyang. Major ions and stable carbon isotope composition of dissolved inorganic carbon (δ¹³C_DIC) were analyzed to understand the biogeochemical processes. The water chemistry was dominated by Ca²⁺, Mg²⁺, HCO₃^-, and SO₄^2-, which mainly derived from the dissolution of carbonate rocks (limestone and dolomite) and oxidation of sulfide. The groundwater was defined as of the HCO₃-Ca Mg and HCO₃·SO₄-Ca·Mg type, according to its hydrochemical characteristics. Results suggested that hydrochemical concentrations changed quickly, in response to rainfall events. The fast response revealed that karst groundwater was easily impacted by rainfall and anthropogenic inputs according to temporal variation of water chemistry. The distribution of DIC (dissolved inorganic carbon) and δ¹³C_DIC showed that DIC is mainly sourced from soil CO₂(g) influx and carbonate dissolution. δ¹³C_DIC and major ions ratios suggested that carbonate minerals were dissolved by H₂SO₄ at groundwater in wooded area, contributing an important source for DIC due to the slight enrichment of heavy δ¹³C_DIC. More negative δ¹³C_DIC values were observed after rainfall reflected the fact that soil CO₂(g) and organic carbon oxidation influxes accounted for a large share during DIC formation. Various δ¹³C_DIC and hydrochemical patterns were observed under various land use and human activity conditions. Meanwhile, relative high nitrate loads were found in groundwater after rainfall, suggesting high anthropogenic inputs following rainwater as having side effects on water quality. This study suggests that water chemistry and isotopic proof provide a better understanding of water quality and carbon dynamics responding to rainfall events in the karst groundwater systems.

1. Introduction

Karst aquifers are important for supplying a large amount of fresh water for the whole world [1]. The karst landscape covers nearly 15% of the Earth’s land area, providing drinking water for 1/4 of the world’s population [2]. Groundwater is important for drinking and irrigation in the karst area, therefore, understanding the biogeochemical processes in karst areas is extremely significant [3,4,5,6]. Groundwater is an important part in the karstic ecosystem, but it is particularly sensitive to environmental changes and human activities [5,7,8]. With the acceleration of urbanization and industrialization, the eco-environment in karst areas is susceptible to being affected [9,10].

Stable carbon isotope (δ¹³C) and water chemistry parameters, can be acted as important tools to understand the biogeochemical processes in groundwater systems [5,6]. Karstic water is one of the most important groundwater resources supplying drinking water for more than 100 million people in Southwest China [8,11]. The aquifers provide a natural storage and movement for groundwater. In the studied region, the soil layers are thin and lack enough filtration for contamination. With the development of karst folding, a fault, and a fracture under the complex hydrogeological conditions, the contaminants on the earth’s surface could seep into aquifers directly or indirectly [9]. The special structure would result in changing the chemical composition drastically and leading to the deterioration of groundwater quality. Groundwater is a “hidden reservoir” of pollutants, and it is difficult to monitor. Springs are natural outcrops of groundwater and main drain [6,12]. The contaminants in aquifers will eventually show up in springs. Therefore, springs naturally become the monitoring medium for contaminants and the hydrogeochemical processes of groundwater in karst aquifers can be fully understood.

Hydrochemical composition is the basis for probing sources, runoff, and the eco-environment, and is important content for assessing groundwater resources [13]. It is of much guiding significance for utilizing and managing groundwater resources in the drainage basin, and the related protection of the eco-environment and construction [14,15]. The δ¹³C values of different carbon sources are differentiated obviously, and the environmental changes of groundwater would influence biogeochemical cycles of carbon and isotopic composition [4,5,6]. The average δ¹³C being added to the SO₄²⁻-rich groundwater is −9.4‰, while the average δ¹³C being added is −11.6‰ for the HCO₃-rich groundwater, due to dedolomitization and soil CO₂(g) dissolution in groundwater [16]. The δ¹³C not only discern the sources of DIC in groundwater, but also identify the sources of contamination in groundwater as well as the biogeochemical processes, which influence the migration and fate of contaminants [5,10]. It was pointed out that the heaviest δ¹³C values of -8.5‰ in an aquifer containing gypsum, lowest δ¹³C values of −14.4‰ in an aquifer contaminated by organic pollutants [5]. The hydrochemical and isotopic detection of spring provides the method to obtain the information on recharge of groundwater, formation lithology, water-rock interaction, and land-use type that greatly impact the quality of ground water [5,6,11].

