Vertical Divergence Characteristics of Dissolved Inorganic Carbon and Influencing Factors in a Karst Deep-Water Reservoir, Southwest China

Abstract

In deep karst reservoirs, the internal environment is complex, and thermal stratification isnot the only factor controlling the vertical distribution of the DIC concentration. Previous studies have not fully understood the migration and transformation of DIC in a deep-water reservoir. In this study, a deep-water reservoir in southwest China was chosen, and the spatial and temporal characteristics of the DIC concentration, pCO₂, δ¹³C_DIC value, and SIc were investigated. It was found that the Pingzhai Reservoir is a double temperature leapfrog reservoir. The DIC concentration, pCO₂, Sic, and δ¹³C_DICvalues showed annual cycle variation. During the thermal stratification phase, the DIC concentration, pCO₂, Sic, and δ¹³C_DICvalues were significantly different between the surface layer and the lower layer. However, during the mixing and mixed phases, the differences were not significant. The vertical divergence of the DIC in the Pingzhai Reservoir was influenced by the subtemperate layer, human activities, and sources. The formation of the subtemperate layer was due to the submerged flow formed when river water enteredthe reservoir, which provides a channel for DIC from the river to enter the lower layer of the reservoir. Human activities increased the solubility of carbonate rocks in the reservoir, and the source of DIC was one of the factors contributing to the concentration stratification of DIC in the reservoir.

1. Introduction

Reservoirs play a significant role in regulating the climate [1,2] and changing engineering water scarcity [3]. The construction of reservoirs is a method of solvingwater scarcity in karst areas. As of 2020, there were 98,566 reservoirs in China of which 53% werelarge reservoirs and 58% weremedium reservoirs inkarst areas (http://www.mwr.gov.cn/sj/) (accessed on 30 June 2023). Theconstruction of reservoirs has altered the continuity of rivers [4], affected the forms of biogenic elements [1,5] andthe ecological environment [6], and changed the structure of water bodies, hydrological features, and community structure of aquatic organisms [7,8]. The mechanism of reservoircarbon cycling is different from that of river carbon cycling. Therefore, the study of reservoir carbon cycling has becomea focus of carbon sink calculation and carbon balance studies [9,10].

Reservoirs play an important role in the exchange of matterand energy at the water–atmosphere interface [11], water–rock interface, and water–biological interface [12], as well as incarbon cycling and accumulation in watersheds [13], and they also play an important role in altering the carbon fluxes and pathways in reverie ecosystems [14]. In recent years, research on carbon cycling reservoirs has produced many results, for example, the exploration of computational methods and models for reservoir carbon fluxes [15,16], the assessment of the carbon sink effects of river–reservoir systems [17,18], the study ofthe effects of thermal stratification, biological pumps, and chemical stratification on carbon cycling [14,19,20]. Reservoirs carbon transport models have confirmed that complex biogeochemical changes occur within reservoirs [14,20], including the conversion of inorganic carbon to organic carbon by photosynthesis of aquatic organisms [21], the degradation and mineralization of organic matter to produce inorganic carbon [22], bacteria producing CH₄ under anaerobic or sulphuricconditions [23,24,25], and the degassing of dissolved inorganic carbon (DIC) to produce CO₂ [26,27]. However, carbon exchange is limited by many factors [18], such as water residence time [28,29], water level [6], thermal stratification [30], chemical stratification [31], and reservoir age [32]. The above findings supported the construction of a reservoir carbon cycle model. However, the applicability of this carbon cycle model in karst areas is limited because of the vulnerability of carbon transport in karst areas to human activities and other factors [33].

DIC in reservoirs is from rock weathering, soil carbon loss [1,5], aquatic respiration [12], and water–air exchange [11]. The DIC concentration in a karst reservoir is significantly higher than in non-karst areas [17]. Aquatic plants can autotrophically convert DIC into organic carbon [21], some of which is permanently stored in the substrate and some of which is released into the atmosphere as CO₂ or CH₄ [22,34]. Because of the influence of the environment, DIC degassed to produce CO₂ and the carbonate precipitation DIC concentration is also important for the calculation of CO₂ fluxes at the water–gas interface. This suggests that the DIC is important for the carbon transport among the hydrosphere and the biosphere, atmosphere, and lithosphere. It is useful to better understand the carbon cycle and carbon balance in karst basins if studying the migration and transformation of the DIC.

