The Wuding River Basin, covering an area of approximately 30,000 km², is located in the northern part of Shaanxi Province, China (108°27′39″–110°34′22″E, 37°02′31″–38°55′52″N). The main channel of this water system, the Wuding River, originates from the northern foothills of Baiyu Mountain in Dingbian County, Shaanxi Province, and extends for 491 km, eventually merging into the Yellow River near Hekou Village in Qingjian County, Yulin city, Shaanxi Province, and becoming a first-class tributary of the Yellow River as a main stream. The main tributaries in the basin are the Yuxi River, the Lu River, the Dali River, the Huaining River, the Hailiutu River, the Nalin River, and the Heimutouchuan River. The basin has an arid and semiarid monsoon climate, with an average annual temperature of 9.5°C and an average annual precipitation of 409.1 mm, of which approximately 74% occurs from June to September, and the average annual evapotranspiration is 1,200 mm (Figure 1, data sourced from http://eia-data.com/) (HE et al., 2025). The Wuding River Basin is divided into upstream and downstream watersheds based on the direction of the river’s flow. The upstream region of the watershed mainly flows through the wind-sand grassland area, while the downstream mainly flows the loess hill and gully area. In this study, on the basis of remote sensing data (https://zenodo.org/records/12779975) from August 2023, a buffer zone with a radius of 5 km was created around each sampling point via ArcGIS version 10.2, and the areas of different land use types, such as cropland, forest, and grassland, were analysed and quantified within each buffer zone. The land use in the study area is dominated by grassland, cropland, and urban land, which account for 74.7%, 23.2%, and 1.1% of the total watershed area, respectively (Figure 2).

In this study, 21 sampling sites (WD1–WD21) were deployed along the mainstem and tributaries of the Wuding River in August 2024; sampling sites WD1–WD11 were located in the downstream area, whereas WD12–WD21 were located in the upstream area (Figure 1). At each sampling site, two sediment subsamples were collected using a stainless steel Peterson grab sampler within 20–30 m north and south of the central location. These subsamples were combined in a sterilized self-sealing bag to form a single representative composite sediment sample per site. From each composite sample, three aliquots were subsequently prepared: one aliquot was refrigerated at −4°C for physicochemical parameter analysis, a second aliquot was frozen at −20°C for organic carbon isotope determination, and the third aliquot was frozen at −80°C for microbial sequencing analysis.

Stable carbon and nitrogen isotope three-terminal elemental Bayesian Monte Carlo (Bayesian Markov chain Monte Carlo (MCMC)) models can be used to quantify the relative contributions of potential sources to DOC. The δ¹³C‒δ¹⁵ N isotope mixing equations read as Equation 2 (Li et al., 2020; Yan et al., 2023):   13 C   12 C m i x t u r e = ∑ i = 1 n x i K i C   13 C   12 C i / ∑ i = 1 n x i K i C   15 N   14 N m i x t u r e = ∑ i = 1 n x i K i N   15 N   14 N i / ∑ i = 1 n x i K i N ∑ i = 1 n x i = 1 (2) Sources of δ¹³C and δ¹⁵N end-member values for DOC were selected based on the major land-use types within and around the Wuding River Basin and previous research in the field, including: terrestrial plant/soil sources, plankton sources, and sewage sources. The mean values of δ¹³C and δ¹⁵N measured in samples from each source were defined as the end-member values (Cai et al., 2004; Liu and Xing, 2012; Yang et al., 2023). The δ¹³C and δ¹⁵N isotope ratios of terrestrial source end elements were −25.74‰ ± 0.40‰ and 11‰ ± 1.46‰, respectively (Rao et al., 2017); whereas the δ¹³C and δ¹⁵N isotope ratios of zooplankton end elements were −23.7‰ ± 0.96‰ and 7‰ ± 0.95‰, respectively (Gu et al., 2017; Yan et al., 2023); and the δ¹⁵N isotope values of domestic sewage sources δ¹³C: −25.32‰ ± 1.00‰, δ¹⁵N: 10‰ ± 0.28‰ (Machiwa, 2010). δ¹⁵N isotope values were extracted by kriging interpolation and smoothing on the basis of data from the August 2023 contemporaneous study of Xu et al. (2023) (Xu et al., 2023).

The R language program package MixSIAR 3.6.3 was used to quantify the contributions of sources of DOC in the sediments of the Wuding River Basin. Statistical analyses, including correlation analysis, t-test, and principal component analysis (PCA), were performed via Origin 2024 software (Microcal Software, Inc., Northampton, MA). Graphs for this study were plotted via ArcGIS 10.5 and Origin 2024b software.