In order to clarify the regional recharge source for groundwater and hydraulic connections in a karst aquifer, we provide reasonable and effective suggestions for management and utilization of groundwater resources. The main objectives of this study are: (1) to analyze the hydrochemical characteristics responding to rainfall in this area, determine the subsequent reactions on the carbonate evolution, and further probe the hydrogeochemical evolutionary process in groundwater, thereby comprehensively understanding the impact of human activities on karst aquifer; (2) to understand the DIC sources in fresh groundwater and document the spatial and temporal characteristics for chemical components and stable carbon isotopes in Guiyang, Southwest China.

2. Materials and Methods

2.1. Study Area

The city of Guiyang is situated at 106°07´E-107°17´E longitude and 26°11´N-26°55´N latitude, in the centre of Guizhou Province, Southwest China. It covers an area of 2960 km², with an average elevation of around 1400 m, characteristic of plateau hilly landforms [5,10]. Guiyang is located close to the boundary, separating the Changjiang River and Pearl River basins, occupying a wide karstic valley basin in the central part of Guizhou Province. Permian and Triassic carbonate rocks are widely distributed in the region; this kind of geology took an area over 70%, the lithology is mainly dolomite and limestone, and part of it contains organic- and sulfide-rich limestone [11]. Groundwater is rich in the terrain with karst development. The Guiyang basin is restricted and controlled by the combination of the topographic features of folding, a fault, and fracture structure planes, forming a karst underground water network model and groundwater accumulation area. Woodland, arable land, and urban land are the main land-use patterns within the sampling region. The climate in Guiyang city is subtropical with an average temperature of 16 °C and annual precipitation of about 1200 mm. A monsoonal climate often results in high precipitation during summer and much less during winter at the studied region. Rainfall is concentrated from May to August and the dominant recharge to ground water is dependent on rain water [8,9].

2.2. Sampling and Analytical Method

Ground water samples were collected in August, 2011, corresponding to the high-flow seasons. Four sampling sites were selected situated at ZZ (Zhen Zhu, name of site) where it is newly urbanized with some cropland. NN (Nai Niu, name of site) is partially forested with some residents in the immediate surroundings, who cultivate crops and raise cows. ZZ and NN are gravity springs, with a mean discharge of 41.2 L/s and 1 L/s [11], respectively. DJ (Dong Jiao, name of site) is artesian spring having a discharge of 1571 L/s and located in a large area of woodland at the mountain zone [11]. DH (Di Hua, name of site) is a well with a discharge of 6.3 L/s and sits in the centre of the city (Figure 1). The Nanming River is the trunk stream, and the water flow direction is from southwest to northeast (Figure 1). Groundwater samples were collected every day in one month, and in total, over one hundred samples were collected. Meanwhile, the daily data of rainfall were collected synchronously during the sampling period through website of local meteorological bureau (http://gz.cma.gov.cn/). Temperature, electrical conductivity, and pH were measured in situ. Alkalinity was measured by titration with 0.02M HCl within 12 h. Major cations (Mg²⁺, Ca²⁺, Na⁺, and K⁺) were analyzed by atomic absorption spectrometry once samples were filtered and acidified, and anions (SO₄²⁻, NO₃⁻, and Cl⁻) were analyzed by high-performance liquid chromatography.