Many researchers have started to study the migration and transformation of the DIC in karst reservoirs [20,35]. They have confirmed that thermal stratification is one of the main factors controlling changes in the DIC [36,37]. Current studies on DIC in karst reservoirs have mostly focused on the characteristics of the surface distribution and influencing factors. However, most reservoirs in karst areas are built according to the terrain and dammed in deep river valleys, so the water depth is mostly more than 30 m. These reservoirs are prone to thermal and chemical stratification in summer. The internal environment of these reservoirs is also more complex, with the water pressure, wind, and hydrodynamics all potentially leading to the exchange of the upper and lower water bodies, even during periods of thermal stratification. Therefore, a study of the characteristics of the vertical distribution and influencing factors of DIC in karst deep-water reservoirs is necessary for an in-depth study of the reservoir carbon cycle and carbon balance. In this study, 63 vertical profile samples were collected from the thermally stratified period to the mixed period in a deep-water reservoir in southwest China [38,39]. The spatial and temporal variation of the DIC concentration, partial pressure of CO₂ (pCO₂), and calcite saturation index (SIc) were analyzed. The purpose wasto investigate the divergence characteristics of the DIC in karst deep-water reservoirs in the vertical direction, influencing factors, and path of the DIC in river water into the reservoir.

2. Materials and Methods

2.1. Study Area

The study area is located at the Pingzhai Reservoir, a humid subtropical monsoon climate zone (105°17′57′′ E~105°25′50′′ E, 26°28′27′′ N~26°34′28′′ N), which is in Guizhou Province, southwest China (Figure 1a). The rainy season in the area is from May to October. The reservoir has a catchment area of 3249 km² of which 74.30% is a karst area. The main water source is the Sancha River, followed by the Shuigong, Zhangwei, Baishui, and Hujiarivers. The reservoir capacity is 1.09 × 10⁹ m³, and there is an underground reservoir with a capacity of 311 million m³, accounting for 29% of the total capacity of Pingzhai Reservoir [40]. In the catchment area, the main human activities are agriculture, residential life, and industrial activities, and the mineral resources are coal and pyrite. The main land use types in the region are cropland, woodland, grassland, building land, water, and unused land (Figure 1c). The reservoir is a cyclic thermally stratified reservoir. It is in the mixed homogeneous phase from January to April, in the thermally stratified phase from May to October, in the thermally stratified stable phase from July to September, and in the mixing phase from November to December [40].

2.2. Sample Collection and Analysis

According to the hydraulic connection and tributary distribution, NY, SG, and PZ were selected as the sampling sites (Figure 1b). The water temperature (T_water) and dissolved oxygen (DO) concentrations were measured every two or three months, from July 2019 to July 2021. Other indicators were measured every two months, from September 2019 to January 2020. During the sampling period, the minimum depth of the reservoir was 65 m, the maximum depth was 101 m, and the average depth was 87 m. Therefore, samples were collected at depths of 0.1, 10, 20, 30, 40, 50, and 60 m.

The study measured the T_water (±0.01 °C), pH (±0.01), electrical conductivity (EC; ±1 mS cm⁻¹), DO concentration (±0.01 mg·L⁻¹), chlorophyll a (chl-a) concentration, DIC isotopes (δ¹³C_DIC; measured values in thousandths (‰), analytical error < 0.2‰), the concentration of anions and cations (Ca²⁺, Mg²⁺, Na⁺, K⁺, HCO₃⁻, CO₃²⁻, SO₄²⁻, NO₃⁻, and Cl⁻). T_water, DO, EC, and pH were measured using a Multi 3430 portable water quality analyzer (WTW Co. Ltd., Germany). The HCO₃⁻ and CO₃²⁻ concentrations were determined using acid–base kits (Aquamerck, Germany) (±0.01 mg·L⁻¹). The samples for the anion and cation analysis were filtered through 0.45 µm cellulose acetate membrane and stored in HDPE bottles. Nitric acid was added to the samples for the cation analysis to bring the pH ≤ 2.0. The samples of δ¹³C_DIC were spiked with 3–5 drops of HgCl₂ solutions and stored in a polyethylene bottle without air bubbles. The cations were analyzed withICP-EMS (relative error < 2%). The anions were analyzed usingan ion spectrometer (IS90). The δ¹³C_DICwasmeasured with agas isotope mass spectrometer (MAT-252). A separate water sample of 500 mL was collected and stored in a brown glass bottle with the addition of magnesium carbonate suspension and kept at 4°C in the dark; then, the chl-a concentration was measured with a spectrophotometer.

