Introduction

The weathering of rocks containing silicate and carbonate minerals consumes CO2, which is an important inorganic carbon sink in the global carbon cycle1. An entire rock weathering process includes in situ weathering in the crust and off-site weathering during sediment transport. Atmospheric and aquatic CO2 is consumed by the weathering of silicate and carbonate minerals to form soluble bicarbonates in solution, which are transported into the sea through rivers. Some of the CO2 in the bicarbonates is deposited as Ca–MgCO3 (biological or chemical deposition) in the sea and is called the “Substantial Inorganic Carbon Sink”, and some of it is released back into the atmosphere and can be called the “Nonsubstantial Inorganic Carbon Sink”2. The chemical equations for Ca–Mg ions in calcium and magnesium silicates to consume and release CO2 during the weathering and deposition processes are expressed as follows:

$$\left( {{\text{Ca}}{-}{\text{Mg}}} \right)_{2} {\text{SiO}}_{4} + 4{\text{CO}}_{2} + 4{\text{H}}_{2} {\text{O}} \to \left( {{\text{Ca}}{-}{\text{Mg}}} \right)_{2} \left( {{\text{HCO}}_{3} } \right)_{4} + {\text{H}}_{4} {\text{SiO}}_{4}$$
(1)
$$\left( {{\text{Ca}}{-}{\text{Mg}}} \right)_{2} \left( {{\text{HCO}}_{3} } \right)_{4} \to 2\left( {{\text{Ca}}{-}{\text{Mg}}} \right){\text{CO}}_{3} + 2{\text{CO}}_{2} \uparrow + {\text{H}}_{4} {\text{SiO}}_{4}$$
(2)

In Eq. (2), the CO2 deposited via (Ca–Mg)CO3 is a substantial carbon sink (SCS), and the CO2 that is released is a nonsubstantial carbon sink (NSCS). In this manuscript, all the carbon sink terms refer to inorganic carbon sinks.

The chemical equation of absorbing and releasing CO2 by the Ca–Mg carbonate compounds during the weathering and deposition processes is expressed as follows.

$$\left( {{\text{Ca}}{-}{\text{Mg}}} \right){\text{CO}}_{3} + {\text{CO}}_{2} + {\text{H}}_{2} {\text{O}} \to \left( {{\text{Ca}}{-}{\text{Mg}}} \right)\left( {{\text{HCO}}_{3} } \right)_{2} \to \left( {{\text{Ca}}{-}{\text{Mg}}} \right){\text{CO}}_{3} \downarrow + {\text{CO}}_{2} \uparrow + {\text{H}}_{2} {\text{O}}$$
(3)

The CO2 consumed by carbonate weathering is a nonsubstantial carbon sink (NSCS), and it is released after entering the sea through rivers.

Suchet and Probst concluded that the CO2 consumption by rock weathering was mainly affected by the flow rate on the rock surface, atmospheric temperature and rock type3. They analyzed the data of surface runoff and major dissolved elements in 232 monolithologic basins in France and established a GEM-CO2 model to estimate the amount of atmospheric CO2 consumed by rock weathering based on the empirical relationships. The model was considered one of the main achievements of the international geological correlation program (IGCP) 404 project. Gaillardet et al. collected the hydrochemical data of 60 major rivers around the world and calculated that the global silicate weathering carbon sink amount was 1.4 × 108 t C per year and the carbonate weathering carbon sink amount was 1.48 × 108 t C per year through an inversion model4,5. It has been largely reported that calcium and magnesium minerals in sediment can dissolve CO2 and function as carbon sinks during river sediment transport, but the effects of human activities (such as damming) on the potential carbon sink capacity during sediment transport have not been reported until now6,7,8,9,10,11. Therefore, the objectives of this research are to (1) elucidate the mineral compositions and transport dynamics (offsite rock weathering) in the Yangtze River, (2) estimate the potential carbon sink capacity of suspended sediment transported in the Yangtze River, and (3) evaluate the effects of damming on the potential carbon sink capacity of suspended sediment.