For the determination of δ¹³C_DIC, samples were filtered through 0.45 μm cellulose acetate filter paper in order to avoid the potential influence of carbonate mineral particles. Some water samples were selected for isotopic analysis due to limitations of cost and time in this study. Using the method of Li et al. (2005) [5], 10 mL water sample was injected into glass bottles that were prefilled with 85% phosphoric acid and a magnetic stir bar by syringe. Then CO₂ was extracted and purified into a vacuum line. Finally, the purified CO₂ was transferred cryogenically into a tube welded for isotopic measurements. Carbon isotope ratios were determined on a Finnigan MAT 252 mass spectrometer (Thermo Finnigan GmbH Inc., Bremen, Germany) and reported in the δ notation relative to Vienna Pee Dee Belemnite (PDB) in per mill, (see Equation (1)):

δ_sample(‰) = ((R_sample ÷ R_VPDB)−1) × 1000

(1)

where R is the ¹³C/¹²C atomic ratio. Routine δ¹³C_DIC has an overall precision of 0.1‰. 10% of duplicate samples were measured, and the results showed that the differences were lower than the measurement accuracy.

3. Results

3.1. Hydrochemical Characteristics of Groundwater

The pH values of groundwater ranged from 6.0–8.0, which reflected the influence of the soil water with high CO₂ concentrations. Under the observed pH conditions, bicarbonate (HCO₃⁻) was the dominant DIC species [9,17,18] and the ion abundance was related to the dissolution equilibrium of the carbonate minerals. The average TDS (Total dissolved solids) values were 564 mg·L⁻¹ for groundwater in summer [11]. The daily rainfall ranged from 0–89.6 mm, and averaged at 11.99 mm.

The major ion concentrations of samples are listed in Table S1. The charge balance errors between cations and anions in groundwater were equivalent to or less than 5%. It was obvious that Ca²⁺ and Mg²⁺ were the dominant cations, and HCO₃⁻ and SO₄²⁻ were the dominant anions (Figure 2), which accounted for 96% and 92.5% of cation and anion in milliequivalents (mEq), respectively. As illustrated in Figure 2, hydrochemical types of spring water in Guiyang were predominantly HCO₃–Ca·Mg and HCO₃·SO₄–Ca·Mg. The chemical compositions of groundwater were characterized by a anion sequence of HCO₃⁻ > SO₄²⁻ > NO₃⁻ + Cl⁻, accounting for 74%, 18%, and 7% of the total anions, respectively. The NO₃^- level of groundwater at the residential district was higher than that of woodland. NO₃⁻ was the most concentrated at DH with the proportion of 5% in total anions, with the absolute average amounts of 32.41 mg·L⁻¹. Similar to NO₃⁻, SO₄²⁻ was the highest at DH, with a range from 114.4–188.9 mg·L⁻¹ (averaging 144.7 ± 16.6 mg·L⁻¹). The SO₄²⁻ concentrations averaged at 25.9 ± 2.4, 33.35 ± 0.27, and 31.51 ± 4.1 mg·L⁻¹ at ZZ, NN, and DJ, with ranges of 20.45–35.42, 20.73–50.13, and 23.86–38.01 mg·L⁻¹, respectively.

