Chemical Weathering and CO2 Consumption Inferred from Riverine Water Chemistry in the Xi River Drainage, South China

Hydrochemistry and strontium isotope data were analysed in water samples from the Xi River Drainage system to reveal the spatial and seasonal variations in chemical weathering, associated CO2 consumption fluxes, and their control factors. The main ions were Ca2+, Mg2+, and HCO3−, which are characteristic of a drainage system on carbonate-dominated bedrock. The dissolved loads were derived from four major end-member reservoirs: silicate, limestone, dolomite, and atmosphere. The silicate weathering rates (SWRs) increased downstream from 0.03 t/km2/year to 2.37 t/km2/year. The carbonate weathering rates (CWRs) increased from 2.14 t/km2/year in the upper reaches, to 32.65 t/km2/year in the middle reaches, and then decreased to 23.20 t/km2/year in the lower reaches. The SWR values were 281.38 and 113.65 kg/km2/month during the high- and low-water periods, respectively. The CWR values were 2456.72 and 1409.32 kg/km2/month, respectively. The limestone weathering rates were 2042.74 and 1222.38 kg/km2/month, respectively. The dolomite weathering rates were 413.98 and 186.94 kg/km2/month, respectively. Spatial and seasonal variations in chemical weathering were controlled mainly by lithology, vegetation, and climate (temperature, water discharge, and precipitation). The CO2 consumption flux by chemical weathering was estimated at 189.79 × 109 mol/year, with 156.37 × 109 and 33.42 × 109 mol/year for carbonate and silicate weathering, respectively. The CO2 fluxes by chemical weathering are substantially influenced by sulfuric acid in the system. The CO2 flux produced by sulfuric acid weathering was estimated at 30.00 × 109 mol/year in the basin. Therefore, the Xi River Basin is a CO2 sink with a net consumption of CO2 flux of 3.42 × 109 mol/year.


Introduction
Chemical weathering plays an important role in surface processes that link the rock cycle of the solid earth to the hydrological cycle of rivers, oceans, and the atmosphere. Rock weathering under the influence of dissolved carbonic acid consumes atmospheric CO 2 and produces HCO 3 − and CO 3 2− , which are eventually discharged into the sea where they are absorbed in the marine carbonate buffer system. The global carbon cycle, in concert with topography development on the continents, plays an important role in regulating global climate, as topography provides a fresh weatherable surface and enhances the consumption of the greenhouse gas CO 2 .
The chemical weathering of terrestrial silicate rocks constitutes a significant carbon sink in global biogeochemical cycles on geological time scales [1][2][3][4][5][6]. The CO 2 consumption flux by silicate rock weathering controls the long-term global carbon cycle time scales of millions of years [7][8][9][10]. This consumed CO 2 is chemically locked in marine sediments and cannot be easily released back into the atmosphere on short timescales. Silicate rock weathering,

Geology
The rocks exposed in the Xi River Basin date from the Precambrian to Quaternary era ( Figure 1). Karstified limestone landscapes are widely distributed in the upper reaches; Permian to Triassic carbonate rocks cover an area of 155,000 km 2 , accounting for 44% of the total Xi River drainage basin. Coal deposits interbedded with carbonates are rich in sulfides. Precambrian metamorphic rocks and magmatic rocks can be found in the lower reaches. The dominant magmatic rocks are associated with the Mesozoic Yanshanian granite suite and include intermediate-acidic intrusive and extrusive rocks. A smaller component of magmatic rocks is that they are Permo-Triassic and early Paleozoic in age. Jurassic clastic sedimentary rocks are scattered in the middle reaches of the basin.

Climate and Human Activities
The Xi River Basin has a humid subtropical climate with an average yearly temperature of 14-22 °C [10,22]. The mean annual precipitation is 1470 mm. The mean annual evaporation varies from 900 to 1600 mm [10].

Geology
The rocks exposed in the Xi River Basin date from the Precambrian to Quaternary era ( Figure 1). Karstified limestone landscapes are widely distributed in the upper reaches; Permian to Triassic carbonate rocks cover an area of 155,000 km 2 , accounting for 44% of the total Xi River drainage basin. Coal deposits interbedded with carbonates are rich in sulfides. Precambrian metamorphic rocks and magmatic rocks can be found in the lower reaches. The dominant magmatic rocks are associated with the Mesozoic Yanshanian granite suite and include intermediate-acidic intrusive and extrusive rocks. A smaller component of magmatic rocks is that they are Permo-Triassic and early Paleozoic in age. Jurassic clastic sedimentary rocks are scattered in the middle reaches of the basin.

Climate and Human Activities
The Xi River Basin has a humid subtropical climate with an average yearly temperature of 14-22 • C [10,22]. The mean annual precipitation is 1470 mm. The mean annual evaporation varies from 900 to 1600 mm [10].
By the end of 2020, the Xi River Basin was home to approximately 66.5 million people. There are rich agricultural lands and mineral resources in the upper reaches. The agricultural area, including dry land and paddy fields, is approximately 40,000 km 2 in Guangxi Province. Crops are grown in dry land with various types of fertilizers and/or pesticides. Paddy fields are favorable for rice with relatively simple fertilizers and/or pesticides. The Hechi region in the upper course is an important nonferrous metal mining area in South China. Arsenic reserves in the Nandan region account for 19% of global reserves. There is an important lead-zinc mining area in Wuxuan located in the middle reaches. Most mineral resources in the Xi River Basin are currently being mined, although some resources are in the post-mining and mine remediation stages. 4