2.3. Data Processing

The α indicator method was to determine the type of reservoir water temperature stratification [41].

$$\alpha =\frac{Total inlet flow}{Total reservoir capacity}$$

If α < 10, the reservoir is a stable stratification type; if 10 ≤ α ≤ 20, it is a transitional type; if α > 20, it is a mixed type.

In water, pCO₂ is the factor controlling the CO₂ exchange at the water–gas interface [42]. In this study, pCO2 was calculated with Henry’s law:

$$p{\mathrm{CO}}_2=\left[H_{2}C{\mathrm{O}}_3^{\ast }\right]{\mathrm{K}}_{{\mathrm{CO}}_2}=\mathsf{\alpha } \left({\mathrm{H}}^{+}\right).\mathsf{\alpha } \left({\mathrm{HCO}}_3^{-}\right){ \mathrm{K}}_{{\mathrm{CO}}_2}\times {\mathrm{K}}_1$$

In Equation, α is the ionic activity, K_CO2 is Henry’s constant, and K₁ is the first dissociation constant.

SIc is the indicator to determine the state of reaction between water and carbonate rock. SIc = 0, Sic < 0, and Sic > 0 denote dissolution equilibrium, precipitation, and erosion of carbonate rocks, respectively [43]. In this study, SIc calculated by Phreeqc3.1.7:

$$SIc=\mathrm{Log}\left({\mathrm{IAP}/\mathrm{K}}_{\mathrm{cal}}\right)$$

In this equation, IAP is ionic activity, and K_cal is the equilibrium constant of calcite dissolution.

3. Results

3.1. Characteristics of T_water, DO Changes, and Reservoir Thermal Stratification Patterns

In the Pingzhai Reservoir, T_water decreased with the water depth. The highest T_water was 27.60 °C at the surface layer, and the lowest T_water was 10.75 °C at the 60 m depth. The T_water varied significantly from the surface layer to 20 min depth. However, it changed slightly at depths of 50 m and below. The T_water in the upper layer showed an annual cycle of variation. The T_water was lowest in January, then warmed up to a maximum in July, and then gradually decreased (Figure 2a–c). The cyclical variation of the T_water in Pingzhai Reservoir is mainly influenced by solar radiation.

The variability of DO concentration was more complex than that of T_water and did not follow an annual cycle. The DO concentrations were higher in the upper layer than in the lower layer and increased in the surface layer between January and July but decreased significantly at depths of 10 m and 20 m. The lowest DO concentrations occurred in May and were less than 2.0 mg/L (Figure 2d–f). This is consistent with the climate trends in the Pingzhai Reservoir basin and the distribution characteristics of algae in the reservoir.

The values of α were calculated with the data collected at the Yangchang Hydrological Station from 2014 to 2019. The total inflow of Pingzhai Reservoir was 1348 million m³/a, the capacity was 1.09 × 10⁹ m³. The result shows that α was 1.19 and was less than 10. Therefore, the Pingzhai Reservoir was stably stratified. Combined with the variation of the T_water, thermal stratification in the Pingzhai Reservoir occurs from April to September, while the upper and lower T_water is consistent in January. Using a temperature difference of 0.2 °C/m to assess the thermal stratification [43], the thermocline occurs below the surface to a depth of 10 m. However, from April to November there is a temperature difference of more than 0.2 °C/m sat at a depth of 20 to 50 m. It makes the Pingzhai Reservoir a double thermocline reservoir.

3.2. Distribution Characteristics of DIC and Vertical Variation of δ¹³C_DIC

The DIC is composed of H₂CO₃, HCO₃⁻, and CO₃²⁻. In the Pingzhai Reservoir, the pH was 7.50–9.01, where approximately 98% of the DIC wasHCO₃⁻, and the HCO₃⁻ concentration was often used to characterize the DIC concentration in the water column [44]. In the Pingzhai Reservoir, the DIC concentration was 76.25–204.35 mg. The mean concentration was 168.43 mg L⁻¹, which is consistent with the DIC concentration in other karst areas [42,45]. Vertically, the DIC concentration increased with the water depth from the surface layer to the 40 m depth. From the 40 m to 60 m depth, the DIC concentration varied slightly. The analysis of the coefficient of variation showed that the DIC concentration varied greatly in the upper layer and less in the lower layer, especially in the 50 m and 60 m depths where the coefficient of variation was the lowest. The analysis of the coefficients of the variation for the different phases at the NY, PZ, and SG sites showed that the variation of the DIC concentration was thermal stratification phase > mixing phase > mixed phase. In the thermal stratification phase, the spatial variation of the DIC concentration was SG > PZ > NY. In the mixing and mixed phases, the spatial variation of the DIC concentration was SG > NY > PZ in (Figure 3a).