HCO₃⁻ was relative high at NN and DH, with averages of 66.52 ± 1.97 and 65.80 ± 0.92 mg C·L⁻¹, whereas 55.69 ± 0.73 and 26.96 ± 1.01 mg C·L⁻¹ were found at ZZ and DJ, respectively. Cl^- occupied the highest and lowest levels at DH and DJ, averaging 29.02 ± 2.47 and 1.60 ± 0.35 mg·L⁻¹, respectively. Cl⁻ at ZZ and NN fell into the ranges of 3.5–5.8 and 3.47–5.64 mg·L⁻¹, averaging 4.24 ± 0.46 and 4.42 ± 0.58 mg·L⁻¹. For cations, the average concentrations of Ca²⁺ (64%) and Mg²⁺ (32%) were much higher than those of K⁺ and Na⁺ (4%). As the most abundant cation species, Ca²⁺ and Mg²⁺ at DH averaged 108.28 ± 10.06 and 41.95 ± 1.66 mg·L⁻¹, respectively. Followed by NN, Ca²⁺, and Mg²⁺ have mean value of 70.58 ± 5.19 and 31.43 ± 0.68 mg·L⁻¹, respectively. Ca²⁺ and Mg²⁺ have a range of 56.83–66.33 mg·L⁻¹ and 25.88–29.28 mg·L⁻¹ at ZZ, with average values of 61.57 ± 2.34 mg·L⁻¹ and 27.66 ± 0.63 mg·L⁻¹, respectively. Levels of 47.18–56.35 mg·L⁻¹ for Ca²⁺ and 3.38–4.38 mg·L⁻¹ for Mg²⁺ were observed, with averages of 52.86 ± 2.34 and 3.83 ± 0.29 mg·L⁻¹ at DJ, respectively. Na⁺ and K⁺ were the lower cationic species in the total cations and the gross of Na⁺ + K⁺ ranged from 18.86–22.43 mg·L⁻¹, 3.61–9.99 mg·L⁻¹, 2.71–5.85 mg·L⁻¹, and 1.4–2.13 mg·L⁻¹ at DH, NN, ZZ, and DJ, with mean values of 20.84 ± 1.03, 5.4 ± 1.68, 3.02 ± 0.58, and 1.66 ± 0.17 mg·L⁻¹, respectively.

3.2. DIC and δ¹³C_DIC

DIC concentrations were relative high at NN and DH, ranging from 60.91–68.4 and 62.9–66.96 mg C·L⁻¹, with averages of 66.52 ± 1.97 and 65.8 ± 0.92 mg C·L⁻¹, respectively. Narrow ranges of 52.3–56.54 and 25.31–28.84 mg C·L⁻¹ with averages of 55.69 ± 0.73 and 26.96 ± 1.01 mg C·L⁻¹ were observed at ZZ and DJ. The δ¹³C_DIC averaged −9.84‰ ± 0.05‰, −10.83‰ ± 0.19‰, −10.35‰ ± 0.13‰, and −11.30‰ ± 0.21‰ at ZZ, NN, DJ, and DH, and they correspondingly ranged from −9.97‰ to −9.73‰, −11.05‰ to −10.45‰, −10.49‰ to −10.17‰, and −11.56‰ to −10.93‰, respectively.

4. Discussion

4.1. Characteristics of Groundwater Chemistry and Water–Rock Reactions

The molar ratio of Mg²⁺/Ca²⁺ is a good indictor to reflect the lithology of aquifers that groundwater flows through [4,9]. When this value is much higher than 0.85, it indicates the dominant presence of dolomite sources. If this ratio was between 0.01 and 0.26, hydrochemical components in groundwater were mainly derived from limestone sources [9]. In this study, the averaged Mg²⁺/Ca²⁺ ratio was 0.71, much higher than 0.26 but lower than 0.85 for ZZ, NN, and DH, indicating that the hydrochemical components were mainly from the dissolution of both dolomite and limestone (Figure 3a). This ratio was much lower than 0.26 at DJ, with an average of 0.12, suggesting a limestone source at this sampling site. In addition, the molar ratios of Mg²⁺/Ca²⁺ at ZZ, NN, and DH varied in a narrow range, from 0.6–0.8. Whereas SO₄²⁻ varied in a large range, the molar ratio at DJ was much lower (Figure 3b). The result indicated that gypsum should not be considered as the main source of SO₄²⁻ and Ca²⁺. SO₄²⁻ in groundwater can be sourced from oxidation of sulfide minerals, which are widely distributed in coal-containing strata [11]. In addition, acid rain events often occur in Guiyang, characterized by a high H₂SO₄ content, and this is probably also an important source of SO₄²⁻ [19].