Sampling and Analysis Methods
The high-water period of Xi River drainage lasts from April to September, whereas the low-water period lasts from October until March. Water samples covering 30 sites were collected in July and October 2019. The samples were collected in the main stream, first-level, and second-level tributaries of the Xi River drainage (Figure 1).
Water samples were collected at a depth of 0.3 m. The samples were filtered through 0.45 µm cellulose acetate lipid membranes and then stored in polyethylene bottles rinsed 3-4 times with water. The water samples were acidified with ultrapure nitric acid to a pH < 2 to prevent algal growth. All samples were stored and refrigerated at 4 • C for further analysis.
The major ion concentrations (K + , Na + , Ca 2+ , Mg 2+ , Cl − , SO 4 2− , NO 3 − ) were determined via an ion chromatograph (IC 925) with an analysis error of less than 5%. The accuracy of the cation and anion concentrations was 0.001 mg/L. The concentrations of HCO 3 − were determined via an ultraviolet spectrophotometer. The concentrations of SiO 2 were measured using the silicomolybdic yellow colorimetric method with an accuracy of 0.01 mg/L. Ion analyses were carried out at China University of Geosciences. Strontium was purified by extraction chromatography. Then, strontium was converted to nitrate and dissolved in a nitrous solution for measurement. The strontium isotope compositions were measured via a Thermofisher Neptune MC-ICPMS at Vrije University Amsterdam. Table 1 shows the basic parameters, major ions, and Sr isotope data in Xi River water. The Xi River waters had pH values of 7.45 to 8.51 with an average of 8.00 during the high-water period, whereas the pH values ranged from 7.36 to 8.48 with a mean of 7.87 during the low-water period, which is consistent with water in contact with limestone and dolomite. The water samples were slightly alkaline. The high pH values reflect the importance of the dissolution of carbonate in the drainage. The water coming from typical carbonate sediment has a high pH value, which is fairly constant at approximately 8.3, as is to be expected for water in equilibrium with dissolved CaCO 3 and atmospheric CO 2 . Lower values appeared in rivers exposed to silicate rock where the carbonate buffering is less effective. Water samples in equilibrium with atmospheric CO 2 and no carbonate buffering have pH's of ca 5.3. Areas with anthropogenic atmospheric SO 2 and limited carbonate buffering also have lower pH values and, in the absence of carbonate buffering, much lower values.

Major Ions
The concentrations of total dissolved solids (TDS) widely varied from 49.84 to 305.61 mg/L during the high-water period (average: 193.22 mg/L) and 82.7~330.5 mg/L during the lowwater period (average: 204.5 mg/L), similar to the Mackenzie River, Yalong River, and Han River draining carbonate-dominated regions [23][24][25][26][27]. Seasonal variation is related to the dilution effect of water discharge. However, the ratio of the dilution and the increase in water discharge was not 1. During the low-water period, a portion of the exposed surface in the basin may be out of contact with river water and therefore be unavailable for water-rock interaction. In contrast, during the high-water period, more surface area is in contact with river water, and chemical weathering is enhanced [8,28,29]. The TDS concentrations decreased downstream along the main trunk ( Figure 2). Total cation concentrations (TZ + = K + + Na + + 2Ca 2+ + 2Mg 2+ ) ranged from 571 to 4015 µeq/L and from 801 to 4437 µeq/L during the high-and low-water periods, which is within the range of variation of the world's 61 largest rivers [11]. The total anion concentrations (TZ − = HCO 3 − + 2SO 4 2− + NO 3 − + Cl − ) ranged from 543 to 3822 µeq/L and from 1018 to 4164 µeq/L during the high-and low-water periods, respectively. The TZ + of most samples was slightly higher than the TZ − for the normalized ionic charge balance values [NICB = (TZ + − TZ − )/(TZ + + TZ − ) × 100%] within ±5%. The slight imbalance may be attributed to unanalysed organic complex matter [8,30].   The predominant cation in Xi River water was Ca 2+ , which composed over 50% of the total cations, followed by Mg 2+ , and then Na + + K + . The Ca 2+ concentrations were 157~1580 μmol/L and 254~1703 μmol/L during high-and low-water periods, respectively. The Mg 2+ concentrations were 70~420 and 38~495 μmol/L, respectively. HCO3 − was the main anion, whose concentrations ranged from 408 to 3095 μmol/L and from 764 to 3250 μmol/L during highand low-water periods, respectively. The second major anion was SO4 2− (46~346 and 55~536 μmol/L, respectively). The Cl − concentrations were 31~181 μmol/L and 39~799 μmol/L, respectively. The NO3 − concentrations were 0~202 and 0~445 μmol/L. Most of the major ions show distinct seasonal variations, with low contents during the high-water   3 show the characteristics of the major ion compositions. The predominant cation in Xi River water was Ca 2+ , which composed over 50% of the total cations, followed by Mg 2+ , and then Na + + K + . The Ca 2+ concentrations were 157~1580 µmol/L and 254~1703 µmol/L during high-and low-water periods, respectively. The Mg 2+ concentrations were 70~420 and 38~495 µmol/L, respectively. HCO 3 − was the main anion, whose concentrations ranged from 408 to 3095 µmol/L and from 764 to 3250 µmol/L during high-and low-water periods, respectively. The second major anion was SO 4 2− (46~346 and 55~536 µmol/L, respectively). The Cl − concentrations were 31~181 µmol/L and 39~799 µmol/L, respectively. The NO 3 − concentrations were 0~202 and 0~445 µmol/L.
Most of the major ions show distinct seasonal variations, with low contents during the high-water period and high contents during the low-water period. Ca 2+ , Mg 2+ , and HCO 3 − predominantly originate from chemical weathering [30,31]. The concentrations of Ca 2+ , Mg 2+ , and HCO 3 − decrease downstream along the trunk ( Figure 2). Spatial variations in ion concentrations are caused by lithologic distribution. The wide carbonate terrain, especially the karst topography in the upper course, provides advantageous conditions for carbonate chemical weathering. The lower course inherits dissolved loads from the upper course when flowing through the silicate terrain. The silicate weathering rate is much slower than that of carbonate under the same conditions, indicating that Ca 2+ , Mg 2+ , and HCO 3 − transported in the upper reaches are much more abundant than Na + + K + originating from silicate weathering [30,32]. Thus, all samples were of the HCO 3 -Ca/Mg type, showing that the water chemistry was dominated by carbonate weathering. period and high contents during the low-water period. Ca 2+ , Mg 2+ , and HCO3 − predominantly originate from chemical weathering [30,31]. The concentrations of Ca 2+ , Mg 2+ , and HCO3 − decrease downstream along the trunk ( Figure 2). Spatial variations in ion concentrations are caused by lithologic distribution. The wide carbonate terrain, especially the karst topography in the upper course, provides advantageous conditions for carbonate chemical weathering. The lower course inherits dissolved loads from the upper course when flowing through the silicate terrain. The silicate weathering rate is much slower than that of carbonate under the same conditions, indicating that Ca 2+ , Mg 2+ , and HCO3 − transported in the upper reaches are much more abundant than Na + + K + originating from silicate weathering [30,32]. Thus, all samples were of the HCO3 -Ca/Mg type, showing that the water chemistry was dominated by carbonate weathering.