In the Pingzhai Reservoir, the δ¹³C_DIC values ranged from −3.454‰ to −14.857‰. The mean δ¹³C_DIC values in the thermal stratification phase, mixing phase, and mixed phases were −10.726‰, −10.254‰, and −10.751‰, respectively. Only the surface δ¹³C_DIC value of the thermal stratification phase was greater than −8‰, while the rest were less than −9‰. In the 10 m–60 m depth range, the δ¹³C_DIC values ranged from −9‰ to −12‰ (Figure 3b). A statistical analysis of the δ¹³C_DIC values also showed that there was a significant difference between the surface and lower layers in the thermal stratification phase, while there was no significant difference between the mixing and mixed phases. The changing trends between the δ¹³C_DIC values and T_water indicated that the lower the T_water, the more the δ¹³C_DIC values were overweighted (Figure 3b,c).

The Ca²⁺, Mg²⁺, and DIC are indicators of carbonate rock weathering. The K⁺ and Na⁺ are indicators of silicate rock weathering. The Cl⁻, SO₄²⁻, and NO₃⁻ are indicators of human activities. The Chl-a is an indicator of aquatic plants. The pH and DO are indicators of the water environment. The correlation analysis showed that carbonate weathering and silicate weathering occurred simultaneously. The salinity effect caused increased solubility of ions in the reservoir and an increased ion concentration. Human activities also affected the pH and hence the weathering of the carbonate rocks. As algal are the main actors in carbon transport from the biosphere to the atmosphere and hydrosphere, they are highly significantly correlated with DIC, while pH and DO affect algal abundance, on the one hand, and carbon exchange at the water–air interface on the other (

Table 1. Correlation analysis of each index in Pingzhai Reservoir.

	Ca2+	K+	Mg2+	Na+	Cl−	NO3−	SO42−	DIC	pH	δ13CDIC	DO	Chl-a
Ca2+	1
K+	0.573 **	1
Mg2+	0.470 **	0.932 **	1
Na+	0.288 *	0.810 **	0.845 **	1
Cl-	0.209	0.488 **	0.523 **	0.684 **	1
NO3-	0.576 **	0.271	0.165	−0.133	0.098	1
SO42-	0.408 **	0.538 **	0.607 **	0.586 **	0.894 **	0.362 **	1
DIC	0.363 **	0.120	0.175	0.038	0.171	0.336 *	0.365 **	1
pH	−0.227	0.036	0.149	0.244	0.002	−0.363 **	−0.065	−0.425 **	1
δ13CDIC	−0.684 **	−0.436 **	−0.325 *	−0.046	−0.074	−0.684 **	−0.267	−0.272	0.463 **	1
DO	−0.091	0.054	0.095	0.077	−0.172	−0.096	−0.166	−0.310 *	0.863 **	0.214	1
Chl-a	0.151	−0.063	−0.058	−0.185	0.003	0.211	0.092	0.489 **	−0.152	−0.231	0.045	1

* p <0.05, ** p <0.01.

). This shows that the DIC concentration in the reservoir was influenced by rock weathering, human activities, aquatics, and the nature of the water column.

However, the stratified correlation analysis showed that the DIC concentrations varied at different depths. The surface DIC concentration was influenced by carbonate weathering and human activities but also by the temperature and atmospheric CO₂ concentrations. The DIC concentration at the 30 m depth was influenced by carbonate and silicate weathering and human activities. The DIC concentrations at the 10 m, 50 m, and 60 m DIC concentrations were not significantly affected. The DIC concentration at the 20 m and 40 m depths were influenced by Na⁺ and water temperature, respectively (

Table 2. Correlation analysis between DIC and water chemistry and environmental factors at different depths.