The Ca²⁺ and Mg²⁺ concentrations were much higher at DH than those found in other sites, because of the employment of exogenous acid increasing carbonate dissolution in summer. Moreover, domestic sewage discharge, sulfuric acid rain, etc., resulted in increased H⁺ and accelerated carbonate weathering [20,21]. As elucidated in Figure 3c, (Mg²⁺ + Ca²⁺)/(HCO₃⁻ + SO₄²⁻) was basically around 1, implying that dissolution of carbonate and S-containing minerals co-affected the hydrochemical compositions of groundwater in Guiyang. It is in agreement with previous studies in Southwestern China [22,23]. This ratio occasionally (from 17–28 August) slightly exceeded 1 at DH, which means that this system needed other anions to balance the excess Ca²⁺ and Mg²⁺, especially since this ratio rose to 1.14 at 23rd and 24th of August. Another possible reason for balance would be high nitrate from nitrification [21] due to high human activities [8] in studied sites. Major cations and anions have relative low concentrations in this study when compared to data at these fours sits at the dry season [5], which suggested dilution processes play an important role in water chemistry at karst ground water.

Both H₂CO₃ and H₂SO₄ can play a part in minerals’ dissolution according to reaction formula (Equation (2)) [22,24]. Figure 4 shows the variations of (Ca²⁺ + Mg²⁺)/HCO₃⁻ vs. SO₄²⁻/HCO₃^- equivalent ratios for the water samples. The transformation mechanism for them was interpreted as follows:

Ca_xMg_1−xCO₃ + H₂SO₄ + H₂CO₃ = 3xCa²⁺ + 3(1 − x)Mg²⁺ + SO₄^2- + 4HCO₃^-

(2)

The chemical reactions would be most likely responsible for the water chemistry. For water samples with unified (Ca²⁺ + Mg²⁺)/HCO₃⁻ and low SO₄²⁻/HCO₃⁻ ratios, carbonate dissolution by H₂CO₃ dominated the mineral/water interaction. With the increasing SO₄²⁻/HCO₃⁻ ratios, the (Ca²⁺ + Mg²⁺)/HCO₃⁻ ratio also increased, and more SO₄²⁻ was needed to balance the excess Ca²⁺ and Mg²⁺ in the water samples.

When carbonate dissolution by both H₂CO₃ and H₂SO₄, water should have a SO₄²⁻/HCO₃⁻ ratio of 0.5 and a (Ca²⁺ + Mg²⁺)/HCO₃⁻ ratio of 1.5, as shown by Figure 4. When both SO₄²⁻/HCO₃⁻ and (Ca²⁺ + Mg²⁺)/HCO₃⁻ ratios continued to increase, dissolution of gypsum was needed to balance the negative and positive ions. The covariation of (Ca²⁺ + Mg²⁺)/HCO₃⁻ and SO₄²⁻/HCO₃⁻ ratios can also be interpreted as binary mixing of HCO₃⁻ type and SO₄²⁻ type water, revealing that water chemistry originated mainly from carbonate dissolution by H₂CO₃ and H₂SO₄.

The proportion of elements can be used to investigate the water–rock interaction and the chemical evolution of groundwater [9]. Generally, studying on the affinity of water chemistry throws light on the geochemical characteristics of the aquifer. The incorporation of industrial and urban NaCl and potash fertilizers, manure and N-containing fertilizers used in agricultural production into aquifers will lead to the increase of K⁺ and Na⁺ in groundwater. As shown in Figure 5, the sewage samples at DH showed low K⁺/Na⁺ ratios, meaning that sewage was strongly enriched in Na⁺. The relationship between Na⁺/HCO₃^- and K⁺/HCO₃^- at ZZ and NN indicated that groundwater has received significant agricultural input. The correlation between Na⁺/HCO₃⁻ and K⁺/HCO₃⁻ was not observed at DJ and DH, and the characteristic of the hydrochemical components will be explained below.