Materials and methods

This study analyzed the chemical element and mineral composition data of suspended sediments at 25 sites and 13 sampling points in tributaries from the upper reaches to the estuary of the Yangtze River from 2003 to 2007 according to Ding Tiping's publication12. Twenty-five sampling sites in the mainstream and 13 sites in the tributaries of the Yangtze River were sampled 4 times (July 2003, April 2004, July 2005 and July 2007) during 2003–2007, as shown in Fig. 1 and Table 1. The variations in the CaO, MgO (during 4 campaigns), calcite and dolomite contents (one campaign in July 2005) in suspended sediment of the Yangtze River along the river are shown in Fig. 2.

Figure 1
figure 1

The map of Yangtze River Basin and the sampling points (Cited from Ding et al.12). Note: The sketch of geological map in the study region was redigitized from Ding’s publication using ARCGIS 10.4 (www.esri.com/en-us/store/overview).

Full size image
Table 1 Sampling points of suspended sediment in the main stream and tributaries of the Yangtze River.
Full size table
Figure 2
figure 2

The variations of CaO, MgO, calcite and dolomite contents in suspended sediment along the mainstream of the Yangtze River.

Full size image

The suspended sediments were not measured at some sampling sites in April 2004 because the sediment concentration in water was too low to be collected in the dry season of April 2004. Furthermore, the sediment yields in the flood seasons of May–October account for more than 75% of the yield for a year; therefore, the mean values of the chemical and mineral contributions of the three campaigns in July were used for the following calculation. However, the value in July 2005 at Zhenjiang was irregular and eliminated from the calculation.

The carbon sink effects of silicate and carbonate mineral weathering during the sediment transport processes were estimated based on the variation in the CaO, MgO, calcite and dolomite contents in suspended sediments along the mainstream of the Yangtze River. In this study, we defined the CO2 consumed per unit weight by the dissolution of CaO + MgO in silicate and carbonate minerals of suspended sediment to bicarbonate forms as the total carbon sink (TCS) capacity of suspended sediment (C1, t/t). In more detail, the molar masses per unit mass of CaO, MgO, calcite (CaCO3) and dolomite (CaCO3+MgCO3) were calculated based on the contents of CaO, MgO, calcite and dolomite in the measured sediment samples at each site, and the respective carbon sink capacities (which consumed CO2 per ton during sediment transport) were calculated based on the same molar mass of Ca/Mg. According to the definition and Eq. (2) mentioned above, the CO2 consumed by the weathering of Ca/Mg carbonate minerals is a nonsubstantial carbon sink (NSCS Cc); this yields approximately one mole of CO2 from one mole of weathered CaCO3, which is released again when carried into the oceans. Moreover, from Eq. (1), we infer that one mole of the CO2 consumed by the weathering of one mole Ca/Mg silicate minerals is a substantial carbon sink (SCS, Cs1) and the other mole is a nonsubstantial carbon sink (NSCS, Cs2):

$${\text{C}}_{1} = {\text{C}}_{2} + {\text{C}}_{3}$$
(4)
$${\text{C}}_{2} = {\text{Cc}}_{{\text{t}}} + {\text{C}}_{{{\text{s}}2}}$$
(5)
$${\text{C}}_{3} = {\text{C}}_{{{\text{s}}1}}$$
(6)

where C1 is the TCS capacity (t/t); C2 is the NSCS capacity (t/t); C3 is the SCS capacity (t/t); Cc is the NSCS capacity of Ca/Mg carbonate minerals (t/t); Cs1 is the SCS capacity of calcium-magnesium silicate minerals (t/t); and Cs2 is the NSCS capacity of Ca/Mg silicate minerals (t/t).

In fact, the SCS capacity was calculated by subtracting the NSCS capacity (based on the molar mass of sequestered CO2 by weathering of calcite and dolomite) from the TCS capacity (based on the molar mass of sequestered CO2 by weathering of all minerals containing CaO/MgO) in this study. Subsequently, the influence of sediment yield changes on the carbon sink contributed by off-site rock weathering in the Yangtze River was preliminarily evaluated via an “ideal mainstream segment” method, especially for comparing the two periods of preimpoundment (1956–2000) and postimpoundment (2003–2017) of the Three Gorges Reservoir (TGR), which was gradually impounded from 2003 to 2008.