The SS and physicochemical properties of the river sediments in the Wuding River Basin exhibited spatial heterogeneity (Table 1). The overall variation in the SS content in the river was in the range of 5–247056 mg/L, with significantly higher SS concentrations downstream than upstream (Supplementary Figure S1). Sediment pH exhibited alkaline to strongly alkaline conditions (range: 8.26–8.94; mean ± SD: 8.63 ± 0.18, n = 21), with significantly higher values observed upstream (8.70 ± 0.14) than downstream (8.58 ± 0.20) (p > 0.05; Figure 3A; Supplementary Table S1). The TOC content ranged from 0.32 to 4.29 g/kg, the mean value of the upstream TOC content was 0.90 g/kg, the mean value of the downstream TOC content was 2.30 g/kg, the TOC content increased by 155.6% from upstream to downstream, and there was a significant difference between the upstream and downstream values (p < 0.05) (Figure 3B; Supplementary Table S1). The mean TP concentration was 0.36 g/kg upstream and 0.48 g/kg downstream, with coefficients of variation of 47.22% (upstream) and 14.58% (downstream), respectively, and the spatial distribution showed a trend towards downstream enrichment, where the TP concentration increased by 33.3% from upstream to downstream (Figure 3C; Supplementary Table S1). The range of AP concentrations was 1.14–10.56 mg/kg, and the mean AP concentration was 2.30 mg/kg upstream and 6.66 mg/kg downstream, with significant differences (p < 0.001) between upstream and downstream AP (Figure 3D; Supplementary Table S1). The mean TN concentration was 0.07 g/kg upstream and 0.26 g/kg downstream, and the TN concentration increased by 271.4% from upstream to downstream, with a significant difference between upstream and downstream (p < 0.001) (Figure 3E; Supplementary Table S1). The mean AN content both upstream and downstream was 0.03 g/kg, and there was no statistically significant difference between upstream and downstream (p > 0.05) (Figure 3F; Supplementary Table S1). In brief, AP, TN and TOC were significantly different upstream and downstream.

The DOC content of river sediments in the Wuding River Basin ranged from 27.78 mg/kg to 77.45 mg/kg (mean 53.06 ± 15.78 mg/kg). The mean DOC content was 39.35 mg/kg upstream and 65.53 mg/kg downstream, and the DOC content increased by 66.53% from upstream to downstream; the difference in the DOC content between upstream and downstream areas was significant (p < 0.001) (Figure 4).

The C/N values, δ¹³C and δ¹⁵N compositions of river sediments in the Wuding River Basin ranged from 6.42 to 46.38, −27.62–23.06, and 1.04‰–11.40‰, respectively. The C/N values ranged from 2.12 to 29.22 (mean 9.84) upstream and from 1.60 to 4.21 (mean 2.71) downstream; the δ¹³C ratio ranged from −23.06–27.62‰ (mean 6.58‰ ± 1.22‰) upstream and δ¹³C ranged from −24.04 to −26.21‰ (mean −25.11‰ ± 0.68‰); the δ¹⁵N ratio ranged from 1.04‰ to 10.36‰ (mean 6.85‰ ± 3.07‰) upstream; and the δ¹⁵N ratio ranged from 8.80‰ to 11.40‰ (mean 9.86‰ ± 0.79‰) downstream (Table 1).

In this study, TOC and TN were significantly positively correlated (r = 0.78, p < 0.001, Spearman’s rank correlation coefficient) (Figure 5A), and the strong correlation between DOC and TN may indirectly indicate the stability of source singularity or the mixing ratio. The C/N values for the sediments in the Wuding River Basin ranged from 7.12 to 43.80 (mean 14.05), with upstream C/N values ranging from 6.42 to 46.38 (mean 20.12) and downstream C/N values ranging from 6.88 to 10.38 (mean 8.46); in this study, the mean C/N value was 13.77, indicating a mixed source. There was a significant negative correlation between δ¹³C and C/N (Figure 5B), and in general, the δ¹³C values of terrestrial higher plants (approximately −25 to −30‰) were lower than those of aquatic organisms (approximately −20% to −25%). With increasing input of organic matter from terrestrial sources, the TOC content increases, and the proportion of terrestrial materials in sediments increases, resulting in lower δ¹³C values. There was a positive correlation between δ¹⁵N and C/N (Figure 5C), and the sediment C/N value and δ¹⁵N composition gradually increased from upstream to downstream of the Wuding River sediments.