4.2. Response of Groundwater to Rainfall Events and Evolution of Water Quality

Cl⁻ is conservative, showing an increasing trend from the upper reaches of the Nanming River due to anthropogenic inputs [11]. Halite and marine sources were not the main sources of Cl⁻, but from agricultural fertilizers, domestic sewage, animal manure, and Cl₂ disinfection treatment of tap water [25]. Figure 6 shows the temporal variations of major ions during the sampling period. The ions, which were the result of anthropogenic inputs, included NO₃⁻, Cl⁻, Na⁺, K⁺, and performed consistently with that of SO₄²⁻ at ZZ, NN, and DH (R² = 0.75, p ≦ 0.01), with an exception at DJ, which was covered with a large area of vegetation (Figure 6). Once the concentration of SO₄²⁻ exceeds 100 mg·L⁻¹ (average level 144.8 mg·L⁻¹), this means that the anthropogenic input of SO₄²⁻ is from the sewage effluent rather than the natural sources at DH. SO₄²⁻ at ZZ, NN, and DH exhibited obvious anticorrelations with HCO₃⁻ (R² = 0.72, p ≦ 0.01), which implied that H₂SO₄ played an important role in DIC production, mediating in mineral weathering such as the oxidation of S-containing coal strata or organic matter in sewage effluent [26]. Contrarily, SO₄²⁻ at DJ varied correspondingly to HCO₃^- (R² = 0.43, p ≦ 0.01), due to the carbonate dissolution by H₂SO₄ and H₂CO₃ in the region with less anthropogenic input that led to SO₄²⁻ and HCO₃⁻ production of similar variation trends (Figure 6). This viewpoint further confirmed the finding that SO₄²⁻ originated from the decomposition of manure, fertilizer, and municipal sewage in groundwater samples at ZZ, NN, and DH. DJ is situated in woodland and has not been influenced by industrial and residential effluent. However, it is an exception at DJ, in that HCO₃⁻ and SO₄²⁻ exhibited a consistent trend and without any correlations to other ions, demonstrating that SO₄²⁻ had a significant natural source from mineral weathering according to reaction formula (Equation (2)).

Atmospheric precipitation is the main source of regional groundwater, and surface water is hydraulically closely linked to groundwater [25]. It has been reported that the average concentrations of major ions in summer were generally lower than other seasons in Guiyang groundwater, due to the influx of infiltrated rain with low solutes contents [11]. As illustrated in Figure 6, rainfall influenced the ionic compositions in that HCO₃⁻ decreased abruptly with the heavy rain at 16th, August. Obviously, HCO₃⁻ from natural sources was acutely diluted by the heavy rainfall due to the large rain exchange flux, and impacted by the removal of the total quantity from weathered rocks [5]. However, frequent rainfall did not change ion concentration largely in the following days due to adaption of water–rock reaction in high rainfall periods. Whereas SO₄²⁻, K⁺, Na⁺, and NO₃⁻ increased quickly under heavy rain, similar to the findings of other study [27] that this group of ions were from anthropogenic discharge, washed and infiltrated into river water and elevated significantly during rainfall events. Although SO₄²⁻ at DJ originated similarly from mineral dissolution, which was not as susceptible to being affected by heavy rain as HCO₃⁻, it is probably due to occasional manure and fertilizer decomposition, or else a contribution from acid rain.

The relationship between ions and meteoric water could be used to constrain the sources of contaminants and their conversions. The hydrogeological features of karst areas permit the rapid penetration of pollutants into groundwater, so karst ground waters are more vulnerable to being polluted from human activities than other types of aquifers [25]. Groundwater in karst areas has been increasingly polluted by penetration of pollutants, e.g., uncontrolled urban sewage effluent and intensive utilization of fertilizers on cultivated land [15,25]. For HCO₃⁻, Ca²⁺, and Mg²⁺, there is a relative low ion content in rainwater compared with that in groundwater [19]. The infiltrated meteoric water diluted and eroded the concentrations of HCO₃⁻, Ca²⁺, and Mg²⁺, but NO₃⁻, SO₄²⁻, K⁺, and Na⁺ were eluted out and leached into groundwater mainly due to anthropogenic input from agricultural fertilizer, industrial waste, and domestic sewage following the rainfall [21]. Thus, relative high flow during the rainfall periods can lead to rapid leaching of surficial fertilizers/manure into the groundwater system through conduits due to weak buffering capacity in the studied area. Of course, the different response to rainfall in various groundwater sites suggested different role of high heterogeneity of karst system and extent of human activities as well as multi biogeochemical processes.