Strontium Isotopes
Strontium concentrations in the Xi River drainage ranged from 0.418 to 2.915 μmol/L in the high-water period (average: 1.090 μmol/L) and from 0.308 to 4.975 μmol/L in the low-water period (average: 1.260 μmol/L), higher than the average of 0.89 μmol/L calculated for the world rivers [33]. The spatial variation in strontium concentrations in the dissolved load of the Xi River drainage followed that of Ca 2+ , Mg 2+ , and HCO3 − . The 87 Sr/ 86 Sr ratios in the Xi River drainage varied from 0.7079 to 0.7157 in the high-water period (average: 0.7108) and from 0.7079 to 0.7165 in the low-water period (average: 0.7112),

Strontium Isotopes
Strontium concentrations in the Xi River drainage ranged from 0.418 to 2.915 µmol/L in the high-water period (average: 1.090 µmol/L) and from 0.308 to 4.975 µmol/L in the low-water period (average: 1.260 µmol/L), higher than the average of 0.89 µmol/L calculated for the world rivers [33]. The spatial variation in strontium concentrations in the dissolved load of the Xi River drainage followed that of Ca 2+ , Mg 2+ , and HCO 3 − . The 87 Sr/ 86 Sr ratios in the Xi River drainage varied from 0.7079 to 0.7157 in the highwater period (average: 0.7108) and from 0.7079 to 0.7165 in the low-water period (average: 0.7112), which is lower than the global average value of rivers (0.7119). The 87 Sr/ 86 Sr ratios increased downstream along the main trunk. A quarter of the samples had 87 Sr/ 86 Sr ratios ranging from 0.707 to 0.709, corresponding to Phanerozoic marine carbonate values (0.7065~0.709) [34]. The variable strontium concentrations and 87 Sr/ 86 Sr ratios clearly reflect the different types of exposed rocks in the Xi River Basin. The upper reaches that drain a region of carbonate rocks have high strontium concentrations and low 87 Sr/ 86 Sr ratios (0.707~0.709), whereas the lower reaches draining a region of clastic sedimentary, magmatic rocks, and metamorphic rocks have lower strontium concentrations and higher 87 Sr/ 86 Sr ratios (0.708~0.910) ( Table 2). Figure 4 shows a positive relation with 87 Sr/ 86 Sr and 1/Sr in the dissolved solutes, indicating that the mixture of strontium originating from carbonate and silicate rocks leads to the observed variation in 87 Sr/ 86 Sr ratios for soluble strontium.  [34]. The variable strontium concentrations and 87 Sr/ 86 Sr ratios clearly reflect the different types of exposed rocks in the Xi River Basin. The upper reaches that drain a region of carbonate rocks have high strontium concentrations and low 87 Sr/ 86 Sr ratios (0.707~0.709), whereas the lower reaches draining a region of clastic sedimentary, magmatic rocks, and metamorphic rocks have lower strontium concentrations and higher 87 Sr/ 86 Sr ratios (0.708~0.910) ( Table 2). Figure 4 shows a positive relation with 87 Sr/ 86 Sr and 1/Sr in the dissolved solutes, indicating that the mixture of strontium originating from carbonate and silicate rocks leads to the observed variation in 87 Sr/ 86 Sr ratios for soluble strontium.

Sources of Dissolved Loads
The solutes are derived mainly from atmospheric deposition, anthropogenic contamination, and chemical weathering.

Sources of Dissolved Loads
The solutes are derived mainly from atmospheric deposition, anthropogenic contamination, and chemical weathering.

Atmospheric Input
Chloride is the most frequently used proxy to assess the atmospheric input to the chemical composition of dissolved matter in river water [2,27,36,37]. Chloride does not participate in biogeochemical cycling and is comparatively conservative [38]. Chloride concentrations are too low to be detected in rocks apart from evaporates. The annual chloride concentrations of precipitation in Guiyang and Lei Gong Shan were 10.1 µmol/L and 7 µmol/L, respectively [16,39]. The median chloride concentration in rainwater was 8 µmol/L in Guiyang [40]. Xu and Liu (2010) and Jiang et al. (2018) showed that Cl − concentrations from the precipitation for the Xi River were 17.2 and 10.4 µmol/L, respectively [19,26]. Han et al. (2010b) reported an average chloride concentration in the atmosphere of 5.2 µmol/L for Maolan [41]. The atmospheric input of chloride into the river can be calculated by multiplying the evapotranspiration factor by the annual average chloride concentration of precipitation in the basin. The molar ratios of Na + /Cl − , K + /Cl − , Ca 2+ /Cl − , and Mg 2+ /Cl − in rainwaters were 0.46 ± 0.24, 0.69 ± 0.35, 2.03 ± 1.14, and 0.30 ± 0.16, respectively.
Na-normalized molar ratios of the ions in the precipitation are shown in Table 2. Previous studies reported a minimal contribution of sea salt to water chemistry [17,21]. The average pH of the rainwaters in the Xi River Basin was lower than 4.5, indicating a serious acid rain deposition problem [42]. Most acid rainwaters were SO 4 2− -type because of sulfur-rich coal combustion [39]. The SO 4 2− /Na + molar ratios in rainwater varied from 1.2 to 24 in the Xi River drainage, implying S enrichment of precipitation relative to sea salt (SO 4 2− /Na + = 0.06) [22]. Han  in the water body originated from rainwater [42]. The NO 3 − /Na + ratios in precipitation ranged from 0.52 to 12.2 in the drainage, pointing to the anthropogenic contribution to the atmosphere (NO 3 − /Na + = 0 in sea salt) [22]. The ion concentrations and ratios in rainwater in the city were much higher than those in the countryside, indicating a greater anthropogenic contribution of NOx to the atmosphere near cites.