Results and discussions

CaO, MgO, calcite and dolomite contents of suspended sediments in the Yangtze River and its main tributaries

Figure 2 shows a declining tendency of the CaO + MgO and calcite + dolomite of suspended sediments in the mainstream of the Yangtze River from upstream to downstream. The total CaO + MgO contents along the Yangtze River were as follows: Tuotuo River, 16.33%; Yibin, 11.43%; Cuntan, 10.35%; Yichang (below the Three Gorges Dam), 6.17%; Wuhan Industrial Port, 6.61%; Datong, 4.80%; and Wusongkou, 4.60%. The total calcite + dolomite contents also decreased along the river as follows: Tuotuo River, 16.8%; Yibin, 9.1%; Cuntan, 6.2%; Yichang, 4.1%; Wuhan Industrial Port, 7.4%; Datong, 4.2%; and Wusongkou, 1.5%. The three stations closest to the estuary were dominated by dolomite and free of calcite.

Due to the dramatic terrain changes in the upper reaches of the Yangtze River, strong gravity erosion and physical weathering, the CaO + MgO and calcite + dolomite contents were high in the mainstream section of the Jinsha River above Yibin and decreased from the headwater to the estuary, which clearly illustrated the dissolution of calcium-magnesium silicate and carbonate minerals during the process of sediment transport in the river. The declining rates of CaO + MgO and calcite + dolomite contents in the upper reaches of the Yangtze River above Yichang were 0.12%/100 km and 0.29%/100 km, respectively, while the rates below Yichang were 0.09%/100 km and 0.15%/100 km, respectively. The declining rate of calcite and dolomite contents was higher than that of CaO + MgO contents, which indicated that carbonate minerals were more likely to be dissolved than calcium-magnesium silicate minerals during sediment transport. Because the upstream reaches had greater slope gradients, faster flow velocities and consequently higher mineral dissolution rates, the CaO + MgO and calcite + dolomite contents of suspended sediments in the upper reaches had a higher declining rate than those in the middle and lower reaches.

The CaO + MgO and calcite + dolomite contents of the suspended sediment in the main tributaries of the Yangtze River were as follows: Minjiang River: 9.75% and 6.3%; Jialing River: 5.40% and 5.5%; Wujiang River: 10.87%; Xiangjiang River: 4.75%; Hanjiang River: 4.41%; and Ganjiang River: 2.87% (only CaO + MgO contents and no calcite + dolomite content data for the last four tributaries). Except for the Wujiang River, CaO + MgO contents and calcite + dolomite contents in the Minjiang River were higher than those in other tributaries and close to the Jinsha River (Yibin site) because the Minjiang River basin had similar environments for erosion and sediment transport to the Jinsha River Basin. The CaO + MgO contents of suspended sediments in the Wujiang River were quite different between July 2003 and July 2007, i.e., 6.67% and 15.06%, respectively. The relatively high contents might be due to its widespread distribution of carbonate rock.

Carbon sink capacity during suspended sediment transport in the Yangtze River mainstream

According to Eqs. (1)–(6), the calculated TCS capacity (C1) decreased gradually from the headwater to the estuary (Fig. 3) in the following order: Tuotuo River, 0.271 t/t; Cuntan, 0.151 t/t; Yichang, 0.117 t/t; Wuhan Industrial Port, 0.127 t/t; Datong, 0.092 t/t; and Wusongkou, 0.091 t/t (Table 2). As CaO + MgO contents decreased, so did the TCS capacities. This result verified that CO2 was consumed by the dissolution of Ca–Mg minerals during sediment transport from upstream to downstream and that a carbon sink function existed. The TCS capacities at Cuntan and Wusongkou were 0.151 t/t and 0.091 t/t, respectively. A total of 0.060 tons of CO2 per ton of suspended sediment was dissolved during transport from Cuntan to the sea.

Figure 3
figure 3

The variation of carbon sink capacities of suspended sediment in the Yangtze River.

Full size image
Table 2 The variation of carbon sink capacity and potential of suspended sediment along the Yangtze main stream.
Full size table

The SCS capacities (C3) of silicate minerals in the sediment were in the range of 0.027–0.047 t/t and had little variation, except for the Tuotuo River (0.061 t/t). The NSCS capacity (C2) was consistent with the variation in the TCS capacity, and both had a gradual decreasing tendency from the headwater to the estuary (Fig. 3), as follows: Tuotuo River, 0.210 t/t; Cuntan, 0.104 t/t; Yichang, 0.078 t/t; Wuhan Industrial Port, 0.097 t/t; Datong, 0.065 t/t; and Wusongkou, 0.051 t/t. The silicate carbon sink capacity (SCS) was smaller and showed a smaller reduction than the NSCS along the Yangtze River, mainly due to the slower dissolution rate of silicate minerals or greater contribution of silicate rock clastics from the watersheds in the middle and lower reaches, which also limited CO2 consumption in comparison with carbonate minerals. Obviously, the decrease in TCS capacity from upstream to downstream was due to the intense dissolution of carbonate minerals in the Yangtze River.

