Multiple spectral comparison of dissolved organic matter in the drainage basin of a reservoir in Northeast China: Implication for the interaction among organic matter, iron, and phosphorus

Abstract

Dissolved organic matter (DOM) plays a major role in ecological systems, affecting the fate and transportation of iron (Fe) and phosphorus (P). To better understand the geochemical cycling of these components, soil and sediment samples were collected around a reservoir downstream of a typical temperate forest in Northeast China. The DOM fractions from these soils, river, and reservoir sediments were extracted and then characterized by spectroscopic techniques. Comparative characterization data showed that the DOM pool of the Xishan Reservoir was partly autochthonous and derived from runoff and deposition of material in terrestrial ecosystems upstream. The upper reaches of the reservoir had significantly lower total Fe (TFe) content in the DOM extracts than those found in the reservoir (p < 0.05). Within the DOM, TFe was correlated with the amino acid tryptophan (p < 0.01). There was also a strong positive correlation between total P (TP) concentrations in DOM and tyrosine (p < 0.01). Organic P (P_o) comprised most of the DOM TP, and was related to dissolved organic carbon (DOC) content and the amino acid tyrosine (p < 0.01). The interaction among DOM, Fe, and P appears to be due to complexation with tryptophan (Fe) and tyrosine (P). This suggests that the formation of Fe-DOM-P would be produced more readily than DOM-Fe-P complexes under optimal conditions. The interaction among DOM, Fe, and P can promote the coordinated migration, transformation, and ultimate fate of components that are complex with DOM from riverine and reservoir ecosystems, ultimately leading to accumulation within a reservoir and transport to downstream regions when reservoir dams are released. Reservoir dams can effectively intercept DOM and minerals prevent its flow downstream; however, it is important to understand the co-cycling of DOM, Fe and P within reservoirs, downstream rivers, and ultimately oceans. The involvement of amino acid components of DOM, tyrosine and tryptophan, in DOM complexation is an issue that requires further study.

Graphical abstract

Highlights

• •The interaction among DOM, Fe, and P was investigated in a watershed system.
• •Interaction among DOM and Fe appeared to be due to complexation with tryptophan.
• •Interaction among DOM and P appeared to be due to complexation with tyrosine.
• •The interaction of DOM, Fe, and P generally leads to the formation of Fe-DOM-P complexes.
• •Interactions between organic and mineral components promote the coordinated migration and transformation of DOM, Fe, and P.

1. Introduction

Dissolved organic matter (DOM) is one of the largest pools of carbon (C) in rivers, reservoirs and lakes [1], with dissolved organic carbon (DOC) in DOM accounting for approximately 50% of the total C in inland water bodies [2]. Thus, DOM plays an important role in the C cycle in terrestrial and aquatic ecosystems, both regionally and globally [3]. DOM consists of a range of organic molecules comprising oxygen (O)-containing acid functional groups (e.g., carboxyls and phenols), as well as relatively less abundant nitrogen (N)- and sulfur (S)-containing functional groups. All these functional groups display strong binding affinities to trace metals, including iron (Fe) [4]. DOM is considered an essential carrier in ecosystems due to its binding to nutrients, such as phosphorus (P), and heavy metals, including Fe, which alters the solubility, bioavailability, mobility, distribution, and ultimate fate of organic and inorganic substances in aquatic and terrestrial systems [5].

Phosphorus is an essential nutrient required for life, however, it can also contribute to the eutrophication of lakes, rivers, and other aquatic systems [6]. Aquatic sediments tend to be enriched in P and can maintain eutrophic (i.e., nutrient-enriched) conditions of lakes or reservoirs for long periods (years, decades, centuries) after the exogenous P load is effectively controlled (Zhu et al., 2018). P exists in sediments by binding to aluminum (Al) oxides and the interiors of Fe and calcium (Ca) oxides to form complexes that vary in motility and bioavailability [7]. Among these, Fe–P is readily dissolved under alkaline conditions and is converted to dissolved orthophosphate by redox reactions in the environment [8]. Furthermore, organic P (P_o) can be converted into inorganic P (P_i) by various enzymatic and physicochemical hydrolysis reactions [9]. Fe is a crucial component of biogeochemical cycling and energy transfer in living systems and is closely associated with DOM and P [5]. Fe(III), in particular, is insoluble in water, has a neutral pH, and can be precipitated as various Fe(III)(hydr) oxides. Lalonde, Mucci, Ouellet and Gélinas [10] found that 22% of organic C was directly bound to Fe(hydr)oxides in sediments. These Fe (hydr)oxides can fix organic C through adsorption, co-precipitation, complexation, and other processes. The Fe mineral adsorption and electrons transfer capacity of DOM with different molecular weight fractions were significantly different during the Fe reduction process. The oxygen-rich and highly aromatic high molecular DOM exhibited stronger adsorption affinity onto Fe minerals than other DOM fractions [11].