	Surface Layer	10 m Depth	20 m Depth	30 m Depth	40 m Depth	50 m Depth	60 m Depth
Ca2+	0.885 **	0.273	0.4	0.706 *	0.188	0.371	0.491
Mg2+	0.627	0.078	0.633	0.832 **	−0.58	−0.381	0.253
K+	0.545	−0.035	0.293	0.691 *	−0.654	−0.455	0.229
Na+	0.411	0.099	0.680 *	0.836 **	−0.593	−0.428	0.256
Cl−	0.138	−0.079	0.366	0.888 **	−0.26	−0.333	−0.325
NO3−	0.915 **	−0.172	−0.279	−0.14	0.317	−0.286	−0.215
SO42−	0.299	−0.035	0.534	0.837 **	0.089	−0.373	−0.191
Twater	−0.989 **	−0.539	−0.597	−0.903 **	−0.760 *	0.01	−0.005
DO	−0.671 *	0.211	0.445	−0.291	−0.551	−0.432	−0.344
pH	−0.780 *	0.063	0.506	0.417	0.555	−0.187	0.073
chl-a	−0.662	0.105	0.17	−0.581	−0.447	0.289	−0.518
Humidity	−0.328	-	-	-	-	-	-
Wind speed	0.048	-	-	-	-	-	-
Air temperature	−0.811 **	-	-	-	-	-	-
Barometric pressure	0.499	-	-	-	-	-	-
Atmospheric CO2 concentration	0.879 **	-	-	-	-	-	-

* p < 0.05, ** p < 0.01.

).

3.3. Characteristics of SIc and pCO₂Variation in the Pingzhai Reservoir

SIc is an indicator of the weathering tendency of carbonate rocks. When Sic > 0, carbonate rocks are precipitating; when Sic < 0, carbonate rocks are dissolving; when SIc = 0, the dissolution reaction of carbonate rocks reaches equilibrium. In the Pingzhai Reservoir, SIc ranged from −0.09 to 0.90. SIc were high in the surface layer and decreased with the water depth. The SIc were positive for all phases, sites, and depths, except for depths below 20 m at the PZ site during the thermal stratification phase, depths below 50 m at the NY site during the mixing phase, and depths below 50 m at theSG site during the mixed phase. Most of the time, the carbonate concentration was supersaturated (Figure 3d). The correlation analysis showed that a SIc above 30 m depth was significantly positively correlated with pH, Na⁺, DO, and Mg²⁺ (p < 0.01) and significantly negatively correlated with T_water (p < 0.01). Below 30 m depth, SIc was significantly positively correlated with pH, Ca²⁺, Na⁺, Mg²⁺, K⁺, SO₄²⁻, and EC (p < 0.01 or p < 0.05) and negatively correlated with chl-a (p < 0.05) (

Table 3. Correlation analysis between SIc and water chemistry and environmental factors at different depths.

	Surface Layer	10 m Depth	20 m Depth	30 m Depth	40 m Depth	50 m Depth	60 m Depth
Ca2+	0.373	0.468	0.519	0.705 *	0.464	0.268	0.921 **
Mg2+	0.51	0.043	0.826 **	0.875 **	0.628	0.832 **	0.878 **
K+	0.645	0.009	0.532	0.778*	0.541	0.799 **	0.830 **
Na+	0.733 *	0.204	0.895 **	0.902 **	0.599	0.764 *	0.824 **
Cl−	0.594	−0.406	0.245	0.869 **	−0.085	0.525	−0.603
NO3−	0.258	−0.505	−0.467	0.082	−0.157	0.613	−0.224
SO42−	0.578	−0.377	0.307	0.832 **	0.409	0.307	−0.374
Twater	0.014	−0.889 **	−0.911 **	−0.694 *	−0.618	0.192	−0.253
DO	0.681 *	0.792 *	0.889 **	0.461	−0.06	0.116	−0.181
EC	0.105	0.006	0.489	0.802 **	0.268	−0.161	0.172
pH	0.585	0.897 **	0.924 **	0.933 **	0.804 **	0.175	0.877 **
chl-a	−0.204	−0.585	−0.136	−0.682 *	−0.695 *	0.1	−0.712 *
Humidity	−0.533	-	-	-	-	-	-
Wind speed	−0.263	-	-	-	-	-	-
Air temperature	0.347	-	-	-	-	-	-
Barometric pressure	−0.596	-	-	-	-	-	-
Atmospheric CO2 concentration	−0.271	-	-	-	-	-	-

* p< 0.05, ** p< 0.01.