4.3. DIC Sources and the Controlling Factors in Groundwater

In the groundwater, the sources of DIC are from (1) the dissolution and dissociation of CO₂(g), mainly from root respiration and soil organic matter decay, which is infiltrated in subsurface water systems, and (2) the dissolution of carbonates by organic acids excreted by plant roots, and carbonic acid produced by the dissolution of the CO₂ in water [13,28,29]. The dissolution of carbonates (calcite and dolomite) is a major source of Ca²⁺ and Mg²⁺, which also generate DIC in groundwater. These are expressed as the following reactions (Equation (3)) and (Equation (4)).

CaCO₃ + H₂O+CO₂(g) ⇿ Ca²⁺ + HCO₃⁻

(3)

CaMg(CO₃)₂ + 2H₂O + 2CO₂ ⇿ Ca²⁺ + Mg²⁺ + 4HCO₃⁻

(4)

The DIC sources are also likely to be organic residuals that are degraded by microbes, which will generate DIC to the water [18,26]. A large of aerobic and anaerobic populations predominates the downgradient from the pit water contaminated source [26]. As CO₂(g) is generated by the microbiological degradation of hydrocarbons, which added to the CO₂(g) generated from the organic matter respiration in the root zone, then the soil CO₂ dissolves carbonates to produce DIC in the aquifer. However, hydrocarbon microbial mineralization should not be an important source of DIC [16]. Guiyang suffered from acid rain, which contains high sulfate [19]. Since sulfide minerals oxidation in the coal-containing strata [9], sulfuric acid plays a significant role in dissolving the carbonate rocks in this study area, producing DIC in groundwater.

Soil CO₂(g), from respiration of C-3 plant, dissolving in groundwater produces DIC, with a δ¹³C_DIC value of −23‰ [4,30]. It is reported that Triassic and Permian limestone and dolomite are the most widely distributed strata in the studied area, with δ¹³C values ranging from −1.8–4.8‰ [31]. Supposing that δ¹³C only originated from soil CO₂ influx and carbonate rock weathering, the δ¹³C value should be −11‰ [4,23]. δ¹³C values were much higher than −11‰ at ZZ, NN, and DJ, whereas it fluctuated around −11‰ or even lower at DH. The δ¹³C values for all samples were lower than −9‰, suggesting the contribution of soil CO₂ influx. DIC generally has more negative δ¹³C value comparing to the data in the dry season from previous study [5], indicating DIC in the wet season has a relatively enriched biogenic origin relative to the dry season.

In this study, HCO₃^- is more than 95% of DIC based on water chemistry and pH. Milliequivalent ratios (MER) of (Ca²⁺ + Mg²⁺)/HCO₃⁻ of the groundwater in Guiyang were in the range of 1–2, averaging 1.16, 1.11, 1.32, and 1.63 at ZZ, NN, DJ, and DH, respectively. The SO₄²⁻/HCO₃⁻ ratios averaged 0.12, 0.13, 0.29, and 0.55 at ZZ, NN, DJ, and DH, respectively. The relationship between (Ca²⁺ + Mg²⁺)/HCO₃⁻ and SO₄²⁻/HCO₃⁻ indicated that H₂SO₄ acted a role in the mineral weathering. In addition, HCO₃^- normalized SO₄²⁻ value decreased in the rank of DH > DJ > NN > ZZ, suggesting that H₂SO₄ was decreasingly involved in carbonate weathering, as elucidated in Figure 4 [6]. Samples at DH showed a relatively high SO₄²⁻/HCO₃⁻ ratio and the lowest δ¹³C value (Figure 7). These results corresponded to the discussion of previous study [11], showing that since SO₄²⁻ was mainly derived from oxidation of sulfide minerals or mineralization of organic sulfur. In this study, groundwater would be characterized by a high SO₄²⁻/HCO₃⁻ ratio, but δ¹³C-depleted DIC value. The carbonate evolution with the average δ¹³C_DIC of −11.3‰ at DH was undergoing in a closed system from carbonate weathering by soil CO₂(g) [16]. The δ¹³C of −9.8‰ at ZZ was similar to the δ¹³C of −9.4‰ in SO₄²⁻-rich water, both due to dedolomitization referred to previous study [16], that high δ¹³C values were relative to the result of the other sampling sites. [NO₃⁻]/[HCO₃⁻] averaged 0.03, 0.05, 0.03, and 0.1, likewise, displayed the same variation trend with (Ca²⁺+Mg²⁺)/HCO₃⁻ at NN and DH, resulting in the lower δ¹³C values at NN and DH (Figure 7). It was inferred that carbonate weathering could be driven by nitric acid. However, NO₃^- might play a minor role in shifting δ¹³C-depleted values due to oxidization of organic matter in sewage effluent at site of DH.