Numerous research efforts have focused on the binary relationship between DOM and Fe, Fe and P, as well as DOM and P [8,12,13]. Natural Fe-rich colloids typically contain various amounts of DOC and DOC-rich waters typically also are enriched in Fe [14]. The most common formation of complex organic-rich-Fe structures occurs when DOM is co-precipitated with Fe. In contrast, DOM can also be adsorbed on Fe oxides, where organic molecules are ‘glued’ together by Fe ions or nanophases of Fe oxides [10]. DOM can also increase the oxidation rate of Fe (II) by O₂ [15]. During sediment suspension in shallow and other lakes, Fe (II) is oxidized to Fe(III) under anaerobic conditions [16]. A large quantity of P is fixed by Fe(III) in the formation of Fe(OOH)–P complexes or as precipitates on the surface of sediments [17]. Both of these mechanisms of P retention inhibit the release of P into the surrounding ecosystem. In addition, P release can be controlled by microbially-mediated OM decomposition, which is largely a function of temperature and the microbial community present [18,19]. Organic P can represent the main pool of potentially mobile P in some sediments where OM decomposition is occurring [9]. However, selected P_o compounds exhibit different affinities and stabilities towards aquatic sediments that may delay or expedite their degradation. The specific fractions of P found in a system are somewhat a function of the DOM present. For instance, humic acid (HA) and fulvic acid (FA) are potent chelating agents that stabilize P binding by cations like Ca, Fe, and Al [20]. HA can also help increase the supply of bioavailable P by buffering soil pH, promoting soil aggregates formation, and competing with P for adsorption sites [21].

To date, research on the ternary interaction of DOM, Fe, and P is scant with no real conclusions drawn. Humic-metal-phosphate complexes can decrease phosphate fixation in soils and increase plant growth and phosphate uptake [22]. Tao, Du, Deng, Huang, Leng, Ma and Wang [23] suggest that two different processes between Fe(III) (oxyhydr) oxides and OM contribute to the high concentrations of TDP in groundwater, especially in alluvial-lacustrine plains. This information is important for understanding the coupling and co-cycling of DOM, Fe, and P in aquatic environments, including freshwater lakes [19,[24], [25], [26]]. Phosphorus can form associations with dissolved and particulate OM through polyvalent cation bridging [19,24,27]. This strong interaction among DOM, Fe, and P, such as P_o makes it likely to influence the enzymatic P_o hydrolysis and sediment P bioavailability [19,26]. Long-term P_o retention in soils and sediments is considered a source of Legacy P with the propensity to contaminate downstream systems later in time. The addition of metal ions can initiate the formation of ternary DOM-metal-P complexes in natural waters [28]. Specifically, the mobile fraction of HA is frequently attributed to the system's DOM pool [1]. Thus, HA can serve as a typical DOM alternative or proxy compound in previous research. HA can adsorb, exchange and complex with minerals such as Fe, Al, and Ca oxides in the environment, with the release of P and other anions [29]. Furthermore, HA could interact with Fe and P to form HA-Fe-P complexes, which may impact geochemical P cycling in lakes [26].

Along the river continuum, dammed reservoirs alter DOM concentrations, nutrient flows, and complex speciation. This scenario is likely to influence the interaction among DOM, Fe, and P in riverine systems. For example, Kao, Mohamed, Sorichetti, Niederkorn, Van Cappellen and Parsons [30] reported that the Fanshawe Reservoir in the United Kingdom is a significant P sink for the Thames River and regulates the timing and speciation of P in the system. Nearly half of the reservoirs worldwide retain 80% or more of the inflowing sediment [31]. The Xishan Reservoir is a typical large artificial reservoir in Northeast China, with extensive vegetation coverage, little human intervention and abundant natural OM and nutrients. In particular, the background value of Fe is high in the drainage area of the Xishan Reservoir. Therefore, the Xishan Reservoir provides a pilot for the study of DOM, Fe, and P interactions. We hypothesized that the reservoir itself intercepts OM derived from large forests in the Xiaoxing'an Mountains and is important for the migration and transformation of DOM, Fe, and P in these forest-river-reservoir systems. Based on this assumption, the main objectives of this study were: (1) to characterize DOM in soil samples and sediments from rivers and reservoirs, (2) to analyze the interaction among OM, P, and Fe by the construction of dammed reservoirs, and (3) to improve understanding on the transformation of Fe and P, together with DOM, after the construction of artificially dammed reservoirs.

2. Materials and methods

2.1. Study sites and sampling

The Xishan Reservoir is a typical artificial reservoir in NE China, located at the middle reaches of the Yichun River and south of the Xiaoxing'an Mountains (Fig. 1). As an important river in NE China, part of the Yichun River originates in the Xiaoxing'an Mountains, passes through the Songhua River, and finally flows into the Heilongjiang and Reka Amur Rivers until it reaches the Tatar Strait that separates China from Russia. Completed in 2009, the Xishan Reservoir is large and covers an area of 11.5 km² with three major river systems in the basin: the Dangshi, the Cuiluan, and the Yao Rivers. The reservoir catchment area is 1613 km². The upper Xiaoxing'an Mountains are predominantly forested woodland with low many small trees and shrubs, accounting for 98% of the watershed area. The area has high natural fertility, and the concentrations of OM and humus are more prominent.

Fig. 1.

Location of the Xishan Reservoir and the sampling sites.

In late May and early June 2021 (flood season), a detailed survey of the Xishan reservoir basin was conducted, which included 21 study sites (Fig. 1). Sediment and water samples were collected from river channels, reservoir areas, and soils near major upstream sources. From these sites, sites S1 to S8 were located in the reservoir itself, while sites R1 to R6, R7 to R11, and R12 to R14 were located in the upper reaches of the Yao, Cuiluan, and Dangshi Rivers, respectively. Water samples were collected by a plexiglass water sampler placed 20 cm underneath the water surface. Surface sediment samples (Approximately 0–10 cm depth) were collected by a grab sampler. Soil and sediment samples were transported to a laboratory in airtight plastic bags under cold storage. Then, the samples were lyophilized, ground to powder, and stored at 0 and -20 °C until further processing.