An “ideal mainstream segment” refers to a segment where there was no sediment input from the tributaries or the amount of sediment supply from the tributaries was equal to the amount of sedimentation in the segment (meaning that the suspended sediment yields at the inlet and outlet were similar), and the calcium and magnesium mineral contents of suspended sediments in the tributaries and mainstream were similar. The carbon sinks via CO2 consumption by dissolution of calcium and magnesium minerals in the sediment transport process can be expressed as follows:

$${\text{Wt}}_{{{\text{h}}1 - 2}} = {\text{Ws}}_{{{\text{h}}1 - 2}} \times \left( {{\text{C}}_{{{\text{h}}1}}{-}{\text{C}}_{{{\text{h}}2}} } \right)$$
(7)

where Wth1-2 is the consumed CO2 via the dissolution of calcium and magnesium minerals in the segment (h1–h2) (104 t/yr); WSh1-2 is the mean suspended sediment transported at the segment (h1–h2) (104 t/yr); Ch1 is the carbon sink capacity of the suspended sediment at the inlet of the segment (t/t); and Ch2 is the carbon sink capacity of the suspended sediment at the outlet of the segment (t/t).

The Yangtze River has a drainage area of 1785 × 106 km2, in which the upper reaches above Yichang have an area of 1.05 × 106 km2. The average sediment yield from 1956 to 2000 was 5.01 × 108 t/yr at Yichang, while it was 4.33 × 108 t/yr at Datong, with a drainage area of 1.705 × 106 km2. In addition, it was 4.39 × 108 t/yr at Cuntan13 with a drainage area of 0.867 × 106 km2. Although there is a large drainage area of the river segment between Cuntan and Datong (0.838 × 106 km2), the sediment yields at Cuntan and Datong were very similar, namely, 4.39 × 108 t/yr, and 4.33 × 108 t/yr, respectively, because sediment deposition in the channels of the segment offset the sediment supply from the tributaries. The contents of CaO and MgO in the suspended sediment in the tributaries (Hanjiang River, Ganjiang River and Xiangjiang River) in the middle and low reaches below the TGR dam were nearly the same as the values in the mainstream of the Yangtze River (Hankou station)12. Thus, we regarded the mainstream segment between Cuntan and Datong as an ideal mainstream segment, with Cuntan being upstream of the TGR dam, which made the evaluation of the damming effects available. The CaO + MgO contents of suspended sediment samples for the three campaigns (July 2003, July 2005 and July 2007) and the calcite + dolomite contents of the sediment (July 2005) at the two sites (Table 2) were used to calculate differences in TCS, NSCS and SCS capacities for the segment. Taking the mean sediment yield of the two sites during the period from 1956 to 2000 for reference, the annual net TCS, NSCS and SCS between the two sites were 26.45 × 106 tons of CO2, 17.51 × 106 tons of CO2 and 8.94 × 106 tons of CO2, respectively. After 2001, due to hydropower exploration (especially the TGR project), ecological mitigation and soil conservation, the sediment yields at the two sites largely decreased and have stabilized since 2006. By comparison to the period before 2000, the sediment yields at Cuntan and Datong decreased by 72.4% and 71.6%, respectively, during the period of 2006–201914. Due to the reduction in sediment yields, the annual net TCS, NSCS and SCS in the segment decreased by 18.52 × 106 tons of CO2, 12.24 × 106 tons of CO2 and 6.28 × 106 tons of CO2, respectively (Table 3).