As shown in Figure 7, NO₃⁻ played an important role in HCO₃⁻ dynamics only at NN, reflecting on the anticorrelation between them (R² = 0.72, p ≦ 0.01). It was postulated that the transformation from organic carbon to DIC occurred with NO₃⁻-mediating in the redox reactions, shifting negative δ¹³C_DIC values. Probably this could be taken as a plausible explanation for the lower δ¹³C value at NN than that at DJ. Utilization of manure and nitrogenous fertilizer produces H⁺ during their oxidation, and sulfides in coal seams are oxidized into sulfuric acid, both of which promote the weathering of carbonate rocks and result in high NO₃⁻ and SO₄²⁻ concentrations and δ¹³C-enriched DIC values in karst waters [21,23,25]; therefore, nitric acid and sulfuric acid were also the driving factors for mineral dissolution and heavy δ¹³C. In addition, mineral dissolution and organic matter decomposition caused the δ¹³C values to fluctuate [17,32]; δ¹³C values deviated off −11‰ toward a more positive value, significantly negatively correlated to NO₃⁻/HCO₃⁻ and SO₄²⁻/HCO₃⁻ ratios at NN (R² = 0.61, p ≦ 0.01), explained by the above elaboration that carbonate dissolution was significantly shifted by H₂CO₃, H₂SO₄, and HNO₃, along with organic matter being oxidized into DIC by SO₄²⁻ and NO₃⁻. This phenomenon was not observed at ZZ, DJ, and DH, which further demonstrated that carbonate weathering by SO₄²⁻ and NO₃⁻ and organic carbon transformation were not the only factors influencing the δ¹³C_DIC. The δ¹³C_DIC value of groundwater in Guiyang is mainly influenced by anthropogenic inputs following by carbonate rock weathering and soil CO₂(g) influx. δ¹³C_DIC decreased with the increasing quotients of SO₄²⁻ and NO₃⁻ vs. HCO₃⁻ at these sampling sites, indicating that H₂SO₄ and HNO₃ were not only involved in carbonate mineral dissolution, but also in organic matter oxidation. In addition, soil CO₂(g) influx shift DIC with δ¹³C-depleted values.

5. Conclusions

This study investigated temporal and spatial variations of water chemistry at the rainy season in a typical karst groundwater system, SW China. Groundwater had a high content of Ca²⁺, Mg²⁺, HCO₃⁻, and SO₄²⁻ in the study region, which accounted for more than 90% of total ions and were largely derived from dissolution of carbonate rocks. The chemical composition of ground water was mainly controlled by rock dissolution. Anthropogenic inputs into the groundwater mainly include SO₄²⁻, Cl⁻, and NO₃⁻, according to the chemical composition of municipal sewage in Guiyang. Domestic and industrial waste was the main contaminant sources of groundwater in urban areas, while agricultural fertilizers and the transformation products polluted groundwater in suburbs.

The study showed that carbonate evolution in freshwater aquifers could be traced by investigating the major ions and δ¹³C_DIC. The coupled analysis of δ¹³C_DIC and hydrochemical parameters is an effective approach to explore the biogeochemical processes of carbon and trace the sources of groundwater pollutants in karst areas. Groundwater sampled from the city center (DH) showed the highest ionic level and lowest δ¹³C values relative to the other samples due to undergoing different carbonate evolution.