2.2. Extraction of dissolved organic matter

Aliquots of sampled soil or sediment were weighed into centrifuge tubes and 1 mol L⁻¹ KCl was added at a ratio of 1:10 (mass/volume). The mixtures were shaken at a rate of 200 r min^-1 for 24 h at 20 °C. The mixtures were then centrifuged, with supernatants (extracts) removed, filtered through 0.45 μm nitrocellulose acetate filters, and stored at 4 °C in the dark until use.

DOC concentrations in extracts (mg C L⁻¹) were determined with a multi-N/C 2100 analyzer (Jena, Germany). Concentrations of TP in extracts were determined after digestion with potassium persulfate (K₂S₂O₈) at 121 °C for 30 min. Concentrations of Pi in the extracts were determined directly by the molybdenum blue/ascorbic acid method [32]. Concentrations of P_o were then determined as the difference between the concentrations of TP and P_i. Concentrations of extractable Fe (TFe) were analyzed by an ICP‒OES (ICPE-9800 apparatus, Shimadzu, Kyoto, Japan).

2.3. Instrumental characterization of DOM

2.3.1. UV–vis absorbance spectrophotometry

UV–vis absorption spectra of DOM were collected between the wavelengths of 200 and 900 nm (at 1 nm intervals) using an Agilent 8453 UV–vis spectrophotometer (Agilent Technologies, CA). A Milli-Q water sample was used as the blank reference. The absorption spectra were baseline corrected by subtracting the mean absorbance between 690 and 700 nm [33]. Sample DOM was generally characterized with several indicators, including UV absorbance at 254 nm (UV₂₅₄), and some specific spectroscopic indicators based on wavelength ratios, such as spectral absorption ratios (A₂₅₀/A₃₆₅ and A₂₅₃/A₂₀₃) [34].

2.3.2. DOM fluorescence

Three-dimensional fluorescence excitation−emission matrix (3D EEM) spectra were measured with an F-7000 fluorescence spectrometer with a 700 V xenon lamp (Hitachi High-Technologies, Tokyo, Japan). The inner filter effect was corrected using the absorption spectra [35]. The blank correction was performed daily by subtracting the EEM spectra of Milli-Q water. Normalization was performed relative to the area of the Raman peak of Milli-Q water [36]. PARAFAC analysis was applied to group sample components using the DOMFluor v.1.7 Toolbox combined with MATLAB R2010a. Based on EEM-PARAFAC analyses, a variety of indices were obtained to quantify differences in DOM fluorescence properties, including the fluorescence index (FI), humification index (HIX), and biological index (BIX). Protein-like substances, including tyrosine-like and tryptophan-like components, may serve as indicators of bioavailable DOM [37].

2.3.3. Orbitrap Fusion mass spectrometer

Extracted DOM was concentrated and desalted using solid-phase extraction (SPE) immediately after filtration to minimize hydrolysis and other sample alterations before analysis. Solid-phase extraction procedures were adapted from Dittmar, Koch, Hertkorn and Kattner [38]. The concentrated, de-salted samples were analyzed by an Orbitrap Fusion mass spectrometer (Thermo Scientific, San Jose, CA). Electrospray ionization (ESI) was used in negative ionization mode. The DOM samples were diluted to a DOC concentration of 50 mg L^-1 and injected into the ESI source at a rate of 180 mL h^-1. The ESI parameters were optimized as follows: negative ion spray voltage, 2600 V; sheath gas (Arb), 5; aux gas (Arb), 2; sweep gas (Arb), 0.1; ion transfer tube temperature, 300 °C; and vaporizer temperature, 20 °C. Full-scan data were acquired in negative mode at a resolving power R of 5000,000. A scan range of m/z 150–1000 was chosen. The automatic gain control (AGC) was set at 5.0e⁵. The injection time was set to 100 ms, and the S-lens RF level was set to 60.

2.4. Statistical analyses

The changes in EEM fluorescence and UV absorbance of DOM were evaluated by the nonparametric Mann‒Whitney U test. Principal component analysis (PCA) was used to identify the possible sources and components of DOM. These analyses were performed using SPSS 18.0 for Windows. The results were considered statistically significant at a confidence level of 95% (p < 0.05).

3. Results

3.1. Characteristics of DOM, phosphorus and iron in the water

Concentrations of DOC in the overlying Xishan reservoir water varied from 6.2 to 10.0 mg L⁻¹ (Table 1). The concentrations of TFe and P were highest in the center of the reservoir at sites S7 and S8. The TFe concentrations in the reservoir water varied from 0.3 to 1.5 mg L⁻¹, with an average of 0.5 mg L⁻¹. Concentrations of TP in the reservoir water varied from 0.1 to 0.2 mg L⁻¹. TP concentration was the highest at S7 near the Dangshi River estuary. The concentration of SRP in the water was negligible (0–0.009 mg L⁻¹). Therefore, TP mainly existed in the form of particulate P or P_o in the water. The water pH was between 6.9 and 7.1.

Table 1.

Physical and chemical parameters of overlying water in the Xishan Reservior.

Sites	DOC	TFe	TP	SRPa	pH
	mg L−1				–
S1	7.85	0.50	0.17	0.009	6.9
S4	6.23	0.25	0.11	-b	7.1
S5	9.98	0.30	0.15	0.001	7.1
S6	9.58	0.34	0.15	0.006	6.9
S7	9.88	1.53	0.19	–	6.9
S8	9.64	0.28	0.16	–	7.0

^aSRP is soluble reactive phosphorus.