).

pCO₂ is an indicator thatdetermines the direction of carbon transport at the water–air interface and calculates exchange fluxes. In the Pingzhai Reservoir, pCO₂ increased with the water depth. Its changing tendency was opposite to that of T_water and the same as that of DIC. The response of pCO₂ to the water depth lagged behind that of DIC. The surface pCO₂ ranged from 4.74 Pa to 61.05 Pa (Figure 4), while the atmospheric pCO₂ ranged from 25.2 Pa to 30.6 Pa. The atmospheric pCO₂ was larger than the surface pCO₂ in the thermal stratification phase but smaller than the surface pCO₂ in the mixing and mixed phases. The correlation analysis showed that the surface pCO₂ was significantly negatively correlated with T_water, pH, DO, and air temperature (p < 0.01) and positively correlated with atmospheric pCO₂ (p < 0.01) and barometric pressure (p < 0.05). Below the surface, pCO₂ was not influenced by atmospheric parameters. At different depths, different factors were significantly correlated with pCO₂ (

Table 4. Correlation analysis between pCO₂ and water chemistry and environmental factors at different depths.

	Surface Layer	10 m Depth	20 m Depth	30 m Depth	40 m Depth	50 m Depth	60 m Depth
Ca2+	0.547	0.017	−0.198	−0.269	−0.146	−0.338	−0.505
Mg2+	0.125	0.285	−0.508	−0.546	−0.368	0.047	−0.646
K+	−0.002	0.239	−0.305	−0.393	−0.241	0.109	−0.587
Na+	−0.119	0.147	−0.625	−0.622	−0.358	0.174	−0.541
Cl−	−0.353	0.3	0.037	−0.526	−0.083	0.153	0.525
NO3−	0.566	0.337	0.618	0.165	−0.099	0.117	0.246
SO42−	−0.225	0.357	0.207	−0.443	−0.418	−0.424	0.433
Twater	−0.803 **	0.567	0.816 **	0.509	0.898 **	0.692 *	0.505
DO	−0.894 **	−0.791 *	−0.928 **	−0.659	0.056	0.391	0.035
EC	0.614	0.383	0.017	−0.412	−0.432	−0.472	0.042
pH	−0.962 **	−0.906 **	−0.909 **	−0.972 **	−0.987 **	−0.981 **	−0.962 **
chl-a	−0.556	0.394	0.44	0.704*	0.905 **	0.793 *	0.534
Humidity	−0.059	-	-	-	-	-	-
Wind speed	0.294	-	-	-	-	-	-
Air temperature	−0.851 **	-	-	-	-	-	-
Barometric pressure	0.796 *	-	-	-	-	-	-
Atmospheric CO2 concentration	0.867 **	-	-	-	-	-	-

* p <0.05, ** p <0.01.

).

There were also spatial and temporal differences in the pCO₂ (Figure 4). At the surface, pCO₂ was generally NY < SG < PZ. However, below the surface, it was SG < PZ < NY. Apart from the surface layer, pCO₂ also showed some temporal patterns of variation. At the NY and PZ sites, pCO₂ generally showed a thermal stratification phase > mixing phase > mixed phase, while SG showed no clear temporal pattern. The maximum pCO₂ at the PZ and NY sites during the thermal stratification phase was at a depth of 40 m, while at the SG site, it was at a depth of 50 m. The maximum pCO₂ was at a depth of 50 m for the NY and PZ sites and 60 m for the SG site during the mixing and mixed phases.

3.4. DIC Sources in Pingzhai Reservoir

Ca²⁺ was the most dominant cation in Pingzhai Reservoir, accounting for 57.7% to 82.53% of the total cation concentration. SO₄²⁻ was the second most dominant anion, accounting for 20.05% to 41.04% of the total anion concentration. The ratios of (Ca²⁺+Mg²⁺)/(HCO₃⁻+SO₄²⁻) and (Ca²⁺+Mg²⁺)/(HCO₃⁻) were both approximately 1 (Figure 5), with significant correlation coefficients of 0.71 and 0.69, respectively. The results indicate that carbonate weathering controlled the water chemistry and DIC correlated with Ca²⁺, Mg²⁺, and SO₄²⁻(Figure 6).