Anthropogenic Inputs
River water pollution as a result of human activities enters the river water through atmospheric contributions and human emissions, including industrial sewage, fertilizer, and pesticide residues. The characteristics of the Xi River Basin reflect variable natural and anthropogenic processes over a large region and a wide east-west span. There are rich agricultural and mineral resources in the upper reaches, whereas there are high population densities with high urbanization in the lower reaches [22,43]. PO 4 3− , NO 3 − , K + , and Cl − mainly originate from agricultural fertilizers and industrial sewage. Figure 2 shows the spatial variations in Cl − , NO 3 − , and SO 4 2− concentrations in the Xi River Basin. Chloride concentrations do not gradually increase with decreasing distances from the sea, suggesting dominant sources other than the influx of marine aerosols, such as halite from evaporite sources and anthropogenic inputs. Halite has not been recorded in the basin, which suggests that recorded Cl − and SO 4 2− originate from other sources. Spatial variations in Cl − indicate that the excess chlorine over atmospheric contributions originate from human activity and are balanced by Na + . Figure 5 shows a positive correlation between the SO 4 2− /Na + and NO 3 − /Na + molar ratios in the Xi River drainage. This relationship indicates that sulfate and nitrate share a common source, possibly anthropogenic. SO 4 2− is derived from gypsum dissolution, sulfide oxidation, and acid deposition. The concentration of NO 3 − is characterized by a large variation range. NO 3 − in river water mainly originates from nitrogen fertilizers used for agriculture. The minor source of nitrate is precipitation. Previous studies reported that gypsum-bearing evaporates were distributed in the Nanpan and Beipan Rivers located in the upper courses [22]. The other source of SO 4 2− in the water body is sulfide oxidation, since coal-bearing sedimentary rocks containing pyrite FeS 2 are widespread in the Xi River Basin. The positive SO 4 2− /Na + and NO 3− /Na + molar ratios would point to the combustion of S-bearing coal as a significant source of both atmospheric SO 2 and NOx. Basin. The positive SO4 2− /Na + and NO 3− /Na + molar ratios would point to the combustion of S-bearing coal as a significant source of both atmospheric SO2 and NOx.

Chemical Weathering Inputs
Stoichiometric analyses are often used to trace sources of ions dissolved in water bodies. The (Na + + K + )/Cl − equivalent ratios in most samples are larger than one, which indicates that Na + and K + mainly originate from sodium and potassium aluminosilicate weathering rather than evaporite weathering (Figure 6a). The excesses Cl − to K + + Na + of Samples X8 (He River) and X15 (Liu River) during the low-water period are derived from residential and industrial wastes ( This fact suggests that water chemistry is mainly derived from carbonate weathering by carbonic and sulfuric acid (Figure 6b-d). The dissolution of sulfate evaporites (such as gypsum) may be another source. It was estimated that approximately 58.8 µmol/L of riverine SO 4 2− in the upper and middle Xi River was derived from the dissolution of sulfate evaporites [28]. (Ca 2+ + Mg 2+ )/(HCO 3 − + SO 4 2− ) equivalent ratios in some samples are lower than one, indicating that extra (HCO 3 − + SO 4 2− ) is derived from silicate weathering (Figure 6d). The features of the silicate end-member are low Ca 2+ /Na + ratios of 0.01-0.56, Mg 2+ /Na + ratios of 0-0.68, HCO 3 − /Na + ratios of 1-3, and high 87 Sr/ 86 Sr ratios of 0.708-0.910 (Table 2) [26]. The features of the carbonate end-member are high Ca 2+ /Na + , Mg 2+ /Na + , and HCO 3 − /Na + ratios of 30-70, 12-28, and 60-140, respectively, and low 87 Sr/ 86 Sr ratios of 0.707-0.709 (Table 2) [26].
The content of strontium, whose chemical properties are stable, is remarkably different in diverse sources. As the strontium isotope ratio is unaffected by material fractionation, dilution, and evaporation effects, the riverine strontium isotopic composition ( 87 Sr/ 86 Sr) directly reflects the weathering of various source rocks. As carbonates tend to be low in 87 Rb and high in total strontium, 87 Sr/ 86 Sr ratios of carbonates are commonly interpreted as reflecting the compositions of marine strontium at the time of deposition of carbonate. In contrast, crystalline rocks are often higher in 87 Rb and lower in total strontium, and here, the 87 Sr/ 86 Sr ratios are a function of Rb content and age. The relationships between Mg 2+ /Ca 2+ and Na + /Ca 2+ molar ratios and between 87 Sr/ 86 Sr ratios and Mg 2+ /Ca 2+ molar ratios (after correction for atmospheric input based on Han et al., 2010b) in Xi River water indicate different mixing trends of three end-members, including silicate, limestone, and dolomite (Figures 7 and 8) [41]. In almost all large rivers in the world, chemical weathering contains limestone, dolomite, and silicate weathering. Table 3 shows the ion ratios and strontium isotopic data of silicate, limestone, and dolomite end-members [44].  The content of strontium, whose chemical properties are stable, is remarkably different in diverse sources. As the strontium isotope ratio is unaffected by material fractionation, dilution, and evaporation effects, the riverine strontium isotopic composition ( 87 Sr/ 86 Sr) directly reflects the weathering of various source rocks. As carbonates tend to be low in 87 Rb and high in total strontium, 87 Sr/ 86 Sr ratios of carbonates are commonly interpreted as reflecting the compositions of marine strontium at the time of deposition of            8 show that water chemistry is mainly controlled by limestone weathering, whereas these results cannot identify any anthropogenic activities. Figure 9 shows the relationship between 87 Sr/ 86 Sr ratios and HCO 3 − /(HCO 3 − + SO 4 2− ) equivalent ratios in Xi River water, indicating information on silicate, limestone, and dolomite weathering. The equivalent ratio of HCO 3 − /(HCO 3 − + SO 4 2− ) is greater than 0.7 for all samples in this study, whereas the ratio lower than 0.7 in some samples indicates the influence of anthropogenic activity (Figure 10) [10].  8 show that water chemistry is mainly controlled by limestone weathering, whereas these results cannot identify any anthropogenic activities. Figure 9 shows the relationship between 87 Sr/ 86 Sr ratios and HCO3 − /(HCO3 − + SO4 2− ) equivalent ratios in Xi River water, indicating information on silicate, limestone, and dolomite weathering. The equivalent ratio of HCO3 − /(HCO3 − + SO4 2− ) is greater than 0.7 for all samples in this study, whereas the ratio lower than 0.7 in some samples indicates the influence of anthropogenic activity ( Figure 10) [10].    8 show that water chemistry is mainly controlled by limestone weathering, whereas these results cannot identify any anthropogenic activities. Figure 9 shows the relationship between 87 Sr/ 86 Sr ratios and HCO3 − /(HCO3 − + SO4 2− ) equivalent ratios in Xi River water, indicating information on silicate, limestone, and dolomite weathering. The equivalent ratio of HCO3 − /(HCO3 − + SO4 2− ) is greater than 0.7 for all samples in this study, whereas the ratio lower than 0.7 in some samples indicates the influence of anthropogenic activity ( Figure 10) [10].