^bThe value is lower than the detection limit.

There was little difference in the fluorescence characteristics of the water from different sites in the reservoir (Fig. S1). Observed peaks A and C, which represent humus-like fluorescent substances [39] were found in various concentrations. The fluorescence intensities of DOM in the λ_Em = 230–310 nm and λ_Ex = 380–460 nm regions were notable, in comparison with the fluorescence intensities of other bands. The fluorescence characteristics of DOM peak A and peak C were mainly related to the fluorescence characteristics of terrigenous humus. These results indicate that the DOM in the overlying water is derived from terrigenous humus that was transported into the reservoir via the river's transport of naturally decomposing litter vegetation upstream and during runoff from terrestrial sources through runoff.

3.2. Concentrations of water-extractable DOC, iron, and P

Concentrations of DOC in sediment extracts ranged from 4.7 to 20.2 mg L⁻¹, with an average of 11.7 mg L⁻¹ (Fig. 2). DOC in the three upstream rivers varied greatly, ranging from 1.0 to 37.9 mg L⁻¹, with an average of 10.5 mg L⁻¹(Fig. 2c), slightly lower than DOC in reservoir sediments (Fig. 2f). The highest DOC was 20.2 mg L⁻¹ at S8 in the central area of the Xishan Reservoir. The DOC at sites S3 and S4 in the Cuiluan River estuary were both 14.0 mg L⁻¹. The lowest concentration of DOC was 4.7 mg L⁻¹ at site S6. The DOC at the sites R1, R7 and R12 were higher than others, and DOC at R1 was the highest, reaching 37.9 mg L⁻¹. These results may be related to the wide, large catchment area of the Yao River Basin. In general, DOC in reservoir sediments was higher than in the upstream area.

Fig. 2.

Distribution characteristics of TP (a, d), TFe (b, e), and DOC (c, f) concentrations in DOM from the Xishan Reservoir (d, e, f) and its upstream region (a, b, c).

Concentrations of TFe in reservoir DOM ranged widely from 0.4 to 18.5 mg L⁻¹, with an average of 4.4 mg L⁻¹(Fig. 2e). This TFe was significantly higher than the value of 0.2 mg L⁻¹ found in the three upstream rivers sampled (Fig. 2b). The lowest TFe was at site S3, which was only 0.38 mg L⁻¹. The highest concentration was found in the middle of the reservoir at site S8. This result aligns with the distribution of DOC in the reservoir. Concentrations of TFe generally increased from the upstream rivers to the reservoir. TP concentrations in reservoir samples ranged from 0.10 to 0.30 mg L⁻¹, with an average of 0.18 mg L⁻¹ (Fig. 2d). This was higher than the average concentration of 0.16 mg L⁻¹ in upstream TP (Fig. 2a). The TP was the highest at site R1, reaching 0.73 mg L⁻¹ (Fig. S1). In all samples, P_i accounted for only a small proportion of TP (Table S1). Thus, P_o is the main component of TP in DOM from this freshwater system and could account for 84% of TP. Concentrations of DOC and P_o were correlated (p < 0.05) and had similar spatial distribution characteristics. These results indicate that the source and transformation of P_o in the drainage basin are closely related to OM.

3.3. Characterization of DOM by UV–vis absorbance

All the UV–vis absorbance spectra of collected DOM decreased monotonically with increasing wavelength (Fig. S2), similar to those noted for humic substances (HS) from many environments [40]. The UV₂₅₄ value at the mouth of the Yao River into the reservoir was the highest at site S1 (Fig. 3 and Fig. S2). The average UV₂₅₄ value of reservoir sediment DOM was approximately 0.03, which was higher than the average value (0.02) of UV₂₅₄ in the DOM samples from the upstream area. The average value of A₂₅₃/A₂₀₃ of upstream DOM samples was approximately 0.06, which was slightly lower than the value of 0.07 found in reservoir samples. These results may indicate that upstream DOM contains more carbonyl, carboxyl, and hydroxyl groups, as well as lipids, and condensed aromatic rings than DOM from the reservoir [41]. Some degradation of DOM may have occurred in the reservoir. The average reservoir value of A₂₅₀/A₃₆₅ was 3.0, which was less than the 3.6 value found upstream. Thus, the aromaticity and high molecular weight of reservoir DOM were likely greater than those found in upstream DOM [34].

Fig. 3.

Optical indicators in the Xishan Reservoir and its upstream region (RA: reservoir; UP: upstream area).

3.4. Fluorescence characteristics of DOM

Based on 3D EEM spectra of DOM from sediments and soils from the Xishan Reservoir (Fig. S3), DOM mainly contained humus-like components (Peak C) and protein-like components (Peak B1 and Peak B2). The protein-like component at peaks B1 and B2 present in the 270–290 (225–240)/300–315 nm band of the Ex/Em spectra reflect biodegradable tryptophan substances, which are related to the aromatic acid structure in DOM [39]. Based on the PARAFAC model, five fluorescent components (C1, C2, C3, C4, and C5) were obtained and tested by fissure semi-analysis and residual analysis, which proved that the model was valid (Fig. S4) [42]. The main fluorescent peaks of these analyzed DOM are detailed in Table S2. Among them, C1 and C3 belong to terrigenous humus fluorescence components, which mainly reflect the fluorescence characteristics of shortwave humus, primarily imported from terrigenous sources and combined with authigenic OM in the reservoir water. The fluorescent components C2, C4 and C5 belong to typical protein-like fluorescence peaks, which are mainly tyrosine-like fluorescence groups [39].