Table 3 Reduction in the carbon sink caused by the reduction in sediment transport in the reach between Cuntan Station and Datong Station after 2006.
Full size table

Carbon sink capacities of the global rivers and their implications

The amount of deposited sediments of the Three Gorges Reservoir (TGR) from June 2003 to December 2017 was 1.669 × 109 tons, and the average sedimentation rate was 114.5 × 106 t/yr15. According to the differences in the TCS, NSCS and SCS capacities of the suspended sediments between Cuntan and Datong, the losses of annual TCS, NSCS and SCS by sedimentation in the TGR were estimated to be approximately 6.756 × 106 tons of CO2, 4.466 × 106 tons of CO2 and 2.290 × 106 tons of CO2 × 106 tons of CO2, respectively. The power generation of the Three Gorges Hydropower Station exceeded 100 × 109 kW/h in 2018, equivalent to saving 31.9 × 106 tons of standard coal and reducing 85.80 × 106 tons of CO2 emissions16. The reduction in inorganic carbon sinks from the sedimentation in the TGR was equivalent only to a limited amount (TCS for 7.9% and SCS for 2.7%) of the reduced CO2 emissions by the Three Gorges Hydropower Station. Moreover, the sediment deposited by the TGR could bury and store vast quantities of organic carbon6, and particulate organic carbon (POC) contents in the TGR in recent reports17,18,19 varied from 0.26 to 9.2%. An average of 1.5% of POC in buried sediment was used to estimate annual buried organic carbon, and the relatively permanent sedimentation of organic carbon was equivalent to 6.30 × 106 tons of CO2 sequestration. The losses of annual TCS and SCS via silicate weathering by the TGR project could also offset 107.28% and 36.36% of its annual CO2 sequestration (6.30 × 106 tons) via permanent sedimentation, respectively.

From a global perspective, 4462 rivers with basin areas of more than 100 km2 showed that the current annual sediment flux of global rivers into the sea was 12.61 × 109 tons20. Taking the preimpoundment period TCS of 0.060 t/t for the stream segment between Cuntan and Wusongkou (the mouth of the Yangtze River) into consideration, the total inorganic carbon sink amount of 7.57 × 108 tons of CO2 was derived from global rivers, which is equivalent to 71.6% of the total inorganic carbon sink of global rock weathering (1.056 × 109 tons of CO2), with weathered silicate being more than 26% of the total weathered rocks4,6,10. The enhanced silicate rock weathering (ERW) strategy proposed by Beerling et al.1 would create a higher annual SCS, reaching 2 × 109 tons of sequestered CO2. To achieve this goal, there is no doubt that the inorganic carbon sink amount contributed by the dissolution of calcium and magnesium minerals in the processes of river sediment transport accounted for a great portion of the carbon sink amount via global rock weathering. The collision and abrasion of river sediments combined with stirring and mixing could promote the dissolution of minerals during sediment transport processes (off-site weathering of the rock). Therefore, it was suggested that the dissolution rate of off-site rock weathering was higher than that of in situ weathering. In comparison to the periods with limited anthropogenic influences, the global sediment fluxes to the sea decreased by approximately 10%20, and the corresponding total carbon sink loss in a year was estimated to be 0.757 × 109 tons of CO2, which was still less than the amount of CO2 emission reduction contributed by hydropower exploration and the associated buried organic carbon per year.

Conclusions

The rock weathering processes included in situ weathering during the development of weathering crust and off-site weathering during sediment transport. The dissolution of silicate and carbonate minerals in river sediment consumed CO2 in river water. From June 2003 to December 2017, the average annual sedimentation of the Three Gorges Reservoir (TGR) was 114.5 × 106 tons, and the related reduction in inorganic carbon sinks from sedimentation by the TGR was only equivalent to a limited amount (7.9% for the TCS and 2.7% for the SCS) of the reduced CO2 emissions by the Three Gorges Hydropower Station. The TCS of global rivers was estimated as 757 × 106 tons, which is equivalent to 71.6% of the TCS by global rock weathering with 1.06 × 109 tons of sequestered CO2, among which the SCS was more than one quarter of the TCS. The dissolution of calcium and magnesium minerals during river sediment transport accounted for a great portion of the carbon sink amount by global rock weathering. The collision and abrasion of river sediments combined with stirring and mixing could promote the dissolution of minerals during sediment transport processes (off-site weathering of the rock). Therefore, it is suggested that more attention should be given to dissolution processes and rates of off-site rock weathering, which might be more significant than in situ weathering, especially in the upper reaches of large rivers.