In the Wujiang River basin, the δ¹³C_DIC from soil organic matter formation ranged from −16‰ to −19‰ [46], from carbonate rock weathering ranged from −2‰ to 2‰ [45], and from atmospheric CO₂ dissolved in water was about 0~2.5‰ [38]. In the Pingzhai Reservoir, the δ¹³C_DICranged from −3.454‰ to −14.857‰. Therefore, based on the water chemistry and isotopic analysis, the DIC was mainly derived from soil organic matter and carbonate rock weathering. Based on the ¹³C mass balance principle [47], approximately 60.44% of the DIC was derived from soil organic matter, 14.78% from carbonate rock and 14.78% from atmosphere. However, the source of DIC varies spatially and seasonally. The NY site was 64.40% from soil organic matter and with the soil source being 5–6 percentage points higher than other points; DIC in the thermal stratification phase had approximately 33.93% from soil and 33.04% from carbonate rock and 33.03%from atmosphere.

4. Discussion

4.1. Mechanism of Subthermocline Formation and Its Influence on the Vertical Distribution of DIC

Studies have shown that thermal stratification is the dominant factor leading to the vertical differentiation of DIC in deep karst reservoir [42]. There are two thermal layers in the Pingzhai Reservoir, the surface to 10 m depth is the main thermal layer, which is about 10 m thick. Between the 30 m depth and 50 m depth, there is a secondary thermal layer the thickness of which varies between 10 and 20 m. The formation mechanism of the main thermocline is solar radiation, the same as that of the other reservoirs in the Wujiang watershed. But this influence decreases below 10 m. However, the formation mechanism of the secondary thermocline is not clear [48]. Previous studies have shown that the thermocline impedes material exchange, and phytoplankton from the surface layer of the reservoir cannot reach the lower layer [30]. However, in the Pingzhai Reservoir, chl-a was detected in the lower layer and affected pCO₂ and Sic below a 30 m depth (Table 3 and Table 4). The DIC concentration, pCO₂, and SIc at a 30 m depth was significantly correlated with Ca²⁺, Mg²⁺, pH, and chl-a (Table 2, Table 3 and Table 4). During the thermally stratified phase, the tributaries of the Pingzhai Reservoir had the largest volumes and flow rates. Because of the inertial flow, the river forms a mainstem below the surface with a higher flow velocity than the surface layer [49]. In addition, the main flow is from west to east, and it is easier for the river to form a subduction below the reservoir surface under the Coriolis forces [49,50]. δ¹³C_DICanalysis shows that the δ¹³C_DICin the reservoir at 30 m to 50 m depth is closer to the δ¹³C_DIC from the soil. The formation of the subtemperate layer at 50 m is due to the formation of the para-temperate layer at 50 m caused by the downward flow of river water into the reservoir. This result further confirms that at the Pingzhai Reservoir, the river water forms a subduction into the reservoir below the surface layer due to a combination of inertial action and Coriolis forces as it enters the reservoir. The material in the river water also enters the lower reservoir layer with the submerged flow. Thus, although thermal stratification controls the vertical partitioning characteristics of DIC in Pingzhai Reservoir, the variability of river DIC in the vertical direction is more complex due to changes in the mixed diffusion pattern of river water entering the reservoir. The sampling interval of 10 m in this study was not sufficiently accurate in determining the thickness of the subthermal layer and the hydraulic characteristics of the submerged flow, and further research is needed.

4.2. Source Influences the Vertical Difference in DIC Concentration

In this study, the DIC concentration was significantly correlated with the DIC source fraction (

Table 5. Correlation analysis of the DIC, SO₄²⁻, and NO₃⁻concentrations with the source proportions.

	DIC (mg/L)	SO42− (mg/L)	NO3− (mg/L)
Sources of carbonate rock	−0.548 **	−0.335	−0.492 **
Sources of soil organic matter	0.548 **	0.335	0.492 **
Sources of air	−0.548	−0.335	−0.492

** p < 0.01.

), indicating that the DIC source influenced the variation of DIC concentration. During the thermal stratification phase, the percentage of DIC from carbonate rocks and atmospheric was high, the δ¹³C_DIC value in the surface layer was about −8‰, while the DIC was readily available to aquatic plants for photosynthesis [51]. The DIC from soil was negative and more likely to enter the lower layer of the reservoir with the submerged flow, so the DIC concentration in the surface layer was significantly lower than that in the subsurface layer. During the mixing and mixed phases, DIC from the soil was dominant. It did not have the advantage of being used by photosynthesis [52]. Therefore the DIC concentration from the river did not change in the reservoir except for mixing, and no stratification of DIC concentration occurred. Some studies show that the surface T_water was higher than the lower layer in winter, the density of surface water was greater than the lower layer and the sinking of the upper layer of water under the action of gravity was the main mechanism of reservoir overtopping and consistent DIC concentration in reservoir [50]. However, in the Pingzhai Reservoir, the surface T_water exceeds 10 °C in autumn and winter, and the surface T_water is still higher than the lower layer, so the density of the upper layer is lower than that of the lower layer. Therefore, in the Pingzhai Reservoir, the reflux gradient flow is the main mechanism of the water mixing in the upper and lower layers of the reservoir [53,54], and the gradual decrease in the pCO2 in the reservoir from the bottom to the surface layers (Figure 4) confirms the existence of this mechanism. Thus, the different sources of DIC influence whether DIC concentration stratification occurs in the reservoir.