Fluorescence indicators, including FI, HIX, and BIX, were also analyzed to characterize DOM components (Table S3). The DOM FI values ranged from 1.8 to 2.8, with an average of 2.1 in upstream soils or river sediments. The average FI value of reservoir sediment DOM was 2.4, higher than found upstream. The FI values of sediment and soil samples varied marginally among sampling regions, and all were approximately 2.0. The DOM BIX values ranged from 0.7 to 1.3 (Fig. 3), with an average of 0.9 in the reservoir and upstream. Among them, BIX ranged from 0.26 to 1.1, with an average of 1.1 for reservoir DOM. For upstream DOM, the average BIX was 0.88, and the minimum was 0.84, as observed at the Cuiluan River. HIX is an indicator of the extent of complexity and humification of an environmental sample [43]. The DOM HIX values ranged from 0.26 to 4.09, with an average of 1.16 (Fig. 3 and Table 2). The average HIX value in the reservoir was 0.76, which was lower than that found upstream (HIX was 1.30). The HIX of the Yao River was approximately twice that of the reservoir itself, which presents as having weak humus-like characteristics [44]. The HIX values were mostly <1.5, which indicates that the degree of DOM humification in soils or sediments was low in this freshwater system.

Table 2.

Summary of the variation range of fluorescence intensity of five components in sediment and soil samples from the Xishan Reservoir.

	value	Upstream	Surface sediments	Soil samples
C1	range	0.09–0.56	0.12–0.20	0.15–0.36
	average	0.24	0.16	0.29
C2	range	0.00–0.75	0.09–0.24	0.30–0.91
	range	0.32	0.16	0.70
C3	range	0.01–1.50	0.04–0.12	0.09–0.21
	average	0.21	0.09	0.18
C4	range	0.28–0.74	0.31–0.66	0.53–0.75
	average	0.57	0.50	0.60
C5	range	0.04–0.23	0.06–0.76	0.08–0.13
	average	0.11	0.20	0.12

Fluorescence intensity was reported as the maximum fluorescence (Fmax), which is the unique value of each DOM component corresponding with the relative amount of a particular fluorescing component [42]. Fluorescence intensity can indirectly reflect DOM concentrations (Fig. 4a and Table S4) [33]. The average value of the total relative fluorescence intensity of upstream DOM was 0.29, higher than the observed value of 0.22 in reservoir sediments. There was little difference in the DOM between the Xishan Reservoir and the upstream area, which was consistent with the observed DOC data. Protein-like component C4 accounted for a large proportion of the total DOM, with a relative proportion near 39 (Fig. 4b). Component C2 accounted for approximately 23% of total F_max. The relative proportions of humus-like components C1 and C3 were low, averaging 16.% and 12%, respectively. Overall, protein-like components, including C2, C4 and C5, comprised the majority of DOM in this study, with an average proportion of 72%, which exceeded that of DOM humus-like components.

Fig. 4.

DOM fluorescence intensity (a) and relative proportion (b) of different components along the sampling sites in the Xishan Reservoir and its upstream area (the site of the DOM samples is the same).

3.5. Characterization of DOM by Orbitrap MS

Based on UV–vis and 3D EEM spectral data, six typical DOM extracts were selected for further molecular composition analyses with an Orbitrap Fusion MS (Thermo Scientific, San Jose, CA). The DOM responded well in ESI mode, where molecular weights were primarily measured in the 200–450 Da range, with mass centers of approximately 300 Da (Fig. S5). The number of different molecular weights determined from DOM at sites S4 and S8 were 3639 and 5445, respectively. At site R5-S in the upper reaches of the Yao River, 8872 molecular formulas were obtained. Located in the Cuiluan River at the R11-S and R10 sites, 7710 and 7136 molecular formulas were obtained, respectively. The Dangshi River at the R14-S site had approximately 4656 molecular formulas. More formulas were detected in the Cuiluan River sample. In general, compounds identified at m/z 360 of the six DOM samples were mainly comprised of CHO compounds with various amounts of oxygen (O₆–O₁₁) (Fig. S6). The relative abundance of O₉ and O₁₀-containing DOM was the highest at sites S4 and R5-S. The relative abundance of S-containing DOM was the highest at site S8 in the Xishan Reservoir, and the relative abundance of nitrogen (N)-containing DOM was highest at site R5-S. The identification degree of the Dangshi River 14-S site was low, with only a low abundance of CHO compounds. Heteroatomic DOM types were primarily CHO, CHON, CHOS, and CHONS. Fig. 5 presents a normalized relative abundance and distribution of these compounds for the reservoir system under study. Nitrogen-containing CHO (CHON) compounds were dominant among the four heteroatomic DOM compounds in these samples, with a relative contribution of 23–43%, while CHONS compounds were found to have the lowest relative percentage for this freshwater system.

Fig. 5.

Relative percentage of assigned molecular formulas based on atomic compositions.