Spatial heterogeneity was observed in the allocation of DOC sources to sediments from the Wuding River Basin. A three-terminal element mixing model was constructed with typical δ¹³C and δ¹⁵N compositions to quantify the relative contributions of three potential sources of DOC (terrestrial sources, sewage sources and planktonic sources). Figure 6 shows the results of the Monte Carlo modelling of the proportional contribution of DOC sources. In the upstream region, terrestrial sources contribute 35% ± 0.011%, sewage sources 34% ± 0.052% and planktonic sources 39% ± 0.030%. In contrast, in the downstream region, the Terrestrial contribution increased significantly to 54% ± 0.034%. The sewage contribution decreased to 31% ± 0.018%, and the plankton contribution decreased to 18% ± 0.022%. Although the endogenous contribution remained high in the river system (34%–31%), the terrestrial contribution gradually increased as the river moved from upstream to downstream (Figure 6).

In this study, it was found that the terrestrial contributions was significantly higher in the downstream (55.7%) than in the upstream (12.7%) (Figure 7; Supplementary Table S2). Although the land use within the 5 km buffer zone of the whole Wuding River basin is dominated by grassland and cultivated land, and there is no significant difference in the proportion of cultivated land coverage between the upstream and downstream (Figure 2), the downstream is more directly adjacent to the intensive agricultural activities and urban areas. This spatial proximity means that downstream receiving sites are more susceptible to surface runoff input from these areas of high human activity. This spatial proximity means that the downstream is more susceptible to surface runoff input from these areas of high human activity. The contribution of terrestrial organic matter is significantly enhanced by fertilizer, sewage, and associated fulic acid exogenous organic carbon carried by surface runoff (Lambert et al., 2017; Varma et al., 2022). Wu et al. (2023) supported this view by finding that the concentration of biodegradable organic matter (BDOM) in urbanized areas was significantly higher than that in natural areas, and that the proportion of agricultural and urban land was positively correlated with the fluorescence fraction of terrestrial sources (Wu et al., 2023). It follows that the intensity of human activity, rather than the simple proportion of LU types, through enhanced surface runoff input, is the key factor driving the significantly higher contribution of downstream terrigenic sources. In addition to anthropogenic inputs, differences in DOC sources between upstream and downstream may also be influenced by other factors, such as potential differences in the specific species composition of C3 plants in vegetation communities, soil organic matter turnover processes, and changes in river hydrodynamic conditions such as water velocity through sediments and residence time of organic matter. For example, the δ¹³C values of sediments in the Wuding River basin fall within the typical range of C3 plants (−29.5‰ to −24.4‰) in both the upper (−23.06 to −27.62‰) and lower (−24.04 to −26.21‰) reaches (Zhuang et al., 2023). These factors may work together to result in a more significant terrigenous organic matter signal in downstream sediments.

Sewage sources contributed more to the upstream (48%) than to the downstream (20.8%). The mean δ¹⁵N composition of the upstream was 6.85‰ (Supplementary Table S2), and the sources were usually more complex in the downstream of the river than in the upstream, e.g., the degree of humification and aromaticity of CDOM were significantly higher in the tributaries of the Qinghai Lake Basin than in the lake itself, with the dominant role of terrestrial source input in the downstream (Yu X.-Q. et al., 2022). In addition, Zhang et al. who studied the analysis of organic matter sources in sediments from Qingpu District, Yangtze River Delta, also showed that the contribution of soil organic matter was as high as 96.8%, diluting the proportion of sewage sources (Zhang Z.-B. et al., 2022).

The contribution of organic matter from planktonic sources was greater upstream (approximately 39.3%) than downstream (approximately 23.4%) (Supplementary Table S2). The greater contribution of the plankton source (39.3%) in the upper reaches was due to two main aspects: first, the plankton source was associated with the δ¹³C value, and when the productivity of the water body was relatively high, the plankton consumed DOC rapidly, resulting in weakened fractionation and high δ¹³C, as the mean value of δ¹³C in the upper reaches of the present study was −24.52‰ higher than that in the lower reaches of the study, which was −25.11‰. Second, the geomorphological type of the upstream area is a windy sandy beach area with flat topography and slow flow, and the slow flowing waters of the river can increase the light intensity (Walks, 2007). In addition, the water quality is better, which can provide a favourable environment for the growth of plankton (Wen et al., 2020). In the downstream area, the reduced contribution of phytoplankton sources was influenced mainly by the input of organic matter from terrestrial sources. Moreover, decreased productivity and negative δ¹³C (mean −25.11‰) due to nutrient dilution or light limitation may have overlapped with C3 plant signals from terrestrial sources, which presently led to the reduced contribution of planktonic biogenic sources.

The strong sediment transport characteristics of the Wuding River Basin lead to highly dynamic carbon exchange at the sediment‒water interface. With a high sediment background, DOC regulates its source‒sink function by altering environmental conditions and carbon burial efficiency, which in turn affects carbon sink potential of the basin. Therefore, in order to better understand the role of DOC in regulating the health of river water quality, it is necessary to consider the driving factors that influence the burial and degradation processes of DOC. In this study, PCA analysis showed that DOC was negatively correlated with pH, C/N, and δ¹³C, and positively correlated with AP, TN, and TOC (Figure 7A). The generated heatmap also showed that DOC was significantly correlated with pH, TC, AP, TP, AN, TN, C/N, δ¹³C, δ¹⁵N, and other indicators (Figure 7B).

In this study, a total of 20 carbon cycle reaction pathways were identified and predicted via the DiTing tool (CX et al., 2021). Among them, the major contributor was the TCA cycle. Overall, the downstream carbon cycle potential was greater than the upstream carbon cycle potential, such as the TCA cycle, Wood–Ljungdahl pathway (WL), glycolysis (glycolysis), and organic acid metabolism pathways (Figure 8).

In this study, DOC was found to be correlated with the Wood–Ljungdahl pathway, the methanogen pathway, the rTCA cycle, and fermentation to lactate in the carbon cycle (Supplementary Figure S2). Among them, DOC in upstream sediments was significantly and positively correlated with methanogenesis (p < 0.001) (Figure 9A). This occurred for two main reasons: first, grassland apoplastic sediments are rich in easily decomposable carbohydrates, and their rapid decomposition can release a large amount of DOC in a short period (Wang et al., 2025); second, the high porosity of sandy sediments facilitates the migration and local enrichment of DOC (Zhan et al., 2023). Second, owing to the flat topography of windy sandy beach terrain, low surface water mobility, and slow pore water exchange in the sediment, a stable anaerobic layer is easily formed, and the presence of DOC in this environment can provide available substrate and reducing power for functional genes that act as methanogens, driving the process of CO₂ reduction to CH₄ (Sun et al., 2021; Zhou et al., 2019). In this anaerobic microenvironment, the concomitant decomposition of DOC and subsequent release of CO₂ drive carbon fixation via the Wood–Ljungdahl pathway. Therefore, DOC in upstream sediments was also significantly positively correlated (p < 0.001) with Wood–Ljungdahl (known as the reduced acetyl coenzyme A pathway, which is a key pathway for carbon fixation in the carbon cycle) (Figure 9A). Downstream, DOC in the sediments was significantly (p < 0.001) negatively correlated with the rTCA cycle (Figure 9B). This may be because high DOC concentrations provide abundant substrates (McDonough et al., 2020) for heterotrophic microorganisms (which rely on readily available organic carbon) to rapidly acquire energy through catabolism (e.g., glycolysis, fermentation) (Liao et al., 2024) without relying on the carbon fixation process of the rTCA cycle.

In summary, DOC is a key factor regulating carbon cycle pathways. DOC supports anaerobic carbon fixation and methanogenesis, realizing carbon regeneration; downstream, human interference leads to excessive DOC, breaking the autotrophic-heterotrophic equilibrium and decreasing the efficiency of carbon fixation; and watershed management needs to pay attention to the coupling of DOC and microbial functionality.

(1) This study only collected spatial scale samples at a single time node and failed to explore the trend of DOC content changes at different time scales. Future studies should comprehensively consider the changing patterns of DOC content at spatial and temporal scales to more fully understand the behavioural characteristics of DOC changes in different seasons and their impacts.

On the basis of the results of this study, the main sources of sediment DOC in the Wuding River Basin can be identified. To effectively control sediment DOC and increase the safety of river water quality, the following recommendations are proposed on the basis of the conclusions of this study: (1) Combine with soil and water conservation measures to reduce sediment inputs, and monitor the DOC dynamics in sediments to warn of water quality risks. (2) DOC input can be reduced by controlling agricultural surface pollution and restoring vegetation buffers to protect the anaerobic microenvironments in natural river sections.