The surface layer is denser than the lower layer, and the sinking of the upper layer is the main mechanism for the reservoir overturning [50]. However, in the Pingzhai Reservoir, the surface T_water is more than 10°C in autumn and winter, and the T_water in the surface layer is still higher than that in the lower layer, so the density of the upper layer is lower than that of the lower layer. The overturning of the reservoir cannot be achieved. Therefore, the difference in the DIC sources affects the DIC concentration in the vertical direction.

4.3. Impact of Human Activities on DIC Concentration in the Reservoir

There are many human activities in the upstream watershed of the Pingzhai Reservoir, including industrial activities related to pyrite and coal, agricultural activities related to farming, and residential production and living. In the reservoir, SO₄²⁻/Na⁺ ranged from 2.11 to 7.00 and NO₃⁻/Na⁺ ranged from 0.27 to 1.74 (

Table 6. Statistical table of major ion concentration ratios.

	Cl−/Na+	SO42−/Na+	NO3−/Na+
Min. value	0.164	2.116	0.276
Max. value	0.782	7.001	1.743
Mean value	0.491	4.320	0.870
Standard deviation	0.098	0.858	0.276
Coefficient of variation	0.200	0.199	0.317
Critical Value of significant (t-test)	0.516	4.536	0.940
Critical value of highly significant (t-test)	0.525	4.607	0.963

), both between agricultural and industrial activity input end elements [55]. This indicated that human activities affected the water chemistry. SO₄²⁻ was the second most abundant anion in the Pingzhai Reservoir, accounting for 20.05% to 41.04% of the total anions. The ratio of (Ca²⁺ + Mg²⁺)/(HCO₃⁻ + SO₄²⁻) was about 1 (Figure 5A). The relationships between SO₄²⁻, NO3⁻, DIC, Ca²⁺, and Mg²⁺ suggested that SO₄²⁻ and NO₃⁻ were involved in the weathering of the carbonate rocks (Table 1).

Studies on the anthropocentric involvement of sulfuric and nitric acids in the weathering of carbonate rocks suggest two mechanisms of involvement, one in the form of acids [56,57], and the other being in a salt effect in the presence of SO₄²⁻/NO₃⁻ which increases the solubility of the carbonate rock, rather than a shift in the weathering equilibrium of the carbonate rock [58]. The pH of the Pingzhai Reservoir is 7.50 to 9.01, so sulfur and nitrogen were imported into the reservoir as ions rather than acids. In the reservoir area, the influx of SO₄²⁻and NO₃⁻ increased the dissociation of carbonate rocks, which supersaturated the Sic of Pingzhai Reservoir (Figure 3d).

The supersaturated DIC concentration was susceptible to change, especially in the surface layer in the reservoir by climatic and hydrological changes, such as aquatic absorbed DIC to generate O₂, the DIC degassed to produce CO₂ discharge into the air. This further explains that the DIC concentrations in the lower layer are higher than that in the subsurface layer. The ratios ofSO₄²⁻/Na⁺ and NO₃⁻/Na⁺ in the surface layer were higher than those in the lower layer, indicating that human activities had greater impact on the water chemistry in the surface layer than in the lower layer [59], affecting the vertical partitioning of DIC in the reservoir.

5. Conclusions

Pingzhai Reservoir is a doublethermocline reservoir. The underflow is the factor leading to the formation of the subthermocline. The concentration of theDIC was lower in the surface layer and higher in the lower layer, while pCO₂ generally increased with depth and SIc gradually decreased with depth. The subthermocline, human activities, and the source of DIC all influence the vertical divergence of DIC concentrations. In particular, the presence of the subtemperate layer allows the DIC in rivers to enter the lower reservoir layer directly through the subtemperate layer channel.