The molecular composition DOM was further analyzed by V–K diagrams (Fig. 6 and Table 3) [33]. This analysis showed that percentages and varieties of lipids, proteins and sugars in reservoir DOM (Fig. 6a and b) were significantly higher than those found upstream (p = 0.005), with the absolute highest found in the center of the reservoir. In contrast, the percentage of condensed substances with aromatic structures had the lowest concentrations and variety. Unsaturated hydrocarbons were present at sites R5-S (Fig. 6c) and R11-S (Fig. 6d) from the Yao and Cuiluan Rivers, respectively. The percentage of lignin content was also higher in DOM from these two rivers. Lignin is a complex biopolymer involved in structural integrity and is typically abundant in terrestrial vascular plants. Lignin is often used to track vegetation shifts [45]. The lignin content at reservoir site S4 was low, which may be due to microbial decomposition or low lignin content in the parent material. Additionally, the percentage of tannins in the DOM samples from site R14-S (Fig. 6e) in the Dangshi River was higher than that from other sites upstream.

Fig. 6.

Van Krevelen diagram of the DOM samples from the Xishan Reservoir (a, b) and its upstream region (d, e,f) analyzed by Orbitrap MS (a: S3; b: S4; c:R5-S; d:R11-S; e: R14-S; f: R10).

Table 3.

Contribution and percentage of molecular formulas in different groups of compounds relative to the total number of molecules.

Group (%)	S4	S8	R5-S	R11-S	R14-S	R10
Lipids	2.06	3.05	2.50	3.43	3.16	2.45
Proteins/amino sugars	25.32	28.52	15.85	17.71	19.22	14.81
Carbohydrates	10.34	15.44	3.76	4.27	4.19	2.99
Unsaturated hydrocarbons	0	0	0.04	0.04	0	0
Lignins	44.47	40.74	52.93	52.93	49.5	55.25
Tannins	10.69	10.69	17.56	15.25	17.6	16.11
Condensed hydrocarbons	1.56	1.56	7.32	6.39	6.33	8.38

4. Discussion

4.1. Characteristics, source, and transformation of DOM

Forest ecosystems produce more leaves and fresh litter than crop and grass ecosystems, which can then leach out this DOC to other systems [46]. Displaced DOC can then be transferred to nearby aquatic ecosystems and influence riverine DOC loads. High DOC concentrations suggest increased catchment-derived allochthonous (i.e., introduced from an alternate source) DOM, which may possess a higher degree of aromaticity, the potential for higher molecular weight compounds, and a lower degree of oxidation than DOM that has already been subjected to biochemical processes that result in lower concentrations of oxygen (O)-containing acidic functional groups, as well as oxo carboxylic and phenolic groups [47]. In the current study, there were no significant differences in DOC concentrations observed between upstream and reservoir DOM (p = 0.745). However, the total DOC concentrations in the combined riverine upstream sites were lower than within the reservoir itself. This result indicated that the reservoir received heterologous DOM from upstream sources.

It has been reported that typical A₂₅₀/A₃₆₅ values of HA are <3.5 and FA has typical values > 3.5 [34]. Most observed A₂₅₀/A₃₆₅ values of DOM from the Xishan Reservoir and its upstream tributaries were less than 3.5. This indicates that the proportion of HA was greater than that of FA in these DOM samples and, in general, they had a high molecular mass. Compared with upstream DOM, reservoir DOM contained more high molecular mass aromatic compounds with unsaturated C–C bonds, which reduces the degradability of these DOM compounds in sediments. The difference between sediment FA and HA contents is likely due to both the internal environment and external inputs. Terrestrial inputs, including surface runoff, increase HA concentrations in reservoir sediments. This work supports the hypothesis that DOM is transformed from upstream sources to downstream reservoir sediments.

Measured fluorescence characteristics show that sediment DOM from the Xishan Reservoir had obvious land-based input characteristics in addition to DOM with high bioavailability following microbial hydrolysis. The FI parameter is often used to characterize sources of HA in DOM [44]. In the current study, most of the DOM FI values were >1.9 in the reservoir, suggesting that input and accumulation of endogenous DOM are related to microbial activities [44]. Compared with FI values of reservoir DOM (average value of 2.35) and its upstream region (average value of 2.00), DOM appeared to be more susceptible to microbial activity during transport from upstream to the reservoir. The BIX values are often used to differentiate DOM from various sources and this study showed that DOM from typical land plants was transferred to soils, and then on to sediments of rivers and reservoirs, with microbial activity occurring throughout the process [48]. Total relative fluorescence intensity values (F_max) were 0.38, 0.29, and 0.22 in the DOM from soils, river sediments, and reservoir sediments, respectively (Table S3). This result indicates that the concentrations of soil DOM were higher than those from other sources tested. The average percentage of protein-like components in the DOM in the river and reservoir were 29%; however, the average percentage of protein-like components in soil DOM was 47%. This was likely because a fraction of protein-like components was degraded by microbial action during the migration of DOM from soils to sediments.

To further explore the complexity of the DOM characteristics, PCA was applied to further analyze the molecular indices derived from the UV–vis and fluorescence spectra (Fig. 7). Principal component 1 (PC1) explained 30% of the total variance, while 23% of the variation was represented by PC2. PC1 was positively associated with the parameters C1, C2, C3, HIX, A₂₅₃/A₂₀₃, and UV₂₅₄. PC1 was negatively correlated with FI and BIX. These results indicate that PC1 was related to terrestrial DOM. Moreover, the positive loadings of PC2 were dominated by BIX and C5. HIX displayed a negative loading on PC2. Thus, PC2 was correlated with DOM aromaticity. These results further indicate that the OM derived from terrestrial plants was the main DOM source to the drainage basin of the Xishan Reservoir.

Fig. 7.