Calculation Methodology
The inversion model, originally developed by Allègre and Lewin (1989), is commonly used to calculate the chemical weathering contribution [45]. For the estimation, we assumed the following: -All potassium was derived from silicate weathering; -Anthropogenic inputs were ignored or classified as atmospheric inputs; -Evaporite (including halite and gypsum) inputs were ignored.
The model was established on the assumption that dissolved loads originated from three end-members, including atmosphere, carbonate, and silicate. According to the contributions of various end-members, the inversion model was based on a series of mass budget equations of Na-normalized ionic molar ratios (X/Na = Ca/Na, Mg/Na, HCO 3 − /Na, Cl/Na and Sr/Na) [Equatioin (1)] and strontium isotopic compositions [Equatioin (2)] of the three end-members (atmosphere, carbonates, and silicates) ( Table 2).
where i represents the end-member (atmosphere, carbonates, and silicates). The α i (Na) represents the mixing proportion of Na in each end-member, and ∑ i α i (Na) = 1. The reason for normalization to Na is that Na + is unaffected by nutrient cycling. K + and SO 4 2− were ignored because they are easily affected by biological activities. The Na-normalized ionic molar ratios and Sr isotopic compositions were used to eliminate the effects of evaporation and discharge [34]. Table 2 shows the chemical compositions of each end-member [10,26,46].
Equations (1) and (2) were weighted by analytical errors of ionic molar ratios and strontium isotopic compositions (10% for elemental concentrations and 0.00002 for 87 Sr/ 86 Sr) to further reduce the error propagation through MATLAB 2022 software. A priori parameters of a series of end-member reservoirs were chosen first. Then, the posterior values that best matched the entire series of model equations were iteratively calculated by the inversion calculation algorithm. A total of 60 samples (30 in the high-water period and 30 in the low-water period) were used in this study. Two hundred model parameters (3i × 60α i (Na) + 20(X/Na) i ) were solved by successive iterations with equations (6 types × 60 samples = 360 mass balance equations and 60 constraint equations). The mixing proportions of other ions in each reservoir (e.g., α i (Ca) and α i (Mg)) were calculated with the posteriori values of α i (Na) for each sample and Na-normalized ionic molar ratios (X/Na) i of the reservoir (Table 4). Table 4 shows that the features of the end-members in the high-and low-water periods were different.
In the Xi River Basin during the high-water period, the percentage contents of the total of cations from the atmosphere, limestone, dolomite, and silicate were 14.10% (0~81.43%), 60.92% (0~84.48%), 9.17% (0~30.56%), and 15.80% (0~43.00%), respectively (Figure 11a). In the low-water period, these values were 15.62% (0~73.53%), 62.56% (14.16~79.63%), 9.25% (0~24.33%), and 12.57% (0.20~47.79%). Most samples had the largest proportion of cations from limestone, indicating that limestone was the major source of cations owing to the distribution of different types of exposed basement rocks. The contributions of silicate weathering in the high-water period were higher than those in the low-water period. The contributions of dolomite and the atmosphere revealed remarkable variations between highand low-water periods. Samples X5, X7, X14, and X15 were characterized by abnormally high contributions of precipitation, indicating the influence of anthropogenic activities. In the Xi River Basin during the high-water period, the percentage contents of the total of cations from the atmosphere, limestone, dolomite, and silicate were 14.10% (0~81.43%), 60.92% (0~84.48%), 9.17% (0~30.56%), and 15.80% (0~43.00%), respectively (Figure 11a). In the low-water period, these values were 15.62% (0~73.53%), 62.56% (14.16~79.63%), 9.25% (0~24.33%), and 12.57% (0.20~47.79%). Most samples had the largest proportion of cations from limestone, indicating that limestone was the major source of cations owing to the distribution of different types of exposed basement rocks. The contributions of silicate weathering in the high-water period were higher than those in the low-water period. The contributions of dolomite and the atmosphere revealed remarkable variations between high-and low-water periods. Samples X5, X7, X14, and X15 were characterized by abnormally high contributions of precipitation, indicating the influence of anthropogenic activities. The contributions of carbonate weathering were relatively high in the upper reaches and decreased downstream along the main stream (Figure 11b). Concomitantly, the contributions of silicate weathering increased downstream along the main stream ( Figure   Figure 11. The fraction of total dissolved cations (α i,Catin = α i,Ca + α i,Na + α i,Mg + α i,K ) from rain, dolomite, limestone, and silicate. (a). The fraction of total dissolved cations in all samples. (b). The fraction of total dissolved cations in each subbasin.
The contributions of carbonate weathering were relatively high in the upper reaches and decreased downstream along the main stream (Figure 11b). Concomitantly, the contributions of silicate weathering increased downstream along the main stream (Figure 11b).
These trends are broadly consistent with the predominant geological features in the basin. In the upper reaches, Permian to Triassic carbonate rocks are widely distributed and contribute to carbonate weathering. Precambrian metamorphic rocks and magmatic rocks in the lower reaches enhance the products of silicate weathering. For first-level tributaries, the ratios of the contribution of carbonate/silicate weathering showed an order of Beipan River > Liu River > Yu River > Gui River > He River (Figure 11b). For second-level tributaries, the You River subbasin had a lower contribution of carbonate weathering and a higher contribution of silicate weathering compared to the Zuo River subbasin (Figure 11b). Compared to the Long River subbasin, the Rong River subbasin had a lower contribution of carbonate weathering and a higher contribution of silicate weathering (Figure 11b). These phenomena were consistent with the order of carbonate/silicate area ratios in these subbasins. The spatial variation in the contribution of limestone weathering was different from that in the contribution of dolomite weathering, indicating the inhomogeneous distribution of limestone and dolomite in carbonate rock.
The contributions of the atmosphere were small in most sample settings, except for Samples X5, X7, X14, and X15. Relatively high contributions were recorded in the middle reaches (Qian and Xun Rivers) (Figure 11b), which represent high precipitation in the middle reaches. During the low-water period, relatively high contributions were also recorded in the upper reaches (Nanpan and Beipan Rivers) (Figure 11b), indicating a shorter reaction time with rocks.