Principal component analysis for optical indices of DOM samples from the Xishan Reservoir and its upstream region (R: reservoir; UP: upstream).

High numbers of C in a compound, which indicate molecular unsaturation, are the primary CHO and CHON elemental components in natural systems [49]. In this study, the large proportion of humic-like compounds found could also be explained by a high proportion of CHO. This result is further confirmed by EEM fluorescence results. Lignin was the most abundant DOM component in samples from this drainage basin (Table 3) and is characterized by high C: N ratios. A large proportion of lignin and other plant-derived OM can be readily flushed into streams and reservoirs during periods of runoff [50]. Furthermore, these compounds are believed to be readily mineralized in streams [51].

4.2. Interaction and co-migration of Fe and P with DOM

To investigate the influence of DOM on the transformation of Fe and P in the Xishan Reservoir, the relationships between Fe and P and the parameters of the DOM samples were further analyzed (Fig. 8). In this study, concentrations of TFe were positively correlated with the C5 components of DOM in the extracts, which indicated that the transformation of Fe was likely closely related to the component tryptophan in the DOM. The stability constant of the Fe–tryptophan complex is 7.6 [52], which is stronger than that of Fe-HA [53]. Thus, it is possible that DOM-Fe complexes in this system were formed by tryptophan components in DOM.

Fig. 8.

Correlations between parameters of DOM, Fe, and P in the DOM samples from soils and sediments in the Xishan Reservoir and its upstream region (* is p ≤ 0.05, ** is p ≤ 0.01).

Additionally, the high tryptophan content of A. ferrooxidans may favor electron transfer and significantly alter the interfacial potential (i.e., charge between the boundary of two compounds) [54]. Moreover, tryptophan could also be complex with Fe (Fe²⁺, Fe³⁺), where Fe³⁺ chelates tryptophan to form a Fe³⁺-tryptophan complex due to the oxidation of Fe²⁺. This reaction creates conditions favorable for the formation of needle-like goethite [54]. While the pH conditions in the current study were neutral, At pH 4 and 12, tryptophan-derived organic ligands induce the formation of unstable Fe, that is, tryptophan coordinates with unstable Fe to adsorb onto the surface of Fe-containing minerals. So organic ligands made from tryptophan can lead to the separation of surface-active Fe complexes to release more Fe [53]. Previous studies have revealed that systems with lower OM content (i.e., C: Fe molar ratio of 0.05) had little effect on the transformation of Fe (hydr)oxides, while other systems with higher OM content (i.e., C: Fe molar ratio of 0.7) can strongly inhibit the transformation of Fe (hydr)oxides [55]. In another study, wetland-derived humic substance (HS) had a higher Fe-binding capacity than plankton-derived HS [56]. The Xishan Reservoir has abundant rainfall and lush vegetation with high molecular mass and more aromatic constituents. The DOM of this reservoir has high O/C and low H/C ratios after high rainfall due to the leaching of terrestrial DOM, where large quantities of OM flow into the reservoir with runoff, combines with Fe, and then are intercepted by the reservoir dam.

Concentrations of TP were positively correlated with DOC levels (Fig. 8). P_o is the main component of TP in the DOM under study, which was also significantly correlated with DOC and tyrosine (p < 0.01). This relationship may indicate that the source of P_o was consistent with the upstream soil, which was the source of the DOM. In addition, a high sediment TP concentration can lead, to increased rates of organic sedimentation and a resulting increase in bacterial activity, which provides the ecosystem with large amounts of DOM and induces ectoenzyme synthesis (i.e., excreted enzymes from microbes and root materials) [57]. Studies have shown that the concentration of OM near the reservoir on the Creuse River in France exceeded that of river sectors by a factor of 7 [58], and OM and P_o were positively correlated during the rainy season in the study (p < 0.01). Altering the moisture and texture of sediments creates conditions where P and Fe complexation is favorable, stimulating the growth of plants and microbes. Consistent with mass spectrometry data, algae-derived DOM contains more Lipid-like compounds, while macrophyte-derived DOM contains more lignin- and tannin-like compounds [59]. Tyrosine-like substances in the environment originate from microbial activities, where algal biomass protein is rich in tyrosine [60]. Thus, the C2 component of DOM in the current study indicates a high nutrient supply in this aquatic environment and had positive effects on both biomass production and assimilation of P. In addition, the P input from upstream sources increased the degree of sediment DOM aromatization to form simple small-molecule DOM substances, such as tyrosine, tryptophan, and protein. DOM-P_o association appears to be due to covalent bonding either at the originating source or under the action of microbial activities during co-migration of DOM and P in the river to reservoir system.

The interaction of DOM with Fe and P was likely due to complexation with the components tryptophan (Fe) and tyrosine (P) in this study. Compared with DOM-Fe-P, complexes of Fe-DOM-P were more likely to be formed within the reservoir via different DOM functional groups. In forest soil, previous studies have found that the higher molecular weight DOM-Fe³⁺-P complex formed by P, HS, and Fe³⁺ was the main mechanism for the migration of P [5]. Based on the Scatchard method, Guardado, Urrutia and Garcia-Mina [61] analyzed laboratory-prepared DOM-Fe-P complexes and found that the binding of phosphate to DOM-Fe complexes has two binding sites with different stabilities. One is high stability, and the other is low stability, where the two binding sites have different maximum binding capacities and apparent stability constants (lgK) under different pH conditions. Although the bioavailability of complex P in the environment is still controversial, most studies believe that P complexation increases the mobility of environmental P [5]. Due to the intrinsic properties of complexed P, soil and sediment DOM-Fe-P can be found at soil-water interfaces such as soil (sediment) and overlying water, particularly in ecologically sensitive areas such as the ebb and flow zones of reservoir areas. During geochemical cycling, P in the form of dissolved DOM-Metal-P complexes in the aquatic environment will be degraded due to environmental factors such as light intensity and pH, with DOM, Fe, and P released upon decomposition. However, for Fe-DOM-P, which is likely more sensitive to DO, the crystal size of the formed complex by co-precipitation of DOM and Fe is small and has weak crystallinity and a high specific surface area [62]. Therefore, Fe (hydr)oxides in the Fe-DOM-P complexes found in the current study are more readily prone to reduction and release and may create continued eutrophic conditions for aquatic systems in the future.