Chemical Weathering Rates
The chemical weathering rates of silicates and carbonates were considered a dynamic process. Table 1 shows that the concentrations of NO 3 − were low or even zero in the Xi River Basin. Hence, we only considered that carbonic and sulfuric acid participated in chemical weathering. The silicate weathering rate (SWR) was defined as the sum of cations from silicate weathering: where α car (Na), α car (Mg), and α car (Ca) mean the proportions of each cation involved in carbonate weathering. The limestone weathering rate (LWR) was defined as the total amount of cations from limestone weathering: LWR = α lim (Na)ΦNa river M(Na) + α lim (Mg)ΦMg river M(Mg) + α lim (Ca)ΦCa river M(Ca) The dolomite weathering rate (DWR) could be calculated as follows: We estimated chemical weathering rates through the surface area, runoff, and discharge of the main stream and tributaries, expressed in t/km 2 /year or kg/km 2 /month. Representative samples from the trunk stream and tributaries were used to calculate the chemical weathering contributions in the Xi River Basin. The results are listed in Table 5. Based on Samples X1 and X2 in the main stream, the SWR values in the Xi River Basin were estimated at 281.38 kg/km 2 /month and 113.65 kg/km 2 /month in the high-and lowwater periods, respectively (Appendix A). The SWR varied from one subbasin to another. For tributaries, the SWR values ranged from 1.72 kg/km 2 /month to 1002.18 kg/km 2 /month and from 2.91 kg/km 2 /month to 492.02 kg/km 2 /month, respectively. The SWR values in the high-water period were 0.59-to 16.32-fold of the values in the low-water period. The CWR values in the basin were estimated at 2456.72 kg/km 2 /month and 1409.32 kg/km 2 /month in the high-and low-water periods, respectively. For tributaries, the CWR varied from 186.66 to 6252.48 kg/km 2 /month and from 170.75 to 2084.92 kg/km 2 / month, respectively. The CWR values in the high-water period were 1.09-to 9.00-fold of the values in the low-water period. The LMR was the major component of the CWR. The LWRs in the basin were estimated at 2042.74 kg/km 2 /month and 1222.38 kg/km 2 /month, respectively. For tributaries, the LWR varied from 131.65 to 5248.98 kg/km 2 /month and from 97.42 to 1699.29 kg/km 2 /month, respectively. The LMR values in the high-water period were 1.07-to 9.00-fold those in the low-water period. The DWR in the basin was estimated at 413.98 kg/km 2 /month and 186.94 kg/km 2 /month, respectively. For tributaries, the DWR varied from 0 to 1003.50 kg/km 2 /month and from 0 to 385.63 kg/km 2 /month, respectively. The DWR values in the high-water period were 1.10-to 4.07-fold greater than the values in the low-water period. Seasonal variations in chemical weathering rates were controlled by multiple parameters, including climate (temperature, water discharge, and precipitation) and so on [8,10,20,31,47]. During the high-water period, the warm and humid climate conditions associated with the Asian monsoon enhanced chemical weathering. Higher temperatures can promote the rapid dissolution of minerals [20]. Furthermore, a warm and humid climate can speed up plant degradation, thereby increasing the intensity of chemical weathering by the release of organic acids [10]. The main element dynamics are dominated by water discharge. Hydrological flushing increases the surface area for water-rock interaction and hence accelerates chemical weathering [31,48].
The SWR and CWR values in the upper reaches were lowest in the Xi River Basin. The upper reaches are characterized by relatively low water discharge, temperature, and precipitation. The SWR and CWR values in the Nanpan River were lower than the values in the Beipan River. This phenomenon can be explained by the higher vegetation cover in the Beipan River basin [35,49]. The SWR value increased from 0.03 t/km 2 /year in the upper reaches to 0.59 t/km 2 /year in the middle reaches. The CWR values increased from 2.14 t/km 2 /year in the upper reaches to 32.65 t/km 2 /year in the middle reaches. These phenomena can be explained by the fact that discharge, temperature, and precipitation increased from the upper to the middle reaches. The SWR increased from 0.59 t/km 2 /year in the middle reaches to 2.37 t/km 2 /year in the lower reaches, whereas the CWR decreased downstream along the main stream with values from 32.65 t/km 2 /year to 23.20 t/km 2 /year. Carbonate/silicate area ratios decreased from the middle to lower reaches. For the firstlevel tributaries, the SWR ranged from 0.12 t/km 2 /year to 8.97 t/km 2 /year, with an order of He River > Gui River > Yu River > Liu River > Beipan River ( Table 5). The CWR ranged from 6.09 t/km 2 /year to 34.12 t/km 2 /year, with an order of He River > Gui River > Liu River > Yu River > Beipan River ( Table 5). The spatial variation in the LWR was consistent with that in the CWR, whereas the spatial variation in the DWR was inconsistent. This phenomenon can be explained by the inhomogeneous distribution of limestone and dolomite in carbonate rock.

CO 2 Consumption Rate
The proportion of sulfuric acid in different rock weathering was an important factor in calculating the CO 2 flux absorbed by the chemical weathering of silicates. When we assumed that all SO 4 2− derived from gypsum coexisting with carbonates, the carbonic acid weathering of silicates (CSW) was equivalent to the CO 2 flux consumed by silicate weathering. Sulfuric acid played a significant role in chemical weathering processes due to anthropogenic activities. There was no CO 2 consumption during the sulfuric acid weathering of silicate. The CSW value was defined as the sum total of cations from the carbonic acid weathering of silicate: CSW = α sil (Na)ΦNa river + α sil (K)ΦK river + 2α sil (Mg)ΦMg river + 2α sil (Ca)ΦCa river − δ * 2ΦSO 4river (7) where δ represents the adjustment coefficient of sulfuric acid with a value from 0 to 1. The proportion of the sulfuric acid weathering of carbonate and silicate was equivalent to the contribution rate of carbonate and silicate to the total dissolved cations. CO 2 consumption involved in carbonate weathering by carbonic acid (CCW) could be expressed as: Furthermore, CO 2 generated by the sulfuric acid weathering of carbonate (SCW) could be expressed as: The corresponding CO 2 production due to the sulfuric acid weathering of limestone (SLW) was expressed as follows: The corresponding CO 2 production due to the sulfuric acid weathering of dolomite (SDW) was expressed as follows: where β represents the adjustment coefficient of limestone in carbonate with values from 0 to 1. β was calculated by the dissolved cation contribution ratio of limestone and dolomite, which could be further applied to calculate the related CO 2 flux of limestone and dolomite. The corresponding CO 2 consumed by the carbonic acid weathering of limestone (CLW) could be calculated as follows: The corresponding CO 2 consumed by the carbonic acid weathering of dolomite (CDW) could be calculated as follows: In this study, the CO 2 flux consumed by chemical weathering in the basin was estimated at 189.79 × 10 9 mol/year based on Samples X1 and X2. The CO 2 fluxes during the high-and low-water periods were 124.03 × 10 9 and 65.76 × 10 9 mol/year, respectively, accounting for 65.35% and 34.65% of the total flux. The water discharge in the basin during the high-water period was 2.63-fold that during the low-water period. The carbon sink was primarily controlled by the water cycle. The contributions of each endmember were slightly different in strength during different periods. The CO 2 fluxes consumed by silicate weathering during the high-and low-water periods were 24.04 × 10 9 mol/year and 9.38 × 10 9 mol/year, respectively, accounting for 12.67% and 4.94% of the total flux. The CO 2 fluxes consumed by carbonate weathering were 99.99 × 10 9 mol/year and 56.38 × 10 9 mol/year, respectively, accounting for 52.68% and 29.71% of the total flux. The CO 2 fluxes consumed by limestone weathering were 79.81 × 10 9 mol/year and 47.67 × 10 9 mol/year, respectively, accounting for 42.05% and 25.11% of the total flux. The CO 2 fluxes consumed by dolomite weathering were 20.18 × 10 9 mol/year and 8.71 × 10 9 mol/year, respectively, accounting for 10.63% and 4.59% of the total flux.
The total CO 2 fluxes consumed by silicate and carbonate weathering in the basin were estimated at 33.42 × 10 9 mol/year and 156.37 × 10 9 mol/year, respectively, accounting for 0.38% and 1.27% of the global CO 2 consumption fluxes (8.7 × 10 12 and 12.3 × 10 12 mol/year [13]) ( Table 5). For first-level tributaries, CO 2 fluxes consumed by chemical weathering ranged from 51.32 × 10 9 mol/year in the Yu River to 2.80 × 10 9 mol/year in the Beipan River. The CO 2 fluxes consumed by chemical weathering in the Beipan, Liu, Yu, Gui, and He River subbasins accounted for 1.47%, 17.03%, 27.04%, 6.96%, and 5.96% of the total CO 2 consumption flux in the Xi River Basin, respectively. The CO 2 fluxes consumed by silicate weathering ranged from 13.34 × 10 9 mol/year in the Yu River to 0.08 × 10 9 mol/year in the Beipan River. The CO 2 fluxes consumed by silicate weathering in the Beipan, Liu, Yu, Gui, and He River subbasins accounted for 0.24%, 22.63%, 39.90%, 11.15%, and 12.76% of the total CO 2 consumption flux, respectively. The CO 2 fluxes consumed by carbonate weathering ranged from 37.99 × 10 9 mol/year in the Yu River to 2.72 × 10 9 mol/year in the Beipan River. The CO 2 fluxes consumed by carbonate weathering in the Beipan, Liu, Yu, Gui, and He River subbasins accounted for 1.74%, 15.84%, 24.29%, 6.07%, and 4.51% of the total CO 2 consumption fluxes, respectively. The CO 2 fluxes consumed by limestone weathering ranged from 35.09 × 10 9 mol/year in the Yu River to 2.02 × 10 9 mol/year in the Beipan River. The CO 2 fluxes consumed by limestone weathering in the Beipan, Liu, Yu, Gui, and He River subbasins accounted for 1.59%, 16.23%, 27.53%, 7.44%, and 4.74% of the total CO 2 consumption fluxes, respectively. The CO 2 fluxes consumed by dolomite weathering ranged from 4.07 × 10 9 mol/year in the Liu River to 0 mol/year in the Gui River. The CO 2 fluxes consumed by dolomite weathering in the Beipan, Liu, Yu, Gui, and He River subbasins accounted for 2.40%, 14.10%, 10.02%, 0%, and 1.01% of the total CO 2 consumption fluxes, respectively.

Sulfuric Acid as Weathering Agent
The carbon budget of chemical weathering consisted of CO 2 consumption by chemical weathering and CO 2 emission by sulfuric acid weathering. The proportions of sulfuric acid weathering were plotted against the proportions of cations from carbonate weathering to research the effects of sulfuric acid weathering on atmospheric CO 2 ( Figure 12) [2,6]. The proportions of cations produced by carbonate weathering ranged from 49.45% to 99.67% ( Figure 12). The proportions of sulfuric acid weathering ranged from 7.26% to 22.42% ( Figure 12). The CO 2 budget by chemical weathering was dramatically influenced by sulfuric acid. The participation of sulfuric acid in carbonate weathering greatly promotes chemical weathering but reduces the CO 2 consumption flux.
Samples in the upper reaches and middle trunk stream were characterized by higher proportions of cations from carbonate weathering and higher proportions of sulfuric acid weathering, indicating a high sulfuric acid weathering contribution to atmospheric CO 2 emissions. The upper and middle reaches are carbon sources on a timescale of 10 7 years. Based on Samples X6 and X20 in the Xun River, the CO 2 flux produced by sulfuric acid weathering was estimated at 36.06 × 10 9 mol/year in the upper and middle reaches. However, the CO 2 flux consumed by silicate weathering was 25.43 × 10 9 mol/year. Therefore, the upper and middle reaches with a widespread distribution of carbonate were net carbon sources on a timescale of 10 7 years with a net released CO 2 flux of 10.63 × 10 9 mol/year. For tributaries, the Beipan River was a carbon source with a net released CO 2 flux of 0.90 × 10 9 mol/year. However, the Yu and Liu Rivers were carbon sinks with net consumed CO 2 fluxes of 8.48 × 10 9 mol/year and 3.08 × 10 9 mol/year, respectively.
Samples in the lower reaches were characterized by lower proportions of cations from carbonate weathering, which acted as carbon sinks ( Figure 12). The Gui and He Rivers were carbon sinks with net consumed CO 2 fluxes of 2.41 × 10 9 mol/year and 2.85 × 10 9 mol/year, respectively. Based on Samples X1 and X2 in the main stream, the CO 2 flux produced by sulfuric acid weathering was estimated at 30.00 × 10 9 mol/year in the Xi River Basin, accounting for 13.65% of the total CO 2 consumption. The CO 2 flux consumed by silicate weathering was 33.42 × 10 9 mol/year. Therefore, the Xi River Basin was a carbon sink with a net consumed CO 2 flux of 3.42 × 10 9 mol/year. Samples in the upper reaches and middle trunk stream were characterized by higher proportions of cations from carbonate weathering and higher proportions of sulfuric acid weathering, indicating a high sulfuric acid weathering contribution to atmospheric CO2 emissions. The upper and middle reaches are carbon sources on a timescale of 10 7 years. Based on Samples X6 and X20 in the Xun River, the CO2 flux produced by sulfuric acid weathering was estimated at 36.06 × 10 9 mol/year in the upper and middle reaches. However, the CO2 flux consumed by silicate weathering was 25.43 × 10 9 mol/year. Therefore, the upper and middle reaches with a widespread distribution of carbonate were net carbon sources on a timescale of 10 7 years with a net released CO2 flux of 10.63 × 10 9 mol/year. For tributaries, the Beipan River was a carbon source with a net released CO2 flux of 0.90 × 10 9 mol/year. However, the Yu and Liu Rivers were carbon sinks with net consumed CO2 fluxes of 8.48 × 10 9 mol/year and 3.08 × 10 9 mol/year, respectively.
Samples in the lower reaches were characterized by lower proportions of cations from carbonate weathering, which acted as carbon sinks ( Figure 12). The Gui and He Rivers were carbon sinks with net consumed CO2 fluxes of 2.41 × 10 9 mol/year and 2.85 × 10 9 mol/year, respectively. Based on Samples X1 and X2 in the main stream, the CO2 flux produced by sulfuric acid weathering was estimated at 30.00 × 10 9 mol/year in the Xi River Basin, accounting for 13.65% of the total CO2 consumption. The CO2 flux consumed by silicate weathering was 33.42 × 10 9 mol/year. Therefore, the Xi River Basin was a carbon sink with a net consumed CO2 flux of 3.42 × 10 9 mol/year.

Conclusions
We present new major ion and Sr isotope ratio data on the chemical evolution of the Xi river in relation to chemical weathering processes in the drainage basin. An inversion model was used to estimate the chemical weathering rates and CO 2 consumption fluxes in the Xi River Basin at monthly and annual scales. The primary conclusions were as follows: 1.
The water in the Xi River drainage is slightly alkaline with average pH values of 8.00 and 7.87 during the high-and low-water periods, respectively. The water was the HCO 3 -Ca/Mg type. The concentrations of Ca 2+ , Mg 2+ , HCO 3 − , and Sr decreased downstream along the main stream of the Xi River, whereas the 87 Sr/ 86 Sr ratios increased downstream. Spatial variations were consistent with the lithologic spatial distribution. Carbonates were most abundant in the upper courses, while more silicates appeared in the lower courses. Most major ion concentrations in the highwater period were in general lower than those in the low-water period. Seasonal variations were dominantly controlled by the water discharge, although a larger area of water-rock interaction could enhance chemical weathering. Variations in chemical weathering rates were controlled by climate (temperature, water discharge, and precipitation), vegetation, and so on. Higher temperatures, increased reactive mineral surface areas, and organic acids can accelerate chemical weathering.

2.
In the Xi River Basin, the SWR value was estimated at 2.37 t/km 2  The SWR values increased from 0.03 t/km 2 /year in the upper reaches to 2.37 t/km 2 / year in the lower reaches. The CWR values increased from 2.14 t/km 2 /year in the upper reaches to 32.65 t/km 2 /year in the middle reaches and then decreased to 23.20 t/km 2 /year in the lower reaches. The chemical weathering rates varied from one subbasin to another. The spatial variations in chemical weathering rates were controlled by lithology, vegetation, climate, and soil conditions. 4.
The CO 2 flux consumed by chemical weathering was 189.79 × 10 9 mol/year in Xi River drainage. The CO 2 fluxes consumed by carbonate and silicate weathering were 156.37 × 10 9 and 33.42 × 10 9 mol/year, respectively, accounting for 1.27% and 0.38% of the global CO 2 consumption fluxes. The CO 2 consumption fluxes by limestone and dolomite weathering were 127.48 × 10 9 and 28.89 × 10 9 mol/year, respectively. Sulfuric acid played a significant role in the CO 2 budget by chemical weathering. The CO 2 flux produced by sulfuric acid weathering was estimated at 30.00 × 10 9 mol/year in the basin. The upper and middle reaches were net carbon sources on a timescale of 10 7 years with a net released CO 2 flux of 10.63 × 10 9 mol/year. However, the Xi River Basin was a CO 2 sink with a net consumed CO 2 flux of 3.42 × 10 9 mol/year.

Data Availability Statement:
The data supporting reported results can be found in the tables in this text and Appendix A.

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