4.3. Environmental significance

Reservoirs act as “in-stream” reactors, impeding the flow of nutrients and increasing the residence time of water in the river. DOM, P, and Fe generated by basin or reservoir autogenesis (i.e., forms in place) will gradually accumulate at the bottom of reservoirs and form stable complexes [17]. These complexes include small inorganic colloids containing Fe and P that are peptized (dispersed in a colloidal form) by surface adsorption of DOM or Fe bonding to P-containing DOM, to form Fe-DOM-P complexes. The alternative is that Fe hydroxides adsorbed to the surface of HS form a bridge between DOM and P to form DOM-Fe-P complexes. However, the interaction of DOM, Fe, and P would be more likely to form Fe-DOM-P complexes rather than DOM-Fe-P complexes in the Xishan Reservoir. It should be noted that Fe is exposed to the surface of DOM during the formation of Fe-DOM-P complexes, which have a strong sensitivity to DO changes at the sediment-water interface [63]. Global warming has changed the physical and chemical environment of lakes or reservoirs, and subsequently, changes in DO have also occurred that result in a cascade of DOM, P, and Fe complexes [64]. With the continuous accumulation of OM in this particular reservoir, as well as increases in reservoir depth, Fe-DOM-P complexes intercepted by the reservoir dam may be further released to the overlying water in the absence of O₂ at the sediment-water interface [65]. Therefore, management of such reservoirs and their organic and nutrient components requires that attention be paid to the DO levels at sediment-water interfaces. These systems can have conditions that are conducive to the long-term burial of C and nutrients, which may assist with preventing water eutrophication and severe short-term hypoxia. DOM, Fe, and P, and their interactions, may impact downstream regions of aquatic systems and may even be of significance in maintaining conditions of global oceans.

5. Conclusion

The interactions of DOM, Fe, and P in the drainage basin of a reservoir system, and the interception of these components by a reservoir dam were discussed in this study. Comparative characterization data showed that the DOM pool of the Xishan Reservoir was partly autochthonous with some components derived from runoff or deposition from the terrestrial ecosystem upstream. The Fe content in the DOM extracts from the upper reaches of the reservoir was significantly lower than in the reservoir area (p < 0.05). The transformation of dissolved Fe with DOM is likely influenced by the component tryptophan in the DOM in this watershed and is likely responsible for the interception of DOM-Fe complexes by the reservoir dam. There was a significant correlation between the component tyrosine and dissolved total P (TP) concentrations in the DOM (r = 0.71, p < 0.01, n = 30). Organic P (P_o) was the main form of TP in the DOM and was correlated with DOC and the component tyrosine (p < 0.01). Thus, tyrosine is likely the key factor in the interaction between DOM and P. For the drainage basin of this riverine system, the interaction among DOM, Fe, and P appears to be largely due to complexation with tryptophan (Fe) and tyrosine (P). The interaction of DOM, Fe, and P is more likely to form Fe-DOM-P complexes rather than DOM-Fe-P complexes. The interaction among DOM, Fe, and P promotes the coordinated migration and transformation of these three components from upstream soil to river sediments, and then to reservoir sediments, where reservoir dams intercept these complexes and they accumulate in the reservoir area. This is likely an important mechanism for DOM and nutrients by dam interception. Complexes of Fe-DOM-P can then be dissolved into the overlying water, with Fe³⁺ reduced to Fe²⁺. Thus, it is important for DO monitoring at the sediment-water interface of reservoirs to prevent the short-term release of DOM, Fe, and P into reservoirs. In addition, controlling and understanding these mechanisms is important for the long-term sequestration of sediment C. More research is needed to better understand the mechanisms between DOM, Fe, and P, with special emphasis on possible complexes of Fe-DOM-P. This type of work will assist with understanding the co-cycling of DOM, Fe and P in soils, rivers, reservoirs, oceans, and other aquatic systems.

Author contribution statement

Juan Jiang; Yuanrong Zhu; Jian Wei: Conceived and designed the experiments; Analyzed and interpreted the data; Contributed reagents, materials, analysis tools or data; Wrote the paper.

Juan Jiang; Xiaojie Bing; Kuo Wang; Huihui Ma; Fan Liu: Performed the experiments; Contributed reagents, materials, analysis tools or data.

Zhongqi He; Jing Ding: Analyzed and interpreted the data; Wrote the paper.

Funding statement

Yuanrong Zhu was supported by National Natural Science Foundation of China [41877380 and 41630645].

Data availability statement

Data included in article/supp. material/referenced in article.

Declaration of interest’s statement

The authors declare no competing interests.

Appendix A. Supplementary data

The following is the Supplementary